Camelia, the Perl 6 bug

IRC log for #bioperl, 2011-03-28

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

All times shown according to UTC.

Time Nick Message
00:35 mmlevitt left #bioperl
00:45 cassj joined #bioperl
00:57 cassj left #bioperl
00:59 drin left #bioperl
01:01 sheenams joined #bioperl
01:20 mmlevitt joined #bioperl
01:31 mmlevitt left #bioperl
01:40 sheenams left #bioperl
04:16 flatwhatson_ left #bioperl
04:41 mmlevitt joined #bioperl
04:52 zenman_ left #bioperl
05:20 Drin joined #bioperl
06:40 mmlevitt left #bioperl
06:54 mmlevitt joined #bioperl
06:57 mmlevitt left #bioperl
07:02 mmlevitt joined #bioperl
07:36 mmlevitt left #bioperl
07:41 mmlevitt joined #bioperl
07:43 mmlevitt does anyone know the secret to compiling with C the kent libraries that Bio::DB::BigFile and Bio::DB::BigWig depend on?
08:08 dbolser mmlevitt: no, I emailed the bioperl mailing list, and so far no reply
08:09 dbolser I also replied to a similar issue on the gmod mailing list, and again no reply :-/
08:09 dbolser if we don't get a reply in 24 hours, I'll re-issue the mails
08:09 dbolser else we could try asking on #perl
08:09 dbolser or hopefully someone in here will reply to you
08:10 mmlevitt dbolser: thanks!
08:10 dbolser mmlevitt: np... I wish I could solve it myself!
08:11 mmlevitt dbolser: I was trying to install GBrowse, and that app apparently depends on these BioPerl modules.
08:11 dbolser mmlevitt: are you sure it depends? not suggests?
08:12 mmlevitt oh, my mistake, suggests.
08:12 dbolser you should be able to get it going without, and they will 'just work' if you add them after
08:12 mmlevitt great!
08:14 mmlevitt dbolser: are there any other bug fixes for GMod BioPerl scripts I could try to sink my teeth into?  Although the experience did familiarize me with general BioPerl, too.
08:14 mmlevitt installing different modules, and such.
08:18 mmlevitt what's really nice is that I've finally got a working cgi-bin directory to serve cgi and Perl pages.
08:27 dbolser mmlevitt: yes, there are many possible bioperl bug fixes that you could look at
08:27 dbolser http://www.bioperl.org/wiki/Bugs
08:28 dbolser http://redmine.open-bio.org/issues/3185
08:28 dbolser (for random example)
08:29 dbolser step 1 would be to write some tests for the bug, and possibly additional tests for the module in general.
08:30 dbolser then work on a bug fix, using the tests to confirm that you didn't break some other part of the code.
08:30 dbolser finally, submit bug fix and tests as suggested on that page
08:30 mmlevitt tests to test method and property functionality of module?
08:32 mmlevitt therefore, I submit bug-fixes to Bio::Tools::SeqStats to the maintainer of Bio::Tools::SeqStats, according to BioPerl.org?
08:33 mmlevitt dbolser:  Ok, I'll try to work on BugFixes on BioPerl, then move to GMOD bugs.  Thank you!
08:37 dbolser mmlevitt: I'd submit to the maintainer, to the bp mailing list, to the bug report and via git
08:38 dbolser yup, if you look under the t directory of bioperl you'll see plenty examples of tests
08:39 dbolser e.g. t/SeqTools/SeqStats.t
08:39 mmlevitt ah!
08:41 dbolser perl SeqTools/SeqStats.t # runs the tests
08:41 dbolser so the first step as I said is to read the bug, and add a test to confirm the bug
08:41 mmlevitt ok.
08:42 dbolser if you cant confirm the bug that way, you can reply that you couldn't recreate teh problem on your system, submit the test as evidence, and move on to the next bug :D
08:44 mmlevitt perfect.  will begin, tomorrow.  thank you, again!
08:44 dbolser np
08:44 dbolser night
08:45 mmlevitt good nite.
12:13 takadonet morning all
12:56 dbolser morning
12:56 andrei_ joined #bioperl
13:07 andrei_ hello
13:07 andrei_ I have this text http://pastebin.com/raw.php?i=81ZtyTnj
13:07 andrei_ how would be the best way to parse it?
13:09 andrei_ if I have the values of aspA    glnA    gltA    glyA    pgm     tkt     uncA and want to find out which St and clonal complex the values that i have correspond to
13:11 dbolser andrei_: sounds like you want a hash
13:11 andrei_ yes I was thinking of that
13:11 andrei_ but should i use the Text::Csv module^
13:11 andrei_ ?
13:11 dbolser I'd just read it in and use split
13:11 andrei_ or simply split?
13:12 andrei_ ok yeah that would be easier
13:29 micrypt joined #bioperl
13:29 micrypt I seem to be getting a webpage back from RemoteBlast.
13:30 micrypt What configuration setting might i be getting wrong?
13:40 dbolser -m
13:44 micrypt dbolser: "-m"?
13:45 dbolser -m 8 = html no?
13:45 dbolser sorry -m7
13:46 cassj joined #bioperl
13:50 micrypt dbolser: Sorry, I'm completely new to this. Do I pass that as part of the parameters map in the Perl script?
13:51 dbolser micrypt: hmm... I haven't run blast via bioperl actually, but that is one of the blast options
13:51 dbolser can you point me at the docs you're using to run remote blast via bioperl
13:52 micrypt dbolser: I'm trying to run the example that comes with the perl-docs: http://doc.bioperl.org/releases/biope​rl-1.4/Bio/Tools/Run/RemoteBlast.html
13:53 dbolser what goes wrong?
13:55 micrypt What I get back is a dump of the http://www.ncbi.nlm.nih.gov/blast/Blast.cgi page.
13:55 dbolser ouch
13:58 dbolser can you find another blast cgi  to run aginst?
13:59 micrypt It's specified by the remoteblast library.
13:59 genehack joined #bioperl
13:59 micrypt I haven't found any docs yet for switching it out.
14:00 dbolser I'm thinking the url may have changed
14:00 micrypt dbolser: Probably, I can't seem to find any mention of that on the web though. :(
14:58 martinsullivan joined #bioperl
14:59 martinsullivan left #bioperl
15:10 genehack left #bioperl
15:12 genehack joined #bioperl
15:21 dbolser micrypt: try the mailing list?
15:21 micrypt I think I might have worked it out.
15:22 dbolser orly?
15:22 micrypt Parameter issue. "The Strange Case of the nOOb"
15:23 dbolser can you fix the doc ... one the equivelent doc on the wiki?
15:24 Drin dbolser: heya
15:24 dbolser hihi
15:25 dbolser 4hours?
15:25 dbolser (just curious)
15:25 Drin 4 hours?
15:26 micrypt dbolser: I'll have a look once I confirm my issues were not strictly due to my ignorance. :)
15:31 dbolser Drin: napping
15:31 dbolser micrypt: ic
15:31 Drin dbolser: ugh. I just ruined my entire sleep schedule
15:31 Drin I slept for 7 hours, to find that my alarm clock was turned off and moved across my room to my bed
15:31 Drin >__>
15:32 Drin anyways, I have gotten kent source compiling, with -fPIC but I can't get BigWig to stop giving me SSL errors
15:36 andrei_ is there a way to use as query in a local blast search only one sequence from a multifasta file?
15:39 dbolser andrei_: from teh cli yup
15:39 dbolser Drin: I'm stuck there too
15:40 dbolser I sent an email to the mailing list, but so far no reply
15:40 dbolser Drin: look in inc/common.mk
15:40 dbolser USE_SSL=1 ??
15:40 Drin dbolser: yea, I tried changing that to 0 and then kent gives me compile error lol
15:41 dbolser I set 'export USE_SSL=1' and it compiled ok
15:41 dbolser I have a feeling it's something tricky like that
15:41 dbolser noone in #perl could help either
15:42 dbolser best guess... kentlib and perl interface are out of sync
15:43 Drin that... would be unfrtunate
15:43 Drin do you know what version of perl your server runs?
15:43 dbolser 5.8.8
15:44 dbolser But the main thing is the version of BiO::DB::BigWig and jkentweb.a or whatever it's called
15:44 Drin which one are you running? That's clearly a big diff from my 5.10
15:44 dbolser 5.8.8 too
15:44 dbolser I'm seriously thinking there should be a better way to do this using perl
15:44 Drin mmm.. well, if your server has it installed then it would seem like if we match your server's versions it should work
15:44 dbolser the whole idea of wig is to efficiently store and 'query' a histogram.
15:45 dbolser was on the point of going into #perl and asking for hints on how to do that
15:45 dbolser Drin: server is x86 in addition
15:46 dbolser do you get a wig2bigwig ahywhere?
15:46 dbolser wigToBigWig.pl
15:48 Drin uhh, well I saw that in the kent source
15:48 Drin but it was a .c I think
15:48 dbolser yeah...
15:48 Drin yea it's in C
15:48 dbolser just wondering how I can find the version of kent that I linked against on the webserver
15:51 Drin mmm
15:54 dbolser 1.03 (bigfile)
15:55 Drin mmm, i have 1.05 bigfile
15:55 Drin wonder if 1.03 is easier to work with
15:56 dbolser anything would be better
15:56 Drin lol truth
15:58 dbolser well... I found kent on the server... nnot sure how to probe version
15:58 rdesfo joined #bioperl
15:58 rdesfo hello
15:58 Drin rdesfo: hiya
15:58 dbolser README says 'This file last updated: $Date: 2010/06/03 16:48:53'
15:59 dbolser hello
16:00 rdesfo has an administrator been selected yet to submit GSOC proposals to?
16:00 Drin 1.03 gives me 'undefined symbol: ERR_print_errors_fp
16:00 Drin '
16:01 Drin >__>
16:01 dbolser I'm going to try building against the kent on the webserver
16:01 Drin sounds good
16:02 dbolser rdesfo: Its Robin Haw or rbuels I believe
16:02 rdesfo dbolser: should I just email them the proposal?
16:02 rbuels dbolser, rdesfo: well, you submit them to the org.
16:02 dbolser Drin: I'm just scared of breaking it!
16:03 dbolser rdesfo: rbuels knows what to do
16:03 rbuels rdesfo: no, you use the submission syste
16:03 rbuels m
16:03 rbuels rdesfo: read the google summer of code guidelines on the main gsoc site
16:04 dbolser if I cp -ra then build... not that'll screw it up
16:05 Drin dbolser: lol how would you break it? you're just downloading some version of BigWig and then building with it pointing to kent src right? and if you cp the kent src then you definitely can't break it
16:05 Drin just don't try to install it xD
16:05 dbolser Drin: the kent source on the webserver is compiled for x86
16:05 dbolser I need to recompile for i686
16:06 dbolser so I'm just making a copy to do that
16:06 Drin dbolser: oh that makes sense
16:06 dbolser should I rsync and cp at the same time... seems it should be more efficient...
16:06 andrei_ can you load, from a seqio object of a multifasta file, a specific sequence and its header to another object?
16:07 Drin i don't know what rsync is, so that's your call
16:07 dbolser andrei_: yup, but to pull out one seq, just use fastacmd or one of the options of blast
16:07 dbolser Drin: rsync is like cp on crack
16:07 andrei_ oh ok thats what i was thinking
16:07 Drin dbolser: anything on crack is always more efficient and better
16:07 dbolser thats true
16:08 dbolser but rsync is pretty magic
16:08 dbolser it allows you to sync two large directories one remote from the other, without comparing the two!
16:08 dbolser (directly)
16:10 Drin nice
16:16 cassj left #bioperl
16:19 pyrimidine joined #bioperl
16:19 dbolser Drin: limited progress...
16:20 dbolser ./Build now builds without warnings, but I see the same errors
16:20 Drin mmm
16:20 Drin is this with BigWig 1.03?
16:20 dbolser undefined symbol: ERR_print_errors_fp
16:20 andrei_ left #bioperl
16:21 dbolser no this is 1.05
16:21 * dbolser tries
16:22 martinsullivan joined #bioperl
16:22 dbolser 1.01, warnings are back and tests error as before
16:22 Drin mmm
16:22 Drin i don't know then..
16:22 dbolser 1.03 ...
16:23 dbolser nope
16:24 dbolser well... still getting no errors in 1.05
16:32 Drin wait. you mean you are gettting errors in 1.05 right?
16:32 Drin you said undefined symbol: ERR_print_errors_fp
16:32 dbolser oh sorry, errors yes, warnings no
16:33 pyrimidine dbolser: what OS?
16:33 * pyrimidine is too lazy to backlog :)
16:34 Drin dbolser: also, i found the wigToBigWig.pl file
16:34 Drin dbolser: it's in the BigWig tarball
16:34 * rbuels welcomes pyrimidine to the lazy club
16:35 * Drin applauds pyrimidine and rbuels xD
16:35 Drin tbh i'm not sure what os dbolser is on
16:35 * pyrimidine is too lazy to wave to rbuels
16:35 dbolser pyrimidine: the backlog goes back days ;-)
16:35 dbolser pyrimidine: centos, i686
16:35 * rbuels listlessly grunts at pyrimidine
16:35 * rbuels chuckles
16:36 rbuels pyrimidine: https://github.com/sheenams/Bio-AGP
16:36 * pyrimidine remains comatose, droooool.....
16:36 rbuels pyrimidine: look at that, total n00b to actual dist in 1 week flat
16:36 * dbolser is adding random linker libs into Build.PL
16:36 rbuels well, probably slightly more
16:36 pyrimidine now it just needs a Build.PL/Makefile.PL
16:36 rbuels yep
16:36 rbuels pyrimidine: the most impressive thing, actually, is the commit history
16:37 rbuels pyrimidine: indicates she understands version control!
16:37 * rbuels is floored
16:37 * rbuels wishes sheenams were here to hear us praising her
16:37 rbuels i'll point her at the log link i guess
16:37 pyrimidine sometimes it's just a matter of getting a concept and not worrying so much about the specifics
16:37 Drin we can just save the backlog for her
16:39 pyrimidine I sometimes think using <pick_your_old_crufty_vcs> taints the git user experience
16:40 rbuels she's really good at learning the specifics on her own, which is *key*
16:40 pyrimidine right
16:41 pyrimidine dbolser: compiling both kent src and Bio-BigFile with -fPIC?
16:42 * pyrimidine recalls that doing so with the kent src tree took some work...
16:42 dbolser pyrimidine: compiling kent src with -fPIC
16:42 dbolser pyrimidine: its easy if you know how
16:42 dbolser I'm just trying to build Bio-BigFile with specific linker flags, but not sure how
16:43 dbolser line 19 of Build.PL has "extra_linker_flags => ["$jk_lib/$LibFile",'-lz']
16:43 pyrimidine huh
16:43 dbolser and I added "extra_linker_flags => ["$jk_lib/$LibFile",'-lz','-lssl','-lcryptox'],"
16:43 dbolser (deliberate mis-spelling to test)
16:43 dbolser not seeing -lssl show up
16:44 pyrimidine could be a Module::Build thing
16:44 pyrimidine not sure
16:44 dbolser yeah...
16:44 * pyrimidine has to go to a meeting
16:45 * dbolser needs ./Build clean
16:47 pyrimidine dbolser: maybe './Build realclean'
16:48 dbolser I'm seeing warnings again
16:48 dbolser GAH!
16:49 dbolser Result: PASS
16:49 dbolser !!!
16:49 * dbolser tries against latest kent libs
16:50 pyrimidine :)
16:50 micrypt left #bioperl
16:50 Drin whaaaaa
16:51 pyrimidine .'/Build clean' doesn't remove everything
16:51 dbolser pyrimidine: it wasn't that causing problems
16:51 pyrimidine what was?
16:52 dbolser how are you supposed to learn?
16:52 dbolser mua ha ah ah!!!
16:52 dbolser its a libs issue
16:52 dbolser one moment
16:52 pyrimidine 32 vs 64?
16:52 dbolser nothing so prosaic
16:52 dbolser extra_linker_flags => ["$jk_lib/$LibFile",'-lz','-lssl','-lcrypto'],
16:52 dbolser added -lssl and -lcrypto
16:53 dbolser ./Build still gives a set of nasty warnings... me pasts
16:53 dbolser https://gist.github.com/890827
16:54 Drin you are money dbolser
16:54 dbolser yes, pay me
16:54 Drin i passed everything
16:54 Drin no warnings either
16:54 dbolser Drin: so hit ./Build realclean in your BigFile dir
16:54 dbolser orly?
16:54 Drin dbolser: got it
16:54 dbolser which version ?
16:54 Drin the -lssl is all that was needed
16:54 Drin 1.05 with newest kent
16:55 dbolser which arch?
16:55 Drin i686
16:55 Drin ubuntu
16:55 dbolser because I was seeing no warnings initially
16:55 dbolser me too
16:55 dbolser weird
16:55 dbolser now ./Build install :D
16:56 martinsullivan left #bioperl
16:57 dbolser Drin: did you -fPIC?
16:58 Drin dbolser: I did on kent source, but not on BigWig
16:58 Drin dbolser: i'm going to try running that example code on the cpan BigWig page
16:58 dbolser just wondering why a) At first installing from source instead of cpan I saw no errors, and b) why you see no errors
16:58 dbolser Drin: cool
16:59 Drin i'm installing from source and not cpan
16:59 Drin so i'm assuming it must be related.. as far as why there are no errors/warnings
16:59 * Drin shrug
16:59 dbolser hehehe
17:13 dbolser Woop!
17:13 dbolser perl -e 'use Bio::DB::BigWig;my $wig  = Bio::DB::BigWig->new(-bigwig=>"intensity_rev.bw"); my @points = $wig->features(-seq_id=>'PGSC0003DMB00​0000507',-start=>314000,-end=>350999); print join("\t", $_->start, $_->end, $_->score), "\n" for @points'
17:13 dbolser Drin: we have access!
17:14 Drin dbolser: woop!
17:15 dbolser heheh... now for debugging ... D:
17:15 * dbolser goes home ;-)
17:17 Drin dbolser: this will be interesting.. I have to get actual test files, or build them.. well, at least BigWig is installed! now I can try to do some actual playing around
17:18 Drin dbolser: good stuff, enjoy home xD
17:19 rdesfo left #bioperl
18:21 ank joined #bioperl
18:46 ank left #bioperl
18:47 ank joined #bioperl
18:47 martinsullivan joined #bioperl
19:05 philsf joined #bioperl
19:30 takadonet pyrimidine: ping
19:38 philsf left #bioperl
19:54 rdesfo joined #bioperl
19:59 zenman_ joined #bioperl
20:06 mmlevitt drin: how did you install the BigFile/BigWig modules?
20:10 pyrimidine mmlevitt: (speaking as dbolser): very carefully
20:10 mmlevitt ah
20:12 pyrimidine mmlevitt: the trick is to make sure to use -fPIC when it's needed, and to use './Build realclean' after changing build options related to XS
20:12 pyrimidine I thnk
20:13 mmlevitt ok. brb.
20:14 pyrimidine not sure why '-lssl' is needed for the linker, maybe dbolser can indicate this at some point
20:23 mmlevitt ?
20:25 mmlevitt what is -fPIC and what is -lssl?  why are these not in the orginal documentation?
20:25 pyrimidine mmlevitt: that is regarding the patch dbolser sent along
20:27 mmlevitt patch for installation?
20:28 pyrimidine mmlevitt: I think so
20:28 pyrimidine -fPIC only is needed for some systems
20:29 pyrimidine one other thing to point out, which isn't obvious (but is mentioned in the README) is setting MACHTYPE
20:29 pyrimidine http://cpansearch.perl.org/sr​c/LDS/Bio-BigFile-1.05/README
20:30 mmlevitt I did that, and after trying to install BigBed, I got the c lib files needed in the kent tree, but all this stuff should be documented.
20:30 pyrimidine huh, funny, -fPIC isn't mentioned in that README, but is mentioned in the samtools one
20:31 pyrimidine mmlevitt: it is documented in the README, it's not in kent's src tree though (which makes it a bit trickier)
20:31 pyrimidine (MACHTYPE, that is)
20:32 mmlevitt I need to take a break.
20:32 pyrimidine compiling perl XS code will make you do that
20:41 mmlevitt and writing pure Perl implementations is the challenge.
20:42 mmlevitt but, not being able to get the C-libraries working with BigWig prevents the progress.
21:08 zenman_ left #bioperl
21:14 martinsullivan left #bioperl
21:41 mmlevitt left #bioperl
21:46 takadonet left #bioperl
21:59 dbolser mmlevitt ...
22:00 dbolser pyrimidine: to clarify, in case you're interested, you need to patch the kent tree by adding '-fPIC' to the CLIBS= section of 'inc/common.mk', that fixes it everywhere else, so you don't need to worry
22:00 dbolser you then need to set -lssl in the libs section of Build.PL
22:01 dbolser :-)
22:01 dbolser pyrimidine: I'm glad you saw my mail to the list...
22:01 dbolser begining to wonder if mails from me are invisible ;-)
22:02 dbolser oh! replies
22:08 dbolser cheers pyrimidine
22:46 deafferret :)
22:46 sheenams joined #bioperl
23:32 rdesfo left #bioperl
23:39 matthew joined #bioperl
23:40 matthew is now known as Guest88749
23:44 ank left #bioperl
23:45 ank joined #bioperl
23:56 mmlevitt joined #bioperl

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