Camelia, the Perl 6 bug

IRC log for #bioclipse, 2007-12-19

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

All times shown according to UTC.

Time Nick Message
08:11 steinbeck joined #bioclipse
08:27 olas joined #bioclipse
08:45 steinbeck left #bioclipse
09:03 EskilA joined #bioclipse
09:42 jonalv joined #bioclipse
09:44 steinbeck joined #bioclipse
09:46 steinbeck left #bioclipse
09:46 steinbeck joined #bioclipse
09:55 masak joined #bioclipse
10:12 egonw joined #bioclipse
10:16 egonw moin
10:17 olas moin egonw
10:19 masak moin
10:35 egonw olas: is 1.2.0 out?
10:37 olas very VERY soon
10:37 olas EskilA is fixing the last on the selection listener bug
10:37 olas should be a matter of minutes
10:37 egonw ah, ok, excellent
10:53 CIA-20 joined #bioclipse
10:58 masak CIA-20: are you one of those that actually print commit logs?
10:59 masak because our last batch of CIA bots have been very disappointing on that front
11:08 egonw masak: I don't think CIA-20 was actually listening ...
11:08 masak egonw: sometimes I talk to bots without any expectation of them answering
11:09 masak it relieves stress
11:09 egonw ah...
11:10 egonw yeah... guess that is Swedish.... NL is very densly populated...
11:10 egonw there is always your neighbor to take that role :)
11:10 masak :)
11:11 masak in Sweden, sometimes a bot is all you've got
11:11 olas egonw: I have a proposal for the QSAR calculation
11:11 egonw olas: please email
11:11 olas Write out the names of the molecules as row headers
11:11 egonw 'name' is rather badly defined...
11:12 olas filename then?
11:12 egonw CDKConstants.TITLE may be used...
11:12 egonw but not always defined...
11:12 olas filename please?
11:12 egonw yeah, filename could be a good backup...
11:12 egonw can do that for 1.2.1
11:12 olas not today?
11:12 olas 5 minutes?
11:12 olas it really ruins the screenshot
11:13 olas where you click the outlier and don't know what molecule it is...
11:13 egonw :)
11:13 olas I think it should be in 1.2.0
11:13 egonw mom, let me fire up Eclipse
11:13 olas great!
11:13 egonw he, I said nothing about coding yet !!
11:23 olas that was implied
11:23 olas holding 1.2.0 for filenames as column headers after descriptopr calculation
11:23 olas and EskilA is fixing axis headers in the chart
11:24 olas the QSAR screenshot will then be complete
11:24 olas :-)
11:25 egonw :)
11:25 egonw no idea why CIA-20 does not post commit messages...
11:26 egonw can't find any weird setting...
11:26 egonw should work
11:26 egonw olas: how do I get the filename of a resource... last bit, so without path?
11:26 egonw I will now check if I can set row labels
11:29 egonw ok, I can...
11:29 egonw now, how the *beep* do those go into a .csv file?
11:29 egonw anyone with a smart answer on that?
11:31 egonw ha, no setRowName() :) funny...
11:31 egonw the getRowName() *is* implemented :)
11:32 egonw brilliant...
11:35 egonw OK, filename is now set as row name
11:35 egonw row label
11:37 egonw Uppsala: now what?
11:37 egonw Uppsala: suggestions on how to have that info stored in the file and be read?
11:39 egonw or shall I make a dirty workaround, and put them in the first column?
11:41 EskilA egonw: if the first entry in a row can't be interpreted as a number it is interpreted as a row label
11:41 egonw OK, sounds reasonable
11:42 egonw what about the number of expected labels then?
11:45 EskilA as many as there are rows?
11:46 egonw ah, sorry.. sure...
11:46 egonw number of column labels, I meant...
11:46 EskilA ah
11:47 EskilA isn't that the same thing, as many as there are columns?
11:47 EskilA shouldn't that be expected?
11:47 egonw yeah, but then the first line has less entries than those who follow
11:48 egonw happy with that?
11:48 EskilA ah
11:48 egonw is R happy about such files?
11:48 EskilA masak and kaskelot is working on that, maybe they know
11:49 * egonw is booting R...
11:51 egonw ah:
11:51 egonw "
11:51 egonw If there is a header and the first row contains one fewer
11:51 egonw field than the number of columns, the first column in the
11:51 egonw input is used for the row names.  Otherwise if 'row.names' is
11:51 egonw missing, the rows are numbered.
11:51 egonw "
11:52 egonw yep, that works :)
11:52 egonw OK, here goes
11:53 EskilA ok, sounds good
11:57 egonw any chance on seeing those row labels as popup for points in the scatter plot?
11:59 EskilA good idea
12:11 egonw oi... this requires some recoding to get it finished...
12:32 egonw bloody index-starts-with-0-or-was-it-1 bugs
12:35 olas :-)
12:35 egonw and you don't want to know how fucked up the code was...
12:35 egonw set(index-1)
12:35 egonw get(index)
12:36 egonw how about that??
12:36 masak oy!
12:36 masak that's bad
12:37 olas who created the fucked-up code?
12:37 * olas want to blame someone
12:38 egonw you moved in your divine plurality again ? :)
12:38 egonw into
12:38 olas haha
12:38 olas indeed
12:49 olas egonw: Kan man ta EPL-kod från något projekt och använda i Bioclipse? Om man släpper resultat som EPL också?
12:49 olas Tänkte på R-editorn i StatEt pluginnet
12:49 olas men den kräver en hel del omkodning för att få in...
12:50 olas oops
12:50 egonw sorry, my Swedish is not that good
12:50 olas soory
12:50 olas right, here it goes again
12:50 olas can you take EPL code from a project and put in Bioclipse (as it's EPL)?
12:50 olas was thinking of the R-editor of the StatEt project
12:51 olas or would that mean breaking their copyright?
12:51 egonw no, I would say that would be fine
12:51 olas ok, good
13:10 olas egonw: are you working on the row label?
13:10 egonw yeah, don't tell me you no longer want it for 1.2.0
13:11 olas great, just wanted to check
13:11 olas I REALLY want it for 1.2.0!
13:13 egonw well, if you so desparately need it: it's done
13:14 olas great
13:14 olas and the commit?
13:14 olas oh, no bot working I assume
13:14 * olas will update
13:15 egonw r4556
13:15 olas cool
13:15 egonw it will require testing!
13:15 egonw I'm sure I have not tried all exotic use cases
13:23 egonw please test
13:25 olas I will
13:26 egonw I have to go to my other desk now...
13:26 egonw will try to be online there...
13:26 olas right
13:26 egonw but likely offline for some 2 hours or so
13:26 olas great job egonw!
13:26 olas ack
13:26 olas let's hope no errors :-)
13:26 egonw sure, they don't call me the miracal worker for nothing...
13:26 egonw oh wait.... they don't
13:27 egonw bbl
13:31 EskilA_ joined #bioclipse
13:44 thomas_ku joined #bioclipse
14:00 steinbeck joined #bioclipse
14:13 egonw now I'm tired...
14:34 EskilA_ joined #bioclipse
14:44 olas egonw: error in Matrix
14:44 olas INFO - Took the first matrix implementation: net.bioclipse.model.JamaMatrix has column headers:  java.util.NoSuchElementException at java.util.Scanner.throwFor(Scanner.java:817) at java.util.Scanner.next(Scanner.java:1317) at net.bioclipse.model.MatrixResource.​insertRow(MatrixResource.java:229)
14:45 olas damn
14:45 olas opens a completely blank matrix editor
14:46 egonw please email me the .csv file
14:46 olas sure
14:46 egonw or put it on a pastebin website
14:46 olas pastebin it is
14:47 olas oh, the matrixEditor erased it
14:47 olas damn
14:48 olas I don't think I can send you a csv...
14:48 olas but I can say how to reproduce
14:48 olas go to chemical-structures folder
14:49 olas select amines
14:49 olas select all in amines
14:49 olas (8 molecules)
14:49 egonw 8??
14:49 olas create >10 descriptors mized from cdk and joelib
14:49 olas yes
14:49 egonw more here
14:49 olas sure, might have an old installation
14:49 olas select all then
14:50 olas does it work for you?
14:50 egonw 32 mols here
14:50 egonw no, it does not
14:50 olas works if I select only CDK descriptors I think...
14:51 olas yup, only error if mixed cdk and joelib
14:51 olas all cdk works
14:51 olas all joelib also works
14:51 olas all joelib + all cdk does not work
14:52 egonw ack
14:52 egonw some counting issue again
15:18 jonalv joined #bioclipse
15:19 egonw olas: what's the time line for 1.2.0 ?
15:19 egonw do I have time to look at that joelib+cdk all bug?
15:19 olas if you fix the descriptor calculations, pretty soon
15:19 olas yes
15:19 egonw do I have time to remove IP calculation from the list of possible descriptors?
15:19 olas waiting for that
15:19 olas sure
15:19 olas there's no rush
15:19 egonw and time for updating to cdk-1.0.x/ as it is now?
15:20 olas how much time do you need?
15:20 egonw but I'd rather not do the latter now, actually...
15:20 egonw on second thought...
15:20 egonw might cause regressions...
15:20 olas would be great to get it released on friday latest IMO
15:20 egonw that's more for 1.2.x
15:20 olas ok
15:20 olas but the descr bug is important
15:20 egonw right...
15:20 egonw will get some coffee and then have a go at that bug
15:42 masak joined #bioclipse
15:43 egonw olas: I think I see the problem...
15:43 egonw some JOElib descriptors return a class BC can't handle...
15:45 olas oh
15:45 olas could we exclude those descriptors?
15:46 egonw yeah, I guess it should...
15:49 olas I mean, fix them is better but probably non-trivial, or what do you say?
15:49 egonw rather non-trivial...
15:49 egonw how to squeeze a matrix into a vector ??
15:59 olas split it into columns and name them descr_1 descr_2 etc
16:12 egonw sure
16:13 egonw unfolding
16:13 egonw not always a good choice
16:30 olas egonw: was this a hard bug?
16:30 egonw it is
16:30 olas aah
16:38 egonw there is some mismatch in counts...
16:38 egonw number of estimated col count with the actual count
16:41 egonw it
16:41 egonw 's a CDK bug...
16:41 egonw crappy state of some QSAR descriptors...
16:54 egonw olas: this'll have to do...
16:54 egonw please try...
16:54 egonw a matrix should show up now...
16:54 egonw but it will have null's...
16:55 egonw because the descriptors say they return more values than they actually do...
16:55 egonw not sure which those are, though...
16:55 egonw but giving up now...
17:05 egonw olas: please verify that your bug is solved
17:05 egonw olas: pint
17:05 egonw olas: ping
17:23 egonw mmm.... no olas
17:24 egonw olas: I will go home in 6 minutes or so...
17:27 egonw bye all
17:29 CIA-20 joined #bioclipse
19:10 egonw_ joined #bioclipse
19:11 egonw__ joined #bioclipse
20:47 olas left #bioclipse

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