From std5 at nyu.edu Wed Feb 1 09:22:27 2012 From: std5 at nyu.edu (Scot Dalton) Date: Wed, 1 Feb 2012 09:22:27 -0500 Subject: [Umlaut-general] Umlaut 3.0 progress update In-Reply-To: <4F2879C4.6040406@jhu.edu> References: <4F205DBA.1070803@jhu.edu> <87384E7A-DDD3-41D5-B036-921EC0B82BD7@nyu.edu> <4F2876AC.5000807@jhu.edu> <14DC76FB-717C-41A8-884B-A7EDEB04F2C9@nyu.edu> <4F2879C4.6040406@jhu.edu> Message-ID: <846216BB-53C4-4CEE-B2F6-7A44D7F687B2@nyu.edu> Hi Jonathan, Do you have any unit tests that you've written for other services? I'll check the wiki as well. Thanks, Scot On Jan 31, 2012, at Jan 31, 6:31 PM, Jonathan Rochkind wrote: > PPS: Rails 3 with bundler is more picky about gem dependencies being explicitly listed. It may just be that you were trying to use a gem (soap4r) that wasn't listed as a gem dependency. You can list it in your local app Gemfile -- or we can list it as a dependency of Umlaut itself, if it's needed for a service plugin we ship wtih Umlaut. For the latter, you'd need to check out a copy of Umlaut and change it's umlaut.gemspec (then once everything is working, commit the changes). > > https://github.com/team-umlaut/umlaut/wiki/Developing > > https://github.com/team-umlaut/umlaut/wiki/Writing-a-service-plugin > > > On 1/31/2012 6:23 PM, Scot Dalton wrote: >> I literally JUST got it up and running and started with the Primo stuff. I'll look at other service, hooks, etc. in a day or so. Trying to get my head around the new architecture right now. >> >> The install was easy once I got familiar with RVM and Passenger Standalone. The Umlaut install itself was very straightforward. Thanks for that! >> >> Re: the SOAP issue, it seems to be an incompatible gem. There are 1.9 ports of the soap4r gem, which I'll be trying out. >> >> Thanks, >> Scot >> >> On Jan 31, 2012, at Jan 31, 6:18 PM, Jonathan Rochkind wrote: >> >>> On 1/31/2012 6:15 PM, Scot Dalton wrote: >>>> Hi Jonathan, >>>> I finally had some time to install Umlaut 3.0. I've got it running as a Passenger Standalone instance using ruby 1.9.3 on a server with several ruby 1.8.7 apps. I'm using rvm to manage the various rubies and would be happy to share my experience in getting it installed. >>> Awesome! Did you have any trouble installing Umlaut 3.0alpha, or was it straightforward? Were my install instructions correct and complete? >>> >>> (Incidentally, Umlaut 3.0alpha does at the moment work perfectly fine on ruby 1.8.7 too). >>> >>> Other than unfinished Primo/aleph stuff, does it so far seem to be working as expected? Any notable features or hooks missing for you to get it to do what your Umlaut 2.x was doing, that you have noticed yet? >>> >>> Thanks! >>> >>> Jonathan >>> >> >> -- >> Scot Dalton >> Phone: (212) 998-2674 >> Web Services >> Division of Libraries >> New York University >> -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From rochkind at jhu.edu Wed Feb 1 11:13:25 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 01 Feb 2012 11:13:25 -0500 Subject: [Umlaut-general] Umlaut 3.0 progress update In-Reply-To: <35737645-6434-45C7-9CA2-C59F10CAD69F@nyu.edu> References: <4F205DBA.1070803@jhu.edu> <87384E7A-DDD3-41D5-B036-921EC0B82BD7@nyu.edu> <4F2876AC.5000807@jhu.edu> <14DC76FB-717C-41A8-884B-A7EDEB04F2C9@nyu.edu> <4F2879C4.6040406@jhu.edu> <35737645-6434-45C7-9CA2-C59F10CAD69F@nyu.edu> Message-ID: <4F2964A5.5070901@jhu.edu> We do want Umlaut to work under both 1.8.7 and 1.9.3 at present -- does the updated SOAP gem work under both, or do you need different gems for diff ruby versions? On 1/31/2012 7:49 PM, Scot Dalton wrote: > An updated SOAP gem for 1.9 in the Gemfile seemed to do the trick. There are discrepancies b/w Hpricot and Nokogiri that I'll have to work out tomorrow. Once the Primo service is actually working, I can deal with the umlaut.gemspec among my other commits. > > Thanks, > Scot > > On Jan 31, 2012, at Jan 31, 6:31 PM, Jonathan Rochkind wrote: > >> PPS: Rails 3 with bundler is more picky about gem dependencies being explicitly listed. It may just be that you were trying to use a gem (soap4r) that wasn't listed as a gem dependency. You can list it in your local app Gemfile -- or we can list it as a dependency of Umlaut itself, if it's needed for a service plugin we ship wtih Umlaut. For the latter, you'd need to check out a copy of Umlaut and change it's umlaut.gemspec (then once everything is working, commit the changes). >> >> https://github.com/team-umlaut/umlaut/wiki/Developing >> >> https://github.com/team-umlaut/umlaut/wiki/Writing-a-service-plugin >> >> >> On 1/31/2012 6:23 PM, Scot Dalton wrote: >>> I literally JUST got it up and running and started with the Primo stuff. I'll look at other service, hooks, etc. in a day or so. Trying to get my head around the new architecture right now. >>> >>> The install was easy once I got familiar with RVM and Passenger Standalone. The Umlaut install itself was very straightforward. Thanks for that! >>> >>> Re: the SOAP issue, it seems to be an incompatible gem. There are 1.9 ports of the soap4r gem, which I'll be trying out. >>> >>> Thanks, >>> Scot >>> >>> On Jan 31, 2012, at Jan 31, 6:18 PM, Jonathan Rochkind wrote: >>> >>>> On 1/31/2012 6:15 PM, Scot Dalton wrote: >>>>> Hi Jonathan, >>>>> I finally had some time to install Umlaut 3.0. I've got it running as a Passenger Standalone instance using ruby 1.9.3 on a server with several ruby 1.8.7 apps. I'm using rvm to manage the various rubies and would be happy to share my experience in getting it installed. >>>> Awesome! Did you have any trouble installing Umlaut 3.0alpha, or was it straightforward? Were my install instructions correct and complete? >>>> >>>> (Incidentally, Umlaut 3.0alpha does at the moment work perfectly fine on ruby 1.8.7 too). >>>> >>>> Other than unfinished Primo/aleph stuff, does it so far seem to be working as expected? Any notable features or hooks missing for you to get it to do what your Umlaut 2.x was doing, that you have noticed yet? >>>> >>>> Thanks! >>>> >>>> Jonathan >>>> >>> -- >>> Scot Dalton >>> Phone: (212) 998-2674 >>> Web Services >>> Division of Libraries >>> New York University >>> > > -- > Scot Dalton > Phone: (212) 998-2674 > Web Services > Division of Libraries > New York University > From rochkind at jhu.edu Wed Feb 1 11:17:44 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 01 Feb 2012 11:17:44 -0500 Subject: [Umlaut-general] Umlaut 3.0 progress update In-Reply-To: <846216BB-53C4-4CEE-B2F6-7A44D7F687B2@nyu.edu> References: <4F205DBA.1070803@jhu.edu> <87384E7A-DDD3-41D5-B036-921EC0B82BD7@nyu.edu> <4F2876AC.5000807@jhu.edu> <14DC76FB-717C-41A8-884B-A7EDEB04F2C9@nyu.edu> <4F2879C4.6040406@jhu.edu> <846216BB-53C4-4CEE-B2F6-7A44D7F687B2@nyu.edu> Message-ID: <4F2965A8.5010704@jhu.edu> I don't. Testing is.... tricky. Haven't completely figured it out. Some notes under "Developing" and "Writing a Service Plugin" on the wiki. If you can't figure out tests, feel free to bail on em, I did. :) Obviously it's better to have tests, but there are some tricks with figuring out how to do em in our environment, I have some notes to that effect on the wiki. On 2/1/2012 9:22 AM, Scot Dalton wrote: > Hi Jonathan, > > Do you have any unit tests that you've written for other services? I'll check the wiki as well. > > Thanks, > Scot > > On Jan 31, 2012, at Jan 31, 6:31 PM, Jonathan Rochkind wrote: > >> PPS: Rails 3 with bundler is more picky about gem dependencies being explicitly listed. It may just be that you were trying to use a gem (soap4r) that wasn't listed as a gem dependency. You can list it in your local app Gemfile -- or we can list it as a dependency of Umlaut itself, if it's needed for a service plugin we ship wtih Umlaut. For the latter, you'd need to check out a copy of Umlaut and change it's umlaut.gemspec (then once everything is working, commit the changes). >> >> https://github.com/team-umlaut/umlaut/wiki/Developing >> >> https://github.com/team-umlaut/umlaut/wiki/Writing-a-service-plugin >> >> >> On 1/31/2012 6:23 PM, Scot Dalton wrote: >>> I literally JUST got it up and running and started with the Primo stuff. I'll look at other service, hooks, etc. in a day or so. Trying to get my head around the new architecture right now. >>> >>> The install was easy once I got familiar with RVM and Passenger Standalone. The Umlaut install itself was very straightforward. Thanks for that! >>> >>> Re: the SOAP issue, it seems to be an incompatible gem. There are 1.9 ports of the soap4r gem, which I'll be trying out. >>> >>> Thanks, >>> Scot >>> >>> On Jan 31, 2012, at Jan 31, 6:18 PM, Jonathan Rochkind wrote: >>> >>>> On 1/31/2012 6:15 PM, Scot Dalton wrote: >>>>> Hi Jonathan, >>>>> I finally had some time to install Umlaut 3.0. I've got it running as a Passenger Standalone instance using ruby 1.9.3 on a server with several ruby 1.8.7 apps. I'm using rvm to manage the various rubies and would be happy to share my experience in getting it installed. >>>> Awesome! Did you have any trouble installing Umlaut 3.0alpha, or was it straightforward? Were my install instructions correct and complete? >>>> >>>> (Incidentally, Umlaut 3.0alpha does at the moment work perfectly fine on ruby 1.8.7 too). >>>> >>>> Other than unfinished Primo/aleph stuff, does it so far seem to be working as expected? Any notable features or hooks missing for you to get it to do what your Umlaut 2.x was doing, that you have noticed yet? >>>> >>>> Thanks! >>>> >>>> Jonathan >>>> >>> -- >>> Scot Dalton >>> Phone: (212) 998-2674 >>> Web Services >>> Division of Libraries >>> New York University >>> > > -- > Scot Dalton > Phone: (212) 998-2674 > Web Services > Division of Libraries > New York University > From std5 at nyu.edu Wed Feb 1 15:24:01 2012 From: std5 at nyu.edu (Scot Dalton) Date: Wed, 1 Feb 2012 15:24:01 -0500 Subject: [Umlaut-general] Umlaut 3.0 progress update In-Reply-To: <4F2965A8.5010704@jhu.edu> References: <4F205DBA.1070803@jhu.edu> <87384E7A-DDD3-41D5-B036-921EC0B82BD7@nyu.edu> <4F2876AC.5000807@jhu.edu> <14DC76FB-717C-41A8-884B-A7EDEB04F2C9@nyu.edu> <4F2879C4.6040406@jhu.edu> <846216BB-53C4-4CEE-B2F6-7A44D7F687B2@nyu.edu> <4F2965A8.5010704@jhu.edu> Message-ID: <901BD443-8290-48A2-8207-4C78B2E9163C@nyu.edu> I pushed some Primo updates back to the umlaut3dev branch. I have a few tests, but haven't pushed them back yet since they have private server access hard coded in them. I'll probably push them back anyway (after a bit of cleanup), since some tests are better than none. Fixtures are a bit weird in the current setup, since Rails is looking for them in the dummy app test folder. I sym linked it locally to get it working but I'm sure there is a better way to do it. I'll start looking at the hooks/customization stuff tomorrow. Re: 1.8.7 vs. 1.9.x, I'll test the soap stuff on both and get back to you. I hope the 1.9 version of the soap gem is backward compatible, but who knows. Thanks, Scot On Feb 1, 2012, at Feb 1, 11:17 AM, Jonathan Rochkind wrote: > I don't. Testing is.... tricky. Haven't completely figured it out. Some notes under "Developing" and "Writing a Service Plugin" on the wiki. > > If you can't figure out tests, feel free to bail on em, I did. :) Obviously it's better to have tests, but there are some tricks with figuring out how to do em in our environment, I have some notes to that effect on the wiki. > > On 2/1/2012 9:22 AM, Scot Dalton wrote: >> Hi Jonathan, >> >> Do you have any unit tests that you've written for other services? I'll check the wiki as well. >> >> Thanks, >> Scot >> >> On Jan 31, 2012, at Jan 31, 6:31 PM, Jonathan Rochkind wrote: >> >>> PPS: Rails 3 with bundler is more picky about gem dependencies being explicitly listed. It may just be that you were trying to use a gem (soap4r) that wasn't listed as a gem dependency. You can list it in your local app Gemfile -- or we can list it as a dependency of Umlaut itself, if it's needed for a service plugin we ship wtih Umlaut. For the latter, you'd need to check out a copy of Umlaut and change it's umlaut.gemspec (then once everything is working, commit the changes). >>> >>> https://github.com/team-umlaut/umlaut/wiki/Developing >>> >>> https://github.com/team-umlaut/umlaut/wiki/Writing-a-service-plugin >>> >>> >>> On 1/31/2012 6:23 PM, Scot Dalton wrote: >>>> I literally JUST got it up and running and started with the Primo stuff. I'll look at other service, hooks, etc. in a day or so. Trying to get my head around the new architecture right now. >>>> >>>> The install was easy once I got familiar with RVM and Passenger Standalone. The Umlaut install itself was very straightforward. Thanks for that! >>>> >>>> Re: the SOAP issue, it seems to be an incompatible gem. There are 1.9 ports of the soap4r gem, which I'll be trying out. >>>> >>>> Thanks, >>>> Scot >>>> >>>> On Jan 31, 2012, at Jan 31, 6:18 PM, Jonathan Rochkind wrote: >>>> >>>>> On 1/31/2012 6:15 PM, Scot Dalton wrote: >>>>>> Hi Jonathan, >>>>>> I finally had some time to install Umlaut 3.0. I've got it running as a Passenger Standalone instance using ruby 1.9.3 on a server with several ruby 1.8.7 apps. I'm using rvm to manage the various rubies and would be happy to share my experience in getting it installed. >>>>> Awesome! Did you have any trouble installing Umlaut 3.0alpha, or was it straightforward? Were my install instructions correct and complete? >>>>> >>>>> (Incidentally, Umlaut 3.0alpha does at the moment work perfectly fine on ruby 1.8.7 too). >>>>> >>>>> Other than unfinished Primo/aleph stuff, does it so far seem to be working as expected? Any notable features or hooks missing for you to get it to do what your Umlaut 2.x was doing, that you have noticed yet? >>>>> >>>>> Thanks! >>>>> >>>>> Jonathan >>>>> >>>> -- >>>> Scot Dalton >>>> Phone: (212) 998-2674 >>>> Web Services >>>> Division of Libraries >>>> New York University >>>> >> >> -- >> Scot Dalton >> Phone: (212) 998-2674 >> Web Services >> Division of Libraries >> New York University >> -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From std5 at nyu.edu Thu Feb 2 10:52:34 2012 From: std5 at nyu.edu (Scot Dalton) Date: Thu, 2 Feb 2012 10:52:34 -0500 Subject: [Umlaut-general] Umlaut Layouts Message-ID: Hi Jonathan, It looks like Umlaut 3.0 shares a layout between the resolve and search controllers. Is it possible to specify different layouts? Thanks, Scot -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From std5 at nyu.edu Fri Feb 3 15:35:35 2012 From: std5 at nyu.edu (Scot Dalton) Date: Fri, 3 Feb 2012 15:35:35 -0500 Subject: [Umlaut-general] aucorp Message-ID: Hi Jonathan, Do you have any issue with my adding aucorp to the author logic when calling referent.to_citation? Thanks, Scot -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From dale.poulter at Vanderbilt.Edu Tue Feb 7 11:02:34 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Tue, 7 Feb 2012 10:02:34 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F0B3CD9.7090809@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> Jonathan, Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. Response: J T 1231 proceeding info:sid/CSA:eric-set-c CSA:eric-set-c 2004 http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ED492558 info:sid/CSA:eric-set-c Reading without Teachers: Literature Circles in an EFL Classroom Jeng-yih T Hsu multi_obj_xml CSA 1 111027614344001 111000147718001 111018614652001 111026921949001 111096046941001 111076678899001 111032274555001 110974985005474 s10783D3C-51A1-11E1-9473-FAA1237FF586 Hsu, Jeng-Yih T 2004 CSA:eric-set-c latin1 J Reading without Teachers: Literature Circles in an EFL Classroom 1 yes <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno ..... From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 1:16 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: Umlaut does pass the full OpenURL to SFX. * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with &sfx.response_type=multi_obj_xml on the end and let's see what it delivers. I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this : http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 That's what Umlaut is going to use. Trying the URL without your own proxy prefix: http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). This means that unfortunately this is a bug/misdesign in SFX and it's API. SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). Jonathan On 1/9/2012 1:38 PM, Poulter, Dale wrote: I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 10:55 AM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Can you reproduce with SFX alone? That'd be the first debugging step I'd take. Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). Jonathan On 1/9/2012 10:48 AM, Poulter, Dale wrote: All, We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom --Dale --------------------------------------- Dale Poulter Automation Coordinator Library Information Technology Services Vanderbilt University 419 21st Avenue South, Room 812 Nashville, TN 37203-2427 (615)343-5388 (615)343-8834 (fax) (615)207-9705 (cell) dale.poulter at vanderbilt.edu _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general -------------- next part -------------- An HTML attachment was scrubbed... URL: From dale.poulter at Vanderbilt.Edu Tue Feb 7 12:32:53 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Tue, 7 Feb 2012 11:32:53 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu>, <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> I took a look and the urls do a kw search into ERIC based on the ED number (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558). The actual url return includes the proxy server information but it works the same. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Tuesday, February 07, 2012 11:10 AM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 11:02 AM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. Response: J T 1231 proceeding info:sid/CSA:eric-set-c CSA:eric-set-c 2004 http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ED492558 info:sid/CSA:eric-set-c Reading without Teachers: Literature Circles in an EFL Classroom Jeng-yih T Hsu multi_obj_xml CSA 1 111027614344001 111000147718001 111018614652001 111026921949001 111096046941001 111076678899001 111032274555001 110974985005474 s10783D3C-51A1-11E1-9473-FAA1237FF586 Hsu, Jeng-Yih T 2004 CSA:eric-set-c latin1 J Reading without Teachers: Literature Circles in an EFL Classroom 1 yes <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno ..... From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 1:16 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: Umlaut does pass the full OpenURL to SFX. * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with &sfx.response_type=multi_obj_xml on the end and let's see what it delivers. I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this : http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 That's what Umlaut is going to use. Trying the URL without your own proxy prefix: http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). This means that unfortunately this is a bug/misdesign in SFX and it's API. SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). Jonathan On 1/9/2012 1:38 PM, Poulter, Dale wrote: I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 10:55 AM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Can you reproduce with SFX alone? That'd be the first debugging step I'd take. Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). Jonathan On 1/9/2012 10:48 AM, Poulter, Dale wrote: All, We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom --Dale --------------------------------------- Dale Poulter Automation Coordinator Library Information Technology Services Vanderbilt University 419 21st Avenue South, Room 812 Nashville, TN 37203-2427 (615)343-5388 (615)343-8834 (fax) (615)207-9705 (cell) dale.poulter at vanderbilt.edu _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Tue Feb 7 12:10:09 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 7 Feb 2012 17:10:09 +0000 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu>, <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 11:02 AM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. Response: J T 1231 proceeding info:sid/CSA:eric-set-c CSA:eric-set-c 2004 http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ED492558 info:sid/CSA:eric-set-c Reading without Teachers: Literature Circles in an EFL Classroom Jeng-yih T Hsu multi_obj_xml CSA 1 111027614344001 111000147718001 111018614652001 111026921949001 111096046941001 111076678899001 111032274555001 110974985005474 s10783D3C-51A1-11E1-9473-FAA1237FF586 Hsu, Jeng-Yih T 2004 CSA:eric-set-c latin1 J Reading without Teachers: Literature Circles in an EFL Classroom 1 yes <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno ?.. From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 1:16 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: Umlaut does pass the full OpenURL to SFX. * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with &sfx.response_type=multi_obj_xml on the end and let's see what it delivers. I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this : http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 That's what Umlaut is going to use. Trying the URL without your own proxy prefix: http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). This means that unfortunately this is a bug/misdesign in SFX and it's API. SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). Jonathan On 1/9/2012 1:38 PM, Poulter, Dale wrote: I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 10:55 AM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Can you reproduce with SFX alone? That'd be the first debugging step I'd take. Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). Jonathan On 1/9/2012 10:48 AM, Poulter, Dale wrote: All, We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom --Dale --------------------------------------- Dale Poulter Automation Coordinator Library Information Technology Services Vanderbilt University 419 21st Avenue South, Room 812 Nashville, TN 37203-2427 (615)343-5388 (615)343-8834 (fax) (615)207-9705 (cell) dale.poulter at vanderbilt.edu _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general From dale.poulter at Vanderbilt.Edu Tue Feb 7 13:35:42 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Tue, 7 Feb 2012 12:35:42 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu>, <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E0134E349C3@its-hcwnem03.ds.Vanderbilt.edu> That is actually the issue, Umlaut does not return an Eric link (http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED493021%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20Together%3A%20Student%20Teacher%20Meet%20in%20Literature%20Circles) , http://umlaut.library.vanderbilt.edu/go/721752 . It is almost as if Umlaut strips the ED number before sending it to SFX. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Tuesday, February 07, 2012 11:10 AM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 11:02 AM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. Response: J T 1231 proceeding info:sid/CSA:eric-set-c CSA:eric-set-c 2004 http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ED492558 info:sid/CSA:eric-set-c Reading without Teachers: Literature Circles in an EFL Classroom Jeng-yih T Hsu multi_obj_xml CSA 1 111027614344001 111000147718001 111018614652001 111026921949001 111096046941001 111076678899001 111032274555001 110974985005474 s10783D3C-51A1-11E1-9473-FAA1237FF586 Hsu, Jeng-Yih T 2004 CSA:eric-set-c latin1 J Reading without Teachers: Literature Circles in an EFL Classroom 1 yes <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno ..... From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 1:16 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: Umlaut does pass the full OpenURL to SFX. * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with &sfx.response_type=multi_obj_xml on the end and let's see what it delivers. I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this : http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 That's what Umlaut is going to use. Trying the URL without your own proxy prefix: http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). This means that unfortunately this is a bug/misdesign in SFX and it's API. SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). Jonathan On 1/9/2012 1:38 PM, Poulter, Dale wrote: I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 10:55 AM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Can you reproduce with SFX alone? That'd be the first debugging step I'd take. Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). Jonathan On 1/9/2012 10:48 AM, Poulter, Dale wrote: All, We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom --Dale --------------------------------------- Dale Poulter Automation Coordinator Library Information Technology Services Vanderbilt University 419 21st Avenue South, Room 812 Nashville, TN 37203-2427 (615)343-5388 (615)343-8834 (fax) (615)207-9705 (cell) dale.poulter at vanderbilt.edu _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Tue Feb 7 13:32:08 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 7 Feb 2012 18:32:08 +0000 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu>, <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> Message-ID: <665DBC51D0250A47B4F9306CE71E5FB752F16C9A@JHEMTEBEX1.win.ad.jhu.edu> What umlaut uses is the URL supplied by Sfx in a element. I'm not sure if you were looking at that? I can investigate again next week when I'm back in the office if I don't have time before then. The Eric I'd being by itself somewhere else in the response doesn't do umlaut any good. Umlaut doesn't know how to build links itself it relies on Sfx for that. For all target links. That's sfxs job. If Sfx is returning a that doesn't match the URL the Sfx native interface sends people to, that's a an Sfx bug. I suppose we could put target-specific workarounds in umlaut to deal with Sfx bugs, but it can start to get messy and hard to maintain quickly . We dont want to put all sorts of linking logic into umlaut, because that stuff changes all the time and is different for each target-- that's why we pay for Sfx. Sfx delivers the proper outgoing link, umlaut just gets it from the Sfx API and uses it. ________________________________________ From: umlaut-general-bounces at rubyforge.org [umlaut-general-bounces at rubyforge.org] on behalf of Jonathan Rochkind [rochkind at jhu.edu] Sent: Tuesday, February 07, 2012 12:10 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 11:02 AM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. Response: J T 1231 proceeding info:sid/CSA:eric-set-c CSA:eric-set-c 2004 http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ED492558 info:sid/CSA:eric-set-c Reading without Teachers: Literature Circles in an EFL Classroom Jeng-yih T Hsu multi_obj_xml CSA 1 111027614344001 111000147718001 111018614652001 111026921949001 111096046941001 111076678899001 111032274555001 110974985005474 s10783D3C-51A1-11E1-9473-FAA1237FF586 Hsu, Jeng-Yih T 2004 CSA:eric-set-c latin1 J Reading without Teachers: Literature Circles in an EFL Classroom 1 yes <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno ?.. From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 1:16 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: Umlaut does pass the full OpenURL to SFX. * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with &sfx.response_type=multi_obj_xml on the end and let's see what it delivers. I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this : http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 That's what Umlaut is going to use. Trying the URL without your own proxy prefix: http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). This means that unfortunately this is a bug/misdesign in SFX and it's API. SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). Jonathan On 1/9/2012 1:38 PM, Poulter, Dale wrote: I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 10:55 AM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Can you reproduce with SFX alone? That'd be the first debugging step I'd take. Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). Jonathan On 1/9/2012 10:48 AM, Poulter, Dale wrote: All, We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom --Dale --------------------------------------- Dale Poulter Automation Coordinator Library Information Technology Services Vanderbilt University 419 21st Avenue South, Room 812 Nashville, TN 37203-2427 (615)343-5388 (615)343-8834 (fax) (615)207-9705 (cell) dale.poulter at vanderbilt.edu _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Tue Feb 7 16:51:04 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 7 Feb 2012 21:51:04 +0000 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E0134E349C3@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu>, <7288D49C0448E840B752320350AFA94E0134E349C3@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <665DBC51D0250A47B4F9306CE71E5FB752F16EA5@JHEMTEBEX1.win.ad.jhu.edu> I probably won't be able too investigate further until I'm back at work next week, feel free please to remind me early next week! ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 1:35 PM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 That is actually the issue, Umlaut does not return an Eric link (http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED493021%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20Together%3A%20Student%20Teacher%20Meet%20in%20Literature%20Circles) , http://umlaut.library.vanderbilt.edu/go/721752 . It is almost as if Umlaut strips the ED number before sending it to SFX. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Tuesday, February 07, 2012 11:10 AM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 11:02 AM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. Response: J T 1231 proceeding info:sid/CSA:eric-set-c CSA:eric-set-c 2004 http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ED492558 info:sid/CSA:eric-set-c Reading without Teachers: Literature Circles in an EFL Classroom Jeng-yih T Hsu multi_obj_xml CSA 1 111027614344001 111000147718001 111018614652001 111026921949001 111096046941001 111076678899001 111032274555001 110974985005474 s10783D3C-51A1-11E1-9473-FAA1237FF586 Hsu, Jeng-Yih T 2004 CSA:eric-set-c latin1 J Reading without Teachers: Literature Circles in an EFL Classroom 1 yes <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno ..... From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 1:16 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: Umlaut does pass the full OpenURL to SFX. * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with &sfx.response_type=multi_obj_xml on the end and let's see what it delivers. I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this : http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 That's what Umlaut is going to use. Trying the URL without your own proxy prefix: http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). This means that unfortunately this is a bug/misdesign in SFX and it's API. SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). Jonathan On 1/9/2012 1:38 PM, Poulter, Dale wrote: I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 10:55 AM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Can you reproduce with SFX alone? That'd be the first debugging step I'd take. Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). Jonathan On 1/9/2012 10:48 AM, Poulter, Dale wrote: All, We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom --Dale --------------------------------------- Dale Poulter Automation Coordinator Library Information Technology Services Vanderbilt University 419 21st Avenue South, Room 812 Nashville, TN 37203-2427 (615)343-5388 (615)343-8834 (fax) (615)207-9705 (cell) dale.poulter at vanderbilt.edu _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Tue Feb 14 10:08:49 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 14 Feb 2012 10:08:49 -0500 Subject: [Umlaut-general] aucorp In-Reply-To: References: Message-ID: <4F3A7901.8050600@jhu.edu> I think that sounds fine. Is aucorp actually part of the official OpenURL spec, or a non-standard thing some people use anyway? Either way, it's fine, actually! On 2/3/2012 3:35 PM, Scot Dalton wrote: > Hi Jonathan, > > Do you have any issue with my adding aucorp to the author logic when calling referent.to_citation? > > Thanks, > Scot > > -- > Scot Dalton > Phone: (212) 998-2674 > Web Services > Division of Libraries > New York University > From rochkind at jhu.edu Tue Feb 14 10:02:36 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 14 Feb 2012 10:02:36 -0500 Subject: [Umlaut-general] Umlaut Layouts In-Reply-To: References: Message-ID: <4F3A778C.8010006@jhu.edu> Sorry, I've been away on vacation. It is possible! Config variables (set in your ResolveController) 'resolve_layout' and 'search_layout'. Should have the name of the local layout you'd like to use. However, consider if you instead might want one layout with some switches for displaying different things in different contexts, for easier maintenance. You can look at the default Umlaut layout to see how it does this in some cases. But if you really want completely seperate ones, yes! On 2/2/2012 10:52 AM, Scot Dalton wrote: > Hi Jonathan, > > It looks like Umlaut 3.0 shares a layout between the resolve and search controllers. Is it possible to specify different layouts? > > > Thanks, > Scot > > -- > Scot Dalton > Phone: (212) 998-2674 > Web Services > Division of Libraries > New York University > From dale.poulter at Vanderbilt.Edu Wed Feb 15 07:59:40 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Wed, 15 Feb 2012 06:59:40 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <665DBC51D0250A47B4F9306CE71E5FB752F16EA5@JHEMTEBEX1.win.ad.jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu>, <7288D49C0448E840B752320350AFA94E0134E349C3@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16EA5@JHEMTEBEX1.win.ad.jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E013504321E@its-hcwnem03.ds.Vanderbilt.edu> Jonathan, Have you had a chance to look at this issue? It seems that umlaut does not recognize the eric number. This seems to be an issue in lib/metadata_helper.rb but I am not sure. Any suggestions will be appreciated. Thanks! -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Tuesday, February 07, 2012 3:51 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 I probably won't be able too investigate further until I'm back at work next week, feel free please to remind me early next week! ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 1:35 PM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 That is actually the issue, Umlaut does not return an Eric link (http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED493021%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20Together%3A%20Student%20Teacher%20Meet%20in%20Literature%20Circles) , http://umlaut.library.vanderbilt.edu/go/721752 . It is almost as if Umlaut strips the ED number before sending it to SFX. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Tuesday, February 07, 2012 11:10 AM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. ________________________________________ From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] Sent: Tuesday, February 07, 2012 11:02 AM To: Jonathan Rochkind Cc: umlaut-general at rubyforge.org Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. Response: J T 1231 proceeding info:sid/CSA:eric-set-c CSA:eric-set-c 2004 http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ED492558 info:sid/CSA:eric-set-c Reading without Teachers: Literature Circles in an EFL Classroom Jeng-yih T Hsu multi_obj_xml CSA 1 111027614344001 111000147718001 111018614652001 111026921949001 111096046941001 111076678899001 111032274555001 110974985005474 s10783D3C-51A1-11E1-9473-FAA1237FF586 Hsu, Jeng-Yih T 2004 CSA:eric-set-c latin1 J Reading without Teachers: Literature Circles in an EFL Classroom 1 yes <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno ..... From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 1:16 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: Umlaut does pass the full OpenURL to SFX. * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with &sfx.response_type=multi_obj_xml on the end and let's see what it delivers. I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this : http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 That's what Umlaut is going to use. Trying the URL without your own proxy prefix: http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). This means that unfortunately this is a bug/misdesign in SFX and it's API. SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). Jonathan On 1/9/2012 1:38 PM, Poulter, Dale wrote: I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, January 09, 2012 10:55 AM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Can you reproduce with SFX alone? That'd be the first debugging step I'd take. Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). Jonathan On 1/9/2012 10:48 AM, Poulter, Dale wrote: All, We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom --Dale --------------------------------------- Dale Poulter Automation Coordinator Library Information Technology Services Vanderbilt University 419 21st Avenue South, Room 812 Nashville, TN 37203-2427 (615)343-5388 (615)343-8834 (fax) (615)207-9705 (cell) dale.poulter at vanderbilt.edu _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general From std5 at nyu.edu Wed Feb 15 09:47:00 2012 From: std5 at nyu.edu (Scot Dalton) Date: Wed, 15 Feb 2012 09:47:00 -0500 Subject: [Umlaut-general] aucorp In-Reply-To: <4F3A7901.8050600@jhu.edu> References: <4F3A7901.8050600@jhu.edu> Message-ID: Great. I pushed my code changes back to umlaut3dev. Not sure if it's part of the official spec, but it seems to be in all OpenURLs coming from our Primo implementation. Thanks, Scot On Feb 14, 2012, at Feb 14, 10:08 AM, Jonathan Rochkind wrote: > I think that sounds fine. Is aucorp actually part of the official OpenURL spec, or a non-standard thing some people use anyway? Either way, it's fine, actually! > > On 2/3/2012 3:35 PM, Scot Dalton wrote: >> Hi Jonathan, >> >> Do you have any issue with my adding aucorp to the author logic when calling referent.to_citation? >> >> Thanks, >> Scot >> >> -- >> Scot Dalton >> Phone: (212) 998-2674 >> Web Services >> Division of Libraries >> New York University >> -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From m.e.phillips at durham.ac.uk Wed Feb 15 10:50:41 2012 From: m.e.phillips at durham.ac.uk (PHILLIPS M.E.) Date: Wed, 15 Feb 2012 15:50:41 -0000 Subject: [Umlaut-general] aucorp In-Reply-To: References: <4F3A7901.8050600@jhu.edu> Message-ID: <1F5DB00D61AF1A479A6F8572FAC9ED8003670990@DURMAIL4.mds.ad.dur.ac.uk> See http://alcme.oclc.org/openurl/servlet/OAIHandler/extension?verb=GetMetad ata&metadataPrefix=mtx&identifier=info:ofi/fmt:kev:mtx:book It's definitely in the OpenURL spec for "book" and "journal" but not for "dissertation". Matthew > -----Original Message----- > From: umlaut-general-bounces at rubyforge.org [mailto:umlaut-general- > bounces at rubyforge.org] On Behalf Of Scot Dalton > Sent: 15 February 2012 14:47 > To: Jonathan Rochkind > Cc: umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] aucorp > > Great. I pushed my code changes back to umlaut3dev. Not sure if it's part of > the official spec, but it seems to be in all OpenURLs coming from our Primo > implementation. > > Thanks, > Scot From rossfsinger at gmail.com Wed Feb 15 11:16:16 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 15 Feb 2012 11:16:16 -0500 Subject: [Umlaut-general] aucorp In-Reply-To: References: <4F3A7901.8050600@jhu.edu> Message-ID: aucorp is legal for both books and journals. http://alcme.oclc.org/openurl/servlet/OAIHandler/extension?verb=GetMetadata&metadataPrefix=mtx&identifier=info:ofi/fmt:kev:mtx:book http://alcme.oclc.org/openurl/servlet/OAIHandler/extension?verb=GetMetadata&metadataPrefix=mtx&identifier=info:ofi/fmt:kev:mtx:journal -Ross. On Wed, Feb 15, 2012 at 9:47 AM, Scot Dalton wrote: > Great. ?I pushed my code changes back to umlaut3dev. ?Not sure if it's part of the official spec, but it seems to be in all OpenURLs coming from our Primo implementation. > > Thanks, > Scot > > On Feb 14, 2012, at Feb 14, 10:08 AM, Jonathan Rochkind wrote: > >> I think that sounds fine. Is aucorp actually part of the official OpenURL spec, or a non-standard thing some people use anyway? Either way, it's fine, actually! >> >> On 2/3/2012 3:35 PM, Scot Dalton wrote: >>> Hi Jonathan, >>> >>> Do you have any issue with my adding aucorp to the author logic when calling referent.to_citation? >>> >>> Thanks, >>> Scot >>> >>> -- >>> Scot Dalton >>> Phone: ?(212) 998-2674 >>> Web Services >>> Division of Libraries >>> New York University >>> > > > -- > Scot Dalton > Phone: ?(212) 998-2674 > Web Services > Division of Libraries > New York University > > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Wed Feb 15 10:52:20 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 15 Feb 2012 10:52:20 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E013504321E@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E349C3@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16EA5@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E013504321E@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3BD4B4.2070708@jhu.edu> Will try to look into this again today! On 2/15/2012 7:59 AM, Poulter, Dale wrote: > Jonathan, > > Have you had a chance to look at this issue? It seems that umlaut does not recognize the eric number. This seems to be an issue in lib/metadata_helper.rb but I am not sure. Any suggestions will be appreciated. Thanks! > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Tuesday, February 07, 2012 3:51 PM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > I probably won't be able too investigate further until I'm back at work next week, feel free please to remind me early next week! > ________________________________________ > From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] > Sent: Tuesday, February 07, 2012 1:35 PM > To: Jonathan Rochkind > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > That is actually the issue, Umlaut does not return an Eric link (http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED493021%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20Together%3A%20Student%20Teacher%20Meet%20in%20Literature%20Circles) , http://umlaut.library.vanderbilt.edu/go/721752 . It is almost as if Umlaut strips the ED number before sending it to SFX. > > > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Tuesday, February 07, 2012 11:10 AM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. > ________________________________________ > From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] > Sent: Tuesday, February 07, 2012 11:02 AM > To: Jonathan Rochkind > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Jonathan, > > Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. > > > Response: > > > > > > J T > > > 1231 > proceeding > > > > > info:sid/CSA:eric-set-c > > > > > CSA:eric-set-c > 2004 > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml > ED492558 > > > info:sid/CSA:eric-set-c > > > Reading without Teachers: Literature Circles in an EFL Classroom > > > > > > > Jeng-yih > > > > > T > > > > > > > > > Hsu > > > multi_obj_xml > CSA > 1 > > > 111027614344001 > 111000147718001 > 111018614652001 > 111026921949001 > 111096046941001 > 111076678899001 > 111032274555001 > 110974985005474 > > > s10783D3C-51A1-11E1-9473-FAA1237FF586 > > > Hsu, Jeng-Yih T > > > 2004 > CSA:eric-set-c > latin1 > > > J > > > Reading without Teachers: Literature Circles in an EFL Classroom > 1 > yes > <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno > > ..... > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 1:16 PM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. > > On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: > Umlaut does pass the full OpenURL to SFX. > > * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. > > * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. > > > > So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with&sfx.response_type=multi_obj_xml on the end and let's see what it delivers. > > I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this: > > http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 > > That's what Umlaut is going to use. Trying the URL without your own proxy prefix: > > http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 > > I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). > > This means that unfortunately this is a bug/misdesign in SFX and it's API. > > SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. > > We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). > > Jonathan > > > On 1/9/2012 1:38 PM, Poulter, Dale wrote: > > I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? > > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 10:55 AM > To: umlaut-general at rubyforge.org > Cc: Poulter, Dale > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > > > Can you reproduce with SFX alone? That'd be the first debugging step I'd take. > > Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? > > If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. > > If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). > > Jonathan > > On 1/9/2012 10:48 AM, Poulter, Dale wrote: > > All, > > > > We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. > > > > > > > > > > http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom > > > > > > > > > > > > > > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom > > > > --Dale > > > > --------------------------------------- > Dale Poulter > Automation Coordinator > > Library Information Technology Services > Vanderbilt University > > 419 21st Avenue South, Room 812 > Nashville, TN 37203-2427 > (615)343-5388 > (615)343-8834 (fax) > (615)207-9705 (cell) > dale.poulter at vanderbilt.edu > > > > > > > > _______________________________________________ > > Umlaut-general mailing list > > Umlaut-general at rubyforge.org > > http://rubyforge.org/mailman/listinfo/umlaut-general > > > > > From rochkind at jhu.edu Wed Feb 15 10:55:00 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 15 Feb 2012 10:55:00 -0500 Subject: [Umlaut-general] aucorp In-Reply-To: References: <4F3A7901.8050600@jhu.edu> Message-ID: <4F3BD554.8080904@jhu.edu> Awesome. Would you mind committing a comment next to the use of aucorp in the code mentioning that it comes from OpenURLs from Primo, whether or not it's part of the official spec, so a future reader knows what's up? On 2/15/2012 9:47 AM, Scot Dalton wrote: > Great. I pushed my code changes back to umlaut3dev. Not sure if it's part of the official spec, but it seems to be in all OpenURLs coming from our Primo implementation. > > Thanks, > Scot > > On Feb 14, 2012, at Feb 14, 10:08 AM, Jonathan Rochkind wrote: > >> I think that sounds fine. Is aucorp actually part of the official OpenURL spec, or a non-standard thing some people use anyway? Either way, it's fine, actually! >> >> On 2/3/2012 3:35 PM, Scot Dalton wrote: >>> Hi Jonathan, >>> >>> Do you have any issue with my adding aucorp to the author logic when calling referent.to_citation? >>> >>> Thanks, >>> Scot >>> >>> -- >>> Scot Dalton >>> Phone: (212) 998-2674 >>> Web Services >>> Division of Libraries >>> New York University >>> > > -- > Scot Dalton > Phone: (212) 998-2674 > Web Services > Division of Libraries > New York University > From rochkind at jhu.edu Wed Feb 15 12:01:55 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 15 Feb 2012 12:01:55 -0500 Subject: [Umlaut-general] aucorp In-Reply-To: <4F3BD554.8080904@jhu.edu> References: <4F3A7901.8050600@jhu.edu> <4F3BD554.8080904@jhu.edu> Message-ID: <4F3BE503.6000909@jhu.edu> Nevermind, now that rsinger confirms it's part of the spec! On 2/15/2012 10:55 AM, Jonathan Rochkind wrote: > Awesome. Would you mind committing a comment next to the use of aucorp > in the code mentioning that it comes from OpenURLs from Primo, whether > or not it's part of the official spec, so a future reader knows what's > up? > > On 2/15/2012 9:47 AM, Scot Dalton wrote: >> Great. I pushed my code changes back to umlaut3dev. Not sure if >> it's part of the official spec, but it seems to be in all OpenURLs >> coming from our Primo implementation. >> >> Thanks, >> Scot >> >> On Feb 14, 2012, at Feb 14, 10:08 AM, Jonathan Rochkind wrote: >> >>> I think that sounds fine. Is aucorp actually part of the official >>> OpenURL spec, or a non-standard thing some people use anyway? Either >>> way, it's fine, actually! >>> >>> On 2/3/2012 3:35 PM, Scot Dalton wrote: >>>> Hi Jonathan, >>>> >>>> Do you have any issue with my adding aucorp to the author logic >>>> when calling referent.to_citation? >>>> >>>> Thanks, >>>> Scot >>>> >>>> -- >>>> Scot Dalton >>>> Phone: (212) 998-2674 >>>> Web Services >>>> Division of Libraries >>>> New York University >>>> >> >> -- >> Scot Dalton >> Phone: (212) 998-2674 >> Web Services >> Division of Libraries >> New York University >> > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Wed Feb 15 15:51:09 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 15 Feb 2012 15:51:09 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3C1ABD.3020606@jhu.edu> Hmm, my SFX does not have an ERIC Full Text target enabled such that it appears for that OpenURL at all (unlike your SFX). Can you tell me how to configure my ERIC Full Text target in my SFX test instance to duplicate your SFX setup, so I can duplicate here? But continuing to debug as best I can without that too, that'd just make it easier. On 2/7/2012 12:32 PM, Poulter, Dale wrote: > I took a look and the urls do a kw search into ERIC based on the ED number (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558). The actual url return includes the proxy server information but it works the same. > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Tuesday, February 07, 2012 11:10 AM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. > ________________________________________ > From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] > Sent: Tuesday, February 07, 2012 11:02 AM > To: Jonathan Rochkind > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Jonathan, > > Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. > > > Response: > > > > > > J T > > > 1231 > proceeding > > > > > info:sid/CSA:eric-set-c > > > > > CSA:eric-set-c > 2004 > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml > ED492558 > > > info:sid/CSA:eric-set-c > > > Reading without Teachers: Literature Circles in an EFL Classroom > > > > > > > Jeng-yih > > > > > T > > > > > > > > > Hsu > > > multi_obj_xml > CSA > 1 > > > 111027614344001 > 111000147718001 > 111018614652001 > 111026921949001 > 111096046941001 > 111076678899001 > 111032274555001 > 110974985005474 > > > s10783D3C-51A1-11E1-9473-FAA1237FF586 > > > Hsu, Jeng-Yih T > > > 2004 > CSA:eric-set-c > latin1 > > > J > > > Reading without Teachers: Literature Circles in an EFL Classroom > 1 > yes > <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno > > ..... > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 1:16 PM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. > > On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: > Umlaut does pass the full OpenURL to SFX. > > * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. > > * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. > > > > So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with&sfx.response_type=multi_obj_xml on the end and let's see what it delivers. > > I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this: > > http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 > > That's what Umlaut is going to use. Trying the URL without your own proxy prefix: > > http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 > > I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). > > This means that unfortunately this is a bug/misdesign in SFX and it's API. > > SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. > > We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). > > Jonathan > > > On 1/9/2012 1:38 PM, Poulter, Dale wrote: > > I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? > > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 10:55 AM > To: umlaut-general at rubyforge.org > Cc: Poulter, Dale > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > > > Can you reproduce with SFX alone? That'd be the first debugging step I'd take. > > Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? > > If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. > > If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). > > Jonathan > > On 1/9/2012 10:48 AM, Poulter, Dale wrote: > > All, > > > > We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. > > > > > > > > > > http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom > > > > > > > > > > > > > > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom > > > > --Dale > > > > --------------------------------------- > Dale Poulter > Automation Coordinator > > Library Information Technology Services > Vanderbilt University > > 419 21st Avenue South, Room 812 > Nashville, TN 37203-2427 > (615)343-5388 > (615)343-8834 (fax) > (615)207-9705 (cell) > dale.poulter at vanderbilt.edu > > > > > > > > _______________________________________________ > > Umlaut-general mailing list > > Umlaut-general at rubyforge.org > > http://rubyforge.org/mailman/listinfo/umlaut-general > > > From rochkind at jhu.edu Wed Feb 15 15:52:09 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 15 Feb 2012 15:52:09 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3C1AF9.4010408@jhu.edu> Or, actually, I realize I can point my test Umlaut right at your SFX! I'll try that for debugging purposes. So nevermind on the previous question. I'll use your test instance at: http://sfx.library.vanderbilt.edu/sfxtst41 as in your example. On 2/7/2012 12:32 PM, Poulter, Dale wrote: > I took a look and the urls do a kw search into ERIC based on the ED number (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558). The actual url return includes the proxy server information but it works the same. > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Tuesday, February 07, 2012 11:10 AM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. > ________________________________________ > From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] > Sent: Tuesday, February 07, 2012 11:02 AM > To: Jonathan Rochkind > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Jonathan, > > Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. > > > Response: > > > > > > J T > > > 1231 > proceeding > > > > > info:sid/CSA:eric-set-c > > > > > CSA:eric-set-c > 2004 > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml > ED492558 > > > info:sid/CSA:eric-set-c > > > Reading without Teachers: Literature Circles in an EFL Classroom > > > > > > > Jeng-yih > > > > > T > > > > > > > > > Hsu > > > multi_obj_xml > CSA > 1 > > > 111027614344001 > 111000147718001 > 111018614652001 > 111026921949001 > 111096046941001 > 111076678899001 > 111032274555001 > 110974985005474 > > > s10783D3C-51A1-11E1-9473-FAA1237FF586 > > > Hsu, Jeng-Yih T > > > 2004 > CSA:eric-set-c > latin1 > > > J > > > Reading without Teachers: Literature Circles in an EFL Classroom > 1 > yes > <AN>ED492558</AN>&<PY>2004</PY>&<AU>Hsu, Jeng-yih Tim</AU> ERIC_FULL_TEXTERIC Full-text111096046941000111096046941001getFullTxtERIC::ERIC_FTurl=http://www.eric.ed.govyesno > > ..... > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 1:16 PM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. > > On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: > Umlaut does pass the full OpenURL to SFX. > > * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. > > * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. > > > > So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with&sfx.response_type=multi_obj_xml on the end and let's see what it delivers. > > I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this: > > http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 > > That's what Umlaut is going to use. Trying the URL without your own proxy prefix: > > http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 > > I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). > > This means that unfortunately this is a bug/misdesign in SFX and it's API. > > SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. > > We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). > > Jonathan > > > On 1/9/2012 1:38 PM, Poulter, Dale wrote: > > I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? > > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 10:55 AM > To: umlaut-general at rubyforge.org > Cc: Poulter, Dale > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > > > Can you reproduce with SFX alone? That'd be the first debugging step I'd take. > > Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? > > If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. > > If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). > > Jonathan > > On 1/9/2012 10:48 AM, Poulter, Dale wrote: > > All, > > > > We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. > > > > > > > > > > http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom > > > > > > > > > > > > > > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom > > > > --Dale > > > > --------------------------------------- > Dale Poulter > Automation Coordinator > > Library Information Technology Services > Vanderbilt University > > 419 21st Avenue South, Room 812 > Nashville, TN 37203-2427 > (615)343-5388 > (615)343-8834 (fax) > (615)207-9705 (cell) > dale.poulter at vanderbilt.edu > > > > > > > > _______________________________________________ > > Umlaut-general mailing list > > Umlaut-general at rubyforge.org > > http://rubyforge.org/mailman/listinfo/umlaut-general > > > From dale.poulter at Vanderbilt.Edu Wed Feb 15 15:54:24 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Wed, 15 Feb 2012 14:54:24 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F3C1AF9.4010408@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> ok, thanks. Please let me know if you need additional information. One of the issues is that SFX does not have a portfolio for ERIC just the service that creates the search url into eric.gov using the ed_number. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Wednesday, February 15, 2012 2:52 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Or, actually, I realize I can point my test Umlaut right at your SFX! I'll try that for debugging purposes. So nevermind on the previous question. I'll use your test instance at: http://sfx.library.vanderbilt.edu/sfxtst41 as in your example. On 2/7/2012 12:32 PM, Poulter, Dale wrote: > I took a look and the urls do a kw search into ERIC based on the ED number (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558). The actual url return includes the proxy server information but it works the same. > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Tuesday, February 07, 2012 11:10 AM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. > ________________________________________ > From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] > Sent: Tuesday, February 07, 2012 11:02 AM > To: Jonathan Rochkind > Cc: umlaut-general at rubyforge.org > Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 > > Jonathan, > > Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. > > > Response: > > > > > > J T > > > 1231 > proceeding > > > > > info:sid/CSA:eric-set-c > > > > > CSA:eric-set-c > 2004 > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml > ED492558 > > > info:sid/CSA:eric-set-c > > > Reading without Teachers: Literature Circles in an EFL Classroom > > > > > > > Jeng-yih > > > > > T > > > > > > > > > Hsu > > > multi_obj_xml > CSA > 1 > > > 111027614344001 > 111000147718001 > 111018614652001 > 111026921949001 > 111096046941001 > 111076678899001 > 111032274555001 > 110974985005474 > > > s10783D3C-51A1-11E1-9473-FAA1237FF586 > > > Hsu, Jeng-Yih T > > > 2004 > CSA:eric-set-c > latin1 > > > J > > > Reading without Teachers: Literature Circles in an EFL Classroom > 1 > yes > key="rft_dat"><AN>ED492558</AN>&<PY>2004</PY& > gt;&<AU>Hsu, Jeng-yih Tim</AU> > > ERIC_FULL_T > EXTERIC > Full-text1110960 > 46941000111096046941001 _id>getFullTxtERIC::ERIC_FT er>url=http://www.eric.ed.govyes oxy>no > > > ..... > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 1:16 PM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. > > On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: > Umlaut does pass the full OpenURL to SFX. > > * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. > > * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. > > > > So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with&sfx.response_type=multi_obj_xml on the end and let's see what it delivers. > > I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this: > > http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/E > RICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw& > ;ERICExtSearch_SearchValue_0=ED492558 > > That's what Umlaut is going to use. Trying the URL without your own proxy prefix: > > http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSe > arch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 > > I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). > > This means that unfortunately this is a bug/misdesign in SFX and it's API. > > SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. > > We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). > > Jonathan > > > On 1/9/2012 1:38 PM, Poulter, Dale wrote: > > I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? > > > > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, January 09, 2012 10:55 AM > To: umlaut-general at rubyforge.org > Cc: Poulter, Dale > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > > > Can you reproduce with SFX alone? That'd be the first debugging step I'd take. > > Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? > > If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. > > If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). > > Jonathan > > On 1/9/2012 10:48 AM, Poulter, Dale wrote: > > All, > > > > We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. > > > > > > > > > > http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3 > CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%2 > 0Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufir > st=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literatu > re%20Circles%20in%20an%20EFL%20Classroom > > > > > > > > > > > > > > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CA > N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20J > eng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst > =Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature > %20Circles%20in%20an%20EFL%20Classroom > > > > --Dale > > > > --------------------------------------- > Dale Poulter > Automation Coordinator > > Library Information Technology Services Vanderbilt University > > 419 21st Avenue South, Room 812 > Nashville, TN 37203-2427 > (615)343-5388 > (615)343-8834 (fax) > (615)207-9705 (cell) > dale.poulter at vanderbilt.edu > > > > > > > > _______________________________________________ > > Umlaut-general mailing list > > Umlaut-general at rubyforge.org > > http://rubyforge.org/mailman/listinfo/umlaut-general > > > From rochkind at jhu.edu Wed Feb 15 16:05:38 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 15 Feb 2012 16:05:38 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3C1E22.7060200@jhu.edu> Ugh, this has something to do with OpenURL 1.0 vs 0.1. The original OpenURL is an "0.1" one. But Umlaut converts it to "1.0" before sending it to SFX. Part of what it does (perhaps erroneously?) is send the original "&pid" as "&rft.pid". rsinger, you're the OpenURL expert, is this an error? In OpenURL 1.0, should "pid" still be a bare "pid" instead of "rft.pid"? Or any light to shed? (Still have trouble figuring out where the heck I'd look in what documentation to verify this). Dale, do you have any idea if the ERIC SFX target is supposed to be supported for both OpenURL 1.0 and 0.1, or if there's anything special you have to do to turn it on for 1.0? I kind of hate OpenURL. Hmm, that might not be the problem though, might be the problem is Umlaut is not sending the "sid" from the original OpenURL onto SFX, which SFX requires to properly resolve this one. I'll continue investigating that one too. Still trying to figure it out. Jonathan On 2/15/2012 3:54 PM, Poulter, Dale wrote: > ok, thanks. Please let me know if you need additional information. One of the issues is that SFX does not have a portfolio for ERIC just the service that creates the search url into eric.gov using the ed_number. > > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Wednesday, February 15, 2012 2:52 PM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > Or, actually, I realize I can point my test Umlaut right at your SFX! > I'll try that for debugging purposes. So nevermind on the previous question. > > I'll use your test instance at: > > http://sfx.library.vanderbilt.edu/sfxtst41 > > as in your example. > > > > On 2/7/2012 12:32 PM, Poulter, Dale wrote: >> I took a look and the urls do a kw search into ERIC based on the ED number (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558). The actual url return includes the proxy server information but it works the same. >> >> -----Original Message----- >> From: Jonathan Rochkind [mailto:rochkind at jhu.edu] >> Sent: Tuesday, February 07, 2012 11:10 AM >> To: Poulter, Dale >> Cc: umlaut-general at rubyforge.org >> Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 >> >> Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. >> ________________________________________ >> From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] >> Sent: Tuesday, February 07, 2012 11:02 AM >> To: Jonathan Rochkind >> Cc: umlaut-general at rubyforge.org >> Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 >> >> Jonathan, >> >> Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. >> >> >> Response: >> >> >> >> >> >> J T >> >> >> 1231 >> proceeding >> >> >> >> >> info:sid/CSA:eric-set-c >> >> >> >> >> CSA:eric-set-c >> 2004 >> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml >> ED492558 >> >> >> info:sid/CSA:eric-set-c >> >> >> Reading without Teachers: Literature Circles in an EFL Classroom >> >> >> >> >> >> >> Jeng-yih >> >> >> >> >> T >> >> >> >> >> >> >> >> >> Hsu >> >> >> multi_obj_xml >> CSA >> 1 >> >> >> 111027614344001 >> 111000147718001 >> 111018614652001 >> 111026921949001 >> 111096046941001 >> 111076678899001 >> 111032274555001 >> 110974985005474 >> >> >> s10783D3C-51A1-11E1-9473-FAA1237FF586 >> >> >> Hsu, Jeng-Yih T >> >> >> 2004 >> CSA:eric-set-c >> latin1 >> >> >> J >> >> >> Reading without Teachers: Literature Circles in an EFL Classroom >> 1 >> yes >> > key="rft_dat"><AN>ED492558</AN>&<PY>2004</PY& >> gt;&<AU>Hsu, Jeng-yih Tim</AU> >> >> ERIC_FULL_T >> EXTERIC >> Full-text1110960 >> 46941000111096046941001> _id>getFullTxtERIC::ERIC_FT> er>url=http://www.eric.ed.govyes> oxy>no >> >> >> ..... >> >> >> From: Jonathan Rochkind [mailto:rochkind at jhu.edu] >> Sent: Monday, January 09, 2012 1:16 PM >> To: Poulter, Dale >> Cc: umlaut-general at rubyforge.org >> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >> >> PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. >> >> On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: >> Umlaut does pass the full OpenURL to SFX. >> >> * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. >> >> * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. >> >> >> >> So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with&sfx.response_type=multi_obj_xml on the end and let's see what it delivers. >> >> I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this: >> >> http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/E >> RICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw& >> ;ERICExtSearch_SearchValue_0=ED492558 >> >> That's what Umlaut is going to use. Trying the URL without your own proxy prefix: >> >> http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSe >> arch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 >> >> I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). >> This means that unfortunately this is a bug/misdesign in SFX and it's API. >> >> SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. >> >> We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). >> >> Jonathan >> >> >> On 1/9/2012 1:38 PM, Poulter, Dale wrote: >> >> I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? >> >> >> >> From: Jonathan Rochkind [mailto:rochkind at jhu.edu] >> Sent: Monday, January 09, 2012 10:55 AM >> To: umlaut-general at rubyforge.org >> Cc: Poulter, Dale >> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >> >> >> >> Can you reproduce with SFX alone? That'd be the first debugging step I'd take. >> >> Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? >> >> If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. >> >> If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). >> >> Jonathan >> >> On 1/9/2012 10:48 AM, Poulter, Dale wrote: >> >> All, >> >> >> >> We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. >> >> >> >> >> >> >> >> >> >> http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3 >> CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%2 >> 0Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufir >> st=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literatu >> re%20Circles%20in%20an%20EFL%20Classroom >> >> >> >> >> >> >> >> >> >> >> >> >> >> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CA >> N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20J >> eng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst >> =Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature >> %20Circles%20in%20an%20EFL%20Classroom >> >> >> >> --Dale >> >> >> >> --------------------------------------- >> Dale Poulter >> Automation Coordinator >> >> Library Information Technology Services Vanderbilt University >> >> 419 21st Avenue South, Room 812 >> Nashville, TN 37203-2427 >> (615)343-5388 >> (615)343-8834 (fax) >> (615)207-9705 (cell) >> dale.poulter at vanderbilt.edu >> >> >> >> >> >> >> >> _______________________________________________ >> >> Umlaut-general mailing list >> >> Umlaut-general at rubyforge.org >> >> http://rubyforge.org/mailman/listinfo/umlaut-general >> >> >> > From rochkind at jhu.edu Wed Feb 15 17:01:34 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 15 Feb 2012 17:01:34 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3C2B3E.4070906@jhu.edu> Okay, yeah, this is kind of a mess. I hate OpenURL. Bare with me, Dale, to understand the problem and possible solutions (none entirely trivial, but some possible). The original OpenURL is an "OpenURL 0.1". The presence of two parameters is enough for SFX to trigger the ERIC link. "sid" (source id, where the OpenURL came from) and "pid" (private id, private non-standard data from that source). So for instance, this is the minimal 0.1 OpenURL which succesfully triggers the ERIC link from your link resolver: http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 But Umlaut ends up sending an OpenURL 1.0 request to SFX. Umlaut actually takes all incoming OpenURLs and translates them into an internal data structure, doesn't even keep track of whether they came in as 1.0 or 0.1, and then translates them back into OpenURL 1.0 to make the API request to SFX. So the first problem is, Umlaut is doing some weird things with the "sid" and "pid" parameters in this round trip to internal data structure and back. (Those params _usually_ don't matter, is why I haven't noticed before). But even when I investigate fixing that, I can't figure out if there _is_ an OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. Here's the closest OpenURL 1.0 equivalent I can come up with (thanks rsinger) http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 But that does NOT properly trigger the ERIC link in SFX. No completely easy fix presents itself. But we have a number of options. So, Dale, at this point, if you could, I think it would be helpful for you to try to get some info from Ex Libris. (A frustrating and time consuming thing to do, which is why I ask you to do it for your problem, heh). * Option 1. I think I remember that SFX gets that "pid" info out with custom source parsers based on "sid". And that these "source parsers" are specific to OpenURL 1.0 or 0.1 So you could ask Ex Libris if there's any way to set up a CSA source parser that will work for OpenURL 1.0 and do the same thing it's doing now for the 0.1 source parser. Adn what the 1.0 OpenURL should look like to be recognized by that source parser. They will say "Why would you want an OpenURL 1.0 source parser for CSA, CSA only sends OpenURL 0.1?" And it'll be confusing. And this would still require some changes to Umlaut, as I've noticed through this debugging that it doesn't handle sid/pid quite right anyway. Fixing 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. * Option 2. Alternately, we could try to hack Umlaut so it can send an OpenURL 0.1 to SFX instead. As a configuration option, always? That might mess other things up. Only if the original OpenURL was 0.1? Umlaut doesn't keep track of this right now, but theoretically it could. Only if the OpenURL has a "pid" in it? That might actually be best, since "pid" as a literal string is OpenURL 0.1-specific so is a signal that the incoming was OpenURL 0.1 -- plus it's really only incoming OpenURLs with 'pid' that need to be sent as OpenURL 0.1 to SFX. This would still require the same fixes to Umlaut's handling of sid/pid mentioned above. * Option 3. I might lean toward this one, actually. The reason we use SFX is for it's knowledge base. That is, for portfolio-based targets, as well as SFX's knowledge of outgoing link syntax. You helpfully (thanks!) mentioned that the ERIC target is not portfolio based. It's just an automatic translation "if ERIC id detected, make a link to ERIC like so." So we could write a new service plugin for Umlaut that handles ERIC id's and outgoing links to ERIC. Activate that in Umlaut, de-activate the ERIC target in SFX. This would require "us" (Umlaut) to take responsibility for maintaining the outgoing ERIC link syntax and changing our code if it changes. But it's not like Ex Libris is very good at that maintenance anyway, and I bet ERIC outgoing link syntax doesn't change much. It would also require "us" to identify all the places an ERIC id might come in in source-specific ways. This example is one of them, a "pid" formatted in a certain way, when "sid" is CSA. Are there other custom places SFX recognizes ERIC id's, from other places? We'd have to figure that out, and copy that code in Umlaut too (taking maintenance responsibility too). But we wouldn't have to deal with SFX's weirdness here, we'd take responsibility for handling this ourselves, which is part of the point of Umlaut, really --- use SFX for it's portfolio-based knowledge base, but do ourselves what we can do ourselves better. (Recognizing in this case it's not really 'better', it's just that SFX's weirdness combined with OpenURL's crappiness puts us in a corner). And someone would have to write the code in Umlaut. Which I _might_ find time to do, or else, Dale, how do you feel about some ruby coding? Or else we could try to find someone else. CONCLUSION So, any thoughts Dale? From rossfsinger at gmail.com Wed Feb 15 18:24:03 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 15 Feb 2012 18:24:03 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F3C1E22.7060200@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1E22.7060200@jhu.edu> Message-ID: <24B8A4BC-1371-4AA3-AD8E-964DF98604DA@gmail.com> I told Jonathan this in IRC, but I'll repeat it here for archival purposes. OpenURL 0.1 sid becomes rfr_id and pid becomes rft_dat. I think if we properly migrate these two key/vals (the rfr_id needs to be an info-uri, for example), this problem will clear up. -Ross. On Feb 15, 2012, at 4:05 PM, Jonathan Rochkind wrote: > Ugh, this has something to do with OpenURL 1.0 vs 0.1. > > The original OpenURL is an "0.1" one. > > But Umlaut converts it to "1.0" before sending it to SFX. > > Part of what it does (perhaps erroneously?) is send the original "&pid" as "&rft.pid". > > rsinger, you're the OpenURL expert, is this an error? In OpenURL 1.0, should "pid" still be a bare "pid" instead of "rft.pid"? Or any light to shed? (Still have trouble figuring out where the heck I'd look in what documentation to verify this). > > Dale, do you have any idea if the ERIC SFX target is supposed to be supported for both OpenURL 1.0 and 0.1, or if there's anything special you have to do to turn it on for 1.0? > > I kind of hate OpenURL. > > Hmm, that might not be the problem though, might be the problem is Umlaut is not sending the "sid" from the original OpenURL onto SFX, which SFX requires to properly resolve this one. I'll continue investigating that one too. > > Still trying to figure it out. > > Jonathan > > > On 2/15/2012 3:54 PM, Poulter, Dale wrote: >> ok, thanks. Please let me know if you need additional information. One of the issues is that SFX does not have a portfolio for ERIC just the service that creates the search url into eric.gov using the ed_number. >> >> >> -----Original Message----- >> From: Jonathan Rochkind [mailto:rochkind at jhu.edu] >> Sent: Wednesday, February 15, 2012 2:52 PM >> To: Poulter, Dale >> Cc: umlaut-general at rubyforge.org >> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >> >> Or, actually, I realize I can point my test Umlaut right at your SFX! >> I'll try that for debugging purposes. So nevermind on the previous question. >> >> I'll use your test instance at: >> >> http://sfx.library.vanderbilt.edu/sfxtst41 >> >> as in your example. >> >> >> >> On 2/7/2012 12:32 PM, Poulter, Dale wrote: >>> I took a look and the urls do a kw search into ERIC based on the ED number (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558). The actual url return includes the proxy server information but it works the same. >>> >>> -----Original Message----- >>> From: Jonathan Rochkind [mailto:rochkind at jhu.edu] >>> Sent: Tuesday, February 07, 2012 11:10 AM >>> To: Poulter, Dale >>> Cc: umlaut-general at rubyforge.org >>> Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 >>> >>> Have you followed the Eric link Sfx returns to see where it takes you? When I tried before, it took me to the wrong place -- the same wrong place umlaut sends you to because that's the URL umlaut is using. Sfx is returning a different ( wrong ) URL in API response compared to what Sfx native interface delivers, that is the bug. Or was when I investigated. >>> ________________________________________ >>> From: Poulter, Dale [dale.poulter at Vanderbilt.Edu] >>> Sent: Tuesday, February 07, 2012 11:02 AM >>> To: Jonathan Rochkind >>> Cc: umlaut-general at rubyforge.org >>> Subject: RE: [Umlaut-general] Strange link issue in Umlaut v2 >>> >>> Jonathan, >>> >>> Thanks again for your help. I am looking back over the email and attempt to lookup the api (http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml ). Based on the response, the system does seem to be returning the ERIC link. Am I missing something? I am happy to followup with Ex Libris but want to fully understand the problem before reporting. Thanks. >>> >>> >>> Response: >>> >>> >>> >>> >>> >>> J T >>> >>> >>> 1231 >>> proceeding >>> >>> >>> >>> >>> info:sid/CSA:eric-set-c >>> >>> >>> >>> >>> CSA:eric-set-c >>> 2004 >>> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA%3Aeric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature%20Circles%20in%20an%20EFL%20Classroom&sfx.response_type=multi_obj_xml >>> ED492558 >>> >>> >>> info:sid/CSA:eric-set-c >>> >>> >>> Reading without Teachers: Literature Circles in an EFL Classroom >>> >>> >>> >>> >>> >>> >>> Jeng-yih >>> >>> >>> >>> >>> T >>> >>> >>> >>> >>> >>> >>> >>> >>> Hsu >>> >>> >>> multi_obj_xml >>> CSA >>> 1 >>> >>> >>> 111027614344001 >>> 111000147718001 >>> 111018614652001 >>> 111026921949001 >>> 111096046941001 >>> 111076678899001 >>> 111032274555001 >>> 110974985005474 >>> >>> >>> s10783D3C-51A1-11E1-9473-FAA1237FF586 >>> >>> >>> Hsu, Jeng-Yih T >>> >>> >>> 2004 >>> CSA:eric-set-c >>> latin1 >>> >>> >>> J >>> >>> >>> Reading without Teachers: Literature Circles in an EFL Classroom >>> 1 >>> yes >>> >> key="rft_dat"><AN>ED492558</AN>&<PY>2004</PY& >>> gt;&<AU>Hsu, Jeng-yih Tim</AU> >>> >>> ERIC_FULL_T >>> EXTERIC >>> Full-text1110960 >>> 46941000111096046941001>> _id>getFullTxtERIC::ERIC_FT>> er>url=http://www.eric.ed.govyes>> oxy>no >>> >>> >>> ..... >>> >>> >>> From: Jonathan Rochkind [mailto:rochkind at jhu.edu] >>> Sent: Monday, January 09, 2012 1:16 PM >>> To: Poulter, Dale >>> Cc: umlaut-general at rubyforge.org >>> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >>> >>> PS: Dale, if you do file this with Ex Libris, can you give me the support number? I want to complain about it too, probably to the SFX product manager. I am very frustrated wtih the way EL keeps adding fancy logic to SFX targets that is not reflected in the API, they are not supporting the API adequately. >>> >>> On 1/9/2012 2:12 PM, Jonathan Rochkind wrote: >>> Umlaut does pass the full OpenURL to SFX. >>> >>> * But Umlaut sometimes modifies the full OpenURL before it passes it to SFX, that could be causing a problem. >>> >>> * Or, it could be a bug in SFX, where the SFX API is not returning the proper URL that the native interface does. >>> >>> >>> >>> So, the second one, bug in SFX, is the easiest to confirm or rule out first. Send the OpenURL exactly to SFX, but with&sfx.response_type=multi_obj_xml on the end and let's see what it delivers. >>> >>> I just did that, to your server. Looking at the XML results, SFX reports an ERIC_FULL_TEXT response with this: >>> >>> http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/E >>> RICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw& >>> ;ERICExtSearch_SearchValue_0=ED492558 >>> >>> That's what Umlaut is going to use. Trying the URL without your own proxy prefix: >>> >>> http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSe >>> arch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 >>> >>> I indeed do NOT get a direct link to Full Text, but instead a link to ERIC home page. (I also indeed verified that your SFX native interface indeed does take you to full text -- or at least to search results with 1 hit, which seems to be the right hit). >>> This means that unfortunately this is a bug/misdesign in SFX and it's API. >>> >>> SFX is not returning the correct URL in the XML API response, to take the user to the same thing that the SFX native interface takes the user too. Umlaut depends on the SFX XML API response, there's nothing Umlaut can do about this. This is unfortunatley not the first time I've run into something like this -- EL seems to keep adding fancy custom per-target logic to SFX, without having this be reflected in the API response. It is very frustrating. Umlaut's use of SFX requires EL to take the API seriously.... sometimes they do, sometimes they don't. >>> >>> We've got to report the bug to Ex Libris, and pray they fix it. Does this email explain enough that you are comfortable filing the support incident with EL, Dale? Makes more sense for you to file it, ideally, since it's on your server that it can currently be reproduced (my server does not currently have KB configured in such a way to get any full text hits for that OpenURL at all). >>> >>> Jonathan >>> >>> >>> On 1/9/2012 1:38 PM, Poulter, Dale wrote: >>> >>> I have been unable to reproduce a failed url in SFX (it always works), but I can reproduce at will in Umlaut. Since the ERIC full text does not have a portfolio as such but just looks for the ED number I wonder if Umlaut passes the full openurl to SFX for evaluation? >>> >>> >>> >>> From: Jonathan Rochkind [mailto:rochkind at jhu.edu] >>> Sent: Monday, January 09, 2012 10:55 AM >>> To: umlaut-general at rubyforge.org >>> Cc: Poulter, Dale >>> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >>> >>> >>> >>> Can you reproduce with SFX alone? That'd be the first debugging step I'd take. >>> >>> Take the OpenURL that the source is sending, and feed it directly to SFX. Does SFX give the right fulltext link? >>> >>> If not, this is purely an SFX issue, and Umlaut is doing the right thing in presenting the same thing as SFX. >>> >>> If yes, SFX does the right thing, then it's either an Umlaut bug/problem, or an SFX API bug problem, more debugging required to figure it out. (I have had problems in the past for some of SFX's fancier stuff where the SFX API does not reveal the same URL as SFX native. Sometimes when I report them, they fix them). >>> >>> Jonathan >>> >>> On 1/9/2012 10:48 AM, Poulter, Dale wrote: >>> >>> All, >>> >>> >>> >>> We have received a report that umlaut no longer provides the ERIC fulltext link for resources based on the ED number in the openurl. The first link below does not present the ERIC full text although it is an active target (requires the ED number which is highlighted). The second link presents the ERIC fulltext link. Has anyone else encountered this issue? Any idea what the issue is? Thanks. >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> http://umlaut.library.vanderbilt.edu/resolve?sid=CSA:eric-set-c&pid=%3 >>> CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%2 >>> 0Jeng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufir >>> st=Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literatu >>> re%20Circles%20in%20an%20EFL%20Classroom >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CA >>> N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20J >>> eng-yih%20Tim%3C%2FAU%3E&date=2004&genre=proceeding&aulast=Hsu&aufirst >>> =Jeng-yih&auinitm=T&title=Reading%20without%20Teachers%3A%20Literature >>> %20Circles%20in%20an%20EFL%20Classroom >>> >>> >>> >>> --Dale >>> >>> >>> >>> --------------------------------------- >>> Dale Poulter >>> Automation Coordinator >>> >>> Library Information Technology Services Vanderbilt University >>> >>> 419 21st Avenue South, Room 812 >>> Nashville, TN 37203-2427 >>> (615)343-5388 >>> (615)343-8834 (fax) >>> (615)207-9705 (cell) >>> dale.poulter at vanderbilt.edu >>> >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> >>> Umlaut-general mailing list >>> >>> Umlaut-general at rubyforge.org >>> >>> http://rubyforge.org/mailman/listinfo/umlaut-general >>> >>> >>> >> > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rossfsinger at gmail.com Wed Feb 15 19:49:25 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 15 Feb 2012 19:49:25 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F3C2B3E.4070906@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> Message-ID: Jonathan, it works fine if you turn the rfr_id into an info-uri http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 so preprend info:sid/ onto the sid value unless it is already an info-uri (as I recall, the OpenURL gem already does something similar with rft_id around dois and pmids, so we should be able to replicate that functionality here). -Ross. On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: > Okay, yeah, this is kind of a mess. I hate OpenURL. ?Bare with me, Dale, to > understand the problem and possible solutions (none entirely trivial, but > some possible). > > The original OpenURL is an "OpenURL 0.1". The presence of two parameters is > enough for SFX to trigger the ERIC link. ?"sid" (source id, where the > OpenURL came from) and "pid" (private id, private non-standard data from > that source). > > So for instance, this is the minimal 0.1 OpenURL which succesfully triggers > the ERIC link from your link resolver: > > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 > > > But Umlaut ends up sending an OpenURL 1.0 request to SFX. ?Umlaut actually > takes all incoming OpenURLs and translates them into an internal data > structure, doesn't even keep track of whether they came in as 1.0 or 0.1, > and then translates them back into OpenURL 1.0 to make the API request to > SFX. > > So the first problem is, Umlaut is doing some weird things with the "sid" > and "pid" parameters in this round trip to internal data structure and back. > (Those params _usually_ don't matter, is why I haven't noticed before). > > But even when I investigate fixing that, I can't figure out if there _is_ an > OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. > > Here's the closest OpenURL 1.0 equivalent I can come up with (thanks > rsinger) > > http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 > > > But that does NOT properly trigger the ERIC link in SFX. > > No completely easy fix presents itself. But we have a number of options. > > So, Dale, at this point, if you could, I think it would be helpful for you > to try to get some info from Ex Libris. (A frustrating and time consuming > thing to do, which is why I ask you to do it for your problem, heh). > > * Option 1. > > I think I remember that SFX gets that "pid" info out with custom source > parsers based on "sid". And that these "source parsers" are specific to > OpenURL 1.0 or 0.1 > > So you could ask Ex Libris if there's any way to set up a CSA source parser > that will work for OpenURL 1.0 and do the same thing it's doing now for the > 0.1 source parser. Adn what the 1.0 OpenURL should look like to be > recognized by that source parser. > > They will say "Why would you want an OpenURL 1.0 source parser for CSA, CSA > only sends OpenURL 0.1?" And it'll be confusing. > > And this would still require some changes to Umlaut, as I've noticed through > this debugging that it doesn't handle sid/pid quite right anyway. Fixing > 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. > > * Option 2. > > Alternately, we could try to hack Umlaut so it can send an OpenURL 0.1 to > SFX instead. As a configuration option, always? That might mess other things > up. Only if the original OpenURL was 0.1? ?Umlaut doesn't keep track of this > right now, but theoretically it could. ?Only if the OpenURL has a "pid" in > it? ?That might actually be best, since "pid" as a literal string is OpenURL > 0.1-specific so is a signal that the incoming was OpenURL 0.1 -- plus it's > really only incoming OpenURLs with 'pid' that need to be sent as OpenURL 0.1 > to SFX. > > This would still require the same fixes to Umlaut's handling of sid/pid > mentioned above. > > * Option 3. > > I might lean toward this one, actually. ?The reason we use SFX is for it's > knowledge base. That is, for portfolio-based targets, as well as SFX's > knowledge of outgoing link syntax. > > You helpfully (thanks!) mentioned that the ERIC target is not portfolio > based. It's just an automatic translation "if ERIC id detected, make a link > to ERIC like so." > > So we could write a new service plugin for Umlaut that handles ERIC id's and > outgoing links to ERIC. Activate that in Umlaut, de-activate the ERIC target > in SFX. > > This would require "us" (Umlaut) to take responsibility for maintaining the > outgoing ERIC link syntax and changing our code if it changes. But it's not > like Ex Libris is very good at that maintenance anyway, and I bet ERIC > outgoing link syntax doesn't change much. > > It would also require "us" to identify all the places an ERIC id might come > in in source-specific ways. This example is one of them, a "pid" formatted > in a certain way, when "sid" is CSA. ?Are there other custom places SFX > recognizes ERIC id's, from other places? ?We'd have to figure that out, and > copy that code in Umlaut too (taking maintenance responsibility too). > > But we wouldn't have to deal with SFX's weirdness here, we'd take > responsibility for handling this ourselves, which is part of the point of > Umlaut, really --- use SFX for it's portfolio-based knowledge base, but do > ourselves what we can do ourselves better. (Recognizing in this case it's > not really 'better', it's just that SFX's weirdness combined with OpenURL's > crappiness puts us in a corner). > > And someone would have to write the code in Umlaut. Which I _might_ find > time to do, or else, Dale, how do you feel about some ruby coding? Or else > we could try to find someone else. > > CONCLUSION > > So, any thoughts Dale? > > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From dale.poulter at Vanderbilt.Edu Wed Feb 15 20:54:21 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Wed, 15 Feb 2012 19:54:21 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> Ross and Jonathan, Thanks. I try this in Umlaut (http://umlaut.library.vanderbilt.edu/resolve?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004) and it did not seem to work. Do I understand correctly that we just need to adapt the metadata_helper to access the ED/EJ numbers similar to the pmid? I did notice that in SFX Ex Libris seems to have included most of the sids but not the eric specific sid. I'll report this as an issue to EL, but am happy to try some coding. -----Original Message----- From: Ross Singer [mailto:rossfsinger at gmail.com] Sent: Wednesday, February 15, 2012 6:49 PM To: umlaut-general at rubyforge.org Cc: Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, it works fine if you turn the rfr_id into an info-uri http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 so preprend info:sid/ onto the sid value unless it is already an info-uri (as I recall, the OpenURL gem already does something similar with rft_id around dois and pmids, so we should be able to replicate that functionality here). -Ross. On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: > Okay, yeah, this is kind of a mess. I hate OpenURL. ?Bare with me, > Dale, to understand the problem and possible solutions (none entirely > trivial, but some possible). > > The original OpenURL is an "OpenURL 0.1". The presence of two > parameters is enough for SFX to trigger the ERIC link. ?"sid" (source > id, where the OpenURL came from) and "pid" (private id, private > non-standard data from that source). > > So for instance, this is the minimal 0.1 OpenURL which succesfully > triggers the ERIC link from your link resolver: > > http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CA > N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20J > eng-yih%20Tim%3C%2FAU%3E&date=2004 > > > But Umlaut ends up sending an OpenURL 1.0 request to SFX. ?Umlaut > actually takes all incoming OpenURLs and translates them into an > internal data structure, doesn't even keep track of whether they came > in as 1.0 or 0.1, and then translates them back into OpenURL 1.0 to > make the API request to SFX. > > So the first problem is, Umlaut is doing some weird things with the "sid" > and "pid" parameters in this round trip to internal data structure and back. > (Those params _usually_ don't matter, is why I haven't noticed before). > > But even when I investigate fixing that, I can't figure out if there > _is_ an OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. > > Here's the closest OpenURL 1.0 equivalent I can come up with (thanks > rsinger) > > http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_d > at=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu > %2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 > > > But that does NOT properly trigger the ERIC link in SFX. > > No completely easy fix presents itself. But we have a number of options. > > So, Dale, at this point, if you could, I think it would be helpful for > you to try to get some info from Ex Libris. (A frustrating and time > consuming thing to do, which is why I ask you to do it for your problem, heh). > > * Option 1. > > I think I remember that SFX gets that "pid" info out with custom > source parsers based on "sid". And that these "source parsers" are > specific to OpenURL 1.0 or 0.1 > > So you could ask Ex Libris if there's any way to set up a CSA source > parser that will work for OpenURL 1.0 and do the same thing it's doing > now for the > 0.1 source parser. Adn what the 1.0 OpenURL should look like to be > recognized by that source parser. > > They will say "Why would you want an OpenURL 1.0 source parser for > CSA, CSA only sends OpenURL 0.1?" And it'll be confusing. > > And this would still require some changes to Umlaut, as I've noticed > through this debugging that it doesn't handle sid/pid quite right > anyway. Fixing 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. > > * Option 2. > > Alternately, we could try to hack Umlaut so it can send an OpenURL 0.1 > to SFX instead. As a configuration option, always? That might mess > other things up. Only if the original OpenURL was 0.1? ?Umlaut doesn't > keep track of this right now, but theoretically it could. ?Only if the > OpenURL has a "pid" in it? ?That might actually be best, since "pid" > as a literal string is OpenURL 0.1-specific so is a signal that the > incoming was OpenURL 0.1 -- plus it's really only incoming OpenURLs > with 'pid' that need to be sent as OpenURL 0.1 to SFX. > > This would still require the same fixes to Umlaut's handling of > sid/pid mentioned above. > > * Option 3. > > I might lean toward this one, actually. ?The reason we use SFX is for > it's knowledge base. That is, for portfolio-based targets, as well as > SFX's knowledge of outgoing link syntax. > > You helpfully (thanks!) mentioned that the ERIC target is not > portfolio based. It's just an automatic translation "if ERIC id > detected, make a link to ERIC like so." > > So we could write a new service plugin for Umlaut that handles ERIC > id's and outgoing links to ERIC. Activate that in Umlaut, de-activate > the ERIC target in SFX. > > This would require "us" (Umlaut) to take responsibility for > maintaining the outgoing ERIC link syntax and changing our code if it > changes. But it's not like Ex Libris is very good at that maintenance > anyway, and I bet ERIC outgoing link syntax doesn't change much. > > It would also require "us" to identify all the places an ERIC id might > come in in source-specific ways. This example is one of them, a "pid" > formatted in a certain way, when "sid" is CSA. ?Are there other custom > places SFX recognizes ERIC id's, from other places? ?We'd have to > figure that out, and copy that code in Umlaut too (taking maintenance responsibility too). > > But we wouldn't have to deal with SFX's weirdness here, we'd take > responsibility for handling this ourselves, which is part of the point > of Umlaut, really --- use SFX for it's portfolio-based knowledge base, > but do ourselves what we can do ourselves better. (Recognizing in this > case it's not really 'better', it's just that SFX's weirdness combined > with OpenURL's crappiness puts us in a corner). > > And someone would have to write the code in Umlaut. Which I _might_ > find time to do, or else, Dale, how do you feel about some ruby > coding? Or else we could try to find someone else. > > CONCLUSION > > So, any thoughts Dale? > > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rossfsinger at gmail.com Wed Feb 15 21:12:43 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 15 Feb 2012 21:12:43 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: Dale, is there any way you can see in your SFX logs what Umlaut sent to SFX? My guess is that it's dropping the rft_dat key (maybe rfr_id, too). Actually, Dale, can you tell me what SFX actually links to with the 0.1 link? Does it link directly to the article? Did the 1.0 link that I made link to the article correctly (I can't tell what happens on the other side of EZProxy)? Thanks, -Ross. On Wed, Feb 15, 2012 at 8:54 PM, Poulter, Dale wrote: > Ross and Jonathan, > > Thanks. ? I try this in Umlaut (http://umlaut.library.vanderbilt.edu/resolve?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004) and it did not seem to work. ?Do I understand correctly that we just need to adapt the metadata_helper to access the ED/EJ numbers similar to the pmid? ?I did notice that in SFX Ex Libris seems to have included most of the sids but not the eric specific sid. ?I'll report this as an issue to EL, but am happy to try some coding. > > > > -----Original Message----- > From: Ross Singer [mailto:rossfsinger at gmail.com] > Sent: Wednesday, February 15, 2012 6:49 PM > To: umlaut-general at rubyforge.org > Cc: Poulter, Dale > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > Jonathan, it works fine if you turn the rfr_id into an info-uri > > http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 > > so preprend info:sid/ onto the sid value unless it is already an info-uri (as I recall, the OpenURL gem already does something similar with rft_id around dois and pmids, so we should be able to replicate that functionality here). > > -Ross. > > On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: >> Okay, yeah, this is kind of a mess. I hate OpenURL. ?Bare with me, >> Dale, to understand the problem and possible solutions (none entirely >> trivial, but some possible). >> >> The original OpenURL is an "OpenURL 0.1". The presence of two >> parameters is enough for SFX to trigger the ERIC link. ?"sid" (source >> id, where the OpenURL came from) and "pid" (private id, private >> non-standard data from that source). >> >> So for instance, this is the minimal 0.1 OpenURL which succesfully >> triggers the ERIC link from your link resolver: >> >> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CA >> N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20J >> eng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> >> But Umlaut ends up sending an OpenURL 1.0 request to SFX. ?Umlaut >> actually takes all incoming OpenURLs and translates them into an >> internal data structure, doesn't even keep track of whether they came >> in as 1.0 or 0.1, and then translates them back into OpenURL 1.0 to >> make the API request to SFX. >> >> So the first problem is, Umlaut is doing some weird things with the "sid" >> and "pid" parameters in this round trip to internal data structure and back. >> (Those params _usually_ don't matter, is why I haven't noticed before). >> >> But even when I investigate fixing that, I can't figure out if there >> _is_ an OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. >> >> Here's the closest OpenURL 1.0 equivalent I can come up with (thanks >> rsinger) >> >> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_d >> at=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu >> %2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> >> But that does NOT properly trigger the ERIC link in SFX. >> >> No completely easy fix presents itself. But we have a number of options. >> >> So, Dale, at this point, if you could, I think it would be helpful for >> you to try to get some info from Ex Libris. (A frustrating and time >> consuming thing to do, which is why I ask you to do it for your problem, heh). >> >> * Option 1. >> >> I think I remember that SFX gets that "pid" info out with custom >> source parsers based on "sid". And that these "source parsers" are >> specific to OpenURL 1.0 or 0.1 >> >> So you could ask Ex Libris if there's any way to set up a CSA source >> parser that will work for OpenURL 1.0 and do the same thing it's doing >> now for the >> 0.1 source parser. Adn what the 1.0 OpenURL should look like to be >> recognized by that source parser. >> >> They will say "Why would you want an OpenURL 1.0 source parser for >> CSA, CSA only sends OpenURL 0.1?" And it'll be confusing. >> >> And this would still require some changes to Umlaut, as I've noticed >> through this debugging that it doesn't handle sid/pid quite right >> anyway. Fixing 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. >> >> * Option 2. >> >> Alternately, we could try to hack Umlaut so it can send an OpenURL 0.1 >> to SFX instead. As a configuration option, always? That might mess >> other things up. Only if the original OpenURL was 0.1? ?Umlaut doesn't >> keep track of this right now, but theoretically it could. ?Only if the >> OpenURL has a "pid" in it? ?That might actually be best, since "pid" >> as a literal string is OpenURL 0.1-specific so is a signal that the >> incoming was OpenURL 0.1 -- plus it's really only incoming OpenURLs >> with 'pid' that need to be sent as OpenURL 0.1 to SFX. >> >> This would still require the same fixes to Umlaut's handling of >> sid/pid mentioned above. >> >> * Option 3. >> >> I might lean toward this one, actually. ?The reason we use SFX is for >> it's knowledge base. That is, for portfolio-based targets, as well as >> SFX's knowledge of outgoing link syntax. >> >> You helpfully (thanks!) mentioned that the ERIC target is not >> portfolio based. It's just an automatic translation "if ERIC id >> detected, make a link to ERIC like so." >> >> So we could write a new service plugin for Umlaut that handles ERIC >> id's and outgoing links to ERIC. Activate that in Umlaut, de-activate >> the ERIC target in SFX. >> >> This would require "us" (Umlaut) to take responsibility for >> maintaining the outgoing ERIC link syntax and changing our code if it >> changes. But it's not like Ex Libris is very good at that maintenance >> anyway, and I bet ERIC outgoing link syntax doesn't change much. >> >> It would also require "us" to identify all the places an ERIC id might >> come in in source-specific ways. This example is one of them, a "pid" >> formatted in a certain way, when "sid" is CSA. ?Are there other custom >> places SFX recognizes ERIC id's, from other places? ?We'd have to >> figure that out, and copy that code in Umlaut too (taking maintenance responsibility too). >> >> But we wouldn't have to deal with SFX's weirdness here, we'd take >> responsibility for handling this ourselves, which is part of the point >> of Umlaut, really --- use SFX for it's portfolio-based knowledge base, >> but do ourselves what we can do ourselves better. (Recognizing in this >> case it's not really 'better', it's just that SFX's weirdness combined >> with OpenURL's crappiness puts us in a corner). >> >> And someone would have to write the code in Umlaut. Which I _might_ >> find time to do, or else, Dale, how do you feel about some ruby >> coding? Or else we could try to find someone else. >> >> CONCLUSION >> >> So, any thoughts Dale? >> >> _______________________________________________ >> Umlaut-general mailing list >> Umlaut-general at rubyforge.org >> http://rubyforge.org/mailman/listinfo/umlaut-general > > From dale.poulter at Vanderbilt.Edu Wed Feb 15 21:29:26 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Wed, 15 Feb 2012 20:29:26 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <7288D49C0448E840B752320350AFA94E0135043402@its-hcwnem03.ds.Vanderbilt.edu> Ross, Here is the entry in the logs (/vu?url_ver=Z39.88-2004&url_ctx_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Actx&ctx_tim=2012-02-15T19%3A16%3A37-06%3A00&ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&ctx_id=&rfr_id=info%3Asid%2FCSA%3Aeric-set-c&req.ip=98.211.26.149&rft.dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C+Jeng-yih+Tim%3C%2FAU%3E&rft.date=2004&rft.genre=journal&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&sfx.response_type=multi_obj_xml) . When SFX processes the link it actually just creates a link into the ERIC site using the ED or EJ number. Here is an example that actually fails to find an item (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED49255) -----Original Message----- From: Ross Singer [mailto:rossfsinger at gmail.com] Sent: Wednesday, February 15, 2012 8:13 PM To: Poulter, Dale Cc: umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Dale, is there any way you can see in your SFX logs what Umlaut sent to SFX? My guess is that it's dropping the rft_dat key (maybe rfr_id, too). Actually, Dale, can you tell me what SFX actually links to with the 0.1 link? Does it link directly to the article? Did the 1.0 link that I made link to the article correctly (I can't tell what happens on the other side of EZProxy)? Thanks, -Ross. On Wed, Feb 15, 2012 at 8:54 PM, Poulter, Dale wrote: > Ross and Jonathan, > > Thanks. ? I try this in Umlaut (http://umlaut.library.vanderbilt.edu/resolve?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004) and it did not seem to work. ?Do I understand correctly that we just need to adapt the metadata_helper to access the ED/EJ numbers similar to the pmid? ?I did notice that in SFX Ex Libris seems to have included most of the sids but not the eric specific sid. ?I'll report this as an issue to EL, but am happy to try some coding. > > > > -----Original Message----- > From: Ross Singer [mailto:rossfsinger at gmail.com] > Sent: Wednesday, February 15, 2012 6:49 PM > To: umlaut-general at rubyforge.org > Cc: Poulter, Dale > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > Jonathan, it works fine if you turn the rfr_id into an info-uri > > http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-se > t-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3 > CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 > > so preprend info:sid/ onto the sid value unless it is already an info-uri (as I recall, the OpenURL gem already does something similar with rft_id around dois and pmids, so we should be able to replicate that functionality here). > > -Ross. > > On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: >> Okay, yeah, this is kind of a mess. I hate OpenURL. ?Bare with me, >> Dale, to understand the problem and possible solutions (none entirely >> trivial, but some possible). >> >> The original OpenURL is an "OpenURL 0.1". The presence of two >> parameters is enough for SFX to trigger the ERIC link. ?"sid" (source >> id, where the OpenURL came from) and "pid" (private id, private >> non-standard data from that source). >> >> So for instance, this is the minimal 0.1 OpenURL which succesfully >> triggers the ERIC link from your link resolver: >> >> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3C >> A >> N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20 >> J >> eng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> >> But Umlaut ends up sending an OpenURL 1.0 request to SFX. ?Umlaut >> actually takes all incoming OpenURLs and translates them into an >> internal data structure, doesn't even keep track of whether they came >> in as 1.0 or 0.1, and then translates them back into OpenURL 1.0 to >> make the API request to SFX. >> >> So the first problem is, Umlaut is doing some weird things with the "sid" >> and "pid" parameters in this round trip to internal data structure and back. >> (Those params _usually_ don't matter, is why I haven't noticed before). >> >> But even when I investigate fixing that, I can't figure out if there >> _is_ an OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. >> >> Here's the closest OpenURL 1.0 equivalent I can come up with (thanks >> rsinger) >> >> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_ >> d >> at=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHs >> u >> %2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> >> But that does NOT properly trigger the ERIC link in SFX. >> >> No completely easy fix presents itself. But we have a number of options. >> >> So, Dale, at this point, if you could, I think it would be helpful >> for you to try to get some info from Ex Libris. (A frustrating and >> time consuming thing to do, which is why I ask you to do it for your problem, heh). >> >> * Option 1. >> >> I think I remember that SFX gets that "pid" info out with custom >> source parsers based on "sid". And that these "source parsers" are >> specific to OpenURL 1.0 or 0.1 >> >> So you could ask Ex Libris if there's any way to set up a CSA source >> parser that will work for OpenURL 1.0 and do the same thing it's >> doing now for the >> 0.1 source parser. Adn what the 1.0 OpenURL should look like to be >> recognized by that source parser. >> >> They will say "Why would you want an OpenURL 1.0 source parser for >> CSA, CSA only sends OpenURL 0.1?" And it'll be confusing. >> >> And this would still require some changes to Umlaut, as I've noticed >> through this debugging that it doesn't handle sid/pid quite right >> anyway. Fixing 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. >> >> * Option 2. >> >> Alternately, we could try to hack Umlaut so it can send an OpenURL >> 0.1 to SFX instead. As a configuration option, always? That might >> mess other things up. Only if the original OpenURL was 0.1? ?Umlaut >> doesn't keep track of this right now, but theoretically it could. ? >> Only if the OpenURL has a "pid" in it? ?That might actually be best, since "pid" >> as a literal string is OpenURL 0.1-specific so is a signal that the >> incoming was OpenURL 0.1 -- plus it's really only incoming OpenURLs >> with 'pid' that need to be sent as OpenURL 0.1 to SFX. >> >> This would still require the same fixes to Umlaut's handling of >> sid/pid mentioned above. >> >> * Option 3. >> >> I might lean toward this one, actually. ?The reason we use SFX is for >> it's knowledge base. That is, for portfolio-based targets, as well as >> SFX's knowledge of outgoing link syntax. >> >> You helpfully (thanks!) mentioned that the ERIC target is not >> portfolio based. It's just an automatic translation "if ERIC id >> detected, make a link to ERIC like so." >> >> So we could write a new service plugin for Umlaut that handles ERIC >> id's and outgoing links to ERIC. Activate that in Umlaut, de-activate >> the ERIC target in SFX. >> >> This would require "us" (Umlaut) to take responsibility for >> maintaining the outgoing ERIC link syntax and changing our code if it >> changes. But it's not like Ex Libris is very good at that maintenance >> anyway, and I bet ERIC outgoing link syntax doesn't change much. >> >> It would also require "us" to identify all the places an ERIC id >> might come in in source-specific ways. This example is one of them, a "pid" >> formatted in a certain way, when "sid" is CSA. ?Are there other >> custom places SFX recognizes ERIC id's, from other places? ?We'd have >> to figure that out, and copy that code in Umlaut too (taking maintenance responsibility too). >> >> But we wouldn't have to deal with SFX's weirdness here, we'd take >> responsibility for handling this ourselves, which is part of the >> point of Umlaut, really --- use SFX for it's portfolio-based >> knowledge base, but do ourselves what we can do ourselves better. >> (Recognizing in this case it's not really 'better', it's just that >> SFX's weirdness combined with OpenURL's crappiness puts us in a corner). >> >> And someone would have to write the code in Umlaut. Which I _might_ >> find time to do, or else, Dale, how do you feel about some ruby >> coding? Or else we could try to find someone else. >> >> CONCLUSION >> >> So, any thoughts Dale? >> >> _______________________________________________ >> Umlaut-general mailing list >> Umlaut-general at rubyforge.org >> http://rubyforge.org/mailman/listinfo/umlaut-general > > From rossfsinger at gmail.com Wed Feb 15 22:10:39 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 15 Feb 2012 22:10:39 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E0135043402@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> <7288D49C0448E840B752320350AFA94E0135043402@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: Ok, it appears to be turning rft_dat into rft.dat, which would be the problem. Not sure why it does this, but given the scarcity of examples that actually use pid/rft_dat, it's not surprising that it's never been noticed before. In fact, I am going to take a guess that this bug is actually in rOpenURL, not Umlaut. -Ross. On Wed, Feb 15, 2012 at 9:29 PM, Poulter, Dale wrote: > Ross, > > Here is the entry in the logs (/vu?url_ver=Z39.88-2004&url_ctx_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Actx&ctx_tim=2012-02-15T19%3A16%3A37-06%3A00&ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&ctx_id=&rfr_id=info%3Asid%2FCSA%3Aeric-set-c&req.ip=98.211.26.149&rft.dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C+Jeng-yih+Tim%3C%2FAU%3E&rft.date=2004&rft.genre=journal&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&sfx.response_type=multi_obj_xml) . > > When SFX processes the link it actually just creates a link into the ERIC site using the ED or EJ number. Here is an example that actually fails to find an item (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED49255) > > > > > > > -----Original Message----- > From: Ross Singer [mailto:rossfsinger at gmail.com] > Sent: Wednesday, February 15, 2012 8:13 PM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > Dale, is there any way you can see in your SFX logs what Umlaut sent to SFX? > > My guess is that it's dropping the rft_dat key (maybe rfr_id, too). > > Actually, Dale, can you tell me what SFX actually links to with the > 0.1 link? ?Does it link directly to the article? > > Did the 1.0 link that I made link to the article correctly (I can't tell what happens on the other side of EZProxy)? > > Thanks, > -Ross. > > On Wed, Feb 15, 2012 at 8:54 PM, Poulter, ? Dale > wrote: >> Ross and Jonathan, >> >> Thanks. ? I try this in Umlaut (http://umlaut.library.vanderbilt.edu/resolve?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004) and it did not seem to work. ?Do I understand correctly that we just need to adapt the metadata_helper to access the ED/EJ numbers similar to the pmid? ?I did notice that in SFX Ex Libris seems to have included most of the sids but not the eric specific sid. ?I'll report this as an issue to EL, but am happy to try some coding. >> >> >> >> -----Original Message----- >> From: Ross Singer [mailto:rossfsinger at gmail.com] >> Sent: Wednesday, February 15, 2012 6:49 PM >> To: umlaut-general at rubyforge.org >> Cc: Poulter, Dale >> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >> >> Jonathan, it works fine if you turn the rfr_id into an info-uri >> >> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-se >> t-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3 >> CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> so preprend info:sid/ onto the sid value unless it is already an info-uri (as I recall, the OpenURL gem already does something similar with rft_id around dois and pmids, so we should be able to replicate that functionality here). >> >> -Ross. >> >> On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: >>> Okay, yeah, this is kind of a mess. I hate OpenURL. ?Bare with me, >>> Dale, to understand the problem and possible solutions (none entirely >>> trivial, but some possible). >>> >>> The original OpenURL is an "OpenURL 0.1". The presence of two >>> parameters is enough for SFX to trigger the ERIC link. ?"sid" (source >>> id, where the OpenURL came from) and "pid" (private id, private >>> non-standard data from that source). >>> >>> So for instance, this is the minimal 0.1 OpenURL which succesfully >>> triggers the ERIC link from your link resolver: >>> >>> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3C >>> A >>> N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20 >>> J >>> eng-yih%20Tim%3C%2FAU%3E&date=2004 >>> >>> >>> But Umlaut ends up sending an OpenURL 1.0 request to SFX. ?Umlaut >>> actually takes all incoming OpenURLs and translates them into an >>> internal data structure, doesn't even keep track of whether they came >>> in as 1.0 or 0.1, and then translates them back into OpenURL 1.0 to >>> make the API request to SFX. >>> >>> So the first problem is, Umlaut is doing some weird things with the "sid" >>> and "pid" parameters in this round trip to internal data structure and back. >>> (Those params _usually_ don't matter, is why I haven't noticed before). >>> >>> But even when I investigate fixing that, I can't figure out if there >>> _is_ an OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. >>> >>> Here's the closest OpenURL 1.0 equivalent I can come up with (thanks >>> rsinger) >>> >>> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_ >>> d >>> at=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHs >>> u >>> %2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >>> >>> >>> But that does NOT properly trigger the ERIC link in SFX. >>> >>> No completely easy fix presents itself. But we have a number of options. >>> >>> So, Dale, at this point, if you could, I think it would be helpful >>> for you to try to get some info from Ex Libris. (A frustrating and >>> time consuming thing to do, which is why I ask you to do it for your problem, heh). >>> >>> * Option 1. >>> >>> I think I remember that SFX gets that "pid" info out with custom >>> source parsers based on "sid". And that these "source parsers" are >>> specific to OpenURL 1.0 or 0.1 >>> >>> So you could ask Ex Libris if there's any way to set up a CSA source >>> parser that will work for OpenURL 1.0 and do the same thing it's >>> doing now for the >>> 0.1 source parser. Adn what the 1.0 OpenURL should look like to be >>> recognized by that source parser. >>> >>> They will say "Why would you want an OpenURL 1.0 source parser for >>> CSA, CSA only sends OpenURL 0.1?" And it'll be confusing. >>> >>> And this would still require some changes to Umlaut, as I've noticed >>> through this debugging that it doesn't handle sid/pid quite right >>> anyway. Fixing 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. >>> >>> * Option 2. >>> >>> Alternately, we could try to hack Umlaut so it can send an OpenURL >>> 0.1 to SFX instead. As a configuration option, always? That might >>> mess other things up. Only if the original OpenURL was 0.1? ?Umlaut >>> doesn't keep track of this right now, but theoretically it could. >>> Only if the OpenURL has a "pid" in it? ?That might actually be best, since "pid" >>> as a literal string is OpenURL 0.1-specific so is a signal that the >>> incoming was OpenURL 0.1 -- plus it's really only incoming OpenURLs >>> with 'pid' that need to be sent as OpenURL 0.1 to SFX. >>> >>> This would still require the same fixes to Umlaut's handling of >>> sid/pid mentioned above. >>> >>> * Option 3. >>> >>> I might lean toward this one, actually. ?The reason we use SFX is for >>> it's knowledge base. That is, for portfolio-based targets, as well as >>> SFX's knowledge of outgoing link syntax. >>> >>> You helpfully (thanks!) mentioned that the ERIC target is not >>> portfolio based. It's just an automatic translation "if ERIC id >>> detected, make a link to ERIC like so." >>> >>> So we could write a new service plugin for Umlaut that handles ERIC >>> id's and outgoing links to ERIC. Activate that in Umlaut, de-activate >>> the ERIC target in SFX. >>> >>> This would require "us" (Umlaut) to take responsibility for >>> maintaining the outgoing ERIC link syntax and changing our code if it >>> changes. But it's not like Ex Libris is very good at that maintenance >>> anyway, and I bet ERIC outgoing link syntax doesn't change much. >>> >>> It would also require "us" to identify all the places an ERIC id >>> might come in in source-specific ways. This example is one of them, a "pid" >>> formatted in a certain way, when "sid" is CSA. ?Are there other >>> custom places SFX recognizes ERIC id's, from other places? ?We'd have >>> to figure that out, and copy that code in Umlaut too (taking maintenance responsibility too). >>> >>> But we wouldn't have to deal with SFX's weirdness here, we'd take >>> responsibility for handling this ourselves, which is part of the >>> point of Umlaut, really --- use SFX for it's portfolio-based >>> knowledge base, but do ourselves what we can do ourselves better. >>> (Recognizing in this case it's not really 'better', it's just that >>> SFX's weirdness combined with OpenURL's crappiness puts us in a corner). >>> >>> And someone would have to write the code in Umlaut. Which I _might_ >>> find time to do, or else, Dale, how do you feel about some ruby >>> coding? Or else we could try to find someone else. >>> >>> CONCLUSION >>> >>> So, any thoughts Dale? >>> >>> _______________________________________________ >>> Umlaut-general mailing list >>> Umlaut-general at rubyforge.org >>> http://rubyforge.org/mailman/listinfo/umlaut-general >> >> > > From rossfsinger at gmail.com Wed Feb 15 22:22:32 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 15 Feb 2012 22:22:32 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> <7288D49C0448E840B752320350AFA94E0135043402@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: Hmm, well, I just looked at rOpenURL, and it doesn't look like it's coming from there (in fact, it looks like it should be properly setting the rfr_id as an info-uri, too), so I'm not sure *where* this bug is coming from. -Ross. On Wed, Feb 15, 2012 at 10:10 PM, Ross Singer wrote: > Ok, it appears to be turning rft_dat into rft.dat, which would be the > problem. ?Not sure why it does this, but given the scarcity of > examples that actually use pid/rft_dat, it's not surprising that it's > never been noticed before. ?In fact, I am going to take a guess that > this bug is actually in rOpenURL, not Umlaut. > > -Ross. > > On Wed, Feb 15, 2012 at 9:29 PM, Poulter, ? Dale > wrote: >> Ross, >> >> Here is the entry in the logs (/vu?url_ver=Z39.88-2004&url_ctx_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Actx&ctx_tim=2012-02-15T19%3A16%3A37-06%3A00&ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&ctx_id=&rfr_id=info%3Asid%2FCSA%3Aeric-set-c&req.ip=98.211.26.149&rft.dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C+Jeng-yih+Tim%3C%2FAU%3E&rft.date=2004&rft.genre=journal&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&sfx.response_type=multi_obj_xml) . >> >> When SFX processes the link it actually just creates a link into the ERIC site using the ED or EJ number. Here is an example that actually fails to find an item (http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED49255) >> >> >> >> >> >> >> -----Original Message----- >> From: Ross Singer [mailto:rossfsinger at gmail.com] >> Sent: Wednesday, February 15, 2012 8:13 PM >> To: Poulter, Dale >> Cc: umlaut-general at rubyforge.org >> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >> >> Dale, is there any way you can see in your SFX logs what Umlaut sent to SFX? >> >> My guess is that it's dropping the rft_dat key (maybe rfr_id, too). >> >> Actually, Dale, can you tell me what SFX actually links to with the >> 0.1 link? ?Does it link directly to the article? >> >> Did the 1.0 link that I made link to the article correctly (I can't tell what happens on the other side of EZProxy)? >> >> Thanks, >> -Ross. >> >> On Wed, Feb 15, 2012 at 8:54 PM, Poulter, ? Dale >> wrote: >>> Ross and Jonathan, >>> >>> Thanks. ? I try this in Umlaut (http://umlaut.library.vanderbilt.edu/resolve?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004) and it did not seem to work. ?Do I understand correctly that we just need to adapt the metadata_helper to access the ED/EJ numbers similar to the pmid? ?I did notice that in SFX Ex Libris seems to have included most of the sids but not the eric specific sid. ?I'll report this as an issue to EL, but am happy to try some coding. >>> >>> >>> >>> -----Original Message----- >>> From: Ross Singer [mailto:rossfsinger at gmail.com] >>> Sent: Wednesday, February 15, 2012 6:49 PM >>> To: umlaut-general at rubyforge.org >>> Cc: Poulter, Dale >>> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >>> >>> Jonathan, it works fine if you turn the rfr_id into an info-uri >>> >>> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-se >>> t-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3 >>> CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >>> >>> so preprend info:sid/ onto the sid value unless it is already an info-uri (as I recall, the OpenURL gem already does something similar with rft_id around dois and pmids, so we should be able to replicate that functionality here). >>> >>> -Ross. >>> >>> On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: >>>> Okay, yeah, this is kind of a mess. I hate OpenURL. ?Bare with me, >>>> Dale, to understand the problem and possible solutions (none entirely >>>> trivial, but some possible). >>>> >>>> The original OpenURL is an "OpenURL 0.1". The presence of two >>>> parameters is enough for SFX to trigger the ERIC link. ?"sid" (source >>>> id, where the OpenURL came from) and "pid" (private id, private >>>> non-standard data from that source). >>>> >>>> So for instance, this is the minimal 0.1 OpenURL which succesfully >>>> triggers the ERIC link from your link resolver: >>>> >>>> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3C >>>> A >>>> N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20 >>>> J >>>> eng-yih%20Tim%3C%2FAU%3E&date=2004 >>>> >>>> >>>> But Umlaut ends up sending an OpenURL 1.0 request to SFX. ?Umlaut >>>> actually takes all incoming OpenURLs and translates them into an >>>> internal data structure, doesn't even keep track of whether they came >>>> in as 1.0 or 0.1, and then translates them back into OpenURL 1.0 to >>>> make the API request to SFX. >>>> >>>> So the first problem is, Umlaut is doing some weird things with the "sid" >>>> and "pid" parameters in this round trip to internal data structure and back. >>>> (Those params _usually_ don't matter, is why I haven't noticed before). >>>> >>>> But even when I investigate fixing that, I can't figure out if there >>>> _is_ an OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. >>>> >>>> Here's the closest OpenURL 1.0 equivalent I can come up with (thanks >>>> rsinger) >>>> >>>> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_ >>>> d >>>> at=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHs >>>> u >>>> %2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >>>> >>>> >>>> But that does NOT properly trigger the ERIC link in SFX. >>>> >>>> No completely easy fix presents itself. But we have a number of options. >>>> >>>> So, Dale, at this point, if you could, I think it would be helpful >>>> for you to try to get some info from Ex Libris. (A frustrating and >>>> time consuming thing to do, which is why I ask you to do it for your problem, heh). >>>> >>>> * Option 1. >>>> >>>> I think I remember that SFX gets that "pid" info out with custom >>>> source parsers based on "sid". And that these "source parsers" are >>>> specific to OpenURL 1.0 or 0.1 >>>> >>>> So you could ask Ex Libris if there's any way to set up a CSA source >>>> parser that will work for OpenURL 1.0 and do the same thing it's >>>> doing now for the >>>> 0.1 source parser. Adn what the 1.0 OpenURL should look like to be >>>> recognized by that source parser. >>>> >>>> They will say "Why would you want an OpenURL 1.0 source parser for >>>> CSA, CSA only sends OpenURL 0.1?" And it'll be confusing. >>>> >>>> And this would still require some changes to Umlaut, as I've noticed >>>> through this debugging that it doesn't handle sid/pid quite right >>>> anyway. Fixing 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. >>>> >>>> * Option 2. >>>> >>>> Alternately, we could try to hack Umlaut so it can send an OpenURL >>>> 0.1 to SFX instead. As a configuration option, always? That might >>>> mess other things up. Only if the original OpenURL was 0.1? ?Umlaut >>>> doesn't keep track of this right now, but theoretically it could. >>>> Only if the OpenURL has a "pid" in it? ?That might actually be best, since "pid" >>>> as a literal string is OpenURL 0.1-specific so is a signal that the >>>> incoming was OpenURL 0.1 -- plus it's really only incoming OpenURLs >>>> with 'pid' that need to be sent as OpenURL 0.1 to SFX. >>>> >>>> This would still require the same fixes to Umlaut's handling of >>>> sid/pid mentioned above. >>>> >>>> * Option 3. >>>> >>>> I might lean toward this one, actually. ?The reason we use SFX is for >>>> it's knowledge base. That is, for portfolio-based targets, as well as >>>> SFX's knowledge of outgoing link syntax. >>>> >>>> You helpfully (thanks!) mentioned that the ERIC target is not >>>> portfolio based. It's just an automatic translation "if ERIC id >>>> detected, make a link to ERIC like so." >>>> >>>> So we could write a new service plugin for Umlaut that handles ERIC >>>> id's and outgoing links to ERIC. Activate that in Umlaut, de-activate >>>> the ERIC target in SFX. >>>> >>>> This would require "us" (Umlaut) to take responsibility for >>>> maintaining the outgoing ERIC link syntax and changing our code if it >>>> changes. But it's not like Ex Libris is very good at that maintenance >>>> anyway, and I bet ERIC outgoing link syntax doesn't change much. >>>> >>>> It would also require "us" to identify all the places an ERIC id >>>> might come in in source-specific ways. This example is one of them, a "pid" >>>> formatted in a certain way, when "sid" is CSA. ?Are there other >>>> custom places SFX recognizes ERIC id's, from other places? ?We'd have >>>> to figure that out, and copy that code in Umlaut too (taking maintenance responsibility too). >>>> >>>> But we wouldn't have to deal with SFX's weirdness here, we'd take >>>> responsibility for handling this ourselves, which is part of the >>>> point of Umlaut, really --- use SFX for it's portfolio-based >>>> knowledge base, but do ourselves what we can do ourselves better. >>>> (Recognizing in this case it's not really 'better', it's just that >>>> SFX's weirdness combined with OpenURL's crappiness puts us in a corner). >>>> >>>> And someone would have to write the code in Umlaut. Which I _might_ >>>> find time to do, or else, Dale, how do you feel about some ruby >>>> coding? Or else we could try to find someone else. >>>> >>>> CONCLUSION >>>> >>>> So, any thoughts Dale? >>>> >>>> _______________________________________________ >>>> Umlaut-general mailing list >>>> Umlaut-general at rubyforge.org >>>> http://rubyforge.org/mailman/listinfo/umlaut-general >>> >>> >> >> From rochkind at jhu.edu Thu Feb 16 10:47:58 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Thu, 16 Feb 2012 10:47:58 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <24B8A4BC-1371-4AA3-AD8E-964DF98604DA@gmail.com> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1E22.7060200@jhu.edu> <24B8A4BC-1371-4AA3-AD8E-964DF98604DA@gmail.com> Message-ID: <4F3D252E.3070309@jhu.edu> On 2/15/2012 6:24 PM, Ross Singer wrote: > I told Jonathan this in IRC, but I'll repeat it here for archival purposes. > > OpenURL 0.1 sid becomes rfr_id and pid becomes rft_dat. > > I think if we properly migrate these two key/vals (the rfr_id needs to be an info-uri, for example), this problem will clear up. I think you may be wrong. Like I said in my message you were responding to (did you read it?), I did try a manually constructured OpenURL in 1.0 style that used rfr_id and rft_dat instead, and it did not succesfully trigger the target link in SFX. It is possible I am not constructing the OpenURL 1.0 right, but I don't know what else to do. But I also, like I also said in the message you are replying to, seem to recall that SFX 'source parser' configuration (the thing that would get the ERIC id out of the "pid" for "sid" of CSA) is written specific to 0.1 and 1.0, so I'm suspecting that since CSA only delivers 0.1 URLs, the SFX "source parser" is specific to that. But the Option 1 I listed below is trying to figure out, probably with Ex Libris support, how to make an OpenURL 1.0 for this CSA supplied ERIC id that SFX will correctly recognize (configuring SFX to do so if neccesary), and making Umlaut supply that. I just do not know how to do the first part of that. But anyone is welcome to try and figure it out, with or without Ex Libris support. Jonathan From rochkind at jhu.edu Thu Feb 16 10:49:26 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Thu, 16 Feb 2012 10:49:26 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> Message-ID: <4F3D2586.4090109@jhu.edu> Aha! This is helpful to know, I got the rfr_id wrong. Okay, awesome, thanks a lot rsinger, sorry for the peevy previous message. (And THIS is why we discuss these issues in public instead of private email, everyone!) I'll see if I can fix up Umlaut to do the right thing there, try to find time today. Or if Dale or anyone else wants to take a stab at it, it _is_ open source! But the code around converting OpenURLs to internal model and back is sadly a bit convoluted, esp in Umlaut 2.x, I admit. On 2/15/2012 7:49 PM, Ross Singer wrote: > Jonathan, it works fine if you turn the rfr_id into an info-uri > > http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 > > so preprend info:sid/ onto the sid value unless it is already an > info-uri (as I recall, the OpenURL gem already does something similar > with rft_id around dois and pmids, so we should be able to replicate > that functionality here). > > -Ross. > > On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: >> Okay, yeah, this is kind of a mess. I hate OpenURL. Bare with me, Dale, to >> understand the problem and possible solutions (none entirely trivial, but >> some possible). >> >> The original OpenURL is an "OpenURL 0.1". The presence of two parameters is >> enough for SFX to trigger the ERIC link. "sid" (source id, where the >> OpenURL came from) and "pid" (private id, private non-standard data from >> that source). >> >> So for instance, this is the minimal 0.1 OpenURL which succesfully triggers >> the ERIC link from your link resolver: >> >> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> >> But Umlaut ends up sending an OpenURL 1.0 request to SFX. Umlaut actually >> takes all incoming OpenURLs and translates them into an internal data >> structure, doesn't even keep track of whether they came in as 1.0 or 0.1, >> and then translates them back into OpenURL 1.0 to make the API request to >> SFX. >> >> So the first problem is, Umlaut is doing some weird things with the "sid" >> and "pid" parameters in this round trip to internal data structure and back. >> (Those params _usually_ don't matter, is why I haven't noticed before). >> >> But even when I investigate fixing that, I can't figure out if there _is_ an >> OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. >> >> Here's the closest OpenURL 1.0 equivalent I can come up with (thanks >> rsinger) >> >> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> >> But that does NOT properly trigger the ERIC link in SFX. >> >> No completely easy fix presents itself. But we have a number of options. >> >> So, Dale, at this point, if you could, I think it would be helpful for you >> to try to get some info from Ex Libris. (A frustrating and time consuming >> thing to do, which is why I ask you to do it for your problem, heh). >> >> * Option 1. >> >> I think I remember that SFX gets that "pid" info out with custom source >> parsers based on "sid". And that these "source parsers" are specific to >> OpenURL 1.0 or 0.1 >> >> So you could ask Ex Libris if there's any way to set up a CSA source parser >> that will work for OpenURL 1.0 and do the same thing it's doing now for the >> 0.1 source parser. Adn what the 1.0 OpenURL should look like to be >> recognized by that source parser. >> >> They will say "Why would you want an OpenURL 1.0 source parser for CSA, CSA >> only sends OpenURL 0.1?" And it'll be confusing. >> >> And this would still require some changes to Umlaut, as I've noticed through >> this debugging that it doesn't handle sid/pid quite right anyway. Fixing >> 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. >> >> * Option 2. >> >> Alternately, we could try to hack Umlaut so it can send an OpenURL 0.1 to >> SFX instead. As a configuration option, always? That might mess other things >> up. Only if the original OpenURL was 0.1? Umlaut doesn't keep track of this >> right now, but theoretically it could. Only if the OpenURL has a "pid" in >> it? That might actually be best, since "pid" as a literal string is OpenURL >> 0.1-specific so is a signal that the incoming was OpenURL 0.1 -- plus it's >> really only incoming OpenURLs with 'pid' that need to be sent as OpenURL 0.1 >> to SFX. >> >> This would still require the same fixes to Umlaut's handling of sid/pid >> mentioned above. >> >> * Option 3. >> >> I might lean toward this one, actually. The reason we use SFX is for it's >> knowledge base. That is, for portfolio-based targets, as well as SFX's >> knowledge of outgoing link syntax. >> >> You helpfully (thanks!) mentioned that the ERIC target is not portfolio >> based. It's just an automatic translation "if ERIC id detected, make a link >> to ERIC like so." >> >> So we could write a new service plugin for Umlaut that handles ERIC id's and >> outgoing links to ERIC. Activate that in Umlaut, de-activate the ERIC target >> in SFX. >> >> This would require "us" (Umlaut) to take responsibility for maintaining the >> outgoing ERIC link syntax and changing our code if it changes. But it's not >> like Ex Libris is very good at that maintenance anyway, and I bet ERIC >> outgoing link syntax doesn't change much. >> >> It would also require "us" to identify all the places an ERIC id might come >> in in source-specific ways. This example is one of them, a "pid" formatted >> in a certain way, when "sid" is CSA. Are there other custom places SFX >> recognizes ERIC id's, from other places? We'd have to figure that out, and >> copy that code in Umlaut too (taking maintenance responsibility too). >> >> But we wouldn't have to deal with SFX's weirdness here, we'd take >> responsibility for handling this ourselves, which is part of the point of >> Umlaut, really --- use SFX for it's portfolio-based knowledge base, but do >> ourselves what we can do ourselves better. (Recognizing in this case it's >> not really 'better', it's just that SFX's weirdness combined with OpenURL's >> crappiness puts us in a corner). >> >> And someone would have to write the code in Umlaut. Which I _might_ find >> time to do, or else, Dale, how do you feel about some ruby coding? Or else >> we could try to find someone else. >> >> CONCLUSION >> >> So, any thoughts Dale? >> >> _______________________________________________ >> Umlaut-general mailing list >> Umlaut-general at rubyforge.org >> http://rubyforge.org/mailman/listinfo/umlaut-general > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Thu Feb 16 10:53:00 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Thu, 16 Feb 2012 10:53:00 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3D265C.7030500@jhu.edu> On 2/15/2012 9:12 PM, Ross Singer wrote: > Dale, is there any way you can see in your SFX logs what Umlaut sent to SFX? > > My guess is that it's dropping the rft_dat key (maybe rfr_id, too). Yes, it is dropping both, I confirmed that on my end, reproducing the scenario. > Actually, Dale, can you tell me what SFX actually links to with the > 0.1 link? Does it link directly to the article? Yes, it links direct to the article, we can try that ourselves with his SFX. > Did the 1.0 link that I made link to the article correctly (I can't > tell what happens on the other side of EZProxy)? It's easy to strip out the EZProxy yourself and find out. Let's verify, with an API query: http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004&sfx.response_type=multi_obj_xml => http://proxy.library.vanderbilt.edu/login?url=http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 => http://www.eric.ed.gov/ERICWebPortal/search/simpleSearch.jsp?ERICExtSearch_SearchType_0=kw&ERICExtSearch_SearchValue_0=ED492558 => Yep, it's a search results page with 1 hit, the article in question, which I think is the closest this target gets to 'article level' links, this is what is expected. So, yeah, thanks so much to rsinger to figuring out what OpenURL 1.0 will succesfully trigger SFX's ERIC behavior for this data -- now it's just up to me/us to get Umlaut to do that, correctly receiving the sid and pid in internal data format, and correctly outputting it in OpenURL 1.0. Jonathan > Thanks, > -Ross. > > On Wed, Feb 15, 2012 at 8:54 PM, Poulter, Dale > wrote: >> Ross and Jonathan, >> >> Thanks. I try this in Umlaut (http://umlaut.library.vanderbilt.edu/resolve?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004) and it did not seem to work. Do I understand correctly that we just need to adapt the metadata_helper to access the ED/EJ numbers similar to the pmid? I did notice that in SFX Ex Libris seems to have included most of the sids but not the eric specific sid. I'll report this as an issue to EL, but am happy to try some coding. >> >> >> >> -----Original Message----- >> From: Ross Singer [mailto:rossfsinger at gmail.com] >> Sent: Wednesday, February 15, 2012 6:49 PM >> To: umlaut-general at rubyforge.org >> Cc: Poulter, Dale >> Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 >> >> Jonathan, it works fine if you turn the rfr_id into an info-uri >> >> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >> >> so preprend info:sid/ onto the sid value unless it is already an info-uri (as I recall, the OpenURL gem already does something similar with rft_id around dois and pmids, so we should be able to replicate that functionality here). >> >> -Ross. >> >> On Wed, Feb 15, 2012 at 5:01 PM, Jonathan Rochkind wrote: >>> Okay, yeah, this is kind of a mess. I hate OpenURL. Bare with me, >>> Dale, to understand the problem and possible solutions (none entirely >>> trivial, but some possible). >>> >>> The original OpenURL is an "OpenURL 0.1". The presence of two >>> parameters is enough for SFX to trigger the ERIC link. "sid" (source >>> id, where the OpenURL came from) and "pid" (private id, private >>> non-standard data from that source). >>> >>> So for instance, this is the minimal 0.1 OpenURL which succesfully >>> triggers the ERIC link from your link resolver: >>> >>> http://sfx.library.vanderbilt.edu/sfxtst41?sid=CSA:eric-set-c&pid=%3CA >>> N%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20J >>> eng-yih%20Tim%3C%2FAU%3E&date=2004 >>> >>> >>> But Umlaut ends up sending an OpenURL 1.0 request to SFX. Umlaut >>> actually takes all incoming OpenURLs and translates them into an >>> internal data structure, doesn't even keep track of whether they came >>> in as 1.0 or 0.1, and then translates them back into OpenURL 1.0 to >>> make the API request to SFX. >>> >>> So the first problem is, Umlaut is doing some weird things with the "sid" >>> and "pid" parameters in this round trip to internal data structure and back. >>> (Those params _usually_ don't matter, is why I haven't noticed before). >>> >>> But even when I investigate fixing that, I can't figure out if there >>> _is_ an OpenURL 1.0 equivalent URL that will trigger the ERIC link in SFX. >>> >>> Here's the closest OpenURL 1.0 equivalent I can come up with (thanks >>> rsinger) >>> >>> http://sfx.library.vanderbilt.edu/sfxtst41?rfr_id=CSA:eric-set-c&rft_d >>> at=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu >>> %2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004 >>> >>> >>> But that does NOT properly trigger the ERIC link in SFX. >>> >>> No completely easy fix presents itself. But we have a number of options. >>> >>> So, Dale, at this point, if you could, I think it would be helpful for >>> you to try to get some info from Ex Libris. (A frustrating and time >>> consuming thing to do, which is why I ask you to do it for your problem, heh). >>> >>> * Option 1. >>> >>> I think I remember that SFX gets that "pid" info out with custom >>> source parsers based on "sid". And that these "source parsers" are >>> specific to OpenURL 1.0 or 0.1 >>> >>> So you could ask Ex Libris if there's any way to set up a CSA source >>> parser that will work for OpenURL 1.0 and do the same thing it's doing >>> now for the >>> 0.1 source parser. Adn what the 1.0 OpenURL should look like to be >>> recognized by that source parser. >>> >>> They will say "Why would you want an OpenURL 1.0 source parser for >>> CSA, CSA only sends OpenURL 0.1?" And it'll be confusing. >>> >>> And this would still require some changes to Umlaut, as I've noticed >>> through this debugging that it doesn't handle sid/pid quite right >>> anyway. Fixing 'pid' in Umlaut may or may not be a pain, I'll try to investigate tomorrow. >>> >>> * Option 2. >>> >>> Alternately, we could try to hack Umlaut so it can send an OpenURL 0.1 >>> to SFX instead. As a configuration option, always? That might mess >>> other things up. Only if the original OpenURL was 0.1? Umlaut doesn't >>> keep track of this right now, but theoretically it could. Only if the >>> OpenURL has a "pid" in it? That might actually be best, since "pid" >>> as a literal string is OpenURL 0.1-specific so is a signal that the >>> incoming was OpenURL 0.1 -- plus it's really only incoming OpenURLs >>> with 'pid' that need to be sent as OpenURL 0.1 to SFX. >>> >>> This would still require the same fixes to Umlaut's handling of >>> sid/pid mentioned above. >>> >>> * Option 3. >>> >>> I might lean toward this one, actually. The reason we use SFX is for >>> it's knowledge base. That is, for portfolio-based targets, as well as >>> SFX's knowledge of outgoing link syntax. >>> >>> You helpfully (thanks!) mentioned that the ERIC target is not >>> portfolio based. It's just an automatic translation "if ERIC id >>> detected, make a link to ERIC like so." >>> >>> So we could write a new service plugin for Umlaut that handles ERIC >>> id's and outgoing links to ERIC. Activate that in Umlaut, de-activate >>> the ERIC target in SFX. >>> >>> This would require "us" (Umlaut) to take responsibility for >>> maintaining the outgoing ERIC link syntax and changing our code if it >>> changes. But it's not like Ex Libris is very good at that maintenance >>> anyway, and I bet ERIC outgoing link syntax doesn't change much. >>> >>> It would also require "us" to identify all the places an ERIC id might >>> come in in source-specific ways. This example is one of them, a "pid" >>> formatted in a certain way, when "sid" is CSA. Are there other custom >>> places SFX recognizes ERIC id's, from other places? We'd have to >>> figure that out, and copy that code in Umlaut too (taking maintenance responsibility too). >>> >>> But we wouldn't have to deal with SFX's weirdness here, we'd take >>> responsibility for handling this ourselves, which is part of the point >>> of Umlaut, really --- use SFX for it's portfolio-based knowledge base, >>> but do ourselves what we can do ourselves better. (Recognizing in this >>> case it's not really 'better', it's just that SFX's weirdness combined >>> with OpenURL's crappiness puts us in a corner). >>> >>> And someone would have to write the code in Umlaut. Which I _might_ >>> find time to do, or else, Dale, how do you feel about some ruby >>> coding? Or else we could try to find someone else. >>> >>> CONCLUSION >>> >>> So, any thoughts Dale? >>> >>> _______________________________________________ >>> Umlaut-general mailing list >>> Umlaut-general at rubyforge.org >>> http://rubyforge.org/mailman/listinfo/umlaut-general >> > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Thu Feb 16 10:54:03 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Thu, 16 Feb 2012 10:54:03 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3D269B.7040306@jhu.edu> On 2/15/2012 8:54 PM, Poulter, Dale wrote: > Ross and Jonathan, > > Thanks. I try this in Umlaut (http://umlaut.library.vanderbilt.edu/resolve?rfr_id=info:sid/CSA:eric-set-c&rft_dat=%3CAN%3EED492558%3C%2FAN%3E%26%3CPY%3E2004%3C%2FPY%3E%26%3CAU%3EHsu%2C%20Jeng-yih%20Tim%3C%2FAU%3E&date=2004) and it did not seem to work. Do I understand correctly that we just need to adapt the metadata_helper to access the ED/EJ numbers similar to the pmid? I did notice that in SFX Ex Libris seems to have included most of the sids but not the eric specific sid. I'll report this as an issue to EL, but am happy to try some coding. No. Thanks to what Ross figured out, we know what we need is to fix Umlaut to correctly pass on the recieved 'sid' and 'pid' values to SFX as the OpenURL 1.0 properly-formatted rfr_id and rft_dat values. From rochkind at jhu.edu Thu Feb 16 10:57:00 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Thu, 16 Feb 2012 10:57:00 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> <7288D49C0448E840B752320350AFA94E0135043402@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F3D274C.4080603@jhu.edu> On 2/15/2012 10:22 PM, Ross Singer wrote: > Hmm, well, I just looked at rOpenURL, and it doesn't look like it's > coming from there (in fact, it looks like it should be properly > setting the rfr_id as an info-uri, too), so I'm not sure *where* this > bug is coming from. I've got some ideas. Umlaut needs to take the incoming OpenURL and convert it to an internal Umlaut data model in the rdbms; then it needs to take that data and convert back to an OpenURL. (That is to say, the Umlaut app doesn't just store a serialized kev or xml openurl in a giant blob, it actually takes the openurl apart into it's own representation, which I'm calling a 'data model' although it's not like it's particularly rigorously specified or anything). The problems are there. But good to know that rOpenURL (the same version I'm using hopefully?) will handle the rfr_id thing properly. rsinger, since you were just looking at this, do you know the proper rOpenURL call to set the sid/rfr_id? How about the pid/rfr_dat? rOpenURL is kind of under-documented, and I have trouble finding the right api sometimes. Jonathan From rossfsinger at gmail.com Thu Feb 16 11:12:13 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Thu, 16 Feb 2012 11:12:13 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F3D274C.4080603@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043401@its-hcwnem03.ds.Vanderbilt.edu> <7288D49C0448E840B752320350AFA94E0135043402@its-hcwnem03.ds.Vanderbilt.edu> <4F3D274C.4080603@jhu.edu> Message-ID: On Thu, Feb 16, 2012 at 10:57 AM, Jonathan Rochkind wrote: > On 2/15/2012 10:22 PM, Ross Singer wrote: >> >> Hmm, well, I just looked at rOpenURL, and it doesn't look like it's >> coming from there (in fact, it looks like it should be properly >> setting the rfr_id as an info-uri, too), so I'm not sure *where* this >> bug is coming from. > > > I've got some ideas. Umlaut needs to take the incoming OpenURL and convert > it to an internal Umlaut data model in the rdbms; then it needs to take that > data and convert back to an OpenURL. ?(That is to say, the Umlaut app > doesn't just store a serialized kev or xml openurl in a giant blob, it > actually takes the openurl apart into it's own representation, which I'm > calling a 'data model' although it's not like it's particularly rigorously > specified or anything). The problems are there. > > But good to know that rOpenURL (the same version I'm using hopefully?) will > handle the rfr_id thing properly. rsinger, since you were just looking at > this, do you know the proper rOpenURL call to set the sid/rfr_id? ?How about > the pid/rfr_dat? ?rOpenURL is kind of under-documented, and I have trouble > finding the right api sometimes. Same version -- branches/metadata_format_refactor Yeah - so assuming the incoming OpenURL is a KEV, the process goes something like this: ContextObject#import_kev -> ContextObject#import_hash When import_hash runs across a "sid" key, it calls: @referrer.set_identifier("info:sid/"+val.to_s) (ContextObjectEntity#set_identifier) When it runs across a "pid", it calls: AHA - the wrong thing! :) I was looking in the wrong place last night about this, I guess. It does, in fact, cause our problem by setting rft.dat. So what we need is for context_object.rb/import_hash to look like: # This is a 0.1 OpenURL. Your mileage may vary on how accurately # this maps. if key == 'id' if value.is_a?(Array) value.each do | id | @referent.add_identifier(id) end else @referent.add_identifier(val) end elsif key == 'sid' @referrer.set_identifier("info:sid/"+val.to_s) elsif key == 'pid' @referent.set_private_data(val.to_s) else @referent.set_metadata(key, val) end Let me see what I can do about fixing this. Sorry! -Ross. > > Jonathan > From rochkind at jhu.edu Thu Feb 16 12:42:00 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Thu, 16 Feb 2012 12:42:00 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F3D2586.4090109@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <4F3D2586.4090109@jhu.edu> Message-ID: <4F3D3FE8.9050907@jhu.edu> Okay, Dale, more on this. Me and Ross (mostly Ross) are trying to fix one problem/bug we noticed in Umlaut with the 'pid' or 'rft_dat' value that must be passed in the OpenURL to SFX. But as I continue to explore and test things, I see that fixing that alone STILL isn't enough to get SFX to recognize the ERIC id properly. For some reason the inclusion of the (perfectly legal and even legally required) "ctx_ver=Z39.88-2004" parameter in the OpenURL Umlaut sends to SFX _also_ somehow prevents SFX from recognize the custom-encoded provided-by-CSA ERIC id. So, basically.... ugh, it's a mess. I'll still keep hacking with it as I have time, but it's unclear if I will have time to actually figure out a fix for you soon. But we'll see. I think I still stand by the three options I mentioned in my previous post, one of which is someone (Dale?) filing a support ticket with EL asking them how to get SFX to recognize an OpenURL 1.0 formatted OpenURL with a CSA-provided ERIC id in it. Jonathan From rochkind at jhu.edu Thu Feb 16 17:31:53 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Thu, 16 Feb 2012 17:31:53 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F3D3FE8.9050907@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <4F3D2586.4090109@jhu.edu> <4F3D3FE8.9050907@jhu.edu> Message-ID: <4F3D83D9.9040800@jhu.edu> Okay, after rsinger fixed up the OpenURL library to handle pid/rft_dat/private_data better... I've made some progress at: a) Fixing up Umlaut to handle private_data better b) Putting a workaround into the SFX plugin so it will send an OpenURL to SFX of a sort that (Dale's test server) SFX will actually recognize the private_data (do need to do some hacks to not send a complete OpenURL 1.0 to SFX in certain cases; the presence of a ctx_ver field identifying as OpenURL 1.0 keeps SFX (or at least Dale's) from recognizing the CSA-provided eric id. sigh.) I think I'll have a solution for you eventually, the path is looking pretty clear. Committed to Umlaut 2.x, then merged to Umlaut 3.x. But not before early next week at the earliest. Really fighting with time here, don't have a lot of time for this, but spending it anyway. (And man, i can't wait until we're all on Umlaut 3.x, so I don't have to fight with svn and pre-refactor Umlaut 2.x and merging all changes into 3.x too. Dale, have you given 3.x a test drive yet?) Jonathan From dale.poulter at Vanderbilt.Edu Thu Feb 16 20:37:46 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Thu, 16 Feb 2012 19:37:46 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <4F3D83D9.9040800@jhu.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <4F3D2586.4090109@jhu.edu> <4F3D3FE8.9050907@jhu.edu> <4F3D83D9.9040800@jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E01350435E3@its-hcwnem03.ds.Vanderbilt.edu> Jonathan, I'll sync up and test. I have done some testing on 3 but encountered a few issues (unable to connect to SFX). I will find time in the next few days to continue testing. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Thursday, February 16, 2012 4:32 PM To: umlaut-general at rubyforge.org; Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Okay, after rsinger fixed up the OpenURL library to handle pid/rft_dat/private_data better... I've made some progress at: a) Fixing up Umlaut to handle private_data better b) Putting a workaround into the SFX plugin so it will send an OpenURL to SFX of a sort that (Dale's test server) SFX will actually recognize the private_data (do need to do some hacks to not send a complete OpenURL 1.0 to SFX in certain cases; the presence of a ctx_ver field identifying as OpenURL 1.0 keeps SFX (or at least Dale's) from recognizing the CSA-provided eric id. sigh.) I think I'll have a solution for you eventually, the path is looking pretty clear. Committed to Umlaut 2.x, then merged to Umlaut 3.x. But not before early next week at the earliest. Really fighting with time here, don't have a lot of time for this, but spending it anyway. (And man, i can't wait until we're all on Umlaut 3.x, so I don't have to fight with svn and pre-refactor Umlaut 2.x and merging all changes into 3.x too. Dale, have you given 3.x a test drive yet?) Jonathan From dale.poulter at Vanderbilt.Edu Sun Feb 19 10:47:15 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Sun, 19 Feb 2012 09:47:15 -0600 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E01350435E3@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <4F3D2586.4090109@jhu.edu> <4F3D3FE8.9050907@jhu.edu> <4F3D83D9.9040800@jhu.edu> <7288D49C0448E840B752320350AFA94E01350435E3@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <7288D49C0448E840B752320350AFA94E01350437E8@its-hcwnem03.ds.Vanderbilt.edu> Jonathan, I encountered an issue when attempting to update the v2 svn. However, after some additional testing and thought, it seems reasonable to push forward and migrate to v3. I had not realized that a method to migrate the permalinks had been posted. Are others already using v3 in production? In my current test instance lf v3 the system does not return any results although the SFX db connection does work (a-z list works). If interested the test url is at http://lib7.library.vanderbilt.edu:3000 . I have also opened an SI with EL to have them review the issues with Eric/CSA (SI 16384-325265 ). -----Original Message----- From: umlaut-general-bounces at rubyforge.org [mailto:umlaut-general-bounces at rubyforge.org] On Behalf Of Poulter, Dale Sent: Thursday, February 16, 2012 7:38 PM To: Jonathan Rochkind; umlaut-general at rubyforge.org Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Jonathan, I'll sync up and test. I have done some testing on 3 but encountered a few issues (unable to connect to SFX). I will find time in the next few days to continue testing. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Thursday, February 16, 2012 4:32 PM To: umlaut-general at rubyforge.org; Poulter, Dale Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 Okay, after rsinger fixed up the OpenURL library to handle pid/rft_dat/private_data better... I've made some progress at: a) Fixing up Umlaut to handle private_data better b) Putting a workaround into the SFX plugin so it will send an OpenURL to SFX of a sort that (Dale's test server) SFX will actually recognize the private_data (do need to do some hacks to not send a complete OpenURL 1.0 to SFX in certain cases; the presence of a ctx_ver field identifying as OpenURL 1.0 keeps SFX (or at least Dale's) from recognizing the CSA-provided eric id. sigh.) I think I'll have a solution for you eventually, the path is looking pretty clear. Committed to Umlaut 2.x, then merged to Umlaut 3.x. But not before early next week at the earliest. Really fighting with time here, don't have a lot of time for this, but spending it anyway. (And man, i can't wait until we're all on Umlaut 3.x, so I don't have to fight with svn and pre-refactor Umlaut 2.x and merging all changes into 3.x too. Dale, have you given 3.x a test drive yet?) Jonathan _______________________________________________ Umlaut-general mailing list Umlaut-general at rubyforge.org http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Mon Feb 20 10:20:30 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Mon, 20 Feb 2012 10:20:30 -0500 Subject: [Umlaut-general] Strange link issue in Umlaut v2 In-Reply-To: <7288D49C0448E840B752320350AFA94E01350437E8@its-hcwnem03.ds.Vanderbilt.edu> References: <7288D49C0448E840B752320350AFA94E012EC64140@its-hcwnem03.ds.Vanderbilt.edu> <4F0B1BC8.4070103@jhu.edu> <7288D49C0448E840B752320350AFA94E012EC641E6@its-hcwnem03.ds.Vanderbilt.edu> <4F0B3C2E.4000103@jhu.edu> <4F0B3CD9.7090809@jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34944@its-hcwnem03.ds.Vanderbilt.edu> <665DBC51D0250A47B4F9306CE71E5FB752F16B73@JHEMTEBEX1.win.ad.jhu.edu> <7288D49C0448E840B752320350AFA94E0134E34990@its-hcwnem03.ds.Vanderbilt.edu> <4F3C1AF9.4010408@jhu.edu> <7288D49C0448E840B752320350AFA94E013504337F@its-hcwnem03.ds.Vanderbilt.edu> <4F3C2B3E.4070906@jhu.edu> <4F3D2586.4090109@jhu.edu> <4F3D3FE8.9050907@jhu.edu> <4F3D83D9.9040800@jhu.edu> <7288D49C0448E840B752320350AFA94E01350435E3@its-hcwnem03.ds.Vanderbilt.edu> <7288D49C0448E840B752320350AFA94E01350437E8@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F4264BE.1010402@jhu.edu> Nobody's using v3 in production yet; I don't expect the app to change significantly before being marked 'stable', but it is currently marked 'alpha'. I won't mark it 'stable' until I personally have it in production, as well as hopefully some others test it. So I appreciate your testing. I am currently working on fixing and adding an SFX workaround for your ERIC CSA issue in Umlaut 2.x, which I'll then merge into 3.x too. On 2/19/2012 10:47 AM, Poulter, Dale wrote: > Jonathan, > > I encountered an issue when attempting to update the v2 svn. However, after some additional testing and thought, it seems reasonable to push forward and migrate to v3. I had not realized that a method to migrate the permalinks had been posted. Are others already using v3 in production? In my current test instance lf v3 the system does not return any results although the SFX db connection does work (a-z list works). If interested the test url is at http://lib7.library.vanderbilt.edu:3000 . > > I have also opened an SI with EL to have them review the issues with Eric/CSA (SI 16384-325265 ). > > > > > -----Original Message----- > From: umlaut-general-bounces at rubyforge.org [mailto:umlaut-general-bounces at rubyforge.org] On Behalf Of Poulter, Dale > Sent: Thursday, February 16, 2012 7:38 PM > To: Jonathan Rochkind; umlaut-general at rubyforge.org > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > Jonathan, > > I'll sync up and test. I have done some testing on 3 but encountered a few issues (unable to connect to SFX). I will find time in the next few days to continue testing. > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Thursday, February 16, 2012 4:32 PM > To: umlaut-general at rubyforge.org; Poulter, Dale > Subject: Re: [Umlaut-general] Strange link issue in Umlaut v2 > > Okay, after rsinger fixed up the OpenURL library to handle pid/rft_dat/private_data better... > > I've made some progress at: > > a) Fixing up Umlaut to handle private_data better > > b) Putting a workaround into the SFX plugin so it will send an OpenURL to SFX of a sort that (Dale's test server) SFX will actually recognize the private_data (do need to do some hacks to not send a complete OpenURL 1.0 to SFX in certain cases; the presence of a ctx_ver field identifying as OpenURL 1.0 keeps SFX (or at least Dale's) from recognizing the CSA-provided eric id. sigh.) > > > I think I'll have a solution for you eventually, the path is looking pretty clear. Committed to Umlaut 2.x, then merged to Umlaut 3.x. > > But not before early next week at the earliest. Really fighting with time here, don't have a lot of time for this, but spending it anyway. > (And man, i can't wait until we're all on Umlaut 3.x, so I don't have to fight with svn and pre-refactor Umlaut 2.x and merging all changes into 3.x too. Dale, have you given 3.x a test drive yet?) > > Jonathan > > > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general > > From rochkind at jhu.edu Mon Feb 20 14:15:03 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Mon, 20 Feb 2012 14:15:03 -0500 Subject: [Umlaut-general] CSA ERIC id and SFX, fix Message-ID: <4F429BB7.9010508@jhu.edu> Phew. OpenURL and SFX are a mess. But. Okay, umlaut 2.x now has a combo of fix and workaround to deal with your CSA-provided eric id, Dale. We had to fix a couple things in the OpenURL gem and in Umlaut itself to handle the "private_data"/pid/rft_dat attribute better, to actually track it properly. Then, we still had trouble getting SFX to recognize the ERIC id in the OpenURL (v 1.0) that SFX would send to it. I still suspect that there's something in SFX config that only recognizes CSA-provided ERIC id's in OpenURL's that are v 0.1 style (an older style of OpenURL), because, after all, that's what CSA sends. But with the Umlaut intermediary, that's not what Umlaut sends when forwarding on the CSA OpenURL. So I still think it might be profitable to follow up with Ex Libris to figure out if there's an easy way to make SFX recognize CSA-provided ERIC id's in an OpenURL 1.0 too. But in the meantime, I've put a hacky hack into the SFX plugin code, so IF it detects the OpenURL was from CSA _and_ has the "private data" component, then it intentionally messes up the OpenURL it sends to SFX just enough to get SFX to recognize it, as determined experimentally. (it needs to blank out the ctx_ver and url_ver keys). This is working in my test against your test SFX. However, it applies to CSA-provided ERIC id's only, if there's some other "private data" from other sources that SFX recognizes, it's possible it's still a problem. And it's possible it will break in the future if SFX changes how it does things. But if you update your Umlaut 2.x checkout from svn, you should have the fix. Because of a "referent caching" feature in Umlaut that I really mean to disable in 3.x, you MAY need to delete some or all rows from your 'referent' table in the database for the fix to take effect with citations you've already looked at in Umlaut. If that happens, I can provide further instructions. The fix has also been merged into the Umlaut 3.x source code. Phew. Hope this helps. Jonathan From dale.poulter at Vanderbilt.Edu Mon Feb 20 19:02:28 2012 From: dale.poulter at Vanderbilt.Edu (Poulter, Dale) Date: Mon, 20 Feb 2012 18:02:28 -0600 Subject: [Umlaut-general] CSA ERIC id and SFX, fix In-Reply-To: <4F429BB7.9010508@jhu.edu> References: <4F429BB7.9010508@jhu.edu> Message-ID: <7288D49C0448E840B752320350AFA94E01350439C8@its-hcwnem03.ds.Vanderbilt.edu> Jonathan, Thank you! I did a quick test after updating (svn update) but did not see any difference in the results. I'll take a look latter tonight or in the morning to determine what I did wrong. I will also let the list know what Ex Libris' response is to the questions. -----Original Message----- From: Jonathan Rochkind [mailto:rochkind at jhu.edu] Sent: Monday, February 20, 2012 1:15 PM To: umlaut-general at rubyforge.org; Poulter, Dale Subject: CSA ERIC id and SFX, fix Phew. OpenURL and SFX are a mess. But. Okay, umlaut 2.x now has a combo of fix and workaround to deal with your CSA-provided eric id, Dale. We had to fix a couple things in the OpenURL gem and in Umlaut itself to handle the "private_data"/pid/rft_dat attribute better, to actually track it properly. Then, we still had trouble getting SFX to recognize the ERIC id in the OpenURL (v 1.0) that SFX would send to it. I still suspect that there's something in SFX config that only recognizes CSA-provided ERIC id's in OpenURL's that are v 0.1 style (an older style of OpenURL), because, after all, that's what CSA sends. But with the Umlaut intermediary, that's not what Umlaut sends when forwarding on the CSA OpenURL. So I still think it might be profitable to follow up with Ex Libris to figure out if there's an easy way to make SFX recognize CSA-provided ERIC id's in an OpenURL 1.0 too. But in the meantime, I've put a hacky hack into the SFX plugin code, so IF it detects the OpenURL was from CSA _and_ has the "private data" component, then it intentionally messes up the OpenURL it sends to SFX just enough to get SFX to recognize it, as determined experimentally. (it needs to blank out the ctx_ver and url_ver keys). This is working in my test against your test SFX. However, it applies to CSA-provided ERIC id's only, if there's some other "private data" from other sources that SFX recognizes, it's possible it's still a problem. And it's possible it will break in the future if SFX changes how it does things. But if you update your Umlaut 2.x checkout from svn, you should have the fix. Because of a "referent caching" feature in Umlaut that I really mean to disable in 3.x, you MAY need to delete some or all rows from your 'referent' table in the database for the fix to take effect with citations you've already looked at in Umlaut. If that happens, I can provide further instructions. The fix has also been merged into the Umlaut 3.x source code. Phew. Hope this helps. Jonathan From rossfsinger at gmail.com Tue Feb 21 08:26:16 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Tue, 21 Feb 2012 08:26:16 +0000 Subject: [Umlaut-general] CSA ERIC id and SFX, fix In-Reply-To: <7288D49C0448E840B752320350AFA94E01350439C8@its-hcwnem03.ds.Vanderbilt.edu> References: <4F429BB7.9010508@jhu.edu> <7288D49C0448E840B752320350AFA94E01350439C8@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <960909F7-2A96-4541-A50B-5DA9EE21F8DB@gmail.com> Dale, Did you try a new article or did you use the same OpenURL from earlier? If the latter, I would definitely suspect that the referent is being cached, which is preventing the fix from showing up (delete the context object from the referents and referent_values tables in Umlaut's MySQL). Otherwise, well, hmm. -Ross. On Feb 21, 2012, at 12:02 AM, Poulter, Dale wrote: > Jonathan, > > Thank you! I did a quick test after updating (svn update) but did not see any difference in the results. I'll take a look latter tonight or in the morning to determine what I did wrong. I will also let the list know what Ex Libris' response is to the questions. > > > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Monday, February 20, 2012 1:15 PM > To: umlaut-general at rubyforge.org; Poulter, Dale > Subject: CSA ERIC id and SFX, fix > > Phew. OpenURL and SFX are a mess. But. Okay, umlaut 2.x now has a combo of fix and workaround to deal with your CSA-provided eric id, Dale. > > We had to fix a couple things in the OpenURL gem and in Umlaut itself to handle the "private_data"/pid/rft_dat attribute better, to actually track it properly. > > Then, we still had trouble getting SFX to recognize the ERIC id in the OpenURL (v 1.0) that SFX would send to it. I still suspect that there's something in SFX config that only recognizes CSA-provided ERIC id's in OpenURL's that are v 0.1 style (an older style of OpenURL), because, after all, that's what CSA sends. But with the Umlaut intermediary, that's not what Umlaut sends when forwarding on the CSA OpenURL. > > So I still think it might be profitable to follow up with Ex Libris to figure out if there's an easy way to make SFX recognize CSA-provided ERIC id's in an OpenURL 1.0 too. > > But in the meantime, I've put a hacky hack into the SFX plugin code, so IF it detects the OpenURL was from CSA _and_ has the "private data" > component, then it intentionally messes up the OpenURL it sends to SFX just enough to get SFX to recognize it, as determined experimentally. > (it needs to blank out the ctx_ver and url_ver keys). > > This is working in my test against your test SFX. However, it applies to CSA-provided ERIC id's only, if there's some other "private data" > from other sources that SFX recognizes, it's possible it's still a problem. And it's possible it will break in the future if SFX changes how it does things. > > But if you update your Umlaut 2.x checkout from svn, you should have the > fix. Because of a "referent caching" feature in Umlaut that I really > mean to disable in 3.x, you MAY need to delete some or all rows from your 'referent' table in the database for the fix to take effect with citations you've already looked at in Umlaut. If that happens, I can provide further instructions. > > The fix has also been merged into the Umlaut 3.x source code. > > Phew. Hope this helps. > > Jonathan > > > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Tue Feb 21 16:36:35 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 21 Feb 2012 11:36:35 -0500 Subject: [Umlaut-general] CSA ERIC id and SFX, fix In-Reply-To: <7288D49C0448E840B752320350AFA94E0135043A84@its-hcwnem03.ds.Vanderbilt.edu> References: <4F429BB7.9010508@jhu.edu> <7288D49C0448E840B752320350AFA94E01350439C8@its-hcwnem03.ds.Vanderbilt.edu> <4F43BB41.2030909@jhu.edu> <7288D49C0448E840B752320350AFA94E0135043A84@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F43C813.6000301@jhu.edu> Try it in a test environment first, but I believe it _ought_ to actually work without ruining permalinks. But ha! I just realized a way to find just the 'offending' referents and delete those. First I'll describe that, then I'll provide some background info on why you probably _could_ delete all referents if you wanted to (since I wrote it before I realized there was an easy way to delete just the 'bad' ones). 1. To find and delete bad ones Look in 'referent_values' database for rows where 'key_name' is 'pid'. There is an index on the key_name column, so this should be pretty quick. Those referent_values rows have a referent_id foreign key into the referents table. Delete the row from the referents table with that id, as well as all the referent_values rows sharing that same parent referent_id. [Permalinks involving those referents should still work, see below]. This may actually get _more_ than just CSA/ERIC referents, but the other ones it gets will also be 'corrupt', need to be re-done with the new umlaut code I committed. 2. Why you could probably delete all of em While the permalinks table does have a foreign key to 'referents' and will use it to look up citation details; it ALSO stores the complete serialized OpenURL as a 'backup', when the referents foreign key can not be found. This was not originally in Umlaut, but I think I added it before Vanderbilt brought up Umlaut, so all your permalinks records should have this 'backup' citation information, and work even with deleted referents. In fact, I believe the current Umlaut 'nightly maintenance' script deletes old referents (but not old permalinks), and the old permalinks keep working even without a valid 'referent', due to the serialized OpenURL 'backup'. I added the serialized OpenURL 'backup' at a point I realized I needed to have Umlaut delete old referents for performance (the table was just too big), and also as part of a gradual move away from the 'cached referents' architecture, as discussed previously. It is _possible_ that if you delete all referents, certain requests _in the middle of being processed_ will have a problem -- but I _think_ even that ought not to happen (or should only happen in very rare cases where things happen _just_ at the wrong time). All Umlaut code at the moment is intended to re-create the Referent for a Request, if the Referent does not exist in the database (and the Request obj in the database contains the incoming OpenURL data to make this possible). That's the idea anyway. No warrantee is provided, test first to your satisfaction. Jonathan On 2/21/2012 11:25 AM, Poulter, Dale wrote: > Taking a look at this, it appears that if I delete the contents of the referents table, I also loose the lookups for the permalinks. > > -----Original Message----- > From: Jonathan Rochkind [mailto:rochkind at jhu.edu] > Sent: Tuesday, February 21, 2012 9:42 AM > To: Poulter, Dale > Cc: umlaut-general at rubyforge.org > Subject: Re: CSA ERIC id and SFX, fix > > On 2/20/2012 7:02 PM, Poulter, Dale wrote: >> Jonathan, >> >> Thank you! I did a quick test after updating (svn update) but did not see any difference in the results. I'll take a look latter tonight or in the morning to determine what I did wrong. I will also let the list know what Ex Libris' response is to the questions. > Okay, due to Umlaut's caching of "Referent" data, you will have to purge rows from your 'referent' table, or purge the entire table. > > I _think_ you can delete the entire contents of the 'referent' table without ill effects -- everything will be created again as/if needed on demand. `Referent.delete_all` from a rails console, or direct sql. > Alternately, you'd have to figure out the id's of the particular citations in the 'referent' table that involve CSA/ERIC data, and delete them individually. This would be a bit tricky, sorry. > > I've been considering eliminating this 'cached Referents' feature for a while, and planned to eliminate it in Umlaut 3.x although I haven't actually done that yet. But this example is yet more motivation. The cached referent feature was originally intended by Ross to make it possible to tell when multiple incoming OpenURL's all refer to the same 'thing', to allow statistical analysis as well as tagging or recommendation systems. These advanced feature were never really implemented, there would be some non-trivial barriers to implementing them effectively even in the existing setup, and meanwhile the existing setup often causes problems when 'bad' data ends up effectively cached, such as in this case. So. Yeah. > > > From rochkind at jhu.edu Tue Feb 21 15:41:53 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 21 Feb 2012 10:41:53 -0500 Subject: [Umlaut-general] CSA ERIC id and SFX, fix In-Reply-To: <7288D49C0448E840B752320350AFA94E01350439C8@its-hcwnem03.ds.Vanderbilt.edu> References: <4F429BB7.9010508@jhu.edu> <7288D49C0448E840B752320350AFA94E01350439C8@its-hcwnem03.ds.Vanderbilt.edu> Message-ID: <4F43BB41.2030909@jhu.edu> On 2/20/2012 7:02 PM, Poulter, Dale wrote: > Jonathan, > > Thank you! I did a quick test after updating (svn update) but did not see any difference in the results. I'll take a look latter tonight or in the morning to determine what I did wrong. I will also let the list know what Ex Libris' response is to the questions. Okay, due to Umlaut's caching of "Referent" data, you will have to purge rows from your 'referent' table, or purge the entire table. I _think_ you can delete the entire contents of the 'referent' table without ill effects -- everything will be created again as/if needed on demand. `Referent.delete_all` from a rails console, or direct sql. Alternately, you'd have to figure out the id's of the particular citations in the 'referent' table that involve CSA/ERIC data, and delete them individually. This would be a bit tricky, sorry. I've been considering eliminating this 'cached Referents' feature for a while, and planned to eliminate it in Umlaut 3.x although I haven't actually done that yet. But this example is yet more motivation. The cached referent feature was originally intended by Ross to make it possible to tell when multiple incoming OpenURL's all refer to the same 'thing', to allow statistical analysis as well as tagging or recommendation systems. These advanced feature were never really implemented, there would be some non-trivial barriers to implementing them effectively even in the existing setup, and meanwhile the existing setup often causes problems when 'bad' data ends up effectively cached, such as in this case. So. Yeah. From rochkind at jhu.edu Wed Feb 22 15:28:19 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 22 Feb 2012 10:28:19 -0500 Subject: [Umlaut-general] umlaut 3.x tests Message-ID: <4F450993.6030908@jhu.edu> Okay, I got the tests to RUN again. But most of em fail. Include all of the umlaut/primo tests you added, Scot. It looks like you've set up the test/dummy application to have aleph and primo services pointing at YOUR aleph and primo, yes? Does a random person like me have access to those when running Umlaut tests, is it even possible for me to run em and have em pass? If yes, cool, if not... we need to figure something else out. Tests that nobody can run and have pass but someone at NYU are no good to include with Umlaut. Alternately, maybe the problem is just something else? SOME of the failing lists have errors like this: No such file or directory - /home/rochkind/umlaut/test/dummy/config/primo.yml Did you forget to include a neccesary file in the test dummy app? (Some tests that I think I probably added not you are failing too for other reasons, the whole test situation is a mess, I realize). Scot, can you figure out why the aleph and primo tests are failing, and either figure out a way to fix em, or punt on the whole thing and mark em so they aren't run for now? (I normally end up punting on the whole thing, getting frustrated after spending too many hours I don't have fighting with the test setup). From rochkind at jhu.edu Wed Feb 22 16:11:18 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 22 Feb 2012 11:11:18 -0500 Subject: [Umlaut-general] umlaut 3.x tests In-Reply-To: <92EF661D-73C4-4A7A-9E43-9082CFE2C57D@nyu.edu> References: <4F450993.6030908@jhu.edu> <92EF661D-73C4-4A7A-9E43-9082CFE2C57D@nyu.edu> Message-ID: <4F4513A6.9030709@jhu.edu> On 2/22/2012 11:08 AM, Scot Dalton wrote: > Yeah, they'll all fail outside of NYU, even if the primo config file is present. I need to remove them, but want to keep them around locally, so I'll need to figure that out. What's the best way to mark them to not run? I have no idea, honestly! This stuff is really frustrating me these days. > Thanks, > Scot > > On Feb 22, 2012, at Feb 22, 10:28 AM, Jonathan Rochkind wrote: > >> Okay, I got the tests to RUN again. >> >> But most of em fail. Include all of the umlaut/primo tests you added, Scot. >> >> It looks like you've set up the test/dummy application to have aleph and primo services pointing at YOUR aleph and primo, yes? Does a random person like me have access to those when running Umlaut tests, is it even possible for me to run em and have em pass? If yes, cool, if not... we need to figure something else out. Tests that nobody can run and have pass but someone at NYU are no good to include with Umlaut. >> >> Alternately, maybe the problem is just something else? SOME of the failing lists have errors like this: >> >> No such file or directory - /home/rochkind/umlaut/test/dummy/config/primo.yml >> >> Did you forget to include a neccesary file in the test dummy app? >> >> (Some tests that I think I probably added not you are failing too for other reasons, the whole test situation is a mess, I realize). >> >> Scot, can you figure out why the aleph and primo tests are failing, and either figure out a way to fix em, or punt on the whole thing and mark em so they aren't run for now? (I normally end up punting on the whole thing, getting frustrated after spending too many hours I don't have fighting with the test setup). > > -- > Scot Dalton > Phone: (212) 998-2674 > Web Services > Division of Libraries > New York University > From rochkind at jhu.edu Wed Feb 22 15:13:58 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 22 Feb 2012 10:13:58 -0500 Subject: [Umlaut-general] umlaut 3.x, automated tests Message-ID: <4F450636.30004@jhu.edu> bah, the umlaut 3.x automated tests don't run anymore. not sure why. Scot, you added some tests, yes? Weird stack trace when i try to run em. don't know if something has changed in my ruby/rails environment (stopped working under rails 3.2 which I upgraded to), or something changed in our test files, or what. sometimes I hate rails. From std5 at nyu.edu Wed Feb 22 16:08:55 2012 From: std5 at nyu.edu (Scot Dalton) Date: Wed, 22 Feb 2012 11:08:55 -0500 Subject: [Umlaut-general] umlaut 3.x tests In-Reply-To: <4F450993.6030908@jhu.edu> References: <4F450993.6030908@jhu.edu> Message-ID: <92EF661D-73C4-4A7A-9E43-9082CFE2C57D@nyu.edu> Yeah, they'll all fail outside of NYU, even if the primo config file is present. I need to remove them, but want to keep them around locally, so I'll need to figure that out. What's the best way to mark them to not run? Thanks, Scot On Feb 22, 2012, at Feb 22, 10:28 AM, Jonathan Rochkind wrote: > Okay, I got the tests to RUN again. > > But most of em fail. Include all of the umlaut/primo tests you added, Scot. > > It looks like you've set up the test/dummy application to have aleph and primo services pointing at YOUR aleph and primo, yes? Does a random person like me have access to those when running Umlaut tests, is it even possible for me to run em and have em pass? If yes, cool, if not... we need to figure something else out. Tests that nobody can run and have pass but someone at NYU are no good to include with Umlaut. > > Alternately, maybe the problem is just something else? SOME of the failing lists have errors like this: > > No such file or directory - /home/rochkind/umlaut/test/dummy/config/primo.yml > > Did you forget to include a neccesary file in the test dummy app? > > (Some tests that I think I probably added not you are failing too for other reasons, the whole test situation is a mess, I realize). > > Scot, can you figure out why the aleph and primo tests are failing, and either figure out a way to fix em, or punt on the whole thing and mark em so they aren't run for now? (I normally end up punting on the whole thing, getting frustrated after spending too many hours I don't have fighting with the test setup). -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From std5 at nyu.edu Wed Feb 22 16:49:40 2012 From: std5 at nyu.edu (Scot Dalton) Date: Wed, 22 Feb 2012 11:49:40 -0500 Subject: [Umlaut-general] umlaut 3.x, automated tests In-Reply-To: <4F450636.30004@jhu.edu> References: <4F450636.30004@jhu.edu> Message-ID: <8CA20FBE-5127-489A-A734-B73A7BB7595F@nyu.edu> I did add some test and I just pushed a change to skip all of my tests. Tests seem to be running on my local machine with rails 3.2.1. On Feb 22, 2012, at Feb 22, 10:13 AM, Jonathan Rochkind wrote: > bah, the umlaut 3.x automated tests don't run anymore. not sure why. Scot, you added some tests, yes? > > Weird stack trace when i try to run em. don't know if something has changed in my ruby/rails environment (stopped working under rails 3.2 which I upgraded to), or something changed in our test files, or what. > > sometimes I hate rails. > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From rochkind at jhu.edu Mon Feb 27 17:00:03 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Mon, 27 Feb 2012 12:00:03 -0500 Subject: [Umlaut-general] umlaut 3.x performance Message-ID: <4F4BB693.80005@jhu.edu> So one thing I'm still a bit worried about is performance/response-time in Umlaut 3.x. Now I was never totally happy with response times in Umlaut 2.x either. Although it was 'good enough', we've been using it here for years and find the response time tolerable. But Umlaut 3.x _seems_ to be even slower to me. Which I do not like. But I've have been having trouble figuring out for sure if that's true, and why. Yeah, I've been messing with ruby_prof, and a buncha other things, but still having trouble figuring out what's going on why. Haven't given up, still trying to figure it out. (In process, I discovered that the limited automated tests I hackily had working for Umlaut 2.x are broken with 3.x's rearchitecture, now trying to get them working again, sigh.) So mostly at this point curious, for anyone that's yet played with Umlaut 3.x, especially if you are a current user of Umlaut 2.x, if you've seen any issues with performance. Or if you haven't yet played with Umlaut 3.x, but plan to, let me know if you see any issues with performance, what you think about it. From std5 at nyu.edu Mon Feb 27 19:56:32 2012 From: std5 at nyu.edu (Scot Dalton) Date: Mon, 27 Feb 2012 14:56:32 -0500 Subject: [Umlaut-general] umlaut 3.x performance In-Reply-To: <4F4BB693.80005@jhu.edu> References: <4F4BB693.80005@jhu.edu> Message-ID: <4B5E141A-EB67-412D-B97D-C9FF516BB93C@nyu.edu> That's interesting. I actually find Umlaut 3.x noticeably faster than Umlaut 2.x. I have no basis for this observation, other than interacting with the two systems. We have them running on the same virtual machine w/ different rubies and passenger running in standalone mode for Umlaut 3.x proxied by Apache, so it's a bit of apples and oranges. Once I find the time, I'll do some benchmarking and see if my observation is true. Thanks, Scot On Feb 27, 2012, at Feb 27, 12:00 PM, Jonathan Rochkind wrote: > So one thing I'm still a bit worried about is performance/response-time in Umlaut 3.x. > > Now I was never totally happy with response times in Umlaut 2.x either. Although it was 'good enough', we've been using it here for years and find the response time tolerable. > > But Umlaut 3.x _seems_ to be even slower to me. Which I do not like. But I've have been having trouble figuring out for sure if that's true, and why. Yeah, I've been messing with ruby_prof, and a buncha other things, but still having trouble figuring out what's going on why. Haven't given up, still trying to figure it out. (In process, I discovered that the limited automated tests I hackily had working for Umlaut 2.x are broken with 3.x's rearchitecture, now trying to get them working again, sigh.) > > So mostly at this point curious, for anyone that's yet played with Umlaut 3.x, especially if you are a current user of Umlaut 2.x, if you've seen any issues with performance. Or if you haven't yet played with Umlaut 3.x, but plan to, let me know if you see any issues with performance, what you think about it. > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general -- Scot Dalton Phone: (212) 998-2674 Web Services Division of Libraries New York University From rochkind at jhu.edu Mon Feb 27 20:03:03 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Mon, 27 Feb 2012 15:03:03 -0500 Subject: [Umlaut-general] umlaut 3.x performance In-Reply-To: <4B5E141A-EB67-412D-B97D-C9FF516BB93C@nyu.edu> References: <4F4BB693.80005@jhu.edu> <4B5E141A-EB67-412D-B97D-C9FF516BB93C@nyu.edu> Message-ID: <4F4BE177.8000707@jhu.edu> Awesome, that's actually good to hear, just that your experience doesn't match mine. Mine could be an idiosyncracy of my environment, or could be a false anecdotal impression. My impression is based on nothing other than interacting with the two systems too, at the moment. Currently trying to whip the existing tests in shape, so theoretically I can add some benchmarked or ruby_prof'd tests too. Although I run into various annoying roadblocks trying either, these days I fight with rails a lot. but one step at a time. Scot, as an aside, I changed the way your nyu-only tests are being suppressed. Took the "skip" of them because: skip isn't supported in test::unit 1.8.7, and I still want to be able to run under 1.8.7 at moment; "skip" resulted in dozens of "skip" messages, making it hard to see rest of test results; no good way for you at NYU to turn em all on with "skip". Now they're all wrapped in a nyu_only_tests method I define in test_helper, which will simply avoid defining all the test method unless env variable NYU_TEST is set. Seems to work in my testing. But feel free to tweak that yet further if there's an even cleaner way to do it you can figure out. Sometimes I really feel like I'm fighting with things, for sure. Jonathan On 2/27/2012 2:56 PM, Scot Dalton wrote: > That's interesting. I actually find Umlaut 3.x noticeably faster than Umlaut 2.x. I have no basis for this observation, other than interacting with the two systems. We have them running on the same virtual machine w/ different rubies and passenger running in standalone mode for Umlaut 3.x proxied by Apache, so it's a bit of apples and oranges. Once I find the time, I'll do some benchmarking and see if my observation is true. > > Thanks, > Scot > > On Feb 27, 2012, at Feb 27, 12:00 PM, Jonathan Rochkind wrote: > >> So one thing I'm still a bit worried about is performance/response-time in Umlaut 3.x. >> >> Now I was never totally happy with response times in Umlaut 2.x either. Although it was 'good enough', we've been using it here for years and find the response time tolerable. >> >> But Umlaut 3.x _seems_ to be even slower to me. Which I do not like. But I've have been having trouble figuring out for sure if that's true, and why. Yeah, I've been messing with ruby_prof, and a buncha other things, but still having trouble figuring out what's going on why. Haven't given up, still trying to figure it out. (In process, I discovered that the limited automated tests I hackily had working for Umlaut 2.x are broken with 3.x's rearchitecture, now trying to get them working again, sigh.) >> >> So mostly at this point curious, for anyone that's yet played with Umlaut 3.x, especially if you are a current user of Umlaut 2.x, if you've seen any issues with performance. Or if you haven't yet played with Umlaut 3.x, but plan to, let me know if you see any issues with performance, what you think about it. >> _______________________________________________ >> Umlaut-general mailing list >> Umlaut-general at rubyforge.org >> http://rubyforge.org/mailman/listinfo/umlaut-general > > -- > Scot Dalton > Phone: (212) 998-2674 > Web Services > Division of Libraries > New York University > From rochkind at jhu.edu Mon Feb 27 22:21:55 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Mon, 27 Feb 2012 17:21:55 -0500 Subject: [Umlaut-general] services and sessions Message-ID: <4F4C0203.7080302@jhu.edu> Umlaut 2.x had a way for services to actually read and write values to the user's session. This is a neat trick, because services run outside of the context of a 'request-response loop', in background threads. Umlaut 2.x used some crazy hacks to do this, looking at some internal API of Rails, to make it work -- but of course just with ActiveRecord-backed sessions (which would be available to the server in between request-response loops, for reading as well as writing), not with cookie-backed session store, which is actually rails default these days. In Umlaut 3.x, none of that code will work. I noticed only when trying to whip our limited automated tests into shape, when one of them was failing -- for good reason, it just doesn't work. I'm not sure it's really tenable to support this functionality, it required crazy hacks in rails 2.x, that I can't figure out how to duplicate in rails 3.x. And it's unfortunate to require ActiveRecord-backed session in a rails 3 world, when everyone wants to use cookie (when this umlaut functionality was first written, the cookie-backed session store didn't even exist yet, I think!). However, searching through the existing umlaut service adapters... as far as I can tell none of them DO use this functionality, none try to read or save anything to the user's session. (They CAN read and save things hung off the current request object, representing a particular openurl request, which of course they do with abandon. Just not things scoped to the entire user session, rather than just the current request). So this functionality is dissapearing from Umlaut 3.x for the moment. However, I suspect that Scot's in-progress stuff to support login, that hasn't made it into Umlaut 3.x yet, relied upon this -- a service needed to know if there's a logged in user, and who. If/when Scot gets to trying to implement this in Umlaut 3.x, I have some ideas for workarounds/alternate implementations that are more robust, and won't actually try to interact with the rails session store from a service (thus still allowing cookie-based rails session store), especially if we just need services to _read_ but not write the session. 1. Store the neccesary session-based information with _each Umlaut request_ when it's created. As the request is available to the service. 2. Pass the neccesary session-based info (or even the entire session hash) into the Collection object, which actually executes services. The Collection object is created in a request-response loop, so has access to the session. 3. If we really needed to, we could implement our own alternate activerecord-based session, parallel to rails, just for certain values that services need to read/write. This wouldn't actually be that bad, it's just a table in our db keyed on session_id. Umlaut _already does_ pay attention to the rails session_id and store some things in the db based off of it -- mainly, to ensure that someone making another request for the same openurl gets the same umlaut Request object, but _only_ if they're coming from the same session. [Anything auth related in Umlaut would have to ensure to delete the rails session when logging out, through documented rails means. This is considered best practice for rails session auth anyway, I think. And would ensure that after a user logged out, her browser session would no longer connect with cached requests that may have info based on the no-longer-valid auth.] Phew. So that's my update. From rochkind at jhu.edu Tue Feb 28 16:30:35 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Tue, 28 Feb 2012 11:30:35 -0500 Subject: [Umlaut-general] umlaut 3.x performance In-Reply-To: <4F4BE177.8000707@jhu.edu> References: <4F4BB693.80005@jhu.edu> <4B5E141A-EB67-412D-B97D-C9FF516BB93C@nyu.edu> <4F4BE177.8000707@jhu.edu> Message-ID: <4F4D012B.9070508@jhu.edu> Yeah, on further playing with it, Umlaut 3.x performance is looking pretty decent, and likely better than Umlaut 0.2x performance. (Still haven't done official rigorous benchmarks, sorry, just a lot of messing around). It looks to me that on an Umlaut configured with only SFX, Umlaut adds on about 0.4-0.7 seconds of extra time to SFX. Much of this (around 0.2-0.4) is actually parsing and dealing with the SFX response (because of the way SFX does things, a double-XML parse is neccesary, as well as some complex xpaths, and some HTML entity substitution. Around 0.2-0.3s is extra Umlaut 'overhead'. This seems, well, not as great as one might like, but tolerable, and I think better than Umlaut 2.x. This is running Rails 3.2 (now required by latest Umlaut's), and ruby 1.9.3, but actually seems about the same in 1.9.2 or even 1.8.7. And if you have other things configured in same foreground service wave with SFX, but they are not as slow as SFX -- indeed, the Umlaut response time seems to be about the same, the multi-threaded service dispatch model is doing what it's supposed to do. I think. (Again, not based on rigorous profiling, having trouble getting that to happen in a way I can make sense of, sorry. Just based on a bunch of logging and messing with things. But this is my impression at the time. Results may vary in the Washington State (just kidding)). Jonathan On 2/27/2012 3:03 PM, Jonathan Rochkind wrote: > Awesome, that's actually good to hear, just that your experience > doesn't match mine. Mine could be an idiosyncracy of my environment, > or could be a false anecdotal impression. My impression is based on > nothing other than interacting with the two systems too, at the moment. > > Currently trying to whip the existing tests in shape, so theoretically > I can add some benchmarked or ruby_prof'd tests too. Although I run > into various annoying roadblocks trying either, these days I fight > with rails a lot. but one step at a time. > > Scot, as an aside, I changed the way your nyu-only tests are being > suppressed. Took the "skip" of them because: skip isn't supported in > test::unit 1.8.7, and I still want to be able to run under 1.8.7 at > moment; "skip" resulted in dozens of "skip" messages, making it hard > to see rest of test results; no good way for you at NYU to turn em all > on with "skip". > > Now they're all wrapped in a nyu_only_tests method I define in > test_helper, which will simply avoid defining all the test method > unless env variable NYU_TEST is set. Seems to work in my testing. But > feel free to tweak that yet further if there's an even cleaner way to > do it you can figure out. Sometimes I really feel like I'm fighting > with things, for sure. > > Jonathan > > On 2/27/2012 2:56 PM, Scot Dalton wrote: >> That's interesting. I actually find Umlaut 3.x noticeably faster >> than Umlaut 2.x. I have no basis for this observation, other than >> interacting with the two systems. We have them running on the same >> virtual machine w/ different rubies and passenger running in >> standalone mode for Umlaut 3.x proxied by Apache, so it's a bit of >> apples and oranges. Once I find the time, I'll do some benchmarking >> and see if my observation is true. >> >> Thanks, >> Scot >> >> On Feb 27, 2012, at Feb 27, 12:00 PM, Jonathan Rochkind wrote: >> >>> So one thing I'm still a bit worried about is >>> performance/response-time in Umlaut 3.x. >>> >>> Now I was never totally happy with response times in Umlaut 2.x >>> either. Although it was 'good enough', we've been using it here for >>> years and find the response time tolerable. >>> >>> But Umlaut 3.x _seems_ to be even slower to me. Which I do not like. >>> But I've have been having trouble figuring out for sure if that's >>> true, and why. Yeah, I've been messing with ruby_prof, and a buncha >>> other things, but still having trouble figuring out what's going on >>> why. Haven't given up, still trying to figure it out. (In process, >>> I discovered that the limited automated tests I hackily had working >>> for Umlaut 2.x are broken with 3.x's rearchitecture, now trying to >>> get them working again, sigh.) >>> >>> So mostly at this point curious, for anyone that's yet played with >>> Umlaut 3.x, especially if you are a current user of Umlaut 2.x, if >>> you've seen any issues with performance. Or if you haven't yet >>> played with Umlaut 3.x, but plan to, let me know if you see any >>> issues with performance, what you think about it. >>> _______________________________________________ >>> Umlaut-general mailing list >>> Umlaut-general at rubyforge.org >>> http://rubyforge.org/mailman/listinfo/umlaut-general >> >> -- >> Scot Dalton >> Phone: (212) 998-2674 >> Web Services >> Division of Libraries >> New York University >> > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Wed Feb 29 14:53:35 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 29 Feb 2012 09:53:35 -0500 Subject: [Umlaut-general] rubyforge svn? Message-ID: <4F4E3BEF.2040800@jhu.edu> The rubyforge svn where Umlaut 2.x is hosted is currently unresponsive to me. I get "svn: Network connection closed unexpectedly" Curious if this is just me or if anyone can confirm this is them too. (I filed a rubyforge ticket, haven't heard back yet). I have a minor tweak I wanted to commit to Umlaut 2.x (restricting size of display of oddly shaped oddly sized cover images fetched from cover image providers), but I can't cause I cant' get the svn. Clearly, we can't move soon enough to Umlaut 3.x with it's github host! (As the Umlaut 2.x, pre-Umlaut-as-a-gem, installation instructions end up being kind of svn-specific, there's no completely straightforward way to just switch umlaut 2.x to git/github without re-analyzing the install/update workflow). Meanwhile, if we end up needing umlaut 2.x changes... I dunno. I wonder if it's just me, if anyone else can succesfully connect to rubyforge umlaut svn for update/commit, maybe I can send you my small patch, and you can commit it to umlaut 2.x for me? Jonathan From rossfsinger at gmail.com Wed Feb 29 16:55:29 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 29 Feb 2012 11:55:29 -0500 Subject: [Umlaut-general] rubyforge svn? In-Reply-To: <4F4E3BEF.2040800@jhu.edu> References: <4F4E3BEF.2040800@jhu.edu> Message-ID: <18F33F5B-5B5E-42FC-BB72-6283D5D2F46A@gmail.com> On Feb 29, 2012, at 9:53 AM, Jonathan Rochkind wrote: > The rubyforge svn where Umlaut 2.x is hosted is currently unresponsive to me. I get "svn: Network connection closed unexpectedly" This sounds more like a problem with your public key. I'm not entirely sure how to test this without committing something, though. -Ross. > > Curious if this is just me or if anyone can confirm this is them too. (I filed a rubyforge ticket, haven't heard back yet). > > I have a minor tweak I wanted to commit to Umlaut 2.x (restricting size of display of oddly shaped oddly sized cover images fetched from cover image providers), but I can't cause I cant' get the svn. > > Clearly, we can't move soon enough to Umlaut 3.x with it's github host! (As the Umlaut 2.x, pre-Umlaut-as-a-gem, installation instructions end up being kind of svn-specific, there's no completely straightforward way to just switch umlaut 2.x to git/github without re-analyzing the install/update workflow). > > Meanwhile, if we end up needing umlaut 2.x changes... I dunno. I wonder if it's just me, if anyone else can succesfully connect to rubyforge umlaut svn for update/commit, maybe I can send you my small patch, and you can commit it to umlaut 2.x for me? > > Jonathan > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Wed Feb 29 17:54:32 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 29 Feb 2012 12:54:32 -0500 Subject: [Umlaut-general] rubyforge svn? In-Reply-To: <18F33F5B-5B5E-42FC-BB72-6283D5D2F46A@gmail.com> References: <4F4E3BEF.2040800@jhu.edu> <18F33F5B-5B5E-42FC-BB72-6283D5D2F46A@gmail.com> Message-ID: <4F4E6658.4060904@jhu.edu> Don't know why my public key would have worked a week ago and stopped working today, not sure how to fix it either. But, I've actually got something to commit! And it's a single file. If I send you the single file new version, wanna commit it for me? On 2/29/2012 11:55 AM, Ross Singer wrote: > On Feb 29, 2012, at 9:53 AM, Jonathan Rochkind wrote: > >> The rubyforge svn where Umlaut 2.x is hosted is currently unresponsive to me. I get "svn: Network connection closed unexpectedly" > This sounds more like a problem with your public key. I'm not entirely sure how to test this without committing something, though. > > -Ross. >> Curious if this is just me or if anyone can confirm this is them too. (I filed a rubyforge ticket, haven't heard back yet). >> >> I have a minor tweak I wanted to commit to Umlaut 2.x (restricting size of display of oddly shaped oddly sized cover images fetched from cover image providers), but I can't cause I cant' get the svn. >> >> Clearly, we can't move soon enough to Umlaut 3.x with it's github host! (As the Umlaut 2.x, pre-Umlaut-as-a-gem, installation instructions end up being kind of svn-specific, there's no completely straightforward way to just switch umlaut 2.x to git/github without re-analyzing the install/update workflow). >> >> Meanwhile, if we end up needing umlaut 2.x changes... I dunno. I wonder if it's just me, if anyone else can succesfully connect to rubyforge umlaut svn for update/commit, maybe I can send you my small patch, and you can commit it to umlaut 2.x for me? >> >> Jonathan >> _______________________________________________ >> Umlaut-general mailing list >> Umlaut-general at rubyforge.org >> http://rubyforge.org/mailman/listinfo/umlaut-general > _______________________________________________ > Umlaut-general mailing list > Umlaut-general at rubyforge.org > http://rubyforge.org/mailman/listinfo/umlaut-general From rossfsinger at gmail.com Wed Feb 29 18:13:53 2012 From: rossfsinger at gmail.com (Ross Singer) Date: Wed, 29 Feb 2012 13:13:53 -0500 Subject: [Umlaut-general] rubyforge svn? In-Reply-To: <4F4E6658.4060904@jhu.edu> References: <4F4E3BEF.2040800@jhu.edu> <18F33F5B-5B5E-42FC-BB72-6283D5D2F46A@gmail.com> <4F4E6658.4060904@jhu.edu> Message-ID: <345DEA5A-EB7B-4D28-A080-5AFD7AB3BE36@gmail.com> Well, I just tried and got the same error. Of course a "problem with your public key" can be on their end too :( Looking at the rubyforge-help list, we are not alone: http://rubyforge.org/forum/forum.php?forum_id=2 -Ross. On Feb 29, 2012, at 12:54 PM, Jonathan Rochkind wrote: > Don't know why my public key would have worked a week ago and stopped working today, not sure how to fix it either. > > But, I've actually got something to commit! And it's a single file. If I send you the single file new version, wanna commit it for me? > > On 2/29/2012 11:55 AM, Ross Singer wrote: >> On Feb 29, 2012, at 9:53 AM, Jonathan Rochkind wrote: >> >>> The rubyforge svn where Umlaut 2.x is hosted is currently unresponsive to me. I get "svn: Network connection closed unexpectedly" >> This sounds more like a problem with your public key. I'm not entirely sure how to test this without committing something, though. >> >> -Ross. >>> Curious if this is just me or if anyone can confirm this is them too. (I filed a rubyforge ticket, haven't heard back yet). >>> >>> I have a minor tweak I wanted to commit to Umlaut 2.x (restricting size of display of oddly shaped oddly sized cover images fetched from cover image providers), but I can't cause I cant' get the svn. >>> >>> Clearly, we can't move soon enough to Umlaut 3.x with it's github host! (As the Umlaut 2.x, pre-Umlaut-as-a-gem, installation instructions end up being kind of svn-specific, there's no completely straightforward way to just switch umlaut 2.x to git/github without re-analyzing the install/update workflow). >>> >>> Meanwhile, if we end up needing umlaut 2.x changes... I dunno. I wonder if it's just me, if anyone else can succesfully connect to rubyforge umlaut svn for update/commit, maybe I can send you my small patch, and you can commit it to umlaut 2.x for me? >>> >>> Jonathan >>> _______________________________________________ >>> Umlaut-general mailing list >>> Umlaut-general at rubyforge.org >>> http://rubyforge.org/mailman/listinfo/umlaut-general >> _______________________________________________ >> Umlaut-general mailing list >> Umlaut-general at rubyforge.org >> http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Wed Feb 29 18:46:01 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 29 Feb 2012 13:46:01 -0500 Subject: [Umlaut-general] rubyforge svn? In-Reply-To: <345DEA5A-EB7B-4D28-A080-5AFD7AB3BE36@gmail.com> References: <4F4E3BEF.2040800@jhu.edu> <18F33F5B-5B5E-42FC-BB72-6283D5D2F46A@gmail.com> <4F4E6658.4060904@jhu.edu> <345DEA5A-EB7B-4D28-A080-5AFD7AB3BE36@gmail.com> Message-ID: <4F4E7269.4080100@jhu.edu> Okay, thanks. I filed a support ticket myself yesterday, we'll see. can't get off rubyforge soon enough, glad the umlaut 3.x on github is in the works. in the meantime, we've got no real great way to update umlaut 2.x, so hopefully they'll get it fixed soon. glad I got the fixes for Dale regarding CSA/EricID/pid in there already. On 2/29/2012 1:13 PM, Ross Singer wrote: > Well, I just tried and got the same error. Of course a "problem with your public key" can be on their end too :( > > Looking at the rubyforge-help list, we are not alone: > > http://rubyforge.org/forum/forum.php?forum_id=2 > > -Ross. > > On Feb 29, 2012, at 12:54 PM, Jonathan Rochkind wrote: > >> Don't know why my public key would have worked a week ago and stopped working today, not sure how to fix it either. >> >> But, I've actually got something to commit! And it's a single file. If I send you the single file new version, wanna commit it for me? >> >> On 2/29/2012 11:55 AM, Ross Singer wrote: >>> On Feb 29, 2012, at 9:53 AM, Jonathan Rochkind wrote: >>> >>>> The rubyforge svn where Umlaut 2.x is hosted is currently unresponsive to me. I get "svn: Network connection closed unexpectedly" >>> This sounds more like a problem with your public key. I'm not entirely sure how to test this without committing something, though. >>> >>> -Ross. >>>> Curious if this is just me or if anyone can confirm this is them too. (I filed a rubyforge ticket, haven't heard back yet). >>>> >>>> I have a minor tweak I wanted to commit to Umlaut 2.x (restricting size of display of oddly shaped oddly sized cover images fetched from cover image providers), but I can't cause I cant' get the svn. >>>> >>>> Clearly, we can't move soon enough to Umlaut 3.x with it's github host! (As the Umlaut 2.x, pre-Umlaut-as-a-gem, installation instructions end up being kind of svn-specific, there's no completely straightforward way to just switch umlaut 2.x to git/github without re-analyzing the install/update workflow). >>>> >>>> Meanwhile, if we end up needing umlaut 2.x changes... I dunno. I wonder if it's just me, if anyone else can succesfully connect to rubyforge umlaut svn for update/commit, maybe I can send you my small patch, and you can commit it to umlaut 2.x for me? >>>> >>>> Jonathan >>>> _______________________________________________ >>>> Umlaut-general mailing list >>>> Umlaut-general at rubyforge.org >>>> http://rubyforge.org/mailman/listinfo/umlaut-general >>> _______________________________________________ >>> Umlaut-general mailing list >>> Umlaut-general at rubyforge.org >>> http://rubyforge.org/mailman/listinfo/umlaut-general From rochkind at jhu.edu Wed Feb 29 18:53:43 2012 From: rochkind at jhu.edu (Jonathan Rochkind) Date: Wed, 29 Feb 2012 13:53:43 -0500 Subject: [Umlaut-general] rubyforge svn? In-Reply-To: <345DEA5A-EB7B-4D28-A080-5AFD7AB3BE36@gmail.com> References: <4F4E3BEF.2040800@jhu.edu> <18F33F5B-5B5E-42FC-BB72-6283D5D2F46A@gmail.com> <4F4E6658.4060904@jhu.edu> <345DEA5A-EB7B-4D28-A080-5AFD7AB3BE36@gmail.com> Message-ID: <4F4E7437.6010609@jhu.edu> Okay, figured out how to fix it. Looks like rubyforge changed their ip address or ssh key or something (assuming it's not actually a man in the middle attack; would be nice if rubyforge would actually warn us through official channels so we'd know it's not a man in the middle attack!). Edit your ~/.ssh/known_hosts file. Remove the line for rubyforge. save. Now svn to rubyforge works again. I filed a support ticket with rubyforge on this yesterday, went to check it today... and my support ticket seems to be gone from the system. What? Got me. Found a support ticket from someone else who gave themselves an answer (no answer from rubyforge support), which led me to this. rubyforge definitely seems to be, um, fading. Jonathan On 2/29/2012 1:13 PM, Ross Singer wrote: > Well, I just tried and got the same error. Of course a "problem with your public key" can be on their end too :( > > Looking at the rubyforge-help list, we are not alone: > > http://rubyforge.org/forum/forum.php?forum_id=2 > > -Ross. > > On Feb 29, 2012, at 12:54 PM, Jonathan Rochkind wrote: > >> Don't know why my public key would have worked a week ago and stopped working today, not sure how to fix it either. >> >> But, I've actually got something to commit! And it's a single file. If I send you the single file new version, wanna commit it for me? >> >> On 2/29/2012 11:55 AM, Ross Singer wrote: >>> On Feb 29, 2012, at 9:53 AM, Jonathan Rochkind wrote: >>> >>>> The rubyforge svn where Umlaut 2.x is hosted is currently unresponsive to me. I get "svn: Network connection closed unexpectedly" >>> This sounds more like a problem with your public key. I'm not entirely sure how to test this without committing something, though. >>> >>> -Ross. >>>> Curious if this is just me or if anyone can confirm this is them too. (I filed a rubyforge ticket, haven't heard back yet). >>>> >>>> I have a minor tweak I wanted to commit to Umlaut 2.x (restricting size of display of oddly shaped oddly sized cover images fetched from cover image providers), but I can't cause I cant' get the svn. >>>> >>>> Clearly, we can't move soon enough to Umlaut 3.x with it's github host! (As the Umlaut 2.x, pre-Umlaut-as-a-gem, installation instructions end up being kind of svn-specific, there's no completely straightforward way to just switch umlaut 2.x to git/github without re-analyzing the install/update workflow). >>>> >>>> Meanwhile, if we end up needing umlaut 2.x changes... I dunno. I wonder if it's just me, if anyone else can succesfully connect to rubyforge umlaut svn for update/commit, maybe I can send you my small patch, and you can commit it to umlaut 2.x for me? >>>> >>>> Jonathan >>>> _______________________________________________ >>>> Umlaut-general mailing list >>>> Umlaut-general at rubyforge.org >>>> http://rubyforge.org/mailman/listinfo/umlaut-general >>> _______________________________________________ >>> Umlaut-general mailing list >>> Umlaut-general at rubyforge.org >>> http://rubyforge.org/mailman/listinfo/umlaut-general