Camelia, the Perl 6 bug

IRC log for #cdk, 2011-05-19

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

All times shown according to UTC.

Time Nick Message
02:15 ThePing joined #cdk
02:15 ThePing left #cdk
05:57 jbrefort joined #cdk
06:42 egonw joined #cdk
06:44 sneumann Hi, yesterday I was about to ask a shapado question, but I might have found the answer myself.
06:44 sneumann I am (still) looking for a way to edit reactions, with the focus on creating the actual data,
06:44 sneumann not getting the perfect layout.
06:45 sneumann I'd like the reaction to be readable by CDK, so CMLreact might be a good data exchange format for this.
06:45 sneumann I had done sth. like this years ago with an (even then) ancient jchempaint:
06:46 sneumann http://msbi.ipb-halle.de/~sne​umann/naringeninFragments.pdf
06:46 sneumann I just realised that the current JCP nightly might be able to do something similar.
06:48 sneumann But it is unable to read the CML files created back then
06:51 Gpox joined #cdk
07:15 sneumann Ack, I can draw a reaction in JCP nightly just fine.
07:16 sneumann Not sure yet how to do "nested" reactions, where a product becomes a rectant in a second reaction
07:18 egonw ah, that I don't know either...
07:18 egonw have you tried making the product a reactant of a second reaction?
07:18 sneumann yes, seems to work
07:18 sneumann you can choose which reaction this is product of
07:19 egonw right
07:19 sneumann Unsure how to name the reaction, the reaction-7623572634 is somewhat confusing
07:19 sneumann seems, the reactions are not nested,
07:19 sneumann and "cleanup" will layout them separately.
07:20 sneumann will check the CML next
07:22 sneumann OK, I get a plain reaction list
07:28 egonw indeed
07:28 egonw Miguel worked on better support of reactions schemes in the CDK
07:28 egonw but that is not available at the JCP editing level, AFAIK
07:44 sneumann Ah, at least the molecule id of the molecule that is both product and reactant is stable
07:44 sneumann so, implicitely there is a hierarchy
07:45 egonw yes
07:45 sneumann but all data (atoms, bonds) are duplicated
07:45 egonw oh...
07:45 egonw yes, allowed, in fact
07:45 egonw not optimal indeed
07:46 egonw the writer does not have an option to have reaction refer to molecules defined elsewhere
07:46 egonw such IOOption could be a nice patch
07:52 sneumann left #cdk
07:53 sneumann joined #cdk
07:55 jonalv joined #cdk
07:55 jonalv zarah: seen maclean?
07:55 zarah maclean was last seen 1 d 20 h 35 m 57 s ago in #cdk saying 'email me?'
08:20 sneumann I think we found the actual workflow for our JCP use case:
08:20 egonw cool
08:21 sneumann Create a CML reaction file with some text editor, where we have as many reactions as our massspec expert (or literature) suggest for a fragmentation reaction
08:22 sneumann import his fragments from a CDX (via SMI conversion by babel), because JCP won't import multiple mols from an SDF
08:22 sneumann but it works in SMI. (Is there a reason ?)
08:22 sneumann Then, assign the products / reactants to the existing reactions we have created beforehand
08:24 egonw SMI?
08:36 egonw left #cdk
08:38 sneumann Bugger, JCP is confused by empty reaction lists
08:38 sneumann error message "Structure does not have bonds or atoms"
08:39 sneumann Sure there are none, but there are also no structures in first place.
08:45 sneumann JCP Trac is throwing all kinds of Warnings / Errors atm.
08:45 sneumann Warning: Can't synchronize with the repository (/nfs/sf-svn/j/jc/jchempaint does not appear to be a Subversion repository.). Look in the Trac log for more information.
08:45 sneumann and /nfs/sf-svn/j/jc/jchempaint does not appear to be a Subversion repository.
08:49 sneumann Ack, I created https://sourceforge.net/apps​/trac/jchempaint/ticket/121 for this
09:08 alchimis1e is now known as Alchimiste
09:08 Alchimiste moin
09:08 zarah saluton Alchimiste
11:16 maclean joined #cdk
11:16 maclean hi jonalv
11:18 jonalv maclean: where can I find the source for signature.AbstractVertexSignature?
11:19 maclean mom
11:19 maclean https://github.com/gilleain/signatures
11:19 maclean https://github.com/gilleain/signatu​res/blob/master/src/main/java/signa​ture/AbstractVertexSignature.java
11:20 jonalv ah, maclean how come this is not part of CDK?
11:21 maclean it's meant to be cdk-independant.
11:21 maclean Indeed, it is chemistry independant.
11:21 maclean That is, there is a CDK implementation of signatures in the o.o.cdk.signature package.
11:22 jonalv yea that was what I was looking at. I wanted to see the base class but I didn't have that source...
11:22 maclean I guess the source jar could be included in the cdk distribution as well..
11:23 jonalv ah, yea that makes sense
11:23 maclean would make it easier for developers to attach source when debugging.
11:23 jonalv one still must know to do it thouhg...
11:24 jonalv :(
11:24 jonalv I want automagic, but I guess that is asking too much
11:24 maclean Hah! No, demand magic, and maybe we'll get it :)
11:25 maclean Certainly, using tomcat from eclipse feels like magic when it auto-reloads servlets when you save changes to the source.
11:25 jonalv yea while we are at it can I get a magic carpet too? :)
11:26 maclean and a pony.
11:45 maclean left #cdk
13:05 jbrefort left #cdk
13:52 Gpox left #cdk
14:08 Conrad left #cdk
14:08 s_wolf left #cdk
14:57 sneumann Anybody knows who is conni75 in the JCP trac ?
15:19 sneumann left #cdk
17:50 jonalv left #cdk
18:32 sneumann joined #cdk
18:38 sneumann left #cdk
20:51 jbrefort joined #cdk
21:26 jbrefort left #cdk
22:13 stain left #cdk
22:18 stain joined #cdk
22:19 stain left #cdk
22:19 stain joined #cdk

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