Camelia, the Perl 6 bug

IRC log for #cdk, 2009-05-21

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

All times shown according to UTC.

Time Nick Message
05:22 egonw joined #cdk
05:28 bag joined #cdk
05:50 rojasm joined #cdk
06:09 rojasm joined #cdk
06:21 jbrefort joined #cdk
07:04 egonw joined #cdk
07:21 rojasm egonw. I have the file 0002.stg01.rxn for testing reactions. And I see with JCP what it contains only one reactant and one product but with more of one structures each one. The RXNreader should extract more of one reactants and products. But If I take a look at the sourcecode I don't see different blocs molfiles but only one-big.
07:21 rojasm that means that this mdl rxn reaction is bad created, isn't it?
07:21 rojasm the 0002.stg01.rxn is in cdk.data.mdl too
07:22 egonw yes, MDL RXN supports about any thing...
07:22 rojasm mmm
07:22 rojasm according this document not
07:23 egonw well, I don't think the format enforces it...
07:23 egonw reactants etc are represented at MDL molfile
07:23 egonw which allow non-connected stuff
07:23 egonw the format had no semantics to enfore those ideas
07:23 egonw which makes it easy for people to do otherwise
07:59 timvdm joined #cdk
08:15 shk3 joined #cdk
08:53 carsten joined #cdk
09:39 jbrefort joined #cdk
11:40 jonalv joined #cdk
11:41 maclean joined #cdk
13:37 maclean egonw : the jcp-primary stuff on pele - is the source/jar download in sync with the version in svn?
13:38 maclean I ask because this guy on the cdk-user list seems to have a version that doesn't have a Renderer method that is in svn, but not in the javadoc
13:39 maclean so I suspect that he downloaded the cdk-jar from that pele page, and it has code from an earlier version...
13:39 egonw Nightly does a svn/git update
13:39 egonw I'll check if the crontab is working properly...
13:39 egonw mom
13:40 maclean ok
13:42 egonw maclean: confirmed...
13:42 egonw set up is broken
13:42 egonw last run was 16 March :(
13:42 maclean ah, right.
13:42 maclean D:
13:43 maclean oh well, at least we know now...
14:44 maclean joined #cdk
14:46 maclean egonw : about your comment on the adjacency matrix post - "have you tried the columns, too" - you're joking, right?
14:46 maclean I leave it as an exercise for the reader as to why this makes no difference...
14:47 maclean :)
14:47 maclean (hint : symmetry)
14:47 egonw sorry, no was not joking...
14:47 egonw will read again, and pay more attention...
14:48 egonw no, I payed attention...
14:48 egonw this is what I see:
14:48 egonw matrix one, on the columns, reads:
14:48 egonw 1 1 2
14:48 egonw matrix two, likewise, reads
14:48 egonw 2 1 1
14:48 maclean 2?
14:49 egonw yes, 1 + 1
14:49 egonw second matrix, first column
14:49 egonw that makes them different in order...
14:49 egonw if uniqueness is defined as:
14:49 egonw increasing order on rows *and* colums, that makes 1 unique, 2 not
14:49 egonw but
14:50 egonw I think you do now show all possible matrices, right?
14:50 egonw now -> not
14:50 maclean all possible matrices with a partial order over the integer representations of their rows as binary numbers
14:51 egonw ok, I think what I do wrong
14:51 maclean so "001" == 1, "110" == 6.
14:51 egonw stupid me
14:51 egonw OK
14:51 maclean but, actually, your scheme might work. hmmm
14:52 maclean if you sum the columns, instead of interpreting them as binary numbers (like the rows), you do get different orders, as you say.
14:52 maclean I'll try it! :)
14:53 maclean the attraction of this method is that (in theory) tests on numbers can be done using bitwise operations on integers, which should be super-fast...
14:55 maclean interesting. the row sums look a lot like integer partitions...
14:56 maclean (except in reverse-standard order :)
14:59 maclean this looks promising, actually!
15:02 rojasm joined #cdk
15:08 rojasm1 joined #cdk
15:42 edrin joined #cdk
16:27 CIA-42 cdk: shk3 * r14594 /cdk/branches/jchempaint-primar​y/src/main/org/openscience/cdk/ (4 files in 3 dirs): Added a controller module for atom-atom mapping and proper scaling of the atom-atom mapping line.
16:51 CIA-42 cdk: shk3 * r14595 /jchempaint/trunk/src/main/​org/openscience/jchempaint/ (10 files in 4 dirs): added gui for atom-atom-mapping; if add to existing reaction is done with just one reactions, dialog no longer pops up

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