Camelia, the Perl 6 bug

IRC log for #cdk, 2011-08-10

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

All times shown according to UTC.

Time Nick Message
04:41 egonw joined #cdk
05:33 jbrefort joined #cdk
06:55 sneumann joined #cdk
07:26 Gpox joined #cdk
07:27 egonw hi Gpox
07:44 s_wolf joined #cdk
07:48 bobbymeds joined #cdk
07:51 egonw bobbymeds: OK, tell me what you like to know/learn
07:52 egonw bobbymeds: do you have the Eclipse git plugin installed?
07:52 egonw http://eclipse.org/egit/
07:52 bobbymeds no, I use the Eclipse "classic" for J2EE application
07:53 egonw which version? 3.6 or 3.7?
07:53 bobbymeds I look at your link
07:53 bobbymeds eclipse indigo, that must be the 3.7
07:54 egonw yeah, I guess so...
07:54 egonw I think egit is part of that update site...
07:54 egonw so, you only need to go to Help → Install new Software
07:54 egonw and pick the Indigo update site
07:54 egonw I think EGit should be listed there
07:55 egonw Gpox: ping
07:55 egonw Gpox: do you happen to know how to use egit to apply a 0001-foo.patch file?
07:55 egonw because I have never done that personally... but should be possible, I think
07:56 egonw bobbymeds: here's an extensive tutorial/user guide for egit: http://wiki.eclipse.org/EGit/User_Guide
07:56 bobbymeds the update tool tells that the link i wrote corresponds to a duplicate
07:57 bobbymeds so I may already have EGIT
07:57 egonw ah :)
07:58 bobbymeds I never applied any patch to some library yet
07:58 bobbymeds :)
07:59 Gpox egonw: http://wiki.eclipse.org/EGit​/User_Guide#Applying_Patches
07:59 bobbymeds let me a moment for EGIT research/install because I do not find it in my eclipse plugins ...
07:59 egonw Gpox: cool... I did not spot it in the TOC yet...
08:02 bobbymeds installation is processing ... :)
08:03 egonw bobbymeds: where are you actually from? you're not English, that I can tell :)
08:03 bobbymeds you heard my accent, don't you ? ;)
08:03 bobbymeds I'm in France
08:04 bobbymeds I develop LIMS
08:05 egonw indeed :)
08:05 egonw ah, ic
08:06 egonw did you see jonalv's (not here yet) Brunn paper?
08:06 egonw very targeted LIMS-like thing
08:06 egonw http://www.ncbi.nlm.nih.gov/pubmed/21599898
08:07 bobbymeds oh thanks, I'll read it during this day :)
08:07 bobbymeds ok, egit is installed !
08:08 egonw ok, check that link from Gpox at 9:59
08:11 bobbymeds one question : what repository do I have to use ? a local folder where to find patches ?
08:13 egonw if you already have a local git copy, you can use that...
08:13 egonw Import... from the File menu in Eclipse
08:15 de_maas joined #cdk
08:16 egonw hi de_maas!
08:17 bobbymeds i created a new git repository. and logically it is empty ... should I have to specify the remote ?
08:19 jonalv joined #cdk
08:20 jonalv joined #cdk
08:21 egonw ah, better to check out one from the remote repos then
08:21 egonw bobbymeds: http://wiki.eclipse.org/EGit/User_Guide#​Starting_from_existing_Git_Repositories
08:22 egonw git://github.com/cdk/cdk.git
08:23 egonw jonalv: bobbymeds makes LIMS systems in France... I pointed him to your Brunn paper
08:25 jonalv egonw: oh cool
08:36 bobbymeds ok, i get it, after some dummy issues ... i'm receiving data
08:37 bobbymeds hi jonalv
08:37 jonalv hi bobbymeds
08:37 bobbymeds I did not read your paper yet but I'll do it after my "git initiation" ;)
08:38 jonalv bobbymeds: ah, yea git is so worth it
08:38 jonalv itäs a bit of a learning curve
08:38 jonalv but you won't regret it
08:39 bobbymeds i do not have any doubt on it ! :)
08:41 bobbymeds i make LIMS for compagnies or national research institutions
08:42 bobbymeds for hith throughput screening
08:42 jonalv oh
08:43 jonalv bobbymeds: and you are using java?
08:43 bobbymeds yes the LIMS i make are java web application
08:44 jonalv okey
08:44 jonalv so you are using CDK then I take it?
08:46 bobbymeds i try to :) . one of our client makes compound screening (some other one make antibody screenings)
08:46 bobbymeds and i test cdk for sdf file import
08:47 jonalv yea I was about to say been there done that
08:47 jonalv but in fact I never did do that cause it turned out that wasn's so important after all
08:48 jonalv The current Brunn users don't really care about the structures, they just work with labels for substances...
08:49 jonalv but reading plate posistions from SDF files seems like a non trivial thing consdiering there are no standards
08:49 bobbymeds my client wants to imports "mother plates" from a SDF file : so I have to create plates, compound (mass and fomula calculations), concentration in wells ...
08:49 bobbymeds yes, i have this problem too :)
08:49 jonalv yea that feature is on the wishlist for Brunn as well
08:50 jonalv the motehr plates are usually 96 and then they want to import it into 384 well plates
08:50 bobbymeds I already imported 384 plates
08:51 bobbymeds I force the user to select a plate model and thanks to it, I know the number of wells
08:51 bobbymeds plate model are defined by admin users
08:52 jonalv ah, well int this case they want to pick from for 96 well plates into 1 3884 well plate
08:52 jonalv *384
08:52 jonalv (seems my fingers are not hitting the right keys today)
08:52 bobbymeds they can do that but my SDF import tool manages severals plate formats
08:53 jonalv okey, that sounds handy
08:54 bobbymeds concerning plate/position in SDF, most of official chemical libraries have standard format for that
08:54 bobbymeds for the other ones, it is a problem indeed
08:55 bobbymeds and at this time, I do not support SDF files without explicit "plate" and "position" attribute for each compound
08:55 jonalv makes sense
08:57 bobbymeds and for standard format, I made a little swing application that read the SDF file and create a new one with "plate" and "position" extracted from the attribute in the original file
08:57 bobbymeds eganw : it is still downloading ... 70% done
08:58 jonalv bobbymeds: what would "standard format" be in this case?
08:58 bobbymeds "egonw" , pardon :)
08:58 egonw yeah, still here
08:58 egonw ok
08:59 bobbymeds in france there is a standard defined by the "national chemical library"
09:00 bobbymeds it is like that : >  <compound_identifier> (1) ICC114-L-012-B05-
09:00 bobbymeds (compound from Curie Institute)
09:01 bobbymeds in this example, you can read that this compound is in plate 12 at the position B05
09:01 jonalv bobbymeds: oh okey
09:02 bobbymeds but for prestwick it is not the same thing, I had to write a little class for a specific reading
09:03 jonalv yea
09:03 bobbymeds but some other provider, such as necan for example, specify plate and position attributes
09:03 egonw wondering what would happen if bobbymeds would give a demo file to jonalv to testing with Brunn....
09:04 bobbymeds a demo of what ? a SDF file ?
09:04 egonw right
09:04 jonalv yea buit they are never named the same are they? I was thinking of a general interface whre the user pick one attribute to be positions
09:04 jonalv egonw: Brunn does not handle sdf file for the moment
09:04 jonalv egonw: Every substance is just added by name by hand
09:04 egonw bobbymeds: does the SD file contain any content that indicates they use that French standard?
09:05 bobbymeds no , unfortunatly
09:05 bobbymeds but the client knows the provider
09:06 bobbymeds i have several "official" sdf files and can send them to you if you want do some tests
09:07 jonalv I don't really have the time for that for the moment :)
09:08 egonw could perhaps be a bioclipse junior project?
09:09 jonalv egonw: what exactly?
09:10 egonw support for such SD files?
09:10 egonw it sounds to me that BC already has the needed components, and that they just need proper linking
09:10 jonalv egonw: in Brunn? As Brunn does not support sd files at all for the moment there is more than just supporting those sd fiels
09:11 jonalv egonw: but yea all the tools are there I agree with that
09:11 egonw so, how much time would it take you? multiple that by 3 or 4 and you have what it would take someone else to do it
09:11 jonalv difficult
09:12 jonalv but yea I guess to do it well and general could actually be a student project
09:14 bobbymeds egonw : ok, download is over
09:15 egonw cool
09:15 egonw that automatically created a project in eclipse?
09:17 bobbymeds no, there is the git repository with the tree : branches, tags ...
09:17 jonalv are you guys checking out cdk with egit?
09:17 egonw yes
09:17 bobbymeds yes
09:17 * jonalv used egit for the first time yesterday so I know a little...
09:17 egonw bobbymeds is going to apply and review a patch by me
09:17 egonw jonalv: excellent!
09:17 jonalv nice bobbymeds++
09:18 egonw jonalv: a howto would make a very nice blog post!
09:18 bobbymeds :)
09:18 bobbymeds but I'm not familiar with the use of git ...
09:18 bobbymeds :)
09:18 egonw so the egit GUI is perfect
09:18 jonalv egonw: how to check out cdk using egit? Yea sounds like an idea. You mean I should make it?
09:19 egonw should, should... see it as an opportunity to attract people to your scientific work
09:19 egonw it's like a cool poster
09:19 egonw so people will look at your other posters too
09:19 jonalv ah
09:20 egonw and you happen to have blogged about your brunn paper too, right?
09:20 jonalv egonw: no, you know I happen to not blog very much
09:20 egonw :)
09:20 bobbymeds you'll share the glory ;)
09:20 egonw well, that's a task of senior scientists... to educate the young in doing science
09:20 egonw like sending PhDs to conferences :)
09:21 bobbymeds it sounds like personal experiment ;)
09:21 egonw bobbymeds: it kinda is → http://chem-bla-ics.blogspot.com/2010/​09/second-acsboston-talk-teaching.html
09:21 zarah egonw's link is also http://tinyurl.com/3hhjxz3
09:30 bobbymeds ok, I have all the sources in my git repository
09:30 bobbymeds but i don't have a java project
09:32 egonw mmm...
09:32 egonw OK, let me boot Eclipse 3.7 myself
09:32 jonalv bobbymeds: In the 'Git repositories' view right clikc on the cdk repo and choose 'Import Projects...'
09:32 egonw so, no cdk project yet?
09:32 egonw ah... makes sense
09:32 egonw jonalv++
09:32 egonw @karma jonalv
09:32 zarah jonalv has a karma of 275
09:33 jonalv I guess I am just a nice guy? :)
09:33 egonw bobbymeds++ (for coming online for willing to got through all this to review my patch :)
09:34 bobbymeds :)
09:34 egonw bobbymeds: so, if I understand correctly, you did not right click on the Package Explorer and hit Import?
09:36 bobbymeds i'm in the git view. i imported project with "import existing project" in the right click menu and an empty project has been created in my workspace
09:37 bobbymeds ah ok
09:37 bobbymeds now i have the ckd master
09:37 egonw ok, almost good
09:37 egonw because you should pick the cdk-1.4.x branch
09:37 bobbymeds "almost" ?? lol
09:39 bobbymeds i think it's good, i have the cdk-1
09:39 bobbymeds 1.4.1
09:40 bobbymeds and now I apply the patch ?
09:41 egonw OK, good
09:41 egonw let's see
09:41 egonw still got that link from Gpox?
09:41 egonw [08/10/11 09:59] <Gpox> egonw: http://wiki.eclipse.org/EGit​/User_Guide#Applying_Patches
09:42 bobbymeds ok, no error message :)
09:43 jonalv egonw: have you seen the Bioclipse mailing list, seems like a complicated SMILES causing a time out. Do you think you can answer him since SMILES and cdk is you thing? :)
09:43 jonalv I am guessing there isnät much to do when timing out during smiles parsing
09:43 egonw jonalv: yeah, saw it
09:43 egonw already tried online with the CDK:
09:43 egonw http://apps.ideaconsult.net:8080/ambit2/depict?se​arch=CC%28C%29CN1C%28%3DO%29N%28C%29C%28%3DO%29c2​c1nn%28Cc3cccc4ccccc34%29c2c5ccncc5&amp;smarts=
09:43 egonw not sure where the time out comes from...
09:44 egonw but not time for that today
09:44 jonalv egonw: oh isnät it armotaticity as usual?
09:44 egonw jonalv: this is what genettasoft should be doing
09:44 egonw dunno
09:44 jonalv -me hits is right pinky trying to teach it where the "'" character is
09:44 egonw because that link shows the CDK handles the SMILES fine
09:44 jonalv (gah)
09:44 jonalv egonw: really_
09:45 jonalv so it is in fact a time out because something is slow then?
09:46 CIA-90 cdk: Jules Kerssemakers 313-14x-mfcleans * r681cef8 / src/test/org/openscience/cdk/tools/manipul​ator/MolecularFormulaManipulatorTest.java :
09:46 CIA-90 cdk: MolecularFormulaManipulatorTest: Cleaned up javadoc - removed all empty/meaningless javadoc - wrote short oneliner javadoc for the more complicated methods
09:46 CIA-90 cdk: Signed-off-by: Egon Willighagen <egonw@users.sourceforge.net> - http://bit.ly/n12q9S
09:46 CIA-90 cdk: Jules Kerssemakers 313-14x-mfcleans * r9028b3c / src/test/org/openscience/cdk/tools/manipul​ator/MolecularFormulaManipulatorTest.java :
09:46 CIA-90 cdk: MolecularFormulaManipulatorTest: Updated testGenerateOrderEle to match updated method
09:46 CIA-90 cdk: Signed-off-by: Egon Willighagen <egonw@users.sourceforge.net> - http://bit.ly/pawRYa
09:46 egonw yeah, perhaps
09:46 egonw could indeed be bioclipse being slow
09:46 egonw dunno
09:46 jonalv egonw: should I write a reply then_
09:46 egonw needs to be explored
09:46 egonw did you reproduce the problem?
09:46 egonw that's the first step, I would say
09:46 jonalv egonw: I guess I could sen him that link you had there
09:47 egonw yes, helpful
09:47 jonalv hm zarah left?
09:47 egonw oh, crap
09:47 egonw bobbymeds: I just read that link from Gpox
09:47 egonw but it does not support 0001-foo.patch it seems :/
09:47 egonw that sucks
09:48 egonw jonalv: I think it was overwhelmed by your karma :)
09:48 jonalv hehe
09:49 egonw bobbymeds: that means we have to go to the command line ... :/
09:51 bobbymeds i built the cdk project, linked to my test project and it failed with the same error ... so if you want to go to the command line, let's go ... I follow :)
09:52 egonw check where eclipse stored your git clone
09:52 egonw and move the command line to that place
09:52 egonw check if you have git installed, and if so, type:
09:52 egonw git status
09:52 egonw (what OS are you running)
09:53 bobbymeds windows xp
09:55 egonw oi
09:55 egonw no command line
09:55 egonw damn
09:55 egonw jonalv: ping
09:55 jonalv if you install git you get a command line, it is not very nice but might be enough for now
09:55 jonalv egonw: pong?
09:56 egonw jonalv: do you know how to add additional remote repositories with egit?
09:56 egonw then I could push the patch to my own github repos
09:56 jonalv egonw: no, maybe Gpox has an idea_
09:56 egonw and then bobbymeds could get the patch that way
09:56 egonw Gpox: do you know?
09:57 bobbymeds no git command actually
09:57 bobbymeds (sorry for the delay, i' had a call)
09:57 egonw bobbymeds: yeah, forget the command line
09:57 egonw we'll try another approach... one sec
09:58 Gpox in the Gir repository view right click remotes and add one
09:58 egonw bobbymeds: ok, follow Gpox's instructions and add this repos:
09:58 egonw git://github.com/egonw/cdk.git
10:02 bobbymeds let me try ...
10:02 egonw jonalv: yeah, Bioclipse needs a replacement for me to keep the CDK version up to date
10:02 jonalv egonw: yea... I know
10:02 egonw jonalv: that might solve the problem, as it sounds a bit like a too old CDK, hence failing atom type
10:02 jonalv egonw: oh
10:03 jonalv I can't help but think that Gpox is the best candidate consiering he has spent some time in CDK code while doing the 2d editor...
10:05 egonw but Gpox is pretty busy too...
10:05 jonalv egonw: yea I know...
10:06 jonalv egonw: do you have a better idea then?
10:06 egonw originally, I had hoped that Ola and Martin would use their spin off to raise money for this kind of work
10:07 jonalv egonw: I think it is too early, I think they have enough to raise money for themselves
10:07 egonw yes, but it is used for different purposes
10:07 egonw not maintenance, but new development
10:08 egonw that's investment in the future, but doesn't help now
10:08 jonalv egonw: not sure this is the right place to dicuss internal Bioclipse politics but yea Gpox is the only one doing maintencance and hence his name was the only suggestion I had...
10:12 egonw sorry. I did not imply to say that GenettaSoft is doing it wrong...
10:12 egonw aux contraire
10:12 egonw this future investment is very important!
10:12 * jonalv is afk for lunch
10:12 egonw just meant that it's not going to help here :)
10:13 egonw if something is not 'good', that does not imply it is 'bad'
10:13 egonw GenettaSoft++
10:13 egonw great place for open source chem/bioinformatics support
10:13 egonw I am not aware of any politics
10:14 bobbymeds egonw : it's done and i applied the patch
10:14 egonw oh!
10:14 egonw bobbymeds++
10:14 egonw well done!
10:15 egonw next step is signing off, or did that work too?
10:15 bobbymeds however i had the same error about atom type for Ag
10:15 egonw grmph
10:15 egonw then I need to open that molfile and look at the structure in more detail
10:15 bobbymeds i can send it to you
10:16 bobbymeds i posted the part of my code in the forum
10:16 egonw just the file is OK
10:16 bobbymeds i attach the file in the topic
10:16 egonw in fact...
10:16 egonw if you can make it a git patch :)
10:16 egonw that would even be better
10:16 egonw and you'd get some git commit fame :)
10:17 bobbymeds lol
10:17 egonw add the test file (make sure it'll be LGPL) to src/test/data/mol or so
10:17 egonw src/test/data/mdl
10:18 bobbymeds you mak me doing lots of new things today ... ;)
10:18 egonw cool :)
10:18 * egonw likes educating :)
10:21 bobbymeds do you wnat the whole file or do you want me to put only the molecule with Ag ?
10:21 egonw the latter would be fine
10:21 egonw make sure to set your name and email address for the git commit 'Author' field
10:22 egonw see http://wiki.eclipse.org/EGit/U​ser_Guide#Identifying_yourself
10:22 bobbymeds ok, but there are other molecule with Ag inside
10:23 bobbymeds (after the one which failed, so I did not test them ...)
10:23 egonw well, you can always make patches with further test files later
10:24 egonw it's very helpful when people supply that when filing a bug report :)
10:28 bobbymeds ok, it's done
10:29 bobbymeds I put it in data/mdl . file name is MoleculeWithAg
10:29 bobbymeds oh, it's bad, i forgot the .sdf extension, sorry :S
10:29 egonw :)
10:29 egonw http://wiki.eclipse.org/EGit/​User_Guide#Reverting_Changes
10:29 egonw o
10:29 egonw or
10:30 egonw http://wiki.eclipse.org/EGit​/User_Guide#Amending_Commits
10:30 egonw so, in your case
10:30 egonw rename the file, and commit with the 'Amend previous commit' option
10:30 egonw (see that second link)
10:32 bobbymeds it's ok
10:36 bobbymeds i go to my lunch
10:36 bobbymeds "bon appétit" :)
10:37 egonw bien sûr!
10:37 egonw just finished my lunch myself :)
11:16 de_maas joined #cdk
11:24 CIA-90 cdk: Jules Kerssemakers cdk-1.4.x * r9028b3c / src/test/org/openscience/cdk/tools/manipul​ator/MolecularFormulaManipulatorTest.java :
11:24 CIA-90 cdk: MolecularFormulaManipulatorTest: Updated testGenerateOrderEle to match updated method
11:24 CIA-90 cdk: Signed-off-by: Egon Willighagen <egonw@users.sourceforge.net> - http://bit.ly/pawRYa
11:24 CIA-90 cdk: Jules Kerssemakers cdk-1.4.x * r681cef8 / src/test/org/openscience/cdk/tools/manipul​ator/MolecularFormulaManipulatorTest.java :
11:24 CIA-90 cdk: MolecularFormulaManipulatorTest: Cleaned up javadoc - removed all empty/meaningless javadoc - wrote short oneliner javadoc for the more complicated methods
16:04 bobbymeds bye cdk team ! have a good evening :)
16:04 egonw bobbymeds: I'll be around
16:04 egonw let me know when you can wrap up the tutorial
16:05 egonw I was already starting to wodner about that long french lunch :)
16:05 egonw bobbymeds: and thanx for dropping bye!
16:05 bobbymeds i'll look at the news on the blog ;)
16:05 egonw did you send me your patches?
16:05 bobbymeds french meals are an art of living ;)
16:05 egonw I'll need those first
16:06 egonw yeah, I always likes the holidays in the french alps
16:06 * egonw is away for dinner
16:07 bobbymeds oh  I think I didn't understand what to do ... I commited the file to the cdk team to look at the atom type problem with Ag
16:08 bobbymeds i think i'll contact you tomorrow, it will be easier because I don't know what to submit about a patch ... I'm not a chemist and I don't know how to solve the problem and create a patch :S
16:09 bobbymeds ciao !
18:49 sneumann joined #cdk
18:53 sneumann_ joined #cdk
19:27 sneumann__ joined #cdk
20:50 egonw joined #cdk

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