Camelia, the Perl 6 bug

IRC log for #bioclipse, 2009-05-01

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

All times shown according to UTC.

Time Nick Message
06:06 egonw joined #bioclipse
08:10 shk3 joined #bioclipse
08:12 rojasm joined #bioclipse
08:12 egonw rojasm: ping
08:13 rojasm ei egown
08:13 egonw rojasm: did you reply yet on @cdk.threadsafe ?
08:13 rojasm I didn't check the mails yet
08:13 rojasm I do see now
08:47 edrin joined #bioclipse
08:53 shk3 hi all
08:53 zarah privet shk3
08:53 egonw hi shk3
08:53 egonw rojasm: tack
08:53 shk3 hi egonw
08:54 shk3 it looks like the "dynamic columns in sdf viewer" has been done.
08:54 shk3 I that correct?
08:54 shk3 I=Is
08:54 egonw ask Gpox
08:54 egonw but today is a holiday in Sweden
08:54 egonw so not sure he will be online
08:55 egonw the moltable is his and Ola's doing
08:55 egonw I see screenies now and then, but not informed about the exact state
08:55 shk3 ah, sweden's the doing mayday on the proper date
08:56 shk3 did you ever try adding help to a wizard?
08:56 egonw no, never tried that
08:58 shk3 it seems to be possible, but the recommended methods have no effect for me
09:17 CIA-73 bioclipse: shk3 * r10022 /bioclipse2/trunk/plugins/net.bioclipse.cdk.u​i.sdfeditor/src/net/bioclipse/cdk/ui/wizards/ (DotViewerFilter.java SelectFilesWizardPage.java): .x files no longer show up in browser in new sdf viewer
09:20 olass joined #bioclipse
09:20 olass hi
09:20 zarah privet olass
09:20 egonw hi olass
09:20 olass hi egonw
09:21 * egonw has a burned skin
09:21 egonw of too much sun :)
09:21 olass what a nice problem
09:21 olass egonw: did you have a good time yesterday?
09:21 * olass had
09:21 egonw yes
09:21 olass sorry we could not meet up
09:21 egonw the white head thingy was somewhat disappointing :)
09:21 olass things just got so stressed
09:21 shk3 hi olass
09:21 egonw just a very mild waving of the head in the air :)
09:22 olass his shk3
09:22 egonw olass: no worries
09:22 olass egonw: actually, I've never seen it :)
09:22 egonw nothing missed
09:22 shk3 it's labour day, so may I ask a short question?
09:22 egonw shk3: ah no
09:22 olass labour day?
09:22 egonw May 1 seems to be recovery day here
09:22 egonw like in NL
09:22 olass yes, here we are off
09:22 * olass works anyway
09:22 olass shk3: just fire off!
09:23 shk3 ok
09:23 shk3 is it correct that the dynmaic sdf columns-thing has been done?
09:23 shk3 i. e. can I now add a column at runtime to sdf table?
09:24 olass shk3: no
09:24 olass not yet
09:24 shk3 :-(
09:24 shk3 ok
09:24 shk3 thanks
09:24 olass very close I think
09:24 olass but not yet
09:24 olass will be in 2.0
09:24 shk3 no problem
09:24 olass (that is the plan)
09:24 shk3 how are things going altogether?
09:24 olass shk3: can you calculate tanimoto distance between molecules with cdk?
09:24 shk3 I try to fix bugs
09:24 shk3 yes
09:24 olass shk3: it is very much up to Gpox now
09:24 shk3 that's what I want to add
09:25 shk3 choose a molecule and add tanimoto distance to sdf table
09:25 olass shk3: always implement the manager call first with junit tests
09:25 shk3 I can do that
09:25 shk3 will do so
09:25 olass shk3: try to make it general
09:25 shk3 yes
09:25 olass to work in any use case you can imagine
09:25 olass from console, gui, one and many molecules etc
09:26 olass that is how maanager methods should be
09:26 olass calling from gui is just one use case
09:26 shk3 sure
09:26 shk3 planned using the manager
09:26 olass shk3: we must
09:26 shk3 will do it for one and many molecules
09:26 olass all business code must be in manager
09:26 shk3 i think it's not possible as a context menu or so
09:26 olass we cannot have business code in UI components
09:27 shk3 since you need two molecules in any case
09:27 olass shk3: what do you mean?
09:27 shk3 well, you could pop up a wizard
09:27 olass ?
09:27 shk3 well, I thought about a context menu "calculate tanimoto"
09:27 shk3 but that must refer to two molecules
09:27 olass shk3: if you select two molecules in navigator, you can calculate tanimoto between them
09:27 shk3 yes
09:28 shk3 but only for two
09:28 shk3 exactly two, I mean
09:28 olass eclipse supports exactly two selections
09:28 shk3 yes, but
09:28 olass that means the action is enabled for exactly two selected molecules
09:28 shk3 would be more general to
09:28 olass shk3: please write a general manager code
09:28 shk3 choose the reference molecule in a wizard
09:29 shk3 and have the popup on 1 to x molecules, I think
09:29 olass shk3: that is just another single use case
09:29 shk3 the manager is the basic thing anyway
09:29 shk3 sure
09:29 shk3 just updating bc
09:29 olass shk3: maybe you could formulate the many use cases on a wiki page?
09:29 shk3 that takes a while
09:29 olass that way people like me could add use cases
09:29 shk3 will do so
09:30 olass please email devel ML so all can contribute there
09:31 olass egonw: would be great if you could look at this tanimoto use cases/implementations/tests too
09:33 shk3 writing the page...
09:35 shk3 http://wiki.bioclipse.net/inde​x.php?title=Tanimoto_Use_Cases
09:35 zarah shk3's link is also http://tinyurl.com/czjs5l
09:35 shk3 please comment and add
09:37 olass shk3, egonw: what other distances can CDK calculate apart from tanimoto?
09:37 shk3 none, as far as i know
09:37 shk3 but normally these are easy to implement
09:37 shk3 so feel free to suggest
09:38 olass I will write my thoughts on wiki
09:40 shk3 cdk has something called DistanceMoment, which seems to be for 3d. I will make this a manager as well
09:40 shk3 manager method, this is.
09:40 shk3 should both be part of cdk manager, I suppose?
09:45 shk3 I had problems with eclipse versions in manifest files again.
09:45 shk3 shall I commit removal of them?
09:48 egonw CDK can calculate other distances too, but not sure if the have been formalized into separate classes
09:48 egonw distance based on descriptors
09:48 egonw MCSS
09:48 egonw that sort of things
09:59 olass ok, will write up my thoughts on the wiki
09:59 olass egonw: please comment later, will ping you
09:59 egonw ack
10:00 olass shk3: will ping you too :)
10:03 olass egonw: is tanimoto a distance metric?
10:03 olass or a distanceType?
10:04 egonw metric
10:07 olass egonw: please comment on http://wiki.bioclipse.net/inde​x.php?title=Tanimoto_Use_Cases
10:07 zarah olass's link is also http://tinyurl.com/czjs5l
10:07 edrin hi
10:07 zarah saluton edrin
10:07 olass shk3: please see the above link for my suggestions
10:07 olass hi edrin
10:08 * olass will take a shower now, be back soon
10:08 edrin just now updated the xws4j sf site
10:20 egonw joined #bioclipse
10:32 CIA-73 bioclipse: shk3 * r10023 /bioclipse2/trunk/plugins/net.biocli​pse.cdk.ui.sdfeditor/src/net/bioclip​se/cdk/ui/wizards/NewSdfWizard.java: no empty files get created
10:41 olass shk3: egon favoured your suggestion
10:41 olass see wiki page
10:42 olass and please use BioList for implementation and List as exposed interface
10:42 egonw olass: distance measures are not so common that generalizing is worth it
10:42 olass you think so?
10:42 olass it seems to appear all the time in examples I see
10:44 egonw true
10:44 egonw but there is no large set of well established measures
10:44 egonw Tanimoto is one...
10:44 egonw but for the rest, people do something ad hoc
10:48 edrin do you know of any nucleotide prediction service that predicts something that takes some time?
10:49 stain_ any of the EBI services?
10:49 edrin stain_: ok ;)
10:49 edrin any suggestion?
10:50 egonw stain_: but not depending on the impl but on the algorithm
10:50 edrin for a specific one?
10:50 stain_ http://www.ebi.ac.uk/Tools/web​services/services/interproscan
10:50 edrin yes
10:50 egonw impl = implementer
10:50 zarah stain_'s link is also http://tinyurl.com/cdkl5x
10:50 egonw stain_: edrin seeks to reproduce the service using XMPP, I think
10:50 edrin i already have interproscan
10:50 edrin i just tried
10:50 stain_ ah.. because it's a good example of an asynchronous web service
10:50 stain_ which of course there's no common interface for
10:51 edrin took 2 minutes for my protein, that's already really ok for me. now i need some service for nucleotide
10:51 CIA-73 bioclipse: shk3 * r10024 /bioclipse2/trunk/plugins/net.bioclipse.chemoinfo​rmatics/src/net/bioclipse/chemoinformatics/conten​tlabelproviders/MoleculeFileContentProvider.java: The MoleculeFileContentProvider only returns directories which contain molecules (or directories which contain molecules (or directories which contain molcules (put in recursion here))
10:52 edrin stain_: interproscan is cool, it uses pfam
10:52 edrin (and others)
11:17 * egonw is relocation to BMC, to pick up some stuff, make some prints, etc.
11:17 egonw bbl
11:20 edrin1 joined #bioclipse
12:17 egonw joined #bioclipse
12:21 CIA-73 bioclipse: shk3 * r10025 /bioclipse2/trunk/plugins/net​.bioclipse.chemoinformatics/ (4 files in 3 dirs): added a method for testing if a file is a molecule
12:25 CIA-73 bioclipse: shk3 * r10026 /bioclipse2/trunk/plugins/net.bioclipse​.cdk.ui.sdfeditor/src/net/bioclipse/cdk​/ui/wizards/SelectFilesWizardPage.java: finish only active is someting is selected to end up in the file
12:52 olass shk3: why r10025?
12:53 olass we have content types for that
12:54 shk3 yes
12:54 shk3 have a look at the code please
12:54 shk3 it uses content types
12:54 shk3 but it needs to test for more than one
12:55 shk3 so repeating this code isn't too good
12:55 shk3 escpecially since the list needs updating if new
12:55 shk3 molecule content types come up
12:55 shk3 if you have a better way, please go ahead
12:55 olass shk3: no time to look at your code now
12:55 olass please comment excessively in the code
12:55 olass with limitations
12:56 olass and file a bug for me to review
12:56 olass shk3: when is this code used and why?
12:57 olass I have a feling that this should not be needed if declared properly in plugin.xml
12:57 shk3 in the molecule conent type provider
12:57 shk3 sorry, but i am getting annoyed
12:57 olass why can't that use content types?
12:57 olass annoyed?
12:57 shk3 WE DO
12:57 shk3 USE CONTENT TYPES
12:57 shk3 you asked me for that
12:58 shk3 please either take time to look at things
12:58 shk3 or don't comment
12:58 shk3 it's not ok to say I have no time
12:58 shk3 but blame you anyway
12:59 shk3 I spend my time doing the things you asked for
13:00 shk3 the code would be easier if all molecule content types would inherit from the same type
13:01 shk3 but they don't
13:03 shk3 ok
13:03 shk3 sorry for getting angry
13:04 shk3 do we have an example for expanding to a file in the navigator?
13:04 olass no
13:05 shk3 hm
13:05 shk3 ok
13:08 shk3 will file a bug report for the molecule file type detection
13:13 egonw shk3: these patches look good to me
13:13 egonw olass: ping
13:13 egonw shk3: except for the fact that if a file does not exist
13:14 egonw then it needs not throw an exception, just return false
13:15 shk3 ok
13:15 egonw olass: the idea makes sense to me... we do not want to update code all over the place to define what an IMolecule is
13:15 egonw shk3: also, I want to encourage a you to suggest on where we should put final static string for the content types...
13:16 CIA-73 bioclipse: shk3 * r10027 /bioclipse2/trunk/plugins/net.biocli​pse.cdk.ui.sdfeditor/src/net/bioclip​se/cdk/ui/wizards/NewSdfWizard.java: new sdf is revealed
13:16 egonw shk3: I am even tended to say that this class could even be dedicated to supported IMolecule content types
13:17 egonw instead of making the class a swiss knive
13:17 shk3 we could do so
13:17 shk3 sure
13:17 shk3 the strings are not used somewhere else right now
13:17 * edrin1 likes swiss knives
13:17 shk3 leatherman for me
13:17 egonw edrin1: they are really bad for playing soccer, though
13:17 shk3 we could still make them constants
13:18 shk3 I think in the same class then
13:18 egonw shk3: possibly
13:18 egonw I'm thinking about it...
13:18 egonw as we will want a unit test to see of the constants are actually existing content types...
13:19 CIA-73 bioclipse: shk3 * r10028 /bioclipse2/trunk/plugins/net.bioclipse.c​hemoinformatics/src/net/bioclipse/chemoin​formatics/util/ChemoinformaticUtils.java: non-existing file return false
13:19 olass egonw: I'm fine with all solutions where things are duplicated as long as a test will fail when things change in either place
13:19 olass because we have had problems with this before
13:19 olass duplicated code
13:20 olass and changing one broke the other
13:20 egonw umm... the idea here is to remove code duplication
13:20 egonw not introduce it...
13:20 edrin1 wow, it takes already 4 minutes to just delete the binding files for the few services created with bc...
13:20 olass egonw: yes
13:20 olass but hardcoding content types in code is duplicating their use in plugin.xml
13:20 olass we COULD read from the plugin registry each time
13:20 olass or
13:21 olass well
13:21 egonw olass: and then?
13:21 olass maybe that is not a good idea
13:21 egonw the registry will not tell you if something is a SingleMolecule type
13:22 olass hmm, I guess not
13:22 edrin ~6 minutes
13:22 olass and having super types is not good
13:22 olass since xml and text are supertpes
13:22 egonw that would work
13:22 edrin these bio schemata are really big
13:22 egonw but not with a single super type
13:22 egonw see my blog post recently
13:22 shk3 we could have a textBasedMoleculeType as super type
13:23 shk3 and an xmlBasedMoleculeType
13:23 shk3 that would reduce checking to two types
13:23 shk3 I guess
13:23 olass I guess if we do it the way shk3 does it now, there is only one method to update in CDKManager
13:23 olass maybe that is the preferred solution
13:24 egonw what method?
13:24 egonw I don't think we have funciontliaty similiar in the CDKmanager... ?
13:24 olass ?
13:24 olass r10025
13:24 * olass can't look at it now but thought that was the thing
13:25 olass hmm
13:25 egonw it's related but not the same
13:25 olass yes, not in cdk
13:25 olass forgive me
13:25 egonw CDKManager looks up the chem format type
13:25 egonw != content type
13:25 edrin olass: the service "description" for the HIV RT Drug thing is the same as for the "HIV Protease" drug, is this correct=
13:25 edrin ?
13:26 olass probably not
13:26 olass can you file me a bug?
13:27 edrin ok
13:31 egonw mexican flu in denmark
13:33 edrin i think it's in munich already
13:34 edrin afaik they still speak of ~200 infected in mexico. interesting that these 200 infected already ~50-100 (?) foreigners, although there live 20.000.000 people in Mexico city.
13:35 edrin maybe these 200 were all working at the mexico city airport checkin ;)
13:35 egonw and I just read a German woman who has not been to Mexico
13:35 edrin yes
13:36 egonw blah...
13:36 egonw it's a real Fri
13:36 egonw trouble focussing :(
13:36 shk3 I think what's important is
13:36 shk3 that we get out bc2 release before we all die of swine flu
13:36 edrin heh :)
13:37 egonw :)
13:37 shk3 I am busy checking bugs, btw
13:37 shk3 I hope you are too
13:39 edrin egonw: I just killed your CDK service by requesting the schemata
13:39 egonw edrin: I'll check
13:39 CIA-73 bioclipse: shk3 * r10029 /bioclipse2/trunk/plugins/net.bioclipse.sp​ectrum.test/src/net/bioclipse/spectrum/bus​iness/test/SpectrumManagerPluginTest.java: fixes bug #268
13:39 zarah bug #268 | http://tinyurl.com/ccczo6
13:45 shk3 is it ok to make the cdk.ui plugin depend on chemoinformatics plugin?
13:45 shk3 shouldn't be a problem, should it?
13:49 CIA-73 bioclipse: shk3 * r10030 /bioclipse2/trunk/plugins/net.bioclipse.cdk.ui/ (2 files in 2 dirs): the new from smiles wizard did not suggest a file name, which we decided for to be standard, to my understanding.
13:53 egonw shk3: mom, checking
13:57 egonw at first glace it looks ok
13:58 shk3 i think so as well
13:58 shk3 i was afraid of some circular dependency or so
13:58 shk3 but can't see one
13:58 shk3 another question
13:59 shk3 sorry, but I try to get things done today
13:59 shk3 the tests are extending the abstractmanagertest
13:59 shk3 this has some tests
13:59 shk3 which also get executed when the tests are run
13:59 edrin egonw: i killed cdk again by requesting schema for SMILES2CML !
14:00 shk3 testForFooString and alike
14:00 egonw edrin++
14:00 edrin requesting for calculateMass and addExplicitHydrogens worked
14:00 shk3 do these count for the statistics on the stable managers test?
14:00 egonw yes
14:00 edrin egonw: could not test other functions because then it crashed...
14:00 egonw edrin: checking...
14:01 shk3 ok
14:01 shk3 thanks
14:01 edrin (means: the generate2D and 3D were not checked yet...,
14:01 egonw edrin: no stacktrace in the log, but I see:
14:02 egonw XWS4J: [Status] OUTBOUND opened changed to disconnected): <stream:stream xmlns:stream='http://etherx.jabber.org/streams' xmlns='jabber:component:accept' to='ws1.bmc.uu.se'/>
14:02 egonw XWS4J: [Status] INBOUND opened changed to disconnected): <stream:stream xmlns:stream='http://etherx.jabber.org/streams' xmlns='jabber:component:accept' from='localhost' id='4108146241'/>
14:02 egonw XWS4J: SessionLoop.run() - End of Thread.
14:02 egonw XWS4J: SessionLoop.run() - Exception: Connection reset by peer
14:02 zarah egonw's link is also http://tinyurl.com/3bx9zm
14:02 zarah egonw's link is also http://tinyurl.com/3bx9zm
14:04 edrin egonw: in this case i would say that you sent illegal xmpp stanza
14:04 edrin maybe the schemata were not correctly read or are invalid?
14:05 egonw dunno
14:05 egonw nothing changed in that respect...
14:05 edrin i am sure the cdk is disconnected from ejabberd because ejabberd disliked the received XMPP stanza
14:06 egonw yeah, possible...
14:06 egonw mom
14:06 egonw ok, the service is restarted
14:06 egonw please request schema one by one
14:06 egonw ok, still fine
14:06 edrin ok, i will do the SMILE2CML as last
14:07 egonw calmass
14:07 egonw I see a CML schema come bye
14:07 egonw I see a CML schema come by
14:07 egonw and a reset
14:07 egonw disconnect type
14:08 edrin ok, seems the generat2d was ok, but the generate3d crashed it, too
14:08 edrin i guess the generate3Dcoordinate and SMILES2CML are buggy
14:10 egonw it's valid XML
14:10 egonw the schema passed
14:10 edrin hm
14:10 edrin is it the CML schema as for the other services?
14:10 egonw yes, exact same
14:10 edrin very strange...
14:11 edrin but why do only they crash the service
14:11 egonw overload?
14:11 egonw maybe ejabberd thinks it is spammed?
14:11 edrin ok, restart the service and i will only request generate3d
14:12 egonw ok, mom
14:12 egonw ok, got for it
14:13 egonw disco
14:13 egonw crash
14:13 egonw well
14:13 edrin yep
14:13 egonw disconnect
14:13 egonw no schema send to STDOUT debug yet!
14:13 egonw failed with DISCO already!
14:13 edrin I'm pretty sure it's specific for these two functions
14:14 edrin no, disco response arrived here!
14:14 egonw ah, indeed:
14:14 egonw XWS4J: [Stanza] received : <iq from='edrin@ws1.bmc.uu.se/Bioclipse' id='JSO-0.12.5-398' to='cdk.ws1.bmc.uu.se' type='set' xml:lang='de-DE'><command xmlns='http://jabber.org/protocol/commands' action='execute' node='generate3Dcoordinates'><iodata xmlns='urn:xmpp:tmp:io-data' type='io-schemata-get'/></command></iq>
14:14 egonw XWS4J: CommandProcessor.process: processing command
14:14 egonw io-schemata-get
14:14 egonw and no sent after that
14:14 egonw and still disconnect
14:15 egonw so I don't think this has anything to do with the schema
14:15 egonw no <iq> sent to ejabberd yet
14:15 edrin egonw: hm...
14:16 edrin egonw: i would say then the stanza already killed JSO when it tried to send it...
14:16 egonw killed JSO?
14:16 egonw the service is still running, AFAIK
14:16 egonw but no longer connected
14:17 edrin and in consequence the log did not print it because the listener that listens for all sent packets and logs them was not called
14:17 CIA-73 bioclipse: shk3 * r10031 /bioclipse2/trunk/plugins/net.bioclip​se.cdk.ui.sdfeditor/src/net/bioclipse​/cdk/ui/wizards/DotViewerFilter.java: added comments
14:18 egonw edrin: ok
14:18 egonw so, JSO crashes on the schema
14:18 egonw ?
14:18 edrin yes
14:18 egonw but not always
14:18 edrin so it is not the schema?
14:19 edrin how do you read the schema into the stanza?
14:19 edrin is it the same code as for the other functions?
14:20 egonw http://cdk.svn.sourceforge.net/viewvc/cdk​/cdk-xws/trunk/xws-cdk-services/src/org/o​penscience/cdk/xws/services/SomeCDKFuncti​onTwo.java?revision=14570&amp;view=markup
14:20 zarah egonw's link is also http://tinyurl.com/bwvmfo
14:20 egonw yes, same code
14:20 CIA-73 bioclipse: shk3 * r10032 /bioclipse2/trunk/plugins/net.bioclipse.s​pecmol.test/src/net/bioclipse/specmol/bus​iness/test/SpecmolManagerPluginTest.java: all tests working
14:20 * shk3 is getting a coke and will continue bug fixing afterwards
15:02 CIA-73 bioclipse: shk3 * r10033 /bioclipse2/trunk/plugins/net.bioclipse.nmrs​hiftdb/src/net/bioclipse/nmrshiftdb/wizards/ (MessagePage.java PasswordWizardPage.java): tab jumps between texts
15:16 CIA-73 bioclipse: shk3 * r10034 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: removed adding explicit hydrogens (bug #834)
15:16 zarah bug #834 | http://tinyurl.com/d3bnd8
15:41 CIA-73 bioclipse: shk3 * r10035 /bioclipse2/trunk/plugins/net.biocl​ipse.cdk.ui/src/net/bioclipse/cdk/u​i/wizards/NewFromSMILESWizard.java: oops, wrong extension
16:13 rajarshi joined #bioclipse
17:20 shk3 olass: just reassigned the intro cheat sheet image bug to you.
17:20 shk3 please have a look.
17:20 olass thx
17:50 shk3 I am leaving
17:51 shk3 have a nice weekend
17:51 shk3 bye
18:51 zarah joined #bioclipse
19:13 ChanServ joined #bioclipse
20:02 rojasm joined #bioclipse
20:44 edrin joined #bioclipse

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