Camelia, the Perl 6 bug

IRC log for #bioperl, 2009-08-18

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

All times shown according to UTC.

Time Nick Message
00:00 rbuels oh that is in CO
00:00 deafferret so that climb was 9,400 -> 11,600 ... so I survived 2,200 vert... hmm
00:01 rbuels long's peak is listed as 14,259
00:01 deafferret Mills Glacier my ass
00:01 deafferret ya, but I only got to Chasm Lake
00:01 rbuels and then you puked up your guts?
00:01 rbuels lol
00:01 deafferret the summit is a whole other universe of pain
00:02 deafferret no, just you have to start at 4am and haul ass to go from trailhead to summit
00:02 rbuels you should try kilimanjaro, it's not that hard, just really high
00:02 deafferret I think I'd heart attack, collapse on that attempt
00:02 rbuels that's a good test for altitude sickness
00:03 deafferret Heading to rocky mountain national park on Fri. woot!
00:03 rbuels i have to get going......
00:03 deafferret toodles!
00:03 rbuels cool!
00:03 deafferret ya.  gonna lunch w/ dhoss
00:03 rbuels heh that will be fun
00:03 rbuels ok, ttyl
00:04 deafferret thanks for your help
00:04 rbuels no
00:04 rbuels prob
00:04 rbuels &
00:43 nErVe joined #bioperl
08:07 faceface joined #bioperl
08:07 faceface joined #bioperl
11:44 kyanardag_ joined #bioperl
12:13 deafferret rbuels: all my code are borrowed from everywhere.
12:16 deafferret It appears to me that Bio-Tools-HMM/trunk/t/001_basics.t demonstrates more knowledge of how HMMs work than my entire univ possesses.
12:17 deafferret laugh "oops, forgot to pick up the Bio::Tools::HMM deletion.  svn is so cumbersome."
14:26 kyanardag_ joined #bioperl
14:28 kyanardag joined #bioperl
14:30 kyanardag_ joined #bioperl
16:46 nErVe joined #bioperl
18:23 ptl_ joined #bioperl
20:06 nErVe joined #bioperl
21:11 vein joined #bioperl
21:15 vein Is there a particular reason that the following code seems to draw in the whole sequence? I had assumed it wouldn't since the seqobject as stored in the variable $seq would be refreshed each iteration...
21:15 vein foreach my $file (@files) {
21:15 vein my $in = Bio::SeqIO->new( -file => "$file", -format => 'Fasta');
21:15 vein while (my $seq = $in->next_seq()) {
21:15 vein my $siteset = $pwm->search_seq(-seqobj=>$seq, -threshold=>"80%");
21:15 vein print OUT $siteset->GFF();
21:16 vein }
21:17 rbuels vein: each next_seq brings in an entire sequence
21:17 vein duh, so the original sequence needs to be broken up if it itself is large
21:18 * rbuels looks for LargeSeq support
21:18 rbuels there's an object for that
21:18 rbuels ...
21:18 vein ok
21:18 deafferret there is? neato
21:18 * deafferret waits patiently on his pile of pillows for rbuels' return
21:18 rbuels lol
21:18 vein I had this figured out awhile back, and naturally misplaced the original script, my apologies
21:19 vein thanks
21:20 * rbuels is looking
21:20 rbuels looks like you can write a largeseq......
21:20 rbuels ah.
21:20 rbuels you want to use -format => 'largefasta'
21:21 rbuels if you have bigbig sequences
21:21 rbuels vein: largefasta uses tempfiles to store sequence info instead of memory
21:21 rbuels vein: do perldoc Bio::SeqIO::largefasta
21:24 deafferret Bio::SeqIO::largefasta POD is wrong? It says next_seq returns a Bio::Seq object
21:24 rbuels Bio::Seq::LargeSeq isa Bio::Seq
21:24 rbuels but.
21:24 rbuels probably could be clearer
21:24 * rbuels points at deafferret's commit bit
21:25 deafferret by that argument it should say it returns a Bio::Root::Root object  :p
21:25 deafferret or, better yet, a Perl object
21:25 * rbuels paddles deafferret
21:25 * deafferret lools
21:26 * deafferret grumbles about people making him fix problems he finds
21:26 deafferret stupid open source
21:26 rbuels Bio::Seq::LargePrimarySeq is actually what it returns.
21:26 rbuels right now
21:27 rbuels wow, guillermo del toro really loves the spanish revolution
21:27 deafferret see?!?! INLINE POD! Everyone but jstaljich and I are wrong
21:27 rbuels er
21:27 rbuels civil war
21:27 * rbuels de-inlines the pod
21:28 * rbuels actually does not care that much
21:28 * deafferret waggles his kick stick at rbuels
21:28 rbuels hahaha
21:28 rbuels op me.
21:28 deafferret we've been over this. you tell me what to type, I'll copy/paste it
21:28 rbuels type this:  /kick deafferret oooh i got you to do it
21:29 was kicked by deafferret: oooh i got you to do it
21:29 rbuels lol
21:29 deafferret joined #bioperl
21:29 deafferret hey... wait a minute
21:29 rbuels good shit.
21:29 deafferret ChanServ: trust rbuels
21:29 deafferret ChanServ: you sucketh
21:29 deafferret ChanServ: punish rbuels
21:30 deafferret rbuels: I have a largegenbank. Now what do I do?
21:31 deafferret I only have 32MB of RAM
21:31 rbuels my $s = $genbank->to_string;
21:31 rbuels will do what you want
21:31 deafferret and 32MB of disk
21:31 deafferret and this single GenBank file is all sequences ever
21:31 vein hmm, i still get a GET_SEQUENCE: Sequence too long error
21:31 deafferret metagenomics
21:32 rbuels vein: that sounds like an error coming from bioperl
21:32 rbuels could you put the full output in http://gist.github.com/
21:32 rbuels (and paste the gist url here)
21:33 rbuels deafferret: how am i supposed to watch El Espinazo del Diablo with you gabbing
21:33 vein just the error output, correct? sorry, never had to mess with git before.
21:34 rbuels it's not git, it's just a pasting service provided by github
21:34 rbuels same as pastebin.com or whatever
21:34 vein ahh
21:34 vein ok
21:36 rbuels hmmmmmmmm
21:36 vein git://gist.github.com/169993.git
21:36 rbuels GET_SEQUENCE is not from bioperl
21:37 vein aha
21:37 vein ok
21:37 deafferret ...huh? I thought someone was saying that everything is in lib/ now? Am I in the wrong commit repo?  svn+ssh://jhannah@dev.open-bio.org/home/svn-reposit​ories/bioperl/bioperl-live/trunk
21:38 rbuels what is pwm?
21:38 rbuels $pwm i mean
21:38 rbuels deafferret: that was just in bioperl-run i think
21:38 deafferret oh. very well. you may go.
21:38 rbuels deafferret: but when i get a little more street cred i'll move Bio/ into lib and laugh as the flak bounces off my armor plate
21:39 deafferret lol
21:39 deafferret I got your back, cuz
21:39 rbuels fo shizzle.
21:39 vein That refers to a position weight matrix-I'm using the TFBS module collection to look for matches to motifs in the sequences I pull in via bioperl modules
21:39 rbuels ooh.
21:39 rbuels if you take out the my $siteset = $pwm->search_seq(-seqobj=>$seq, -threshold=>"80%"); and print OUT $siteset->GFF(); lines
21:39 rbuels does it work?
21:40 rbuels (if so, that means you error is emanating from TFBS, whatever that is)
21:40 rbuels s/you/your/
21:40 * rbuels googles for TFBS
21:40 rbuels http://tfbs.genereg.net/
21:40 rbuels well look at that.
21:40 rbuels it's as if someone published a bunch of perl modules
21:41 rbuels and they had never heard of the existing infrastructure for publishing perl modules
21:41 rbuels nor did they bother to learn it
21:41 rbuels vein: not talking about you
21:42 vein they kept them separate perhaps because they were written up in a journal?
21:42 rbuels does not matter.
21:42 rbuels journal is orthogonal to this
21:44 deafferret Committed revision 15988.   Fear me.
21:46 vein hmm, I can't use GFF via the LargePrimarySeq package
21:47 vein I might have broken up the files manually last time. I need to do less spring cleaning of files :). I threw away too much.
21:47 rbuels vein: gff isn't a seq)uence format, it's for annotations (primarily
21:47 vein right
21:47 vein ok
21:47 vein never mind, I understand now
21:49 vein I understand it's for annotations, but it's a useful output format when containing other info from the TFBS scripts
21:49 vein er, modules
21:55 deafferret heh. at least PDL is in CPAN
21:56 deafferret oo! dev release 3 days ago. I approve.
21:57 vein well, the code still causes things to file up in memory/swap even if only printing the sequence to an output file.
21:58 vein no gff, just a print out $seq->seq() call
21:58 deafferret here's more protein factor binding site software packages, if you need 'em. A guy just got his PhD and left here doing this stuff: http://biobase.ist.unomaha.edu/med​iawiki/index.php/Special:Allpages
21:59 rbuels deafferret: put your name on them and put them on cpan
21:59 rbuels lol
21:59 rbuels and then write some idiotic paper about them
22:00 pyrimidine joined #bioperl
22:00 * rbuels waves at pyrimidine
22:00 pyrimidine hello
22:00 deafferret rbuels: http://biobase.ist.unomaha.edu/me​diawiki/images/1/17/MyPoster.pdf
22:00 * deafferret waves
22:00 deafferret & foos. back in 10
22:00 pyrimidine just read the backlog, TFBS is (in my mind at least) obsolete
22:01 rbuels how did you read the backlog? you weren't here?
22:01 pyrimidine http://irclog.perlgeek.de/bioperl/today
22:02 * rbuels needs to pay attention to vein's problem
22:02 rbuels vein: are you using the largefasta output format also?
22:05 * rbuels feels spied upon by irclog.perlgeek.de
22:05 rbuels my most secret feelings.....
22:05 pyrimidine big brother is watching...
22:06 rbuels gaaaaah! *it heard that*
22:07 rbuels vein: did your script eat all your memory and buffer-overflow you into oblivion?
22:07 vein almost
22:07 rbuels vein: meaning: did my suggestion work?
22:08 vein I'm trying again with a different write method, but hopefully that won't happen again
22:08 pyrimidine re: LargePrimarySeq, not sure how integrated that is within bioperl.  The problem is most packages call seq(), that by definition pulls in the entire sequence string
22:09 * pyrimidine thinking we need a sequence iterator
22:09 rbuels need so many things.
22:09 pyrimidine ...and so little time  ;)
22:10 vein i seems to be taking the time to write out the temp file, but then it pulls things in too fast or in too large of chunks.
22:10 pyrimidine how much memory on your system?
22:10 rbuels vein:  post your whole script with gist?
22:12 pyrimidine gotta go.  have to beat the rain (biking home)
22:12 * pyrimidine commuting
22:17 vein git://gist.github.com/170020.git
22:17 deafferret 'orthogonal' is tangential to concise communication
22:18 * rbuels paddles deafferret
22:18 rbuels hmm
22:19 deafferret & birthday dinner! yay me!
22:19 deafferret yay! I'm old!  :)
22:19 rbuels deafferret: happy birthday then!
22:19 * rbuels did not know
22:19 * deafferret blushes
22:19 vein happy birthday mate
22:19 rbuels vein:  you're still getting errors with this?
22:19 deafferret you guys got me IRC /say's! you shouldn't have!
22:20 deafferret /say, right?
22:20 deafferret yup
22:20 rbuels ?
22:20 rbuels what?
22:20 deafferret /say is the long way to start typing
22:21 rbuels vein: what actual errors are you getting with this script
22:21 rbuels vein: what are the error messages?
22:21 deafferret &
22:21 rbuels deafferret: have fun
22:21 vein I'm not getting anything now, but i'm not getting any output either in the outputfile...
22:22 vein When I tried to print out the sequences, they would print, but I got massive buffer overflow
22:22 vein now, nothing seems to print out
22:22 rbuels if you call ->seq(), the whole seq goes into memory
22:23 vein which may just mean the compiler will have it print in mass at the end rather than continuaously
22:23 vein right
22:23 rbuels vein:  you are printing to OUT
22:24 vein right, and I'm watching the size of the output file ;)
22:24 rbuels vein: btw, you should use lexical filehandles
22:24 rbuels open my $out, '>', 'output.gff'
22:24 rbuels hmmmmm
22:26 rbuels vein: you might need to close the outfile to get it all to print
22:26 rbuels close OUT;
22:27 rbuels also, the /n in the print is not a newline.
22:27 rbuels that's \n
22:27 vein oops, typo
22:28 vein *sigh*
22:28 vein 36 hours of no sleep will catch up to you
22:28 rbuels go another 12 hours or so and you might start hallucinating
22:28 rbuels that's a good time.
22:29 * rbuels has hallucinated from no sleep
22:38 vein I'm going to have to revisit this later today, I've got an IP that needs to be finished...
22:39 vein thank you very, very much for all your help
22:40 rbuels vein: no problem
22:40 rbuels vein: good luck
22:40 rbuels vein: IP?
22:43 vein Immuno precipitation
22:43 rbuels oh.
22:44 rbuels vein: ok then carry on
22:44 * rbuels wishes vein luck with the colorless-liquids-squirting ;-)
22:48 vein lol
23:44 vein joined #bioperl

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