Camelia, the Perl 6 bug

IRC log for #cdk, 2007-09-29

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

All times shown according to UTC.

Time Nick Message
00:02 CIA-31 cdk: rajarshi * r8985 /trunk/cdk-qa/projects/070928-0001/runjob.sh: Created a job script to get the data and run the QA code
00:05 CIA-31 cdk: rajarshi * r8986 /trunk/cdk-qa/projects/070928-0001/runjob.sh: Some refinements to the job script
00:15 CIA-31 cdk: rajarshi * r8987 /trunk/cdk-qa/projects/070928-0001/runjob.sh: Made the job script a little more robust
03:25 CIA-31 cdk: rajarshi * r8988 /trunk/cdk-qa/projects/070​928-0001/AtomTypeQA.java: Converted the QA code to read SDF rather than SMILES to avoid atom typing while molecule loading
03:34 CIA-31 cdk: rajarshi * r8989 /trunk/cdk-qa/projects/070928-0001/ (pubchem.sdf pubchem.smi): temporarily adding data files for testing
03:42 CIA-31 cdk: rajarshi * r8990 /trunk/cdk-qa/projects/070928-0001/ (AtomTypeQA.java pubchem.sdf pubchem.smi): Removed temp data files. Fixed atom type code to properly identify missing atom types
03:49 CIA-31 cdk: rajarshi * r8991 /trunk/cdk-qa/projects/070928-0001/runjob.sh: Updated run script to print out some summary stats
05:44 jbrefort joined #cdk
07:13 carsten joined #cdk
08:41 rojasm joined #cdk
08:44 egonw joined #cdk
08:45 egonw moin
08:46 rojasm moin egonw
08:46 egonw hi rojasm
08:46 rojasm sorry the last day I had to away
08:46 egonw no problem
08:46 egonw in the end I did not have time thursday afternoon anyway
09:06 egonw I'm actually feeling quite ill...
09:06 egonw so not sure how useful I can be today...
09:06 egonw I was hoping for a short hack session...
09:06 egonw but don't think that's going to work...
09:06 egonw maybe I can just talk you through the code problems?
09:07 rojasm oh sorry I wish you will find better
09:11 egonw shall I talk you through the problems?
09:13 rojasm yes
09:13 egonw http://cheminfo.informatics.indiana.edu/~​rguha/code/java/nightly/junitsummary.html
09:13 egonw choose reaction
09:13 egonw let's start with that one
09:14 egonw http://cheminfo.informatics.indiana.edu/~rguh​a/code/java/nightly/test/result-reaction.html
09:14 egonw most problems come down to:
09:14 egonw org.openscience.cdk.excepti​on.InvalidSmilesException: Cannot percieve atom type for the 1th atom: C
09:14 rojasm yies I see
09:14 egonw or something like that...
09:14 egonw this means that the SMILES uses an atom type not in the new CDK list yet...
09:14 rojasm ok
09:15 egonw and I need your help here...
09:15 egonw because I am not so familiar with reaction intermediates...
09:15 rojasm But this only problem form reaction right
09:15 egonw e.g. somewhere you use a structures with a positive F... something like: [F+]=C-bla
09:16 rojasm yes
09:16 egonw now, I think the new atom type list is actually better suited for what you do...
09:17 egonw but would very much like your opinion on that...
09:17 egonw e.g. the atom types now define the number of lone pairs
09:17 egonw explicitly...
09:17 egonw and the number of pi-bonds...
09:17 egonw I think this should be very useful to your code...
09:17 rojasm ok
09:17 rojasm yes
09:18 rojasm so atom type list defines lone pairs, pi-bons ...
09:18 rojasm and sigma bonds
09:18 egonw yes, and number of neighbors
09:18 rojasm ok
09:18 egonw oh, and charge of course
09:18 egonw and hybridization state
09:18 rojasm neighbors is so the sigma
09:18 rojasm ok
09:18 egonw yes, sort of...
09:19 rojasm may be sould be better then to change the name??
09:19 egonw as in?
09:19 rojasm you said that atom type list knows about schema neighbors ...
09:20 rojasm but it is not better to write sigma
09:20 rojasm bond
09:20 rojasm or it doesnt matter for the computer
09:20 rojasm ok, question about smarts in cdk
09:21 egonw I think #neighbors is ebtter...
09:21 rojasm ok
09:21 egonw counting missing hydrogens is much easier then
09:21 rojasm ok
09:22 rojasm I am updating the cdk
09:23 rojasm in Computerhome
09:30 trakt0r joined #cdk
09:31 trakt0r http://tubeimage.com/viewer.php​?file=x7o3ydsiwmvc6pnvqyd3.jpg
09:33 was kicked by egonw: User terminated!
09:33 trakt0r joined #cdk
09:33 was kicked by egonw: User terminated!
09:35 kloeri joined #cdk
09:40 egonw rojasm: the other major problem is in the qsar IP descriptors...
09:40 egonw see this page: http://cheminfo.informatics.indiana.edu/~rguh​a/code/java/nightly/test/result-qsarBond.html
09:41 rojasm ok
09:41 egonw java.lang.ArrayIndexOutOfBoundsException: -6
09:41 egonw that's the problem...
09:41 egonw has to do with cloning or so...
09:41 egonw leading to a test where an atom is not found (the -6 problem)
09:41 rojasm ok
09:41 egonw becuase it's the atom not from the clone, but from the original...
09:41 egonw or something like that...
09:41 egonw got to hit the couch again...
09:42 egonw will be back later...
09:42 rojasm then I have too principals bugs reactions with smiles and clone stuff in qsar IP
09:42 rojasm ok
09:42 egonw rojasm: yes
09:42 rojasm too => two
09:42 egonw I think that covers 80% of the current issues in trunk/
09:42 egonw cu later
09:42 rojasm ok
09:42 rojasm cu later
10:39 CIA-31 joined #cdk
10:49 rojasm joined #cdk
11:22 CIA-31 joined #cdk
13:41 CIA-31 joined #cdk
14:28 CIA-31 joined #cdk
14:45 CIA-31 jmol: nicove * r8337 /trunk/Jmol/src/org/jmol/translation/ (Jmol/pt_BR.po JmolApplet/pt_BR.po): Update of .po files
16:30 carsten joined #cdk
16:49 CIA-31 joined #cdk
18:09 CIA-31 joined #cdk
19:48 CIA-31 joined #cdk
21:48 CIA-31 joined #cdk
22:46 CIA-31 joined #cdk

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