Camelia, the Perl 6 bug

IRC log for #cdk, 2011-03-10

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

All times shown according to UTC.

Time Nick Message
05:55 egonw joined #cdk
06:25 jbrefort joined #cdk
06:26 sneumann joined #cdk
06:29 egonw left #cdk
06:37 egonw joined #cdk
06:49 bag_ joined #cdk
07:55 Gpox joined #cdk
08:24 egonw left #cdk
08:27 jonalv joined #cdk
09:35 egonw joined #cdk
11:18 egonw left #cdk
12:05 egonw joined #cdk
12:08 maclean joined #cdk
12:14 egonw left #cdk
12:19 egonw joined #cdk
12:26 jonalv left #cdk
13:34 maclean "private Map<String, Map<String, Map<String, Integer>>> map;" - hmmmm.
13:34 egonw hi maclean
13:34 maclean hi
13:34 zarah saluton maclean
13:34 maclean Zarah!
13:34 egonw yeah, that sounds like some idea for:
13:35 maclean ?
13:35 egonw ThisIsSomeClass extends Map<String,Map<String,Integer>>
13:36 egonw or at least use a proper variable name, rahter than 'map'
13:36 maclean bah
13:36 maclean would you prefer mapOfStringToMapOfStringToMapOfStringToInteger?
13:37 maclean But the class idea is possible, yes.
13:37 egonw well...
13:37 egonw I would have hoped for something like
13:37 egonw thinking...
13:38 egonw namePerClientPerCountry
13:38 egonw better:
13:38 egonw turnoverPerClientPerCountryPerPlanet
13:38 maclean well,  Iguess it is a residueNameToBondMap
13:38 maclean like TBD : PB : O3B : 2
13:39 egonw 2 == residueNumber
13:39 egonw ?
13:39 maclean 2 == bond order
13:39 egonw ouch
13:39 maclean (or IBond.Order.ordinal() + 1)
13:39 egonw ouch
13:40 egonw you're hurting me eyes, you are
13:40 maclean Oh pschaw.
13:40 maclean I'm sure it will be fine.
13:40 egonw yeah, already getting better
13:41 maclean It's an unappealing approach, but I don't see much of an alternative. PDB files don't have bond orders.
13:42 maclean I read a paper yesterday about working out atom types (and double bonds etc) from 3D coords, which was interesting.
13:44 maclean doi : 10.1021/ci049915d
13:45 maclean Paul Labute++
13:58 egonw Jmol does that too
13:58 egonw CDK has a copy of that code
13:58 egonw RebondTool
13:58 maclean Oh really? I couldn't find that in the Jmol code/
13:58 maclean Ohhh.....
13:59 maclean Well, ok, yes but no.
13:59 egonw still uses the Jmol atom types
13:59 egonw which has some info on bond lengths
13:59 egonw but, that also defines the limitations
13:59 maclean Right, so (crudely) double bonds are shorter
13:59 egonw yes
14:00 maclean Hmmm. I had forgotten that the RebondTool was the reason why I was assigning types to PDB atoms.
14:02 maclean anyway, just rebooting because apparently Safari updates require this...
14:02 maclean back in a sec
14:02 maclean left #cdk
14:11 maclean joined #cdk
14:48 maclean Heh. Experience the full horror here : https://github.com/gilleain/CDKPDB/bl​ob/HEAD/src/test/BondTypeFactory.java
15:13 maclean Hooray! It (seems to) work(s).
15:18 jonalv joined #cdk
15:18 egonw maclean: .ordinal() is dangerous, as it ties you to a particular impl
15:18 egonw maclean: consider using BondManipulator.destroyBondOder(IBond.Order)
15:20 maclean Hmm. Ok. "destroyBondOrder" is a very odd name, but I have seen it before. Thanks for the advice.
15:21 maclean In any case, it has to be tr...oh wait.
15:22 maclean No, never mind.
15:24 maclean What I'm thinking of doing now is testing the RebondTool against the BondDictionary to see if they disagree for any of the PDB hetdict.
15:24 egonw sounds good
15:33 egonw jonalv: https://sourceforge.net/tracker/?func=detail&amp​;aid=3189297&amp;group_id=20024&amp;atid=320024
15:35 jonalv egonw: thankyou
15:39 jonalv egonw: uhm that's not really a fingerprint though, is it? It throws an UnsupOPExc when trying to get the BitSet...
15:41 egonw ah... well... discuss that with Rajarshi
15:41 egonw you need the other method
15:41 egonw I argued this with him, but had no time to come up with a better patch at the time :/
15:41 egonw use getRawFingerprint()
15:42 jonalv uhm yea... but that looks more like you get the signature...
16:23 sneumann left #cdk
17:18 egonw left #cdk
17:21 sneumann joined #cdk
17:27 maclean left #cdk
17:56 jonalv left #cdk
18:11 egonw joined #cdk
18:50 egonw left #cdk
18:56 egonw joined #cdk
19:04 egonw left #cdk
19:12 Gpox left #cdk
20:26 sneumann left #cdk
21:26 jbrefort left #cdk
23:53 bag_ left #cdk

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