Camelia, the Perl 6 bug

IRC log for #bioclipse, 2010-07-07

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

All times shown according to UTC.

Time Nick Message
05:51 annzi joined #bioclipse
06:42 annzi joined #bioclipse
06:53 Gpox joined #bioclipse
07:19 olass joined #bioclipse
07:20 olass_ joined #bioclipse
08:34 jonalv joined #bioclipse
08:35 jonalv hi
08:35 zarah privet jonalv
08:39 masak joined #bioclipse
09:12 CIA-71 bioclipse.balloon: jonalv 2.4.x * r7a15c45 / (7 files in 7 dirs): set version 2.4.0.RC3 - http://bit.ly/c5hRWO
09:16 CIA-71 bioclipse.ds: jonalv 2.4.x * r2582bac / (14 files in 14 dirs): set version 2.4.0.RC3 - http://bit.ly/cEWUV0
09:16 CIA-71 bioclipse.ds: jonalv 2.4.x * r146ba0a / (2 files in 2 dirs): fixed deps - http://bit.ly/c1ocHQ
09:21 CIA-71 bioclipse.qsar: jonalv 2.4.x * r6fa1937 / (11 files in 11 dirs): set version 2.4.0.RC3 - http://bit.ly/bhgQ7Y
09:21 CIA-71 bioclipse.qsar: jonalv 2.4.x * rd0cb48b / (2 files in 2 dirs): fixed deps - http://bit.ly/9AqJBQ
09:30 CIA-71 bioclipse.statistics: jonalv 2.4.x * rc023bba / (5 files in 5 dirs): set version 2.4.0.RC3 - http://bit.ly/9nQhz7
09:39 edrin joined #bioclipse
09:40 edrin hi
09:40 zarah ni hao edrin
09:41 jonalv hi edrin, so you going to test 2.4.0.RC3 now? :)
09:42 edrin I can't, unfortunately I am at work now
09:42 edrin I can test at the weekend
09:42 jonalv ah right, you can't even download and install a Bioclispe release from work...
09:43 jonalv yea I hope to release sharp version before weekend
09:43 jonalv (feel free to call me naive)
09:43 jonalv :)
09:43 edrin no problem :)
11:54 CIA-71 bioclipse.ds: Ola Spjuth 2.4.x * r79bd64e / (5 files in 2 dirs):
11:54 CIA-71 bioclipse.ds: Various PropertyCalculators fixes:
11:54 CIA-71 bioclipse.ds: * Semi-colon-separated atom signatures instead of comma
11:54 CIA-71 bioclipse.ds: * Height 0-5 instead of 1-5 as PropertyCalculator
11:54 CIA-71 bioclipse.ds: * Set correct id for Molecular Signature in class - http://bit.ly/9Swvw3
12:03 annzi joined #bioclipse
12:41 annzi joined #bioclipse
12:43 jonalv joined #bioclipse
13:14 slyrus joined #bioclipse
13:38 maclean joined #bioclipse
13:38 maclean hi
13:38 zarah hello maclean, you fantastic person you
13:39 maclean Watch it, toaster.
13:39 jonalv hehe
13:42 olass hi maclean
13:42 maclean hi olass
13:42 * olass is reading an email from maclean
13:43 * maclean is debugging
13:44 olass maclean++ //for the explanation
13:44 maclean Sorry though, it should handle aromatics, I forgot about them.
13:44 olass maclean: is it possible for me to do a quick-hack and copy some class into net.bioclipse.signatures plugin to bridge from cdk to signature molecule?
13:45 maclean Uhmmm.
13:45 olass egonw will not be back until a few eeeks, and this is sort of urgent..
13:45 maclean I guess you don't need a lot of the cdk machinery
13:45 maclean like the bits in the various xml files, etc
13:45 maclean so maybe just the contents of the o.o.cdk.signature package
13:45 olass just a translation from IAtomContainer to signature.chemistry.Molecule
13:46 maclean Err.
13:46 olass no?
13:46 maclean Not really.
13:46 olass ok?
13:46 maclean The CDK now has a MoleculeSignature and AtomSignature that take an IAtomContainer.
13:46 olass sorry, cannot use latest CDK
13:47 maclean But they have no relation to the signature.MoleculeSignature
13:47 maclean Oh.
13:47 olass not in Bioclipse yet
13:47 maclean Hmmmm.
13:47 olass I have implemented my own calculations
13:47 olass (with help of Lars)
13:47 olass now I serialize an AC to SDF, and read that with Signature.fromSDF()
13:47 olass that serialization breaks aromaticity
13:47 maclean Ok.
13:47 olass so a translation would suffice for me
13:48 olass is that complicated?
13:48 maclean Probably not.
13:48 maclean You need to take an IAtomContainer that has aromaticity flags, and convert that to a signature.
13:48 olass yup
13:49 olass well, convert that to a signature.chemistry.Molecule would be enough
13:49 maclean Yes, but that class (sig.chem.Mol) has no concept of aromatic...
13:49 maclean IT's really basic.
13:50 olass hmm, so the lib Signatures built with mvn cannot handle aromaticity?
13:50 maclean No
13:51 * olass is confused now
13:51 olass I thought you said it could..
13:52 maclean Nope, sorry. The cdk can handle aromaticity, the signature.chemistry package cannot.
13:52 maclean I could add aromaticity support to signature.chemistry, but it was never intended to be anything but a test package.
13:53 maclean Is the code you are working on in plugins/net.bioclipse.ds.signatures/src/net/biocl​ipse/ds/signatures/prop/calc/AtomSignatures.java ?
13:54 olass correct plugin, yes
13:55 olass SignaturesManager
13:56 maclean Hmm. So in SDF files the bond order is '4' for aromatic - perhaps it would be enough to make that mean aromatic, and work with that.
13:57 olass yes, but CDK does not support writing that
13:57 maclean Oh.
13:57 maclean Yes, you said. I forgot.
13:57 maclean Hmmm.
13:57 olass and I really would like to avoid serializing to SDF every time..
13:57 maclean Indeed.
13:59 maclean So what is needed is both the 'bridge' class (SignatureMoleculeCDKAdapter ? :) and support in the signature.chemistry.Molecule for aromatic.
13:59 maclean So that you could have some function "cdkMoleculeToSignatureMolecule(IAtomContainer) : signature.chemistry.Molecule"
14:00 maclean Or whatever you feel like calling it :)
14:01 olass exactly!
14:02 olass maclean: is this feasible?
14:02 olass and could you help mewith this?
14:02 maclean olass : Certainly. Ok, well I'll make the changes to the signatures package.
14:02 olass thx
14:03 olass and then explain/help with the bridge?
14:03 maclean Then I'll make the adapter class separately (probably just pasebin it) as I don't want to introduce circular dependencies.
14:03 olass super
14:03 olass maclean++
14:03 maclean n/p.
14:49 masak joined #bioclipse
14:55 maclean olass : http://gist.github.com/466796
14:55 maclean A test is at http://gist.github.com/466798
14:56 olass maclean++
14:56 maclean and the commits to the signature repository are in
14:56 olass super
14:56 * olass will test with the now working aromaticity in MolTable in Bioclipse
14:56 maclean ok
14:57 masak maclean: did you call zarah a toaster earlier today? :)
14:57 maclean Not I m'lud.
14:57 masak maclean: should I change the 'fantastic person' greeting?
14:57 maclean >.> <.<
14:57 masak :P
14:58 maclean No, not at all.
14:58 olass zarah: three cheers for maclean
14:58 zarah olass: excuse me?
14:58 olass ?
14:58 masak three cheers for me!
14:58 zarah hooray! hooray! hooray! :)
14:58 masak :)
14:58 olass three cheers for maclean
14:58 zarah hooray! hooray! hooray! :)
14:58 maclean A lazily computed infinite list of cheers for masak
14:59 jonalv zarah: hug olass
14:59 * zarah hugs olass and blushes
14:59 masak maclean: I see you're keeping up with the news :P
14:59 * olass hides
14:59 maclean masak : indeed.
14:59 masak maclean: big release on the 29th. \o/
14:59 maclean cool.
15:00 * masak is OT
15:00 masak @slap masak
15:00 zarah Come on, let's all slap masak
15:00 masak oops.
15:00 * jonalv slaps masak
15:00 masak ow.
15:00 jonalv masak: you brought that upon youself
15:01 * masak rubs cheek
15:01 maclean masak : by the way, can AOP bring mixins to Java?
15:01 * maclean is googling it, also
15:01 * jonalv tries to remember what micins was again
15:02 jonalv |mixins
15:02 masak jonalv: a mixin is a behaviour that you can add to an instance after it's been created.
15:03 jonalv decorator?
15:03 masak where 'behaviour' means 'set of methods'.
15:03 maclean oh, actually, that's probably not what I meant then.
15:03 masak jonalv: well, they are similar.
15:03 masak maclean: re AOP/mixins: I don't know.
15:03 maclean I was thinking of adding methods to a class without inheritance
15:04 masak *maybe* reflection can do it.
15:04 maclean well, mainly I was thinking "why is this so painful", but anyway.
15:04 masak but modifying a class post facto is evil, even in Perl 6.
15:04 masak well, unless you predeclare, of course.
15:04 jonalv maclean: I don't fully understand you use case
15:06 maclean Hmmm. Well, I had an "AbstractBaseEngine" which does some stuff, but the actual engine classes didn't seem like they should be inheriting the behaviour, just acting like a BaseEngine.
15:07 maclean (where an 'engine' matches a pattern to a protein structure).
15:07 maclean I think I am thinking of 'roles'.
15:07 jonalv soudns like an interface to m
15:07 jonalv e
15:08 maclean Hmmm. There is an interface, but also a base class.
15:08 jonalv I think inheritng behavior without acting like sounds like interface and different implementations but it could be that I missunderstand you
15:08 maclean I seem to do that quite a lot.
15:09 maclean Not misunderstand people (though I probably do that too :)
15:10 maclean anyway, will probably get back to it...
15:10 maclean bye o/
15:10 jonalv huh?
15:10 * jonalv neverminds
15:10 * olass neverminds too
15:13 masak zarah: never mind
15:13 * zarah never minds
15:17 olass zarah: do you mind?
15:17 zarah olass: excuse me?
15:17 masak zarah: excuse me?
15:17 zarah masak: you've just exceeded my capabilities :/
17:29 slyrus_ joined #bioclipse

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