Camelia, the Perl 6 bug

IRC log for #cdk, 2008-05-14

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

All times shown according to UTC.

Time Nick Message
03:33 CIA-31 cdk: rajarshi * r10992 /cdk/trunk/src/ (2 files in 2 dirs):
03:33 CIA-31 cdk: Updated code so that the evaluation of matching groups is not done during
03:33 CIA-31 cdk: search. Instead we do it lazily. Also, added a new method to get a list of
03:33 CIA-31 cdk: constraints in the target that matched the query. The bond analog of getting the
03:33 CIA-31 cdk: matching pharmacophore atoms
04:08 CIA-31 cdk: rajarshi * r10993 /cdk/trunk/src/main/org/openscience/cdk/​pharmacophore/PharmacophoreMatcher.java: Fixed a bug which caused the code to not generate the mapping for new matches, after the first succesful match
04:39 egonw joined #cdk
04:42 egonw joined #cdk
04:58 carsten joined #cdk
05:17 jbrefort joined #cdk
06:47 rojasm joined #cdk
07:57 jonalv joined #cdk
08:00 egonw joined #cdk
08:06 jpansanel moin
08:06 jpansanel http://wiki.debian.org/SSLkeys
08:06 egonw yeah... saw that...
08:32 rojasm hi egonw
08:32 CIA-31 cdk: miguelrojasch * r10994 /cdk/trunk/src/test/org/openscience/cdk/to​ols/StructureResonanceGeneratorTest.java: added new test for 1-Propene, 2-chloro-
08:32 rojasm did you check the structures that I put
08:37 masak joined #cdk
08:56 rojasm egonw I have also a list with structures which the smiles parser doesn't work
08:56 rojasm should I put a bug
09:02 egonw rojasm: ok, sorry I did not get back to you...
09:02 rojasm no problem
09:02 egonw for the SMILES, please file bug reports, and add unit tests for each of them trunk/ please
09:03 rojasm also 40 tests
09:03 rojasm I have them in mol file
09:03 egonw that's fine
09:04 rojasm perfect
09:04 egonw what's with the mol files?
09:04 rojasm I mean the structure is coming from mol files
09:09 egonw for the SMILES tests?
09:09 CIA-31 bioclipse: ospjuth * r5623 /bioclipse2/trunk/plugins/ (2 files in 2 dirs): Adopted to structural changes in CDK.
09:10 rojasm Yes! I will make the test taking the structure from the mol files
09:11 egonw rather not
09:11 rojasm mmm
09:11 egonw because then we match the SMILES parser against the MDL molfile reader
09:11 rojasm I see
09:11 egonw better: parse the SMILES, and assert atom count, bond count, etc
09:12 rojasm ok
09:13 jonalv How does smarts differ from smiles?
09:22 CIA-31 bioclipse: ospjuth * r5624 /bioclipse2/trunk/plugins/net​.bioclipse.cdk10.jchempaint/ (META-INF/MANIFEST.MF plugin.xml): Removed dep from cdk10 to cdk.business
09:24 egonw jonalv: smarts is a query language
09:24 egonw so, a SMARTS may say... give me a substructure with a N attached to a carbon or a oxygen
09:25 jonalv so in the mail when you asked for a cdk.fromSmarts mwthod. Does that makes sence?
09:26 jonalv I mean if it is a query language can you create a molecule without having something to query upon?
09:27 egonw rojasm: please make sure to use @cdk.bug in the JavaDoc of the unit tests
09:27 egonw jonalv: right, except that you don't create a molecule, but a query :)
09:29 jonalv egonw: I think you lost me now. What is it you want the mehod cdk.fromSmarts to return?
09:29 rojasm yes egonw
09:30 masak jonalv: a query object
09:30 jonalv to be used in strucutedb? Or what do you want t odo with it?
09:31 egonw jonalv: to define the substructure
09:31 egonw SMARTS is to SQL, what structuredb is to MySQL
09:31 egonw (sort of)
09:33 jonalv egonw: IS there really a need to create a query object? Can't that be hidden from the user?
09:33 egonw sure
09:34 jonalv I thinking something like this: structuredb.smartSearch("mydatabse", "smarts-query")
09:34 jonalv (have I understood smarts now?)
09:34 egonw yes, fine
09:34 egonw yes, you did
09:35 jonalv egonw, I am guessing cdk has a method a nit like SmartChecker.matches("smarts", molecule) somewhere or?
09:36 jonalv (I am guessing it's a bit more complicated though...)
09:36 egonw yes
09:36 egonw no, not really...
09:36 egonw I don't know the exact API... but see cdk.smiles.smarts
09:37 jonalv the tests suite is the best place to start right?
09:37 egonw yes, or google the cdk-user archives
09:37 jonalv oki
09:38 egonw http://cheminfo.informatics.indiana.edu/~​rguha/code/java/nightly/api/org/openscien​ce/cdk/smiles/smarts/SMARTSQueryTool.html
09:38 egonw to be precise
09:42 jonalv egonw, thanks
10:00 thomas_ku joined #cdk
10:01 thomas_ku moin
10:04 jonalv Another thing regarding SMARTS matching. There is no better or worse match right? Either it matches or it doesn't match?
10:04 jonalv One match is not better than another?
10:15 CIA-31 jmol: nicove * r9376 /trunk/Jmol-FAH/ (6 files in 2 dirs): Folding@Home
10:26 CIA-31 jmol: nicove * r9377 /trunk/Jmol-FAH/projects/p4535.xyz.gz: Folding@Home
11:11 CIA-31 jmol: hansonr * r9378 /branches/v11_4/Jmol/src/org/jmol/ (3 files in 2 dirs):
11:11 CIA-31 jmol: 11.4.3 VERSION
11:11 CIA-31 jmol: # bug fix: modelLoader was assigning too much array space for chains
11:11 CIA-31 jmol: # bug fix: rocket positions after translateselected
11:12 CIA-31 jmol: hansonr * r9379 /trunk/Jmol/src/org/jmol/model​setbio/ProteinStructure.java: code
11:27 egonw joined #cdk
11:42 CIA-31 jmol: nicove * r9380 /branches/v11_4/Jmol/src/org/​jmol/viewer/Jmol.properties: Version
11:49 thomas_ku joined #cdk
11:57 CIA-31 bioclipse: jonalv * r5625 /bioclipse2/trunk/plugins/net.bioclip​se.usermanager/src/net/bioclipse/user​manager/business/IUserManager.java: Added params to accountExists publication. Fixes bug #1963125
12:02 jonalv egonw, do you have a moment?
12:22 egonw no, not really
12:43 CIA-31 bioclipse: egonw * r5626 /bioclipse2/trunk/plugins/net.bioclipse​.structuredb/src/net/bioclipse/structur​edb/business/StructuredbManager.java: Added code to count how much work needs to be done
12:49 steinbeck joined #cdk
12:49 masak egonw: not sure I follow the code
12:49 masak but shouldn't there be a read in the innermost if too?
12:50 egonw masak: good point :)
12:50 masak :)
12:51 masak also, you indent funny
12:51 masak indenting blocks is cool, but indenting other things not so much
12:52 egonw he, it's your tab wizard xml
12:52 egonw seriously...
12:52 egonw what??
12:52 masak really? :) I'm flattered
12:52 egonw where is that non-block indent?
12:52 masak we might be seeing different things
12:52 masak but I see
12:52 masak c = counterStread.read()
12:53 masak if (...
12:53 masak etc.
12:53 masak do you indent that way, or is what I'm seeing wrong?
12:59 egonw mmm... I don't typically do that, nor do I see that here
13:00 egonw might be tabs ?
13:00 masak glad to hear it
13:00 masak might be tabs, yes
13:00 egonw mmm... seems that I did not properly install your .xml for tabs :(
13:00 masak never too late to do that :)
13:01 masak I have a few Perl script to expunge tabs from all source files, too
13:01 masak though there might be better ways to do that
13:15 thomas_ku steinbeck around=
13:15 thomas_ku ?
14:38 CIA-31 cdk: miguelrojasch * r10995 /cdk/branches/miguelrojasch-IP/src/ (12 files in 6 dirs): bifurcated the IP in HOSE model and Learning Model
14:54 CIA-31 cdk: miguelrojasch * r10996 /cdk/branches/miguelrojasch-IP/src/ (2 files in 2 dirs): Added the other method to calculate the IP atomic based on desicion tree learning
15:25 CIA-31 cdk: miguelrojasch * r10997 /cdk/branches/miguelrojasch-IP/ (245 files in 57 dirs): Synchronized with trunk the branches.miguelrojasch-IP, Rev=10914:10996
15:46 CIA-31 bioclipse: jonalv * r5627 /bioclipse2/trunk/plugins/ (7 files in 4 dirs):
15:46 CIA-31 bioclipse: Wrote console progressmonitor and hooked it up to addStructuresFromSDFFile in
15:46 CIA-31 bioclipse: structuredb only problem is that it doesn't update continously. Because the
15:46 CIA-31 bioclipse: console uses the main thread for running commands with the main thread doesn't
15:46 CIA-31 bioclipse: get around to updating the console view until after the command is done. It
15:46 CIA-31 bioclipse: should work once we fix that though.
15:59 egonw joined #cdk
16:07 steinbeck joined #cdk
16:09 rojasm joined #cdk
16:10 carsten joined #cdk
18:03 egonw joined #cdk
19:15 CIA-31 cdk: miguelrojasch * r10998 /cdk/branches/miguelrojasch-IP/src/ (6 files in 4 dirs): Added affinity proton prediction descriptor
19:18 egonw cool... close to the 11k-th commit :)
20:15 rojasm egonw around?
20:16 rojasm to calculate the IP from learning process calculating descriptors is easier if I call the descriptor-self
20:16 rojasm than I use the their algorithm but that violate the rules about descriptors
20:17 rojasm should I create a new package for IP
20:18 CIA-31 jmol: nicove * r9381 /trunk/Jmol-FAH/projects/p4534.xyz.gz: Folding@Home
20:18 rojasm with dependence of qsar model
20:29 egonw no, you should ensure that all algorithms that you need (atom/bond properties) are factored out of qsar modules
20:29 egonw you can extract those algorithms from the qsar descriptor code and put them in 'extra' for now...
20:31 egonw have to go now...
20:31 egonw won't have time tomorrow...
20:31 egonw but will have time this weekend/next week (maybe bits on friday)
20:31 egonw file bug reports for the QSAR descriptors which implement new algorithms...
20:31 egonw new = old...
20:31 egonw but not factored out
20:32 egonw if you make one bug report for each descriptor, I could more easily help you refactor bits...
20:32 egonw bye now
21:23 CIA-31 cdk: miguelrojasch * r10999 /cdk/branches/miguelrojasch-IP/src/main/org/o​penscience/cdk/config/data/cdk_atomtypes.xml: Added atomtype Halogen.plus.radical for Cl, Br and I
22:01 CIA-31 jmol: hansonr * r9382 /trunk/Jmol/src/org/jmol/ (7 files in 4 dirs): 11.5.37 version (b)
22:04 edrin joined #cdk
22:58 CIA-31 cdk: miguelrojasch * r11000 /cdk/branches/miguelrojasch-IP/src/test/or​g/openscience/cdk/qsar/descriptors/atomic/ (2 files): added more test
23:00 CIA-31 cdk: miguelrojasch * r11001 /cdk/branches/miguelrojasch-IP/src/ (2 files in 2 dirs): the charge stabilization calculation able for atoms with single electrons too
23:06 CIA-31 cdk: miguelrojasch * r11002 /cdk/branches/miguelrojasch-IP/src/ (2 files in 2 dirs): set the algorithms to extract the values. First steps before the prediction using the decision-tree

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