From: SourceForge.net <no...@so...> - 2010-03-01 02:50:25
|
Patches item #2944080, was opened at 2010-02-01 16:37 Message generated for change (Comment added) made by rajarshi You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=320024&aid=2944080&group_id=20024 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: Needs Review Status: Open Resolution: None Priority: 5 Private: No Submitted By: Lars Carlsson (lc69) Assigned to: Rajarshi Guha (rajarshi) Summary: No result with MCSS Initial Comment: A call to UniversalIsomorphismTester.getSubgraphAtomsMap(mol1, mol2); does not render an MCSS for the structures mol1=CCC(CC)(C(=O)NC(=O)NC(C)=O)Br mol2= CCC(=CC)C(=O)NC(N)=O. They seem to overlap. ---------------------------------------------------------------------- >Comment By: Rajarshi Guha (rajarshi) Date: 2010-02-28 21:50 Message: Patch 2 fails to apply against master ---------------------------------------------------------------------- Comment By: Egon Willighagen (egonw) Date: 2010-02-24 15:10 Message: Attached two patches against CDK master. First adding unit tests; Second to update the JavaDoc to describe more clearly that it maps a substructure onto a graph, and returns that/those mapping(s). ---------------------------------------------------------------------- Comment By: Egon Willighagen (egonw) Date: 2010-02-24 14:50 Message: Peter, yes, the reason my first unit test worked is indeed because I used a different API. I wrote further unit tests, and realized after long thinking and debugging that the getSubgraphAtomsMap(...) map actually *is* different. I have sent a more extensive analysis to the list, but it boils down that this get method requires the second graph to be a substructure of the first. It cannot be used for MCSS. ---------------------------------------------------------------------- Comment By: Peter Odéus (peterodeus) Date: 2010-02-05 04:42 Message: If I'm connecting the dots correctly, the checked in (and working) Junit test (1) uses a different api call compared to the api call described in this bug report. The question is whether the getSubgraphAtomsMap(...) is the preferred api call over a call to getOverlaps(...). And, if so, why? 1: http://pele.farmbio.uu.se/cgi-bin/gitweb.cgi?p=cdk-jchempaint.git;a=blobdiff;f=src/test/org/openscience/cdk/isomorphism/UniversalIsomorphismTesterTest.java;h=f2b9031dfd22ea41257f8051b54032a690c7525b;hp=2f8bcd40fb89a14003485cc1908b9de0d3df3cdf;hb=b0666e97443a15a3ca56e43bf745c3c8b0d04053;hpb=68e696a54d379947c5583de304c25add744f43ec ---------------------------------------------------------------------- Comment By: Egon Willighagen (egonw) Date: 2010-02-01 17:08 Message: Lars, I just checked in master, and it seems to work fine there... the patch is commited, and I sent you it off-line too... Please let us compare code... And, please let me know what CDK version you are using... ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=320024&aid=2944080&group_id=20024 |