Camelia, the Perl 6 bug

IRC log for #bioclipse, 2011-09-02

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

All times shown according to UTC.

Time Nick Message
03:43 egonw joined #bioclipse
06:20 egonw joined #bioclipse
06:25 zaetnick joined #bioclipse
06:47 olass joined #bioclipse
08:32 jonalv joined #bioclipse
08:58 egonw jonalv: pushing a patch to bioclipse.educational *now*
08:58 CIA-120 bioclipse.educational: Egon Willighagen master * r0e43ae2 / (33 files in 5 dirs): Removed pharmbio book specific stuff - http://git.io/k4rQlA
08:58 egonw good
08:58 egonw jonalv++ (for setting that up :)
09:06 olass egonw, jonalv: ping
09:06 * olass is looking at CDK
09:07 egonw olass: I was thinking that we might selectively update parts of the CDK...
09:07 olass is it good practice to have a method declaring that it should return an IChemObject but that always returns an IMolecule?
09:07 egonw one advantage of modularization...
09:07 egonw in particular the core module
09:07 olass egonw: indeed
09:07 jonalv olass: pong
09:07 olass we should discuss with Gpox next week
09:07 egonw that's CDK manager?
09:08 olass no, IteratingMDLReader
09:08 olass it overrides the next() function
09:08 jonalv sounds better to me to return IMolecule if it always does that
09:08 olass I am just thinking if the superclass should say ? extends IChemObject
09:08 olass jonalv: me too
09:09 olass since nextMolecule ALWAYS is an IMolecule, it confused me that the method rerurns an IChemObject
09:09 olass and means I need to cast it
09:09 egonw agreed
09:09 egonw this is a good place to use <T>
09:10 egonw that API is from before generics got introduced into the language
09:10 egonw perfect patch for CDK master
09:10 egonw was thinking about that too already, but no time for it yet
09:10 olass yes
09:10 olass I sort of want to commit but it is soo much easier to just tell you egonw...
09:11 olass but it lowers my commit stats in cdk
09:11 olass don't know how many of these things I've just told you....
09:11 egonw well, and it would be for the fourth CDK paper in some distant future
09:11 * olass was thinking of many suggestions in the past
09:11 olass but yes
09:12 olass you are right about the versions
09:14 egonw well, the work is not in the suggestions, but actually make it happen... suggestions are the easy part
09:14 jonalv bbl
09:17 egonw olass: anyway, you're on the CDK III paper already, not?
09:17 olass egonw: well, to some extent I agree but fixing small fixes is easy but requires having the latest version of CDK checked out, committing, sending patch by SF etc... the overhead for small patches is IMHO a burden but I of course realize it's useful and worth it in the long run.  When you work on a different project, switching to an updated CDK context is not something you do in 10 seconds. Just a comment! :)
09:17 olass egonw: yes, thanks for this
09:18 olass it just gives me even more interest in contributing more :)
09:18 egonw true...
09:18 egonw project management is not easy
09:18 olass indeed
09:18 egonw there are quick fixes, but in the past they resulted too often in very serious regressions
09:18 egonw or: there is no free lunch
09:18 olass completely agree
09:19 olass I think Bioclipse and CDK has been very good for both projects, even though CDK does not depend on Bioclipse but the other way around
09:19 egonw what would be great for the CDK, is someone who'd be running a 'fork' with such experimental small fixes
09:19 olass yes, that could be one idea..
09:19 egonw absolutely... it was great being in Jarl's group being able to work on the CDK so much
09:19 olass and then cherry-pick commits to master?
09:19 egonw right
09:19 olass interesting thought...
09:20 egonw once it is clear nothing serious got broken
09:21 egonw but being "release manager" for two branches is more than enough work for me
09:32 olass understandable
09:45 CIA-120 bioclipse.educational: Egon Willighagen master * rd1214ba / (12 files in 6 dirs): Removed plugins content specific for the PharmBio book: ontologies - http://git.io/WWQ56A
09:52 jonalv joined #bioclipse
09:55 CIA-120 bioclipse.educational: Egon Willighagen master * r0aee0b1 / (22 files in 9 dirs): Renaming stuff from pharmbio to a more general educational - http://git.io/kD-wKg
09:58 CIA-120 bioclipse.educational: Egon Willighagen master * r4f25ab0 / (38 files in 14 dirs): Renamed folders too now - http://git.io/EWU0Eg
10:56 olass_ joined #bioclipse
11:05 egonw olass_, jonalv: http://pele.farmbio.uu.se/jenk​ins/job/bioclipse.educational/
11:10 jonalv joined #bioclipse
13:13 egonw jonalv: can you add this one to the CIA too: de.ipbhalle.msbi ?
13:14 jonalv egonw: what's that?
13:14 sneumann MetFrag
13:14 sneumann It tries to identify a metabolite
13:14 sneumann from its MS2 mass spectrum
13:15 jonalv okey, it's added
13:15 jonalv What part of Bioclipse is using it?
13:16 egonw the metabolomics feature
13:16 egonw well, if I get an understanding with Jenkins, that is
13:30 CIA-120 de.ipbhalle.msbi: Egon Willighagen master * r16afdeb / (2 files in 2 dirs): Some cleanup of meta files - http://git.io/ctuxKQ
13:31 s_wolf joined #bioclipse
13:36 egonw it's blue! it's blue!
13:36 egonw http://pele.farmbio.uu.se/je​nkins/job/de.ipbhalle.msbi/
13:38 zaetnick just a few raindrops, but blue
13:39 egonw ok, next stop: net.bioclipse.spectrum
13:39 egonw sneumann: that will be interesting :)
13:42 egonw https://github.com/bioclipse/biocl​ipse.speclipse/tree/master/plugins
13:47 egonw jonalv: another request... please add to the CIA bioclipse.speclipse
13:48 jonalv egonw: done
13:49 egonw jonalv++
13:49 jonalv egonw: hae you looked aything at that padel code? I am trying to fing that actual fingerprint code... :)
13:49 egonw yeah, I look at it at some point
13:49 zaetnick joined #bioclipse
13:50 egonw jonalv: BTW, I can recommend to make the first commit of your devel branch a commit with the exact code by this author, with his name and email as 'author'
13:50 egonw as we did on that branch on github.com/cdk
13:50 edrin joined #bioclipse
13:51 jonalv egonw: please explain
13:51 egonw one sec
13:53 egonw see the first post Dec 16 commit of the branch: https://github.com/cdk/cdk/commits/dev-1.4.x-padel
13:53 egonw or, directly:
13:53 egonw https://github.com/cdk/cdk/commit/fa2​93501759a21aa747ac74d756b5d6a7190aa30
13:53 egonw that code only and really only dumps the code from the PaDEL software into the CDK tree
13:54 egonw and only after that dump, we started moving it to a better place, etc
13:54 egonw defining Yap Chun Wei <yapchunwei@users.sourceforge.net> as clear original author
13:56 jonalv that's nice, so you dumed everything? Taht commit looks rather small, is that really all?
13:56 egonw no, no
13:56 egonw that's just a dump of the code I was interested at that moment
13:56 egonw the reason:
13:56 egonw cleaning up the code in one go is impossible
13:57 egonw so, bit by bit
13:57 egonw but not a dump of the full PaDEL
13:57 jonalv exactly
13:58 jonalv egonw: but, the article talks about more than one new fingerprint
13:58 egonw yeah, others come directly from CDK
13:59 jonalv no I think it said 7 fingerprints from CDK and 3 others
14:00 jonalv or what have I missed?
14:00 egonw it's been a while
14:00 egonw earlier this week I checked the .xsl he has on the webste
14:00 egonw from that I saw only one new
14:00 jonalv okey
14:00 egonw but paper is probably more accurate :)
14:00 egonw which is good for your commit stats, I guess :)
14:00 jonalv I am having big problems just finding that actual fingerprint to be honest
14:01 egonw did you download the source code?
14:02 jonalv yea
14:02 jonalv I think I have finally found it
14:02 jonalv :)
14:03 samuell joined #bioclipse
14:04 egonw excellent!
14:04 jonalv lol this is simple
14:04 jonalv almost naive
14:04 egonw :)
14:05 jonalv why do substructure seaches using SMARTS?
14:05 jonalv that's like chatting by snail mail, isn't it?
14:06 jonalv egonw: can you give me a SMARTS crash course?
14:08 egonw see section 5.6 in our book :)
14:09 jonalv what does !#1  mean?
14:09 egonw not(hydrogen)
14:11 egonw ! == not
14:11 jonalv I am looking at them and thinking it might be easy to build molecules of them, and use one of the faster fingerpritns to speed up the process
14:11 egonw # == atomic number
14:11 egonw sounds like a good idea
14:11 egonw there are only 4000 of them
14:11 jonalv I feel I know too little SMARTS to do it correctly though
14:12 jonalv I suppose a big dataset could be used to test run on to make sure it's made correctly
14:16 jonalv egonw: I see how they are counting now
14:16 jonalv the three fingerprints not in CDK are, count version of substructure
14:16 jonalv klekotha-roth and count version of klekotha-roth
14:17 egonw ah, ic
14:17 jonalv it's trivial to poert klekotha-roth, but probably possible to make it faster I would think
14:17 jonalv *port
14:17 jonalv as for count versions my fork covers that
14:18 egonw the work lies in unit tests, JavaDoc, PMD warnings
14:18 jonalv count fingerprints support is a bit of a struggle in todays cdk
14:18 egonw yeah, master addresses that, not?
14:18 jonalv PMD warnings?
14:18 egonw well, to some extend
14:19 egonw pmd.sf.net
14:19 jonalv it's just like the pubmed fingerprint I think only many more SMARTS
14:20 jonalv egonw: do we have any indications that this fingerprint performs well and is worth using considering it's speed?
14:20 jonalv *its
14:20 egonw dunno
14:20 egonw no one tested it independently, AFAIK
14:21 egonw but the substructures alone are already interesting...
14:22 jonalv well yea it seems very intresting
14:22 jonalv I think so too
14:24 jonalv it's just that manually converting all those SMARTS to substructure SMILES seems painfull. Do ou know some tool that can help?
14:24 egonw why too SMILES??
14:24 egonw that will *not* work
14:24 egonw there is a combinatorial explosion of SMILES
14:24 jonalv really? I thought that would be easeast consiering they are almost that
14:24 zaetnick joined #bioclipse
14:25 jonalv egonw: no I don't think so cause I don't think they are using the whole power of SMARTS
14:25 jonalv I think they are just substructures represented as SMARTS
14:25 egonw oh, that would be sad
14:26 jonalv egonw: well everything points to that
14:27 jonalv which is why my first impression was that this was naive
14:28 jonalv egonw: because the original article only talks about substructures
14:29 egonw now, a C in SMILES does have a different meaning...
14:29 egonw but that excluded...
14:29 jonalv egonw: quote from the original paper: "The resulting
14:29 jonalv substructures were represented as SMARTS strings (symbols representing
14:29 jonalv the non-hydrogen wild-card ‘[!#1]’ were used where appropriate)."
14:30 CIA-120 bioclipse.ds: Ola Spjuth master * rab825af / plugins/net.bioclipse.ds.sdk/src/net/biocl​ipse/ds/sdk/libsvm/ParallelSignModel.java : removed unused file - http://git.io/lHMeqA
14:30 CIA-120 bioclipse.ds: Ola Spjuth master * rbb8e938 / (10 files in 3 dirs):
14:30 CIA-120 bioclipse.ds: GUI + refactoring of SignatureDatasetFromSDF
14:30 CIA-120 bioclipse.ds: * Refactored to use IDataset implementation
14:30 CIA-120 bioclipse.ds: * Implemented GUI dialog for creating dataset from SDF - http://git.io/cKPm2w
14:30 egonw origin paper is the Klethato one
14:30 egonw ?
14:30 jonalv also the title is:
14:30 jonalv Chemical substructures that enrich for biological activit
14:30 jonalv y
14:30 jonalv so yea it's substructures all right
14:30 jonalv and doing substructure matching using SMARTS seems naive to me
14:31 jonalv egonw: yea the Klekota Roth paper
14:32 egonw well, more accurate than SMILES
14:32 jonalv egonw: so if you can think of an easy way to convert them into real substructures I suggest I do that + speed it up using one of CDKs fingerprinters before doing the actual sub structure match
14:33 olass_ jonalv: ping
14:33 jonalv olass_: pong
14:33 egonw SMARTS are real substructures
14:33 jonalv egonw: you know what I mean
14:33 egonw SMARTS is a substructure language
14:33 egonw no, actually I don't :)
14:33 egonw SMILES is not a substructure lang
14:33 egonw so, I'm lost at what you are getting at :)
14:34 jonalv SMARTS supports general queries like "take this SMILES and add any non aromatic thing over here and any aromatic thing over there"
14:34 jonalv using that to do just plain substructure search is overkill
14:34 egonw that's something else...
14:34 egonw but SMILES does not have 'not-a-hydrogen'
14:35 jonalv egonw: but why would we nned that?
14:35 jonalv oh hang on I see what you missed now
14:35 egonw apparently those Harvard guys were using that
14:35 egonw [09/02/11 16:29] <jonalv> egonw: quote from the original paper: "The resulting
14:35 egonw [09/02/11 16:29] <jonalv> substructures were represented as SMARTS strings (symbols representing
14:35 egonw [09/02/11 16:29] <jonalv> the non-hydrogen wild-card ‘[!#1]’ were used where appropriate)."
14:36 jonalv egonw: okey I admit I suck at corresponding though text
14:36 jonalv let's take it from the top
14:36 olass_ egonw: ping
14:36 olass_ egonw: in bioclipse.cheminformatics, is it OK if I merge 2.5 into master branch?
14:37 egonw yes
14:37 egonw please do
14:37 olass_ ok
14:37 egonw but don't use rebase :)
14:38 jonalv egonw: What I am trying to say, and probably not in any easy to understand way or anything, is that if we can produce the for example ExtendedFingerprint fingerprints for those substructures we can pre filter the actual SMARTS query with a much faster matching and thus speed up the query. However to do that we need the "real substructures" in a format than we can feed to the ExtendedFingerPrinter.
14:39 jonalv In theory we might end up with a decent fingerprint wihtout the SMARTS mathcing but to get the exactly same one we need to do the SMARTS mathcing if the other fingerprinting process sees a hit
14:39 jonalv egonw: did I make a better work this time? It's not so easy for a bear of very little brains
14:40 egonw prefiltering sounds good
14:40 egonw I dunno what the SMARTSes look like
14:40 jonalv egonw: do you understand what I mean now with "real substructures"?
14:41 egonw and how well they can be used for prefiltering
14:41 jonalv like SMILES but with the [!#1] operator
14:42 egonw but why would SMILES matching first give so much boost?
14:42 egonw assuming each SMARTS matches to just one SMILES
14:43 jonalv I don't know exactly how the SMARTS matching code in CDK works
14:43 jonalv but my feeling is that fingerprint matching is much faster than anything it can do
14:44 egonw they are both using the same matching code
14:44 jonalv egonw: hang on, what do you mean with SMILES matching? I suggest using those SMILES to build fingerprints and then do a fingerprint match first. Jst like the classic substrucutre search speedup using fingerprints
14:48 CIA-120 bioclipse.cheminformatics: Ola Spjuth master * r21cedec / plugins/net.bioclipse.chemoinformatics/plugin.xml : Added a general popup menu "Create dataset". - http://git.io/_Xb9Bw
14:50 egonw but both a substructure search with SMILES and SMARTS uses the same algorithm
14:50 egonw I think performance will be comparable
14:50 egonw however, most FPs you prescreen with use *other* algorithms
14:50 egonw that are much faster than substructure searching algorithms
14:51 jonalv egonw: yea I am not suggesting we do a match with that geometry thing but I am suggesting a preescreen using a fingerprint
14:51 sneumann http://theplateisbad.blogspot.com/2010/11​/pgchemtigress-sets-new-world-record.html
14:51 zarah sneumann's link is also http://tinyurl.com/2vzzw4q
14:52 sneumann although it uses OpenBabel under the hood ...
14:53 jonalv sneumann: fancy I suppose :)
14:54 sneumann It makes searching in PubChem with *multiple* substructures and full SQL powers fun
14:54 sneumann http://journal.imbio.de/article.php?aid=157
14:55 sneumann See the SQL on p.9
14:56 jonalv okey, I am confused now
14:57 sneumann questions ?
14:57 jonalv what about that SQL is specific chemistry related
14:57 jonalv it looks like lookups of values to me
14:57 sneumann the <= is the "is substructure" predicate
14:57 sneumann the thing has fingerprinted pubchem
14:57 sneumann and fingerprinted the query structures (fragments here)
14:58 sneumann and postgres (!) does a quick prescan using the fingerprints
14:58 jonalv Oh
14:58 sneumann and only on the remaining onces a "proper" substructure test
14:58 sneumann Essentially, the fingerprints are used as database index
14:58 jonalv so to relate to the thing we were discussing, what fingerprint algorithm is it using?
14:59 sneumann Ernst-Georg Schmid. Database-driven procurement of substances in the researching
14:59 sneumann chemical industry - An algorithmic optimization approach. PhD thesis, Mercator School
14:59 sneumann of Management - Fakult¨at f¨ur Betriebswirtschaftslehre - Technology and Operations Management
14:59 sneumann - Wirtschaftsinformatik und Operations Research, June 2010.
15:00 sneumann He is using 1024 bits of OpenBabel OB2 fingerprinter,
15:00 jonalv sneumann: and by the way, do you know some handy way to convert SMARTS only containing what to me looks like SMILES + [!#1] into something that one could fingerprint? :)
15:01 sneumann plus a set of ~500 substructures selected such that they are most discriminatory for a given training set
15:01 jonalv wow I should probably look into the OpenBabel stuff one day... I know way too little cheminformatics :)
15:01 sneumann http://duepublico.uni-duisburg-essen.de/ser​vlets/DocumentServlet?id=22715&amp;lang=en
15:01 zarah sneumann's link is also http://tinyurl.com/3p8tqow
15:03 sneumann p57 has a schema of the search strategy
15:03 jonalv hm that one is in german though :)
15:04 sneumann He had a nice presentation at CCC2010 in Goslar
15:05 jonalv yea I keep hearing about this Goslar thing
15:05 olass_ sneumann: when is the goslar meeting this year?
15:09 jonalv egonw, sneumann: You people are good at cheminormatics, maybe you could help a bear of very little brain on this friday afternoon (or sneumann you are not on US timezone or so are you?) If I have a SMARTS query representation of a substructure like for example: [!#1][CH]([!#1])[CH]([!#1])[!#1] why can't I just simply remove all [!#1] and get the substructure as SMILES?
15:11 egonw because you'd get false positives
15:12 jonalv egonw: ah but is not a problem as long as I have no false negatives
15:12 jonalv egonw: the false positives will be sorted away by the SMARTS matching in the next step
15:13 egonw yes, the question is if the speed up is more than the downside of having to do two substructure searches
15:13 jonalv egonw: I have a feeling that we are not understanding eachother
15:14 egonw well, I do have fever
15:14 egonw so it would not surprise me :)
15:14 jonalv egonw: maybe you could explain which two substructure searches you are talking about
15:14 egonw whether I have big or small brains... brains-with-fever always wins over some brains, in terms of nonsense
15:14 jonalv egonw: or should we postpone this to another day?
15:14 egonw SMILES + SMARTS
15:15 egonw why otherwise do you want to create SMILES?
15:15 jonalv to build fragments to fingerprint
15:15 sneumann sounds essentiall like that german phd thesis
15:16 jonalv sneumann: yea exactly
15:16 sneumann p153 explains
15:16 sneumann how he is combining 1024 bit "traditional" FP2 from OpenBabel
15:16 jonalv sneumann: it's the same general approach that is used when speeding up substructure searches by pre filtering with fingerprints
15:16 jonalv it's not anything strange
15:16 sneumann with 512 bits from predefined substructure searches
15:16 jonalv it's a school book example
15:17 egonw ah, ok...
15:17 egonw so, SMARTS -> SMILES -> FP ?
15:17 jonalv sneumann: yea he is proabbly doing some sort of fancy fingerprinting. I am was thinking just the astet CDK fingerprint there is
15:17 jonalv egonw: yea
15:17 egonw then you can just remove the [!#1]
15:18 jonalv *fastest
15:19 jonalv egonw: well in that case I think it very possible to port that fingerprint to CDK and to speed it up.
15:19 jonalv maybe we are in the wrong channel btw? :)
15:20 * egonw does not mind
15:20 egonw jonalv: this trick will particularly help, if you group the 4k substructures with the same FP
15:20 egonw or even make a binary search tree
15:20 jonalv egonw: how do you mean?
15:20 egonw well, naively, you'd just create one FP for each substructure
15:21 egonw and match bit by bit
15:21 jonalv yea
15:21 jonalv fail fast
15:21 egonw but if the FPs are the same for more substructures, and there is no hit, you can...
15:21 egonw right
15:21 jonalv oh
15:22 egonw better, even... make a binary tree, and compare FP bit by bit... and rule out multiple substructures based on that FP
15:22 jonalv will that speed up actually make any difference you think?
15:22 egonw because if that bit is not found in the structure, then *all* substructure with the FP set will not match
15:22 egonw dunno
15:22 jonalv true
15:22 egonw interesting study
15:22 sneumann He is using the dictionary.fp3.txt buried in https://github.com/sneumann/pgchem​/blob/master/setup/tigressdata.zip
15:23 jonalv sneumann: you saying someone is doing the exact same thing?
15:24 sneumann could very well be
15:25 sneumann you could ask him for his CCC slides
15:25 sneumann as I don't want to hand them out
15:25 jonalv I don't even know who we are talking about. But I doubt I can use any code or so. Besides I am only ontrested in building a faster version of that fingerprint
15:27 jonalv sneumann: is this person on IRC?
15:27 egonw jonalv: well, that same route could be applied to speed up the PubChemFP too
15:28 sneumann no, haven't seen him on IRC
15:28 jonalv egonw: yea, absolutely, but is anyone actually using it?
15:28 sneumann http://pgfoundry.org/users/ergo70/
15:29 sneumann pgchem @nospam@ tuschehund.de
15:29 jonalv egonw: I have started to look at it as some sort of cute proof of concept thing
15:29 egonw the CDK FPs are used a lot, yes
15:29 jonalv it's slow and outperformed by all other CDK fingerprints
15:29 jonalv is it not?
15:29 jonalv egonw: I mean the pubmed one in particular
15:30 egonw http://www.citeulike.org/user/egonw​/tag/cdk--usespackage--fingerprint
15:30 egonw I don't have more finetuned detail
15:30 jonalv but yea, if this works that one could be speeded up, as well as that other SMARTS based one
15:30 egonw on to *which* FP they use
15:30 jonalv or is it not just substructre perhaps?
15:31 jonalv I mean not all SMARTS can trivially be speeded up like this can they?
15:31 sneumann running off now
15:31 sneumann nice weekend!
15:32 jonalv sneumann: you too
15:32 egonw no, not all
15:32 jonalv well I suppose one could do some sort of fancy regexp finding the largest substructure of a SMARTS and then fingerrint that
15:32 jonalv hm that's complicated stuff though
15:33 jonalv and hard to know how much it actually would help
15:33 jonalv for a given SMARTS
15:33 jonalv I am not too familiar with how SMARTS strings normally look
15:34 jonalv it's and intresting field though
15:35 jonalv it's just that the step from SMARTS to fingerprint must be done correctly
15:35 jonalv of course
15:42 egonw grpmh... cannot get net.bioclipse.bibtex to work :(
15:44 CIA-120 bioclipse.speclipse: Egon Willighagen master * rdce1e3a / (12 files in 10 dirs): Initial run of cleanups: 2.5.0 qualifiers, removed many versioned depedencies, updated for XOM and CMLXOM changes - http://git.io/9yVlrg
15:47 CIA-120 bioclipse.speclipse: Egon Willighagen master * r797f465 / (3 files): Added a plain, Hudson-oriented update site - http://git.io/Sbrazg
15:51 CIA-120 bioclipse.speclipse: Egon Willighagen master * r34374a7 / features/net.bioclipse.speclipse_site/feature.xml : Added the Speclipse feature to the new update site - http://git.io/jLoYbw
16:05 egonw btw, really making great progress iwth the metabolomics stuff...
16:13 CIA-120 bioclipse.speclipse: Egon Willighagen master * re9bd91d / features/net.bioclipse.spe​clipse_feature/feature.xml : Removed two plugins that don't exist anymore, and fixed feature deps - http://git.io/Jag5vw
17:38 CIA-120 bioclipse.speclipse: Egon Willighagen master * r49c58a8 / (4 files in 4 dirs): Removed more traces of net.bioclipse.xom - http://git.io/ic0FAw
17:55 CIA-120 bioclipse.speclipse: Egon Willighagen master * rf7adc21 / plugins/net.bioclipse.bibtex/plugin.xml : Updated content type definition to current Eclipse standards - http://git.io/WpzZNw
17:55 CIA-120 bioclipse.speclipse: Egon Willighagen master * r839b3c9 / (3 files in 3 dirs): Further OSGi updates: removed more version in deps, and some unwated build.props stuff - http://git.io/xvGf1w
18:05 CIA-120 bioclipse.speclipse: Egon Willighagen master * rb8c7349 / plugins/net.bioclipse.bibtex/.classpath : Export the JabRef jar too - http://git.io/_HBSgg
18:23 egonw_ joined #bioclipse
19:01 CIA-120 bioclipse.speclipse: Egon Willighagen master * r4b86507 / plugins/net.bioclipse.bibtex/META-INF/MANIFEST.MF : Make all of JabRef visible - http://git.io/WQFc5w
19:23 CIA-120 bioclipse.speclipse: Egon Willighagen master * ra9c05d2 / (10 files in 7 dirs): Misc updates: for CDK 1.4, missing deps, etc - http://git.io/hsQWqw
19:29 CIA-120 bioclipse.speclipse: Egon Willighagen master * rc8bcd32 / (2 files in 2 dirs): CMLExtractor comes from the net.bioclipse.xom plugin - http://git.io/GGYHBg
19:42 CIA-120 bioclipse.core: Egon Willighagen 2.5.x * ra5f07e1 / features/net.bioclipse.core_feature/feature.xml : Added net.bioclipse.xom to the core feature, needed by bioclipse.speclipse - http://git.io/fsSXNg

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