From schnober at elda.org Fri Oct 1 04:40:57 2010 From: schnober at elda.org (Carsten Schnober) Date: Fri, 01 Oct 2010 10:40:57 +0200 Subject: [BioCatalogue-hackers] Gem problems In-Reply-To: <4CA1F139.7050902@cs.man.ac.uk> References: <4C9C7D07.6010500@elda.org> <4CA0A85E.6050309@cs.man.ac.uk> <4CA0A93B.1060501@cs.man.ac.uk> <4CA1B2DE.7030503@elda.org> <4CA1B490.9080007@cs.man.ac.uk> <4CA1BABC.60304@elda.org> <4CA1BBD4.3050406@cs.man.ac.uk> <4CA1E56F.7090707@cs.man.ac.uk> <4CA1EE3A.9090607@elda.org> <4CA1F139.7050902@cs.man.ac.uk> Message-ID: <4CA59E99.7020300@elda.org> On 28/09/2010 15:44, Jiten Bhagat wrote: Hi, > My recommendation at this stage would be to uninstall Ruby and RubyGems > and reinstall these using the instructions provided on the BioCat set up > wiki page. I've un- and re-installed every involved package and have been following the instructions on the page in detail, no success... As using RVM does not seem practical to me with regard to a later production use, I've now tried and set up Biocatalogue on a Debian server. So far, it seems to be working fine except that I have not yet set up a mail server there and therefore I don't seem to be able to activate an account. Is there a way to avoid the e-mail activation? A full documentation for the configuration would be helpful, anyway, but I could not find one on the homepage. Is there something you could point me at? Thanks again for your great effort! Carsten From jits at cs.man.ac.uk Fri Oct 1 12:46:57 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Fri, 01 Oct 2010 17:46:57 +0100 Subject: [BioCatalogue-hackers] Gem problems In-Reply-To: <4CA59E99.7020300@elda.org> References: <4C9C7D07.6010500@elda.org> <4CA0A85E.6050309@cs.man.ac.uk> <4CA0A93B.1060501@cs.man.ac.uk> <4CA1B2DE.7030503@elda.org> <4CA1B490.9080007@cs.man.ac.uk> <4CA1BABC.60304@elda.org> <4CA1BBD4.3050406@cs.man.ac.uk> <4CA1E56F.7090707@cs.man.ac.uk> <4CA1EE3A.9090607@elda.org> <4CA1F139.7050902@cs.man.ac.uk> <4CA59E99.7020300@elda.org> Message-ID: <4CA61081.5040900@cs.man.ac.uk> Hi Carsten, Carsten Schnober wrote: > On 28/09/2010 15:44, Jiten Bhagat wrote: > > Hi, > > >> My recommendation at this stage would be to uninstall Ruby and RubyGems >> and reinstall these using the instructions provided on the BioCat set up >> wiki page. >> > > I've un- and re-installed every involved package and have been following > the instructions on the page in detail, no success... > Sorry to hear that :-( I'll keep an eye out for this issue in the future and try again to replicate it at some point. > As using RVM does not seem practical to me with regard to a later > production use, I've now tried and set up Biocatalogue on a Debian server. > So far, it seems to be working fine except that I have not yet set up a > mail server there and therefore I don't seem to be able to activate an > account. Is there a way to avoid the e-mail activation? > Oh, that's good that you've got it set up on a Debian server. For email activation you can go into the script console (by running: ruby script/console) and then run something like: User.last.activate! We need to figure out a better way of handling this process and this is on our list of things to do in the future [1]. > A full documentation for the configuration would be helpful, anyway, but > I could not find one on the homepage. Is there something you could point > me at? > Unfortunately this doesn't exist at the moment, and I agree we definitely need it. This will be part of the myBioCatalogue functionality in the future (I'll be replying to your other emails about this very soon :-)). > Thanks again for your great effort! > Pleasure. Shame that we weren't able to get it working on your Ubuntu installation. Let us know if we can help with anything else. Cheers, Jits [1] http://www.mygrid.org.uk/dev/issues/browse/BIOCAT-490 > Carsten > From schnober at elda.org Mon Oct 4 04:45:42 2010 From: schnober at elda.org (Carsten Schnober) Date: Mon, 04 Oct 2010 10:45:42 +0200 Subject: [BioCatalogue-hackers] Addion Soaplab2 services Message-ID: <4CA99436.5080309@elda.org> Hi list, I am opening a new thread with this issue because it's (probably) not related to my previous problems. I have set up a local Biocatalogue server and a Tomcat server providing Soaplab 2 on another machine in the local network. Now I am trying to add the Soaplab services to the Biocatalogue registry by selecting "Register a Service -> Soaplab Server". Then, I am pasting the URL "http://192.168.1.198:8080/soaplab2/services/list?wsdl" (which is perfectly reachable in the local network) into the form. A click on "Preview" gives me a complete list of the services provided by the Soaplab server. However, when I do "Submit", the host is registered as a service provider (named 192-168-1-198), but none of the services is. Any clues? Thanks, Carsten From jits at cs.man.ac.uk Mon Oct 4 05:30:42 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Mon, 04 Oct 2010 10:30:42 +0100 Subject: [BioCatalogue-hackers] Addion Soaplab2 services In-Reply-To: <4CA99436.5080309@elda.org> References: <4CA99436.5080309@elda.org> Message-ID: <4CA99EC2.2070005@cs.man.ac.uk> Hi Carsten. Thanks for reporting this. We'll look into this; it sounds like it could be a bug in the system. Thanks, Jits Carsten Schnober wrote: > Hi list, > I am opening a new thread with this issue because it's (probably) not > related to my previous problems. > I have set up a local Biocatalogue server and a Tomcat server providing > Soaplab 2 on another machine in the local network. Now I am trying to > add the Soaplab services to the Biocatalogue registry by selecting > "Register a Service -> Soaplab Server". Then, I am pasting the URL > "http://192.168.1.198:8080/soaplab2/services/list?wsdl" (which is > perfectly reachable in the local network) into the form. A click on > "Preview" gives me a complete list of the services provided by the > Soaplab server. However, when I do "Submit", the host is registered as a > service provider (named 192-168-1-198), but none of the services is. > Any clues? Thanks, > Carsten > _______________________________________________ > BioCatalogue-hackers mailing list > BioCatalogue-hackers at rubyforge.org > http://rubyforge.org/mailman/listinfo/biocatalogue-hackers > From schnober at elda.org Thu Oct 7 10:30:45 2010 From: schnober at elda.org (Carsten Schnober) Date: Thu, 07 Oct 2010 16:30:45 +0200 Subject: [BioCatalogue-hackers] Customisation Message-ID: <4CADD995.9060709@elda.org> Hi again, I am now trying to customise a local Biocatalogue installation but I am still struggling with understanding the file organisation there. E.g., how can I change the categories? I am able to manually introduce some by adding entries to data/service_categories.yml and to delete some using the server console command Category.delete, but the overall logic is still somewhat fuzzy to me. Could you point out at which point exactly this file is read and when categories are added to the database, please? Is there a 'clean' way to delete categories? Thank you very much, Carsten From jits at cs.man.ac.uk Thu Oct 7 10:46:49 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Thu, 07 Oct 2010 15:46:49 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CADD995.9060709@elda.org> References: <4CADD995.9060709@elda.org> Message-ID: <4CADDD59.6000004@cs.man.ac.uk> Hi Carsten, You're in the right place by editing the data/service_categories.yml file! What happens is... on application startup, the system loads the service_categories.yml file and then *synchronises* the state of these categories with the categories table in the database. This table is then used to power the UI and to assign in annotations (created when categorising services). So if you add, remove or edit these categories in the service_categories.yml file the system will take care of making sure this is reflected in the database, automatically. One VERY important aspect of the service_categories.yml is the IDs in square brackets next to the entries. For example... Protein Sequence Alignment [3] ... indicates the category with internal ID 3. All annotations that reference this category will do so by this ID number. If you then rename this to: Bacteria [3] ... then you are essentially changing the whole meaning of the category. So as long as the IDs are in place, and unique, you can add, remove, edit, reorder (within the tree) as much as you like. Of course, if you have no category annotations yet then you can change it however you want :-) Another thing to watch out for is invalid YAML content. Unfortunately YAML can be quite temperamental at times and really cares deeply about exact whitespace used. Let me know if I can help with anything else. Cheers, Jits Carsten Schnober wrote: > Hi again, > I am now trying to customise a local Biocatalogue installation but I am > still struggling with understanding the file organisation there. E.g., > how can I change the categories? I am able to manually introduce some by > adding entries to data/service_categories.yml and to delete some using > the server console command Category.delete, but the overall logic is > still somewhat fuzzy to me. > Could you point out at which point exactly this file is read and when > categories are added to the database, please? Is there a 'clean' way to > delete categories? > Thank you very much, > Carsten > _______________________________________________ > BioCatalogue-hackers mailing list > BioCatalogue-hackers at rubyforge.org > http://rubyforge.org/mailman/listinfo/biocatalogue-hackers > From jits at cs.man.ac.uk Thu Oct 7 10:49:02 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Thu, 07 Oct 2010 15:49:02 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CADDD59.6000004@cs.man.ac.uk> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> Message-ID: <4CADDDDE.1080209@cs.man.ac.uk> I should mention that in the future we intend to have a single place where you can customise things like this. We are working on making the codebase much more customisable and brandable. Cheers, Jits Jiten Bhagat wrote: > Hi Carsten, > > You're in the right place by editing the data/service_categories.yml file! > > What happens is... on application startup, the system loads the > service_categories.yml file and then *synchronises* the state of these > categories with the categories table in the database. This table is then > used to power the UI and to assign in annotations (created when > categorising services). > > So if you add, remove or edit these categories in the > service_categories.yml file the system will take care of making sure > this is reflected in the database, automatically. > > One VERY important aspect of the service_categories.yml is the IDs in > square brackets next to the entries. For example... > > Protein Sequence Alignment [3] > > ... indicates the category with internal ID 3. All annotations that > reference this category will do so by this ID number. If you then rename > this to: > > Bacteria [3] > > ... then you are essentially changing the whole meaning of the category. > > So as long as the IDs are in place, and unique, you can add, remove, > edit, reorder (within the tree) as much as you like. > > Of course, if you have no category annotations yet then you can change > it however you want :-) > > Another thing to watch out for is invalid YAML content. Unfortunately > YAML can be quite temperamental at times and really cares deeply about > exact whitespace used. > > Let me know if I can help with anything else. > > Cheers, > Jits > > > Carsten Schnober wrote: > >> Hi again, >> I am now trying to customise a local Biocatalogue installation but I am >> still struggling with understanding the file organisation there. E.g., >> how can I change the categories? I am able to manually introduce some by >> adding entries to data/service_categories.yml and to delete some using >> the server console command Category.delete, but the overall logic is >> still somewhat fuzzy to me. >> Could you point out at which point exactly this file is read and when >> categories are added to the database, please? Is there a 'clean' way to >> delete categories? >> Thank you very much, >> Carsten >> _______________________________________________ >> BioCatalogue-hackers mailing list >> BioCatalogue-hackers at rubyforge.org >> http://rubyforge.org/mailman/listinfo/biocatalogue-hackers >> >> > > > From schnober at elda.org Thu Oct 14 06:41:47 2010 From: schnober at elda.org (Carsten Schnober) Date: Thu, 14 Oct 2010 12:41:47 +0200 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CADDD59.6000004@cs.man.ac.uk> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> Message-ID: <4CB6DE6B.8010904@elda.org> On 07/10/2010 16:46, Jiten Bhagat wrote: Hi Jiten, thanks again for your help! > You're in the right place by editing the data/service_categories.yml file! > > What happens is... on application startup, the system loads the > service_categories.yml file and then *synchronises* the state of these > categories with the categories table in the database. This table is then > used to power the UI and to assign in annotations (created when > categorising services). > > So if you add, remove or edit these categories in the > service_categories.yml file the system will take care of making sure > this is reflected in the database, automatically. Adding and editing categories works fine using the data/service_categories.yml file. However, removing categories from the list does not seem to have any impact. Also, the subcategory schemes seem to remain untouched even when changing the Yaml file accordingly. Will I have to delete categories directly on the console and I will I risk doing any harm to the database by doing that? Maybe you could also point me to the Ruby file that implements the categories updates, so I can check out the details myself. Thanks, Carsten From jits at cs.man.ac.uk Thu Oct 14 07:36:43 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Thu, 14 Oct 2010 12:36:43 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CB6DE6B.8010904@elda.org> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> Message-ID: <4CB6EB4B.8080108@cs.man.ac.uk> Hi Carsten, Oh, that's weird and unexpected behaviour. As a temporary measure, feel free to manually delete the categories via the console. But please restart your server after you have done so, because all the Categories are cached in memory once when the app is loaded (to help performance). The file that handles the sync process (and other categorising related stuff) is in: lib/bio_catalogue/categorising.rb (I.e.: http://biocatalogue.rubyforge.org/svn/trunk/lib/bio_catalogue/categorising.rb) ... and it's the self.load_data method that does it. I'll also take a closer look at this file to see if I can spot any bugs. You could also look at the Category model (http://biocatalogue.rubyforge.org/svn/trunk/app/models/category.rb). Would it be possible to see your new categories YAML file? Thanks, Jits Carsten Schnober wrote: > On 07/10/2010 16:46, Jiten Bhagat wrote: > > Hi Jiten, > thanks again for your help! > > >> You're in the right place by editing the data/service_categories.yml file! >> >> What happens is... on application startup, the system loads the >> service_categories.yml file and then *synchronises* the state of these >> categories with the categories table in the database. This table is then >> used to power the UI and to assign in annotations (created when >> categorising services). >> >> So if you add, remove or edit these categories in the >> service_categories.yml file the system will take care of making sure >> this is reflected in the database, automatically. >> > > Adding and editing categories works fine using the > data/service_categories.yml file. However, removing categories from the > list does not seem to have any impact. Also, the subcategory schemes > seem to remain untouched even when changing the Yaml file accordingly. > Will I have to delete categories directly on the console and I will I > risk doing any harm to the database by doing that? > Maybe you could also point me to the Ruby file that implements the > categories updates, so I can check out the details myself. > Thanks, > Carsten > From jits at cs.man.ac.uk Thu Oct 14 07:41:38 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Thu, 14 Oct 2010 12:41:38 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CB6EB4B.8080108@cs.man.ac.uk> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> Message-ID: <4CB6EC72.3080406@cs.man.ac.uk> Ahh, it looks like the load_data method doesn't actually do deletions of categories. I would recommend just... - shutting down your server - deleting all the records in the 'categories' table - starting up your server again This will reset the state of the categories the way you want them. Cheers, Jits Jiten Bhagat wrote: > Hi Carsten, > > Oh, that's weird and unexpected behaviour. > > As a temporary measure, feel free to manually delete the categories via > the console. But please restart your server after you have done so, > because all the Categories are cached in memory once when the app is > loaded (to help performance). > > The file that handles the sync process (and other categorising related > stuff) is in: > > lib/bio_catalogue/categorising.rb > > (I.e.: > http://biocatalogue.rubyforge.org/svn/trunk/lib/bio_catalogue/categorising.rb) > > ... and it's the self.load_data method that does it. > > I'll also take a closer look at this file to see if I can spot any bugs. > > You could also look at the Category model > (http://biocatalogue.rubyforge.org/svn/trunk/app/models/category.rb). > > Would it be possible to see your new categories YAML file? > > Thanks, > Jits > > > Carsten Schnober wrote: > >> On 07/10/2010 16:46, Jiten Bhagat wrote: >> >> Hi Jiten, >> thanks again for your help! >> >> >> >>> You're in the right place by editing the data/service_categories.yml file! >>> >>> What happens is... on application startup, the system loads the >>> service_categories.yml file and then *synchronises* the state of these >>> categories with the categories table in the database. This table is then >>> used to power the UI and to assign in annotations (created when >>> categorising services). >>> >>> So if you add, remove or edit these categories in the >>> service_categories.yml file the system will take care of making sure >>> this is reflected in the database, automatically. >>> >>> >> Adding and editing categories works fine using the >> data/service_categories.yml file. However, removing categories from the >> list does not seem to have any impact. Also, the subcategory schemes >> seem to remain untouched even when changing the Yaml file accordingly. >> Will I have to delete categories directly on the console and I will I >> risk doing any harm to the database by doing that? >> Maybe you could also point me to the Ruby file that implements the >> categories updates, so I can check out the details myself. >> Thanks, >> Carsten >> >> > > > From schnober at elda.org Thu Oct 14 10:50:39 2010 From: schnober at elda.org (Carsten Schnober) Date: Thu, 14 Oct 2010 16:50:39 +0200 Subject: [BioCatalogue-hackers] Monitoring Message-ID: <4CB718BF.8010500@elda.org> Hi list, you'll know me by now. ;-) I keep posting my issues here because I assume that they might be interesting for others as well. I am now trying to set up the monitoring system for my local Biocatalogue installation. Currently, it just seems not to be running at all. I could not figure out whether I need to start any additional services to activate it. Do I? The worker is running, however. Thanks, Carsten From jits at cs.man.ac.uk Mon Oct 18 10:04:21 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Mon, 18 Oct 2010 15:04:21 +0100 Subject: [BioCatalogue-hackers] REST API: features coming up and proposed breaking changes to the JSON outputs Message-ID: <4CBC53E5.7060303@cs.man.ac.uk> Hello all, In 2-3 weeks time we anticipate a big release of the BioCatalogue in which there will be many new API additions, such as: - full read API for REST services. - filtering for the Users and ServiceProviders indexes. - full JSON outputs for all endpoints. - an initial Write API for registering services and submitting annotations. - OAuth support for authentication/authorisation through the API. This will result in some breaking changes to the old JSON outputs for Annotations and Annotation Attributes indexes, to make them more like the index resources in the XML. You can check what the new JSON outputs looks like on the sandbox server: http://sandbox.biocatalogue.org: http://www.biocatalogue.org/annotations.json vs http://sandbox.biocatalogue.org/annotations.json http://www.biocatalogue.org/annotation_attributes.json vs http://sandbox.biocatalogue.org/annotation_attributes.json We hope this does not cause much inconvenience for you. Please let us know if we can help in any way, and feel free to give us feedback/suggestions on this proposed change. Kind regards, Jits From jits at cs.man.ac.uk Mon Oct 18 10:21:16 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Mon, 18 Oct 2010 15:21:16 +0100 Subject: [BioCatalogue-hackers] Monitoring In-Reply-To: <4CB718BF.8010500@elda.org> References: <4CB718BF.8010500@elda.org> Message-ID: <4CBC57DC.9020408@cs.man.ac.uk> Carsten Schnober wrote: > Hi list, > you'll know me by now. ;-) Hi :-) > I keep posting my issues here because I > assume that they might be interesting for others as well. > Please carry on doing so. This will hopefully be useful to others in the future (especially since the archive of these conversations are available via Google searches). > I am now trying to set up the monitoring system for my local > Biocatalogue installation. Currently, it just seems not to be running at > all. I could not figure out whether I need to start any additional > services to activate it. Do I? The worker is running, however. > I think you need to run some additional rake tasks (or submit some other jobs) to set up what exactly should be monitored. This page might be useful for an overview: http://www.biocatalogue.org/wiki/doku.php?id=development:service_testing The table at the end of this page lists a set of jobs that need to be set up / run for the system (incl how often a cron job must run): http://www.biocatalogue.org/wiki/doku.php?id=development:background_processing#current_jobs Hope this gets you started. I've asked Eric to provide you with more detailed info. Let us know how you get on. Cheers, Jits > Thanks, > Carsten > _______________________________________________ > BioCatalogue-hackers mailing list > BioCatalogue-hackers at rubyforge.org > http://rubyforge.org/mailman/listinfo/biocatalogue-hackers > From schnober at elda.org Tue Oct 19 11:45:52 2010 From: schnober at elda.org (Carsten Schnober) Date: Tue, 19 Oct 2010 17:45:52 +0200 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CB6EC72.3080406@cs.man.ac.uk> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> Message-ID: <4CBDBD30.5070005@elda.org> On 14/10/2010 13:41, Jiten Bhagat wrote: > I would recommend just... > - shutting down your server > - deleting all the records in the 'categories' table > - starting up your server again > > This will reset the state of the categories the way you want them. Hi, I have deleted and re-populated the categories in the way described by Jits. However, after having added and categorised some services, I have bumped into a very strange problem. In the services view, some actually unused category tags show numbers >0 in the overview. When I add one of those as a filter, the services list presents me this number of registered services. However, none of these services is actually assigned to the chosen category. Any clue what's going on there? Might this have to do with former sub-categories of the same IDs that have now become main categories? Where does Biocatalogue store the service-category assignments? Regards, Carsten From jits at cs.man.ac.uk Tue Oct 19 12:04:08 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Tue, 19 Oct 2010 17:04:08 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBDBD30.5070005@elda.org> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> Message-ID: <4CBDC178.2060504@cs.man.ac.uk> Hi Carsten, Ahh, this is very likely to do with the former sub-categories, if you had previously categorised some services before the changes. Sorry, I was under the impression you hadn't done any categorising just yet, so I recommended the full deletion. The service-category assignments are stored in the Annotation model / annotations table*. Currently, the "value" of the annotation is the ID of the category assigned, stored as a string and with no referential integrity put in place. The aim of this was to decouple the annotations mechanism from the core of the system and hence why it is crucial to keep the existing IDs of categories intact. However, as you have experienced, there are some major weaknesses to this :-( We're hoping to address this in future work of the Annotations plugin whereby the "value" of an Annotation is an actual entity in the system (so would actually point to the Category object) and not just an implicit reference by a fragile value. * Note that the annotations table will contain all kinds of annotations - tags, descriptions, etc. To find just the categories, first check what the ID is of the "category" AnnotationAttribute, and then find all Annotations with that attribute_id. You can also do this (more easily) in the rails console, eg: > category_annotations = Annotation.with_attribute_name("category") Hope this helps. Cheers, Jits Carsten Schnober wrote: > On 14/10/2010 13:41, Jiten Bhagat wrote: > > >> I would recommend just... >> - shutting down your server >> - deleting all the records in the 'categories' table >> - starting up your server again >> >> This will reset the state of the categories the way you want them. >> > > Hi, > I have deleted and re-populated the categories in the way described by > Jits. However, after having added and categorised some services, I have > bumped into a very strange problem. > > In the services view, some actually unused category tags show numbers >0 > in the overview. When I add one of those as a filter, the services list > presents me this number of registered services. However, none of these > services is actually assigned to the chosen category. > > Any clue what's going on there? Might this have to do with former > sub-categories of the same IDs that have now become main categories? > Where does Biocatalogue store the service-category assignments? > > Regards, > Carsten > From schnober at elda.org Wed Oct 20 05:55:15 2010 From: schnober at elda.org (Carsten Schnober) Date: Wed, 20 Oct 2010 11:55:15 +0200 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBDC178.2060504@cs.man.ac.uk> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> <4CBDC178.2060504@cs.man.ac.uk> Message-ID: <4CBEBC83.5070906@elda.org> On 19/10/2010 18:04, Jiten Bhagat wrote: Hi Jiten, > The service-category assignments are stored in the Annotation model / > annotations table*. > > Currently, the "value" of the annotation is the ID of the category > assigned, stored as a string and with no referential integrity put in > place. Ok, I have found this. However, the table contents seem to be perfect, there is no reference to one of the unused categories. I've thought that some services are wrongly listed under e.g. category 2 because this may previously have been the category of which e.g. category 3 was a sub-category, i.e. some kind of indirect assignment. However, in the categories database, the column 'parent_id' correctly contains NULL for all entries. Is there another place where relations between categories are stored? Or am I possibly on a wrong track and this problem is not cause by previous sub-categories? Regards, Carsten From jits at cs.man.ac.uk Wed Oct 20 06:15:12 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Wed, 20 Oct 2010 11:15:12 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBEBC83.5070906@elda.org> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> <4CBDC178.2060504@cs.man.ac.uk> <4CBEBC83.5070906@elda.org> Message-ID: <4CBEC130.3020809@cs.man.ac.uk> Hi Carsten, Carsten Schnober wrote: > On 19/10/2010 18:04, Jiten Bhagat wrote: > > Hi Jiten, > > >> The service-category assignments are stored in the Annotation model / >> annotations table*. >> >> Currently, the "value" of the annotation is the ID of the category >> assigned, stored as a string and with no referential integrity put in >> place. >> > > Ok, I have found this. However, the table contents seem to be perfect, > there is no reference to one of the unused categories. > > I've thought that some services are wrongly listed under e.g. category 2 > because this may previously have been the category of which e.g. > category 3 was a sub-category, i.e. some kind of indirect assignment. > > However, in the categories database, the column 'parent_id' correctly > contains NULL for all entries. By this, do you mean that your categories are just a flat list at the moment, with no sub categories? (I.e.: it's not a tree, it's just a list?) > Is there another place where relations > between categories are stored? Or am I possibly on a wrong track and > this problem is not cause by previous sub-categories? > It doesn't seem to be a problem with the previous sub-categories (assuming you've verified that the annotations have the correct values). The relations between categories are only stored in the categories table at the moment, via the 'parent_id' field. Then when the system is started up the categories tree is built from this info and stored into memory once. Any subsequent changes to the table, whilst the system is running, won't take into effect until the system is restarted. As an aside... Note that the counts shown in the categories filters take into account child categories, so if I have a category tree like: - a -- b -- c - d ... then the count for 'a' will take into account all services with 'b' and 'c' too. Cheers, Jits > Regards, > Carsten > From schnober at elda.org Wed Oct 20 06:33:01 2010 From: schnober at elda.org (Carsten Schnober) Date: Wed, 20 Oct 2010 12:33:01 +0200 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBEC130.3020809@cs.man.ac.uk> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> <4CBDC178.2060504@cs.man.ac.uk> <4CBEBC83.5070906@elda.org> <4CBEC130.3020809@cs.man.ac.uk> Message-ID: <4CBEC55D.2070008@elda.org> On 20/10/2010 12:15, Jiten Bhagat wrote: >> However, in the categories database, the column 'parent_id' correctly >> contains NULL for all entries. > > By this, do you mean that your categories are just a flat list at the > moment, with no sub categories? (I.e.: it's not a tree, it's just a list?) Exactly. My service_categories.yml only contains entries like this: Named Entity Recognition [1] Language Guessing [2] [...] Actually, after reading the YAML specification, I had the impression the each line should begin with a dash. However, categories are parsed well with this syntax (no dashes) while they are not recognised with preceding dashed. So this should be fine, I guess. I am not a YAML expert, so do you think that this could be related to the categorisation/listing problems? Carsten From jits at cs.man.ac.uk Wed Oct 20 10:38:52 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Wed, 20 Oct 2010 15:38:52 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBEC55D.2070008@elda.org> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> <4CBDC178.2060504@cs.man.ac.uk> <4CBEBC83.5070906@elda.org> <4CBEC130.3020809@cs.man.ac.uk> <4CBEC55D.2070008@elda.org> Message-ID: <4CBEFEFC.80405@cs.man.ac.uk> Hi Carsten, I have a suspicion it could be due to the YAML format (in hindsight, YAML was probably not the best choice of format for this! Even if it is quite readable). I'll try and replicate this locally by having a new instance of the code and a flat YAML categories file. To help with this, could you describe the exact issue you are experiencing again please? Thanks, Jits Carsten Schnober wrote: > On 20/10/2010 12:15, Jiten Bhagat wrote: > > >>> However, in the categories database, the column 'parent_id' correctly >>> contains NULL for all entries. >>> >> By this, do you mean that your categories are just a flat list at the >> moment, with no sub categories? (I.e.: it's not a tree, it's just a list?) >> > > Exactly. My service_categories.yml only contains entries like this: > > Named Entity Recognition [1] > > Language Guessing [2] > > [...] > > Actually, after reading the YAML specification, I had the impression the > each line should begin with a dash. However, categories are parsed well > with this syntax (no dashes) while they are not recognised with > preceding dashed. So this should be fine, I guess. > > I am not a YAML expert, so do you think that this could be related to > the categorisation/listing problems? > > Carsten > From schnober at elda.org Wed Oct 20 11:20:50 2010 From: schnober at elda.org (Carsten Schnober) Date: Wed, 20 Oct 2010 17:20:50 +0200 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBEFEFC.80405@cs.man.ac.uk> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> <4CBDC178.2060504@cs.man.ac.uk> <4CBEBC83.5070906@elda.org> <4CBEC130.3020809@cs.man.ac.uk> <4CBEC55D.2070008@elda.org> <4CBEFEFC.80405@cs.man.ac.uk> Message-ID: <4CBF08D2.5040208@elda.org> On 20/10/2010 16:38, Jiten Bhagat wrote: Hi, > I have a suspicion it could be due to the YAML format (in hindsight, > YAML was probably not the best choice of format for this! Even if it is > quite readable). Well, the YAML file being the cause was just another suspicion of mine while search for the error cause. > I'll try and replicate this locally by having a new instance of the code > and a flat YAML categories file. > > To help with this, could you describe the exact issue you are > experiencing again please? Ok, here's my setup: I had a local installation which was initially populated with the categories defined in the packaged service_categories.yml file. I then deleted all the categories there and changed the tree to a list of the following pattern: --- A [1] B [2] C [3] [...] I then added some services via the web interface and assigned categories to some of them. When I call the services sub-page now, I'll see the 'Service Categories' list in the left-hand frame. Each of its entries contains a number that is supposed to indicate the number of services assigned to that category. When clicking on one of these, all the services assigned to it should be shown in the middle frame. However, in the list, e.g. category A indicates that there are 8 services assigned to A, although in fact none is. When I activate it as a filter, 8 services are shown, but none of them is actually assigned to category A. How is this link made? There are other categories for which the listing/filtering works flawlessly. Regards, Carsten From jits at cs.man.ac.uk Wed Oct 20 11:43:04 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Wed, 20 Oct 2010 16:43:04 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBF08D2.5040208@elda.org> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> <4CBDC178.2060504@cs.man.ac.uk> <4CBEBC83.5070906@elda.org> <4CBEC130.3020809@cs.man.ac.uk> <4CBEC55D.2070008@elda.org> <4CBEFEFC.80405@cs.man.ac.uk> <4CBF08D2.5040208@elda.org> Message-ID: <4CBF0E08.2020200@cs.man.ac.uk> Thanks Carsten. That sounds like it might be a bug in the filtering code then. I will take a look as soon as I can (I am away for the rest of this week so expect me to get back to you early next week). The filtering code lives in the following files (in case you're curious): - lib/bio_catalogue/filtering.rb - lib/bio_catalogue/filtering/services.rb Kind regards, Jits Carsten Schnober wrote: > On 20/10/2010 16:38, Jiten Bhagat wrote: > > Hi, > > >> I have a suspicion it could be due to the YAML format (in hindsight, >> YAML was probably not the best choice of format for this! Even if it is >> quite readable). >> > > Well, the YAML file being the cause was just another suspicion of mine > while search for the error cause. > > >> I'll try and replicate this locally by having a new instance of the code >> and a flat YAML categories file. >> >> To help with this, could you describe the exact issue you are >> experiencing again please? >> > > Ok, here's my setup: I had a local installation which was initially > populated with the categories defined in the packaged > service_categories.yml file. I then deleted all the categories there and > changed the tree to a list of the following pattern: > > --- > A [1] > B [2] > C [3] > [...] > > I then added some services via the web interface and assigned categories > to some of them. When I call the services sub-page now, I'll see the > 'Service Categories' list in the left-hand frame. Each of its entries > contains a number that is supposed to indicate the number of services > assigned to that category. When clicking on one of these, all the > services assigned to it should be shown in the middle frame. > > However, in the list, e.g. category A indicates that there are 8 > services assigned to A, although in fact none is. When I activate it as > a filter, 8 services are shown, but none of them is actually assigned to > category A. > How is this link made? > > There are other categories for which the listing/filtering works flawlessly. > > Regards, > Carsten > From tod.jackson at emory.edu Fri Oct 29 10:47:33 2010 From: tod.jackson at emory.edu (Jackson, Tod) Date: Fri, 29 Oct 2010 10:47:33 -0400 Subject: [BioCatalogue-hackers] Does/will BioCatalogue support wsdl:import Message-ID: Greetings BioCataloguers, My colleagues and I here at Emory University and the Atlanta Clinical & Translational Science Institute are attempting to register some services in our local Emory BioCatalogue deployment. These services are authored using caGrid tools and hosted in various caGrid deployments both internally and externally. I believe many of you are familiar with the caGrid project (http://cagrid.org). It appears as though most, if not all of the WSDL definitions associated to these ?grid services? leverage ?wsdl:import? to fully resolve themselves. After some quick analysis, it does not appear as though BioCatalogue is able to fully resolve service definitions that use wsdl:import. I am able to load the services using soapUI and Taverna however. So, my first question is, am I missing anything regarding BioCatalogue?s support for this? i.e., is there a configuration setting or something that might make this work? Assuming the answer to that is ?no?, my second question is, are there any plans to add support for this in future versions of the application? As always, thanks for any input you might have on this. Best Regards, Tod Jackson ________________________________ This e-mail message (including any attachments) is for the sole use of the intended recipient(s) and may contain confidential and privileged information. If the reader of this message is not the intended recipient, you are hereby notified that any dissemination, distribution or copying of this message (including any attachments) is strictly prohibited. If you have received this message in error, please contact the sender by reply e-mail message and destroy all copies of the original message (including attachments). -------------- next part -------------- An HTML attachment was scrubbed... URL: From jits at cs.man.ac.uk Fri Oct 29 12:05:50 2010 From: jits at cs.man.ac.uk (Jiten Bhagat) Date: Fri, 29 Oct 2010 17:05:50 +0100 Subject: [BioCatalogue-hackers] Customisation In-Reply-To: <4CBF08D2.5040208@elda.org> References: <4CADD995.9060709@elda.org> <4CADDD59.6000004@cs.man.ac.uk> <4CB6DE6B.8010904@elda.org> <4CB6EB4B.8080108@cs.man.ac.uk> <4CB6EC72.3080406@cs.man.ac.uk> <4CBDBD30.5070005@elda.org> <4CBDC178.2060504@cs.man.ac.uk> <4CBEBC83.5070906@elda.org> <4CBEC130.3020809@cs.man.ac.uk> <4CBEC55D.2070008@elda.org> <4CBEFEFC.80405@cs.man.ac.uk> <4CBF08D2.5040208@elda.org> Message-ID: <4CCAF0DE.8030507@cs.man.ac.uk> Hi again Carsten, It occurs to me that this may have been caused by stale data in your memcache instance. If you restart your memcache server with the following command: sudo service memcached restart ... does the problem go away? Cheers, Jits Carsten Schnober wrote: > On 20/10/2010 16:38, Jiten Bhagat wrote: > > Hi, > > >> I have a suspicion it could be due to the YAML format (in hindsight, >> YAML was probably not the best choice of format for this! Even if it is >> quite readable). >> > > Well, the YAML file being the cause was just another suspicion of mine > while search for the error cause. > > >> I'll try and replicate this locally by having a new instance of the code >> and a flat YAML categories file. >> >> To help with this, could you describe the exact issue you are >> experiencing again please? >> > > Ok, here's my setup: I had a local installation which was initially > populated with the categories defined in the packaged > service_categories.yml file. I then deleted all the categories there and > changed the tree to a list of the following pattern: > > --- > A [1] > B [2] > C [3] > [...] > > I then added some services via the web interface and assigned categories > to some of them. When I call the services sub-page now, I'll see the > 'Service Categories' list in the left-hand frame. Each of its entries > contains a number that is supposed to indicate the number of services > assigned to that category. When clicking on one of these, all the > services assigned to it should be shown in the middle frame. > > However, in the list, e.g. category A indicates that there are 8 > services assigned to A, although in fact none is. When I activate it as > a filter, 8 services are shown, but none of them is actually assigned to > category A. > How is this link made? > > There are other categories for which the listing/filtering works flawlessly. > > Regards, > Carsten > From ericnzuo at ebi.ac.uk Fri Oct 29 12:22:16 2010 From: ericnzuo at ebi.ac.uk (Eric Nzuobontane) Date: Fri, 29 Oct 2010 17:22:16 +0100 Subject: [BioCatalogue-hackers] Does/will BioCatalogue support wsdl:import In-Reply-To: References: Message-ID: <4CCAF4B8.1030306@ebi.ac.uk> Hi Tod, I am on my way out now but I will look into this later. One thing I would like to know is if those imports are schema imports and whether they have schema locations. If they do and the schema location relative path, then BioCatalogue may have issues with this as it does not try to resolve relative paths for schema locations. regards, Eric On 29/10/2010 15:47, Jackson, Tod wrote: > Greetings BioCataloguers, > > My colleagues and I here at Emory University and the Atlanta Clinical > & Translational Science Institute are attempting to register some > services in our local Emory BioCatalogue deployment. These services > are authored using caGrid tools and hosted in various caGrid > deployments both internally and externally. I believe many of you are > familiar with the caGrid project (http://cagrid.org). > > It appears as though most, if not all of the WSDL definitions > associated to these ?grid services? leverage ?wsdl:import? to fully > resolve themselves. After some quick analysis, it does not appear as > though BioCatalogue is able to fully resolve service definitions that > use wsdl:import. I am able to load the services using soapUI and > Taverna however. > > So, my first question is, am I missing anything regarding > BioCatalogue?s support for this? i.e., is there a configuration > setting or something that might make this work? Assuming the answer > to that is ?no?, my second question is, are there any plans to add > support for this in future versions of the application? > > As always, thanks for any input you might have on this. > > Best Regards, > Tod Jackson > ------------------------------------------------------------------------ > This e-mail message (including any attachments) is for the sole use of > the intended recipient(s) and may contain confidential and privileged > information. If the reader of this message is not the intended > recipient, you are hereby notified that any dissemination, distribution > or copying of this message (including any attachments) is strictly > prohibited. > > If you have received this message in error, please contact > the sender by reply e-mail message and destroy all copies of the > original message (including attachments). > > > _______________________________________________ > BioCatalogue-hackers mailing list > BioCatalogue-hackers at rubyforge.org > http://rubyforge.org/mailman/listinfo/biocatalogue-hackers -- Eric Nzuobontane ================================== European Bioinformatics Institute Wellcome Trust Genome Campus Hinxton, Cambridge CB10 1SD United Kingdom Tel : +44 1223 492654 email : ericnzuo at ebi.ac.uk ================================== -------------- next part -------------- An HTML attachment was scrubbed... URL: From tod.jackson at emory.edu Fri Oct 29 12:37:33 2010 From: tod.jackson at emory.edu (Jackson, Tod) Date: Fri, 29 Oct 2010 12:37:33 -0400 Subject: [BioCatalogue-hackers] Does/will BioCatalogue support wsdl:import In-Reply-To: <4CCAF4B8.1030306@ebi.ac.uk> Message-ID: Here?s an example just so everything is clear, this is at: http://grid.poc.actsi.org/wsrf/services/DefaultIndexService?wsdl This is the WSDL returned initially (like when you access the above URL from a browser). It?s also what BioCatalogue gets back in it?s initial request (like when you try to preview the service). I?ve tried to ?follow? the import programatically in my local deployment because I didn?t think it looked like it was currently supported. However, as I went further with my mods, it started looking like there was perhaps more to it than I initially thought. I disabled WSDLUTILS_BASE_URI so I could force it through my own local wsdl_parser.rb to get a better handle on how things were behaving. I?m still trying to gather more information as to how ?deep? it ultimately goes but when I follow this first import and merge those results with the first ?wsdl_hash? that gets created from this, I started running into other issues. So, I figured it was best to check with you all before I took it too far. Thanks! Tod On 10/29/10 11:22 AM, "Eric Nzuobontane" wrote: Hi Tod, I am on my way out now but I will look into this later. One thing I would like to know is if those imports are schema imports and whether they have schema locations. If they do and the schema location relative path, then BioCatalogue may have issues with this as it does not try to resolve relative paths for schema locations. regards, Eric On 29/10/2010 15:47, Jackson, Tod wrote: Does/will BioCatalogue support wsdl:import Greetings BioCataloguers, My colleagues and I here at Emory University and the Atlanta Clinical & Translational Science Institute are attempting to register some services in our local Emory BioCatalogue deployment. These services are authored using caGrid tools and hosted in various caGrid deployments both internally and externally. I believe many of you are familiar with the caGrid project (http://cagrid.org). It appears as though most, if not all of the WSDL definitions associated to these ?grid services? leverage ?wsdl:import? to fully resolve themselves. After some quick analysis, it does not appear as though BioCatalogue is able to fully resolve service definitions that use wsdl:import. I am able to load the services using soapUI and Taverna however. So, my first question is, am I missing anything regarding BioCatalogue?s support for this? i.e., is there a configuration setting or something that might make this work? Assuming the answer to that is ?no?, my second question is, are there any plans to add support for this in future versions of the application? As always, thanks for any input you might have on this. Best Regards, Tod Jackson ________________________________ This e-mail message (including any attachments) is for the sole use of the intended recipient(s) and may contain confidential and privileged information. If the reader of this message is not the intended recipient, you are hereby notified that any dissemination, distribution or copying of this message (including any attachments) is strictly prohibited. If you have received this message in error, please contact the sender by reply e-mail message and destroy all copies of the original message (including attachments). _______________________________________________ BioCatalogue-hackers mailing list BioCatalogue-hackers at rubyforge.org http://rubyforge.org/mailman/listinfo/biocatalogue-hackers -------------- next part -------------- An HTML attachment was scrubbed... URL: From tod.jackson at emory.edu Fri Oct 29 13:08:24 2010 From: tod.jackson at emory.edu (Jackson, Tod) Date: Fri, 29 Oct 2010 13:08:24 -0400 Subject: [BioCatalogue-hackers] Does/will BioCatalogue support wsdl:import In-Reply-To: Message-ID: However, once that wsdl:import is processed, here is a snippet from the big_index_bindings.wsdl file (i.e., the WSDL that was imported in the first one): . . This second import does look relative. So, it looks like the first import has an absolute location but the second one doesn?t? Therefore, it can?t ?follow? the imports all the way down?? Does that sound right? Thanks, Tod On 10/29/10 11:37 AM, "Jackson, Tod" wrote: Here?s an example just so everything is clear, this is at: http://grid.poc.actsi.org/wsrf/services/DefaultIndexService?wsdl This is the WSDL returned initially (like when you access the above URL from a browser). It?s also what BioCatalogue gets back in it?s initial request (like when you try to preview the service). I?ve tried to ?follow? the import programatically in my local deployment because I didn?t think it looked like it was currently supported. However, as I went further with my mods, it started looking like there was perhaps more to it than I initially thought. I disabled WSDLUTILS_BASE_URI so I could force it through my own local wsdl_parser.rb to get a better handle on how things were behaving. I?m still trying to gather more information as to how ?deep? it ultimately goes but when I follow this first import and merge those results with the first ?wsdl_hash? that gets created from this, I started running into other issues. So, I figured it was best to check with you all before I took it too far. Thanks! Tod On 10/29/10 11:22 AM, "Eric Nzuobontane" wrote: Hi Tod, I am on my way out now but I will look into this later. One thing I would like to know is if those imports are schema imports and whether they have schema locations. If they do and the schema location relative path, then BioCatalogue may have issues with this as it does not try to resolve relative paths for schema locations. regards, Eric On 29/10/2010 15:47, Jackson, Tod wrote: Does/will BioCatalogue support wsdl:import Greetings BioCataloguers, My colleagues and I here at Emory University and the Atlanta Clinical & Translational Science Institute are attempting to register some services in our local Emory BioCatalogue deployment. These services are authored using caGrid tools and hosted in various caGrid deployments both internally and externally. I believe many of you are familiar with the caGrid project (http://cagrid.org). It appears as though most, if not all of the WSDL definitions associated to these ?grid services? leverage ?wsdl:import? to fully resolve themselves. After some quick analysis, it does not appear as though BioCatalogue is able to fully resolve service definitions that use wsdl:import. I am able to load the services using soapUI and Taverna however. So, my first question is, am I missing anything regarding BioCatalogue?s support for this? i.e., is there a configuration setting or something that might make this work? Assuming the answer to that is ?no?, my second question is, are there any plans to add support for this in future versions of the application? As always, thanks for any input you might have on this. Best Regards, Tod Jackson ________________________________ This e-mail message (including any attachments) is for the sole use of the intended recipient(s) and may contain confidential and privileged information. If the reader of this message is not the intended recipient, you are hereby notified that any dissemination, distribution or copying of this message (including any attachments) is strictly prohibited. If you have received this message in error, please contact the sender by reply e-mail message and destroy all copies of the original message (including attachments). _______________________________________________ BioCatalogue-hackers mailing list BioCatalogue-hackers at rubyforge.org http://rubyforge.org/mailman/listinfo/biocatalogue-hackers -------------- next part -------------- An HTML attachment was scrubbed... URL: