Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-08-23

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

All times shown according to UTC.

Time Nick Message
00:42 deafferret got 'em on the run now, the little blighters
02:32 dnewkirk joined #bioperl
04:35 CIA-95 bioperl-live: Chris Fields master * r21e409f / Bio/DB/EUtilities.pm : add some synopsis info, remove TODO item (GET/POST is now implemented) - http://bit.ly/93v9vS
04:35 CIA-95 bioperl-live: Chris Fields master * rfc50584 / Bio/Tools/EUtilities/Cookie.pm : remove extraneous Cookie class - http://bit.ly/duxJiL
04:35 CIA-95 bioperl-live: Chris Fields master * rfe4fffd / (4 files in 2 dirs): doc updates - http://bit.ly/bpYkGH
04:35 CIA-95 bioperl-live: Chris Fields master * rc0a9696 / (2 files in 2 dirs): more docs - http://bit.ly/a1awTI
04:43 CIA-95 bioperl-live: Chris Fields master * r56d6251 / Bio/Tools/EUtilities/Summary/DocSum.pm : more docs, for esummary - http://bit.ly/a1HZTp
04:43 CIA-95 bioperl-live: Chris Fields master * r7020019 / (9 files in 3 dirs): remove Id tags (strangely not removed from the last round - http://bit.ly/bOCqcP
05:32 jhamilton left #bioperl
05:44 jhamilton joined #bioperl
07:18 melic left #bioperl
07:20 melic joined #bioperl
07:20 kai_ is now known as kai
07:20 kai left #bioperl
07:20 kai joined #bioperl
07:29 kai morning folks
07:31 kai it seems like SLE11
07:32 kai it seems like SLES11 has a version of File::Path that's too old to have "make_path" and "remove_tree"
07:33 kai is there a way to make perl fall back to mkpath and rmtree if make_path and remove_tree is not available?
07:41 dbolser kai: yup, but you need to eval the code
07:41 dbolser I noticed those fails too
07:41 dbolser or you can do some VERSION tricks, but I forget how
07:42 kai my perl-fu isn't quite on that level yet
07:50 kai meh, let's do this differently
07:50 kai I'll patch it out for the packaging on SLES11
07:50 kai but it's kind of silly to go back to a legacy API just because some enterprise distribution ships a too old package
12:44 genehack joined #bioperl
14:55 Yksi joined #bioperl
15:04 Yksi left #bioperl
15:19 perl_splut joined #bioperl
15:34 dnewkirk left #bioperl
15:41 pyrimidine joined #bioperl
15:56 kblin hi pyrimidine
15:58 pyrimidine o/
16:01 kblin pyrimidine: you might need to add a new dependency check to bioperl
16:01 pyrimidine ?!?
16:01 kblin there's two tests that require File::Path of 2.06 or newer
16:02 kblin older distros (e.g. SLES11) ship 2.04, and it's part of the core perl package
16:02 pyrimidine which two?
16:02 kblin er, hang on
16:03 kblin t/SeqIO/mbsout.t and t/SeqIO/msout.t
16:03 kblin the use make_tree and remove_tree instead of mktree and rmtree (or whatever they were called)
16:04 kblin make_path/mkpath, sorry
16:04 pyrimidine ok, those are new.
16:04 kblin it might be possible to use some perl magic to detect if the new names are around and fall back to the old names if not, but that's beyond my perl-fu
16:05 pyrimidine no need if the old names just work
16:05 kblin yeah, my patch is search/replace
16:05 kblin I can push it if you like
16:05 pyrimidine go ahead
16:06 pyrimidine no point in adding a version-based dependency on top of everything else
16:07 pghpgc joined #bioperl
16:08 kblin http://github.com/kblin/bioperl-live/commi​t/b1f171e96a3d7e9d1c48f428f94a2032b03d8f9f
16:09 kblin just to have more than one pair of eyes looking at it
16:10 pghpgc hi...I have a question about piping STDIN using SeqIO
16:10 * pyrimidine running a checkout
16:10 pghpgc I'm getting an error ...please see http://pastebin.com/dqkDtH39
16:13 kblin that looks like some perl stdin handling weirdness
16:14 kblin you need somebody with a stronger perl-fu
16:14 pghpgc i'm working in bash on Cygwin
16:14 pyrimidine kblin: all passed.  do you have direct commit access?
16:14 kblin pyrimidine: yeah
16:14 pyrimidine go ahead and push that, then
16:14 kblin pyrimidine: want me to push that to master right away?
16:14 pyrimidine sure
16:14 kblin ok, will do
16:15 pyrimidine if it isn't, it will fall out with CPAN Testers with older versions
16:15 pyrimidine but I think it's fine
16:15 kblin it passes make test on my SLES11 box and my work box
16:16 pyrimidine pghpgc: set the format for SeqIO
16:16 pyrimidine this is a known issue, you can't reset the file pointer for STDIN
16:17 pghpgc but I'd like to write a general script that recognizes the input format automatically and use my script in a pipeline on the commandline
16:17 pyrimidine when the format guesser is sampling the data stream, it advances the file pointer
16:17 pghpgc oh
16:17 pyrimidine so it can't be seek'd back
16:17 pghpgc so no hack work around?
16:17 pyrimidine this is possibly fixable within GuessSeqFormat
16:18 pyrimidine just hasn't been top priority
16:19 pyrimidine you could possibly slurp the file into a string if they are short enough
16:20 pyrimidine then do: open(my $fh, '<', \$string); my $in = Bio::SeqIO->new(-fh => $fh);
16:20 pghpgc can you create a SeqIO object on a string?
16:20 pghpgc oh
16:20 pghpgc ah ...i'll try that
16:20 pyrimidine IO::String is another way
16:21 CIA-95 bioperl-live: Kai Blin master * raa23a89 / (t/SeqIO/mbsout.t t/SeqIO/msout.t): Fix File::Path calls for older File::Path versions - http://bit.ly/cCyntl
16:21 pyrimidine I think that's supposed to be seekable
16:22 kblin and it's already a BioPerl dep, so you're bound to have it on the system anyway :)
16:22 pyrimidine the nice thing is, we could possibly get rid of that dependency now
16:22 pyrimidine just using the idiom above
16:23 pyrimidine Root::IO directly supports this for input only, actually
16:23 pyrimidine (in master branch)
16:23 pyrimidine so, if using bioperl-live: my $in = Bio::SeqIO->new(-string => $string);
16:24 pyrimidine (though it would be better to use a scalar ref there)
16:24 kblin pyrimidine: oh, btw, one of the suse packagers already picked up a bioperl package for the devel:languages:perl repository. so far it's 1.6.1, but I'm happy to provide the update once you release 1.6.2
16:24 kblin makes my life much easier
16:25 pghpgc pyrimidine: u the man...it worked! ....thank you
16:25 pyrimidine :)
16:25 pyrimidine pghpgc: no problem
17:01 driley joined #bioperl
17:09 deafferret is this dying if none exist a bug or feature?   $feat->get_tag_values('db_xref')   Personally, I'd prefer it return an empty list.
17:12 deafferret next unless ($feat->has_tag('db_xref'));    meh
17:13 deafferret Bio::SeqFeature::Generic docs don't say either way
17:33 dnewkirk joined #bioperl
17:36 deafferret "Do you know of a way to get perhaps PROSITE's entire motif database, and then statistically analyze it for amino acids that are listed commonly in a motif specifically (for example, prolines in the SH3 binding motif), and also amino acids that are commonly grouped together (like S and T in ser/thr kinase phosphorylation motifs)."
17:37 deafferret hahaha... um, ya. i have about 20 seconds to dedicate to that. let me see if I can help
17:39 kblin deafferret: if you're done with that, could you also wash my car and take out the trash?
17:40 deafferret maybe he'll still do the cover page of his own thesis?
17:41 pyrimidine yeah, but he'll want you to type it :)
17:41 deafferret i know enough about motifs to try to avoid them, and know nothing about prosite
17:44 kblin it's kind of handy
17:44 pyrimidine is prosite used that much anymore, over Pfam/SMART/etc?
17:45 kblin pfam user myself, here
17:45 deafferret @clue == 0
17:45 pyrimidine Protein families
17:46 deafferret like (('chicken', 'duck'),('peanuts', 'soy'),('cow', 'pig'))  ?
17:46 deafferret :)
17:48 perl_splut heh, take every protein discovered and do a phylo tree...
17:48 perl_splut see how many supercomputers you can bring to their knees with that problem
17:48 kblin hehe
17:48 kblin and that for an utterly useless result
17:48 perl_splut not necessarily
17:49 perl_splut might help see how proteins have changed as species diverged
17:49 perl_splut really good for things like single-cell organisms that might absorb dna from their surroundings and not just from procreating
17:49 kyanardag joined #bioperl
17:50 kyanardag left #bioperl
17:50 deafferret also, excellent for employment continuance
17:50 kyanardag joined #bioperl
17:51 kblin perl_splut: not totally convinced
17:51 perl_splut true, unless you automate yourself totally out of the job, heheh
17:51 kblin perl_splut: I doubt there's a single "ancestor"
17:52 * kblin goes back to learning about perl OO
17:52 perl_splut if there was, it is long since dead or changed into a few million descendant species
17:53 kblin I mean I doubt things started with a single protein that evolved
17:53 perl_splut I understand that, kblin
17:55 kblin from the architecture side of things, I guess everything that you'd annotate to a sequence would be a SeqFeature, right?
17:57 deafferret http://www.bioperl.org/wik​i/Features_vs._Annotations   "In imprecise "user-centric" terms, a feature is metadata attached to a particular section or fragment of a sequence, while an annotation is metadata attached to the sequence object itself, and so describes something about the entire sequence. "
17:57 kblin ah, right, I mean feature then :)
17:58 deafferret um...   everything "you'd annotate to a sequence" would be a Feature or an Annotation, depending on whether or not that thing has a specific position
17:59 deafferret is my reading of the canon
17:59 kblin I've got a software pipeline that predicts secondary metabolite clusters on genomic data, and I'm trying to come up with a datatype for the modules
17:59 kblin so it's specific to a particular section of the sequence
18:00 deafferret ya, so those are SeqFeatures I think
18:00 kblin modules in this case being secondary metabolite synthesis modules
18:16 pyrimidine just to add a twist to that, SeqFeatures can have Bio::Annotation
18:16 pyrimidine just to confuse you
18:16 pyrimidine http://www.youtube.com/watch?v=B2Je1CEPkUM
18:17 ank joined #bioperl
18:37 perl_splut heh
20:31 pghpgc left #bioperl
20:44 kyanardag left #bioperl
20:47 kyanardag_ joined #bioperl
21:10 pyrimidine left #bioperl
22:06 ank left #bioperl
22:12 ank joined #bioperl
23:11 bag joined #bioperl
23:46 bag left #bioperl

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