Camelia, the Perl 6 bug

IRC log for #bioclipse, 2008-08-27

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
04:37 egonw joined #bioclipse
04:58 thomas_ku joined #bioclipse
05:00 egonw moin
05:00 zarah hi egonw
05:34 Gpox joined #bioclipse
06:51 olas joined #bioclipse
06:51 olas moin
06:51 zarah hi olas
07:30 jonalv joined #bioclipse
07:35 masak joined #bioclipse
07:44 CIA-51 bioclipse: ospjuth * r6749 /bioclipse2/trunk/plugins/net.biocl​ipse.cdk.qsar.test/src/net/bioclips​e/cdk/qsar/test/TestCDKQsar.java: Correctly reference DescriptorImpl, not Descriptor
07:46 CIA-51 bioclipse: ospjuth * r6750 /bioclipse2/trunk/plugins/net.bioclipse.qsar/ (3 files in 2 dirs): Removed category in decriptorImpl. Added convenience methods in manager.
07:46 CIA-51 bioclipse: ospjuth * r6751 /bioclipse2/trunk/plugins/ (3 files in 3 dirs): Removed category for descriptorImpl.
07:51 jonalv hi
07:51 zarah hi jonalv
07:54 masak hi!
07:54 zarah hi masak
07:54 masak hi zarah
07:56 CIA-51 bioclipse: goglepox * r6752 /bioclipse2/trunk/plugins/net.bioclipse.str​ucturedb.tests/tests/net/bioclipse/structur​edb/business/StructuredbManagerTest.java: Workaround for bug in java 1.5 on OS X. Sets thread class loader if null
08:08 maclean joined #bioclipse
08:12 jonalv Gpox++
08:14 maclean Are there updates for the new JCP or something?
08:23 jonalv maclean: sry?
08:23 maclean Oh, I just wondered why Gpox was getting credit, is all.
08:24 maclean I'm curious by nature.
08:24 olas Gpox++
08:24 jonalv he figured out my strange bug with xml reading on OS X
08:24 olas how did you solve it?
08:24 jonalv (I really must test it on Windows now too)
08:27 maclean Oh right, in structuredb this is.
08:28 jonalv yup
08:29 maclean well, it seems to ...build and run on windows.
08:30 jonalv maclean: yea but I wonder how many tests passes...
08:31 maclean I'll download structuredb.tests plugin...
08:40 maclean Hmm. The AllStructuredbTestsSuite didn't work.
08:41 jonalv in what way?
08:42 maclean NullPointerException at org.tigris.subversion.subclipse.core....
08:42 maclean I may be running it wrong, of course.
08:42 jonalv in subclipse?
08:43 jonalv what does subclipse does to do with anything?
08:43 zarah jonalv: C-u C-\ swedish-postfix RET
08:43 jonalv strange...
08:43 maclean I have no idea.
08:43 jonalv zarah: shut up!
08:43 zarah jonalv: :)
08:44 maclean Maybe I'm missing another plugin.
08:46 jonalv maclean: don't know. but if ti compiles you should have everything...
08:46 maclean It does compile (I didn't have core.tests, for the MockIFile, but once I did it did).
08:46 CIA-51 bioclipse: carl_masak * r6753 /bot/trunk/zarah.pl: [zarah] now only annoys jonalv about key mappings on #farmbio
08:46 masak @update
08:47 zarah joined #bioclipse
08:59 edrin joined #bioclipse
09:09 jonalv hm Gpox's fix wasn't bullt proof :(
09:12 maclean :( Oh well - it's better than it was though?
09:15 jonalv well it still doesn't work but we are on it perhaps...
09:25 jonalv okey using the uggliest hack off the year we actually got it somewhat working...
09:26 jonalv but let's just say that switching classloaders of the active thread for each method call that happens to be using cml probably isn't the way t odo this...
09:27 maclean Probably not. :)
09:27 jonalv I don't want to know what happends when we get a thread the needs both this classloader and another... :(
09:28 jonalv @seen egonw
09:28 zarah I have not seen egonw
09:28 masak strange
09:31 masak @seen malic1
09:31 zarah malic1 was last seen 2008-08-27T08:18:51 in #metware saying 'Pepseeker is another BioMart/spectra proteomics system that uses SpectralViewer from PRIDE'
09:45 shk3 joined #bioclipse
09:46 jonalv shk3: hi
09:46 jonalv shk3: do you know where I can get in touch with the cml peole?
09:46 jonalv |people
09:51 olas jonalv: https://lists.sourceforge.ne​t/lists/listinfo/cml-discuss
09:51 jonalv olas: yea my mail is awaitning acceptance... :/
09:52 olas oh
09:52 jonalv would be way better to find one of them on irc...
10:13 * jonalv <~ food
10:17 shk3 jonalv: not sure about irc, but you could mail pmr or jim downing directly
10:36 CIA-51 bioclipse: shk3 * r6754 /bioclipse2/trunk/plugins/net.bioclipse.spectrum/ (13 files in 3 dirs): added a new spectrum wizard
10:45 shk3 does anybody know how to open a file in an editor in bc2?
10:45 shk3 from code, obviously
10:47 edrin hi
10:47 zarah hi edrin
10:57 olas shk3: yes
10:57 olas mom
10:58 olas IUIManager.open(IFile)
10:58 olas it's in plugin net.bioclipse.ui.business
10:58 olas call that manager code instead of IDE.openEditor(IFile)
10:58 olas that means it's recorded
10:59 olas and we want as much business code as possible to go via managers
11:05 shk3 thanks
11:20 shk3 olas:
11:20 shk3 in net.bc.ui.business, nothing is exported
11:20 shk3 is that intended?
11:21 shk3 it needs to be exported to be used, I believe?
11:26 olas yes
11:26 olas please export the packages
11:27 olas get the manager via Activator.getDefault().getUIManager
11:29 shk3 this method returns object - changed it to UIUManager
11:31 shk3 IUIManager, that is
11:33 CIA-51 bioclipse: shk3 * r6755 /bioclipse2/trunk/plugins/​net.bioclipse.ui.business/ (2 files in 2 dirs): export and changed return type of getUIManager
11:34 CIA-51 bioclipse: shk3 * r6756 /bioclipse2/trunk/plugins/net.bioclipse.spectrum/ (4 files in 3 dirs): new file gets opened
12:12 malick joined #bioclipse
12:17 shk3 olas: did anybody find a solution for the menu item visibility problem?
12:18 olas no
12:18 olas please refresh my memory
12:18 olas and I will look at it again
12:21 shk3 net.bc.cdk.ui
12:21 shk3 plugin.xml
12:21 shk3 line 74 ff
12:22 olas mom, in tel
12:24 shk3 91 ff actually
12:24 shk3 the smiles stuff is something different
12:25 egonw joined #bioclipse
13:11 zarah joined #bioclipse
13:18 CIA-51 bioclipse: jonalv * r6757 /bioclipse2/trunk/plugins/​net.bioclipse.structuredb/ (2 files in 2 dirs): Ugly hack not fixing (but at least for some cases) the problems with cml parsing on mac. Hopes for new fixed version of the troublesome library...
13:18 egonw joined #bioclipse
13:22 olas egonw!
13:22 olas help!
13:22 olas did you see my email?
13:22 olas please enlightn me
13:22 egonw yes
13:22 egonw OWL is more recent, but not sure I did and did not convert yet
13:23 egonw there is an algorithm dict, which is used by the QSAR descriptor dict
13:23 egonw being algorithms too
13:23 egonw the former I converted into OWL, not sure if I did the QSAR stuff...
13:23 egonw (of the top of my head)
13:23 olas but the classifications in the descriptors refer to the XML file?
13:23 egonw I'd say, use what the CDK uses
13:23 olas damn
13:23 egonw I know this needs updating, but never got around to that...
13:23 olas it uses the XML
13:24 egonw I can fix it in a day or so..
13:24 egonw but don't have the time right now
13:24 olas and I have spent 3 days setting bioclipse up to use the OWL
13:24 olas :-(
13:24 olas :-(
13:24 egonw well, consider it failed synthesis
13:24 egonw you gained OWL knowledge
13:24 egonw which you'll use later
13:24 olas got no time for this
13:25 olas seriously, is there anything in the net.sf.qsar project thats working or up to date?
13:25 olas don't wanna waste more time
13:25 olas thought I could use it out of the box
13:25 olas but apparantly not
13:25 olas hmm
13:25 egonw I would need to check
13:26 egonw been a while since I looked at it...
13:26 olas grand plans
13:26 egonw yielded a paper
13:27 olas so no maintenance plan from any of the involved?
13:27 olas sad
13:27 olas it's a good idea
13:27 olas ok
13:27 egonw well, it fits with my plans in Uppsala...
13:27 olas right
13:27 egonw this is one reason why I like to move closer to chemoinformatics again
13:27 egonw but right now... still in bioinformatics
13:28 olas I see
13:28 egonw OK, what do you decide?
13:28 egonw a bit of XSLT is easily written to convert the QSAR xml to OWL
13:28 olas update all CDK descriptors to point to OWL namespace
13:28 egonw the one I used in the past should still be in SVN somewhere
13:28 olas cdk descriptors should reference a descriptor implementation
13:29 jonalv /
13:29 olas egonw: please explain this:
13:29 egonw is CDK using the OWL version?
13:29 egonw mom... booting Eclipse...
13:29 olas the descriptor-implementations is an OWL for implementations, right?
13:30 egonw full URL?
13:30 olas the descriptor XML -> XHTML is the one put on the web page
13:30 olas in net.sf.qsar svn
13:30 olas anyway
13:30 egonw checking...
13:30 olas are there two ontologies?
13:30 olas one for descriptors and one for descrptor implementations?
13:31 egonw http://qsar.svn.sourceforge.net/viewvc/qsar/trunk/​bo-dicts/chemoinformatics-algorithms.owl?view=log
13:31 egonw that's the latest
13:31 egonw 23 months
13:31 egonw that's chemoinformatics oriented...
13:31 egonw algorithms, implementations...
13:31 olas that's algorithms
13:31 egonw but not focusing on QSAR algorithms
13:31 olas well
13:31 egonw which would be:
13:31 egonw http://qsar.svn.sourceforge.net/viewvc/qsar/trun​k/qsar-dicts/descriptor-algorithms.owl?view=log
13:32 olas yes
13:32 olas that's the one I use
13:32 olas but if you look in any of CDK Descriptors, they reference the XHTML
13:32 egonw that should be the OWL version of the qsar-descriptor.xml and -metadata.xml in the same folder
13:32 olas let me give you an example
13:33 egonw yeah, that's because I never released the OWL
13:33 olas in owl:
13:33 olas <Descriptor rdf:ID="hBondDonorsDaylight"> <Descriptor rdf:ID="hBondDonorsBoehmKlebe">
13:33 * egonw is now checking if the CDK reads the XML or OWL
13:33 olas on XML: #hBondDonor
13:33 egonw I though Rajarshi updated the code to use OWL... mom
13:34 olas see this: http://paste2.org/p/66371
13:34 olas clearly that descriptor references the web pagehBondDonors, not one of the implementations
13:35 olas should it not point to: http://www.blueobelisk.org/ontologies/chemo​informatics-algorithms/#hBondDonorsDaylight
13:37 olas egonw?
13:37 egonw yes, checking
13:37 olas how do I get the OWL reference from a class?
13:37 egonw not sure why that points to the bo.org website...
13:37 egonw because that dict, is really the chemoinfo algorithm dictionary...
13:39 olas I want to read the OWL, then read CDK and match the impl in CDK to the ones in OWL
13:39 egonw ack
13:39 olas but clearly, there's an additional thing I haven't thought of
13:40 olas the qsar-impl.xml does not cover the general descriptor, but only cares about specific implementations
13:40 olas egonw: time to skype for 5 mins?
13:40 olas just need some answers
13:40 olas got another skype call at 16
13:41 jonalv olas: talking about sjype. I think we have a meeting about 20 mins, correct?
13:41 jonalv ... :)
13:41 olas olas (3:44:41 PM): got another skype call at 16
13:41 jonalv <jonalv> ... :)
13:42 masak <jonalv> <jonalv> ... :)
13:44 egonw http://qsar.svn.sourceforge.net/viewvc/qsar/trunk​/qsar-dicts/descriptor-algorithms.owl?view=markup
14:01 CIA-51 bioclipse: carl_masak * r6758 /bioclipse2/trunk/plugins/net.bioclip​se.jmol/src/net/bioclipse/jmol/views/​outline/JmolContentOutlinePage.java:
14:01 CIA-51 bioclipse: [JmolContentOutlinePage]
14:01 CIA-51 bioclipse: * removed unused imports
14:01 CIA-51 bioclipse: * removed unused fields and local variables
14:01 CIA-51 bioclipse: * fixed whitespace
14:01 CIA-51 bioclipse: * shortened long lines
14:02 maclean yay for code cleanup! masak++
14:05 masak thanks. there are more such files, though, and only one of me...
14:05 maclean I never used to care about unused imports until I started using eclipse :)
14:06 masak I never _caused_ unused imports until I started using Eclipse
14:06 masak well, practically never
14:06 masak but I guess it's rather a question of code size/complexity
14:06 maclean it does help to know which classes a class needs.
14:07 maclean I mean, when reading someone else's class.
14:07 masak above a certain threshold, one gives up and hands the code to the IDE
14:07 masak yes.
14:07 masak I like the feeling of no warnings in a file.
14:07 masak it gives me a sense of control.
14:08 maclean Me too.
14:08 masak maclean++
14:08 maclean It can also help to mark classes that need work.
14:17 CIA-51 bioclipse: jonalv * r6759 /bioclipse2/trunk/plugins/ (11 files in 3 dirs): Created a bunch of dao tests.
14:26 olas egonw: a problem
14:26 olas the owl file in CDK is not parseable
14:27 olas there are many occasions where it complains about mixed XML content without proper rdf:parseType='Literal'
14:27 olas can I just replace all <description> with     <description rdf:parseType='Literal'>
14:27 egonw yes
14:27 olas ok, will try that
14:57 CIA-51 bioclipse: ospjuth * r6760 /bioclipse2/trunk/plugins/net.bioclipse.​qsar/ontology/descriptor-algorithms.owl: Curated and updated ontology. Now in synch with CDK.
14:57 CIA-51 bioclipse: ospjuth * r6761 /bioclipse2/trunk/plugins/net.bioclipse.​qsar/ontology/descriptor-algorithms.owl: Curated and updated ontology. Now in synch with CDK.
14:58 egonw olas++
14:58 egonw read email, looks goog
14:58 egonw doog
14:58 egonw good
14:59 jonalv good dog. nice puppy :)
15:02 CIA-51 bioclipse: ospjuth * r6762 /bioclipse2/trunk/plugins/net.bioclipse.qsar/ (3 files in 2 dirs): Even more updates to ontology and model reading.
15:36 CIA-51 bioclipse: ospjuth * r6763 /bioclipse2/trunk/plugins/net.bioc​lipse.qsar/src/net/bioclipse/qsar/ (4 files in 2 dirs): Updated reading in Jena after ontology cleanup. Now we can use # again everywhere!
15:36 CIA-51 bioclipse: ospjuth * r6764 /bioclipse2/trunk/plugins/net.bioclipse.qsar.test/ (2 files in 2 dirs): Adapted tests for removed methods and changed plugin.xml.
15:36 CIA-51 bioclipse: ospjuth * r6765 /bioclipse2/trunk/plugins/net.biocl​ipse.cdk.qsar.test/src/net/bioclips​e/cdk/qsar/test/TestCDKQsar.java: Updated CDK descriptor tests to test correct methods.
15:37 gilleain joined #bioclipse
15:37 olas clone(maclean)
15:40 gilleain Prepare to catch a CloneNotSupportedException...
15:41 gilleain I recommend that you use my copy constructor like so : copy_of_me = new maclean(gilleain)
15:41 jonalv :)
15:42 CIA-51 bioclipse: jonalv * r6766 /bioclipse2/trunk/plugins/net.bioclipse.s​tructuredb/src/net/bioclipse/structuredb/ (42 files in 9 dirs): [Structuredb] Code cleanup
15:49 maclean joined #bioclipse
16:04 malick joined #bioclipse
16:11 CIA-51 bioclipse: ospjuth * r6767 /bioclipse2/trunk/plugins/net.biocl​ipse.cdk.qsar.test/src/net/bioclips​e/cdk/qsar/test/TestCDKQsar.java: Clearly mark those not in ontology.
16:11 CIA-51 bioclipse: ospjuth * r6768 /bioclipse2/trunk/plugins/net.biocli​pse.qsar.test/src/net/bioclipse/qsar​/test/business/TestQsarManager.java: Verified descriptor categories from new OWL.
16:15 malick left #bioclipse
17:01 edrin left #bioclipse
18:51 egonw joined #bioclipse
21:08 olas left #bioclipse

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary