Camelia, the Perl 6 bug

IRC log for #bioperl, 2011-03-25

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

All times shown according to UTC.

Time Nick Message
00:31 flatwhatson_ joined #bioperl
00:59 kyanardag joined #bioperl
01:07 rdesfo joined #bioperl
02:10 kyanardag left #bioperl
02:32 sheenams joined #bioperl
02:54 rdesfo left #bioperl
02:54 rbuels left #bioperl
02:57 rdesfo joined #bioperl
03:00 rbuels joined #bioperl
04:47 kyanardag joined #bioperl
05:05 rdesfo left #bioperl
05:30 zenman_ joined #bioperl
05:57 zenman_ left #bioperl
06:35 Drin joined #bioperl
06:47 sheenams left #bioperl
07:37 cassj joined #bioperl
08:20 cassj left #bioperl
08:52 Drin_ joined #bioperl
08:52 Drin left #bioperl
09:13 Drin_ left #bioperl
09:44 dbolser mmlevitt quipped: I learned more about BioPerl's assembly formats, and wonder why BSML
09:44 dbolser couldn't be used as the intermediate format for all assembly files.
09:44 dbolser (http://bsml.sourceforge.net/)
09:44 dbolser ?
10:43 dbolser can I write BAM/SAM using Bio::SAM?
10:43 dbolser seems it's read only :(
10:43 dbolser Bio::DB::Sam -- Read SAM/BAM database files
10:45 * dbolser wonders what the most efficient format to write and then convert to SAM would be...
10:45 dbolser long run of Newbler failed to output an ace :(
10:52 dbolser later in that doc...
10:52 dbolser Some operations, such as writing BAM files, are only available through the low-level API.
10:52 svaksha joined #bioperl
10:58 dbolser Bio::DB::Bam->open('my.bam'); # open: No such file or directory
10:59 * dbolser wonders if the information about writing is just wishfull thinking
11:00 dbolser Bio::Samtools, which is separate from bioperl but compatible with it,
11:00 dbolser provides read/write access to SAM and BAM via Heng's C library.
11:03 dbolser Interested parties might look at Bio::Asssembly::IO::sam to see how Lincoln's Bio::DB::Sam (which uses the libbam library directly via XS, also not BioPerl proper but we love it anyway) might be employed.
11:05 dbolser I can't find Bio::Samtools distinct from Bio::DB::Bam
11:06 dbolser http://search.cpan.org/perl​doc?Bio::Assembly::IO::sam # Not found
11:07 dbolser I see pod in the PM
11:08 dbolser "This is a (currently) read-only IO module designed to convert" Gaaa!
12:18 rdesfo1 joined #bioperl
12:45 zenman_ joined #bioperl
13:02 dbolser the sam format seems a bit ambigous
13:03 dbolser I don't have alignment details, so it says cigar should be '*'
13:03 dbolser also, I'm dealing with 'single fragment' alignments, so it says tlen should be 0
13:03 dbolser however, how does it know how long my read is?
13:04 dbolser the sam mailing list is almost as high traffic as bioperl
13:22 * dbolser sends a letter
14:04 rdesfo1 left #bioperl
14:59 zenman_ left #bioperl
15:05 andrei_ joined #bioperl
15:06 andrei_ hello
15:06 andrei_ I have a problem
15:06 andrei_ I have a script, that was running perfectly untill it stopped working :P
15:08 andrei_ this is the error https://gist.github.com/886981
15:09 andrei_ and this is the part of the code where it gives the error https://gist.github.com/886986
15:10 andrei_ I dont know why it does this, as it was working a few minutes ago
15:10 andrei_ also I tested the script on another mac and it run perfectly
15:10 andrei_ on linux also
15:10 andrei_ it's only on this computer
15:11 andrei_ a mac
15:16 rbuels andrei_: well, something has changed with it, definitely, if it used to work :-)
15:18 rbuels andrei_: the error message you gisted doesn't mean anything to me, except that it seems to be mostly Tk (i.e. the gui) modules involved
15:19 rbuels andrei_: if i were you, i would look at script.pl line 585 and see what it's trying to call ->accession_number on, that might give you more information about what is breaking
15:25 andrei_ I have updated the gist with the whole subroutine that does not work         https://gist.github.com/886986
15:26 andrei_ I did not modify anything in the script between the moment when it worked and the one when it stopped working
15:26 andrei_ that is what I find weird
15:30 andrei_ @rbuels: I just made a new script https://gist.github.com/887018 and it gave that error msg
15:30 andrei_ but the file exists and it is a good genbank file it's like it does not want to create the seqio_object
15:50 andrei_ or perhaps that the next_seq is the one that gives the error
15:51 andrei_ is there a way to see if next_seq worked?
16:01 andrei__ joined #bioperl
16:02 andrei_ left #bioperl
16:03 rbuels andrei__: you're writing to that genbank file?  ">./CGYHR025.gbk" means write.
16:04 andrei__ no i'm not writing
16:04 andrei__ I'm reading from it
16:04 rbuels andrei__: then you shouldn't have a > in there.
16:04 andrei__ where did you see it?
16:04 rbuels andrei__: in line 5 of your gist!
16:05 * rbuels boggles
16:06 andrei__ yes that was a mistake I made, but still the same error
16:06 andrei__ it does not create the seq object
16:07 rbuels andrei__: try with a different genbank file?
16:11 rbuels some genbank files have weirdnesses that the bioperl parser has trouble with
16:12 andrei__ yes that is true
16:12 perl_splut and doing the > before the filename might have blanked that particular genbank file
16:12 rbuels andrei__: also, if you were opening that file to write, bioperl may have deleted it.
16:12 rbuels yep
16:12 rbuels perl_splut: jinx
16:12 perl_splut you owe me a soda
16:12 rbuels yeah, blanked is a better word
16:12 andrei__ it worked on one genbak
16:12 andrei__ but on 2 others it did not work
16:13 rbuels andrei__: those files might be blank now, did you look in them?
16:13 andrei__ yeah that one got blanked and i was eandering why :))
16:13 andrei__ so then it is the genbank file that is not good
16:13 rbuels probably
16:19 perl_splut could also be complications from the method used to retrieve the GB file.
16:26 takadonet left #bioperl
16:26 takadonet joined #bioperl
16:27 takadonet left #bioperl
16:27 takadonet joined #bioperl
16:27 andrei__ left #bioperl
16:39 kyanardag left #bioperl
16:45 gawbul joined #bioperl
17:25 kyanardag joined #bioperl
17:29 philsf joined #bioperl
17:29 philsf left #bioperl
17:29 philsf joined #bioperl
17:35 dbolser hello bioperl friends
17:36 dbolser I have the following warning spewing forth from my code, and I don't know where it comes from or if it's serious:
17:36 dbolser Use of uninitialized value in length at /homes/dbolser/perl5/lib/perl5/Bio​/DB/SeqFeature/Store/DBI/mysql.pm line 1060, <I> line 199.
17:38 rbuels dbolser: ah, that comes from line 1060 of Bio::DB::SeqFeature::Store::DBI::mysql.
17:38 dbolser oh right
17:38 * rbuels nods sagely
17:38 rbuels dbolser: one good response would be "thanks, helpie!"
17:39 dbolser screw you pig!
17:40 dbolser so... how can I find which line of my code triggers this other line of code to spew?
17:40 dbolser and why
17:41 * rbuels sighs
17:41 rbuels now you're asking the hard questions.
17:41 dbolser lets take it from the top
17:41 rbuels dbolser: probably seeing what's at that line will give you a clue
17:42 dbolser gimmy Bio/DB/SeqFeature/Store/DBI/mysql.pm line 1060
17:44 dbolser if (length $source_tag)
17:44 dbolser that seems like a weird test to perform
17:45 dbolser shouldn't it be better "if (defined $source_tag)"?
17:45 rbuels https://github.com/bioperl/bioperl-live/blob/ma​ster/Bio/DB/SeqFeature/Store/DBI/mysql.pm#L1060
17:45 dbolser not that I think my script has any blank sources
17:45 rbuels dbolser: well, length and defined are different
17:45 dbolser thanks, botie!
17:45 dbolser rbuels: orly?
17:45 rbuels dbolser: two T's!
17:46 dbolser yes I know...
17:46 rbuels bottie!
17:46 rbuels dbolser: '' is defined, but not length.
17:46 dbolser I know about the difference beween length and defined
17:46 dbolser but ... I don't know why length is being used here when it can clearly be undef
17:46 rbuels probably wasn't undef in his tests
17:46 dbolser can't we just change it and see if any tests fail?
17:47 * dbolser wonders why it's undef now
17:47 * rbuels wonders this too
17:48 * dbolser adds debuging calls to this insane script
17:50 dbolser this bug started recently btw
17:51 dbolser ok, I tracked down the line of my script:
17:51 dbolser https://gist.github.com/887263
17:53 dbolser https://gist.github.com/887264
17:53 dbolser ^^ dumped output of returned results
17:55 dbolser prolly redundant: https://gist.github.com/887272
17:56 rbuels dbolser: OMG the data!!!  I can see yur potato features!  what about the *publication*!?!?!
17:57 * rbuels wrings his hands
17:57 dbolser $scaff and $end  are equal to PGSC0003DMB000000037    2097001
17:57 dbolser swine
17:57 dbolser lol
17:57 * rbuels chuckles
17:57 dbolser forget the data, which one of you put my name on it ;-)
17:58 rbuels that's all lukas
17:58 dbolser that made me laugh I have to say
17:58 * rbuels doesn't touch the maps
17:58 deafferret sudo -u rbuels /say i did
17:58 rbuels bash: /say: No such file or directory
17:58 deafferret bash?  you don't use irssi-shell?  luser
17:59 dbolser deafferret: the monthly genomics 'newsletter' came out with a paragraph saying "A new map was submitted by Dan Bolser!!!"
17:59 deafferret dbolser++
17:59 dbolser how I laughed... when I found out ... that 'private' data was released by me...
17:59 dbolser still ... no harm done
17:59 deafferret :)
18:00 mmlevitt joined #bioperl
18:00 dbolser so we can agree that Bio/DB/SeqFeature/Store/DBI/mysql.pm is broken and needs to be re-written from scratch?
18:01 rbuels dbolser: http://irclog.perlgeek.de/bioperl/2010-10-26
18:01 dbolser rbuels: I'm reading the documentation for _types_sql, but I'm still confused
18:01 * rbuels chuckles that all this is googlable
18:01 dbolser rbuels: are you pointing out that that excange was logged?
18:01 * rbuels nods
18:02 dbolser I need logs... to cover my asses
18:03 dbolser any clue what that sub is for?
18:03 rbuels what sub?
18:03 deafferret sammich
18:03 rbuels (looks like your feature has a source ...)
18:03 deafferret nomnom
18:03 dbolser indeed
18:04 rbuels it's probably doing some other feature
18:04 dbolser further ... I'm seeing lots of those
18:04 dbolser errors
18:05 dbolser rbuels: would moose / catalyst improve this at all?
18:05 dbolser or is it inherently ... messy
18:05 rbuels dbolser: no, that would not help
18:06 dbolser all that sql is hurting my eyes
18:06 rbuels dbolser: try adding a piece of code to your mysql.pm that dies and dumps if the source is undef
18:06 rbuels dbolser: to see what the piece of data actually is that it's choking on
18:06 dbolser dump what?
18:07 dbolser where's drily when you want a bug fixed...
18:07 dbolser driley
18:08 rbuels dbolser: dump the feature that has the undef source
18:08 dbolser ok
18:11 dbolser rbuels: g2g
18:11 dbolser here is the dump
18:12 dbolser meh... thats wrong: https://gist.github.com/887301
18:12 dbolser sorry... g2g
19:07 ank left #bioperl
19:10 mmlevitt left #bioperl
20:04 sheenams joined #bioperl
20:45 dave_messina joined #bioperl
20:46 dave_messina hey all, can anyone suggest a safe and easy way to install Perl 5.12 on OS X 10.6 without clobbering the system's installation of 5.10 ?
20:46 rbuels dave_messina: perlbrew
20:47 rbuels dave_messina: http://search.cpan.org/~gugod/App-perlbrew-0.18/
20:47 dave_messina aha, sweet. thanks, Rob.
20:47 rbuels bp
20:47 rbuels er, well that too
20:47 rbuels but i meant np
20:47 rbuels which stands for "no problem"
20:47 rbuels at least, that's what dbolser told me
20:48 dave_messina either that or nurse practitioner. But hey, whatever floats Dan's boat...
20:49 perl_splut I'd take either as long as it wasn't Nurse Ratchet
20:50 perl_splut sorry, Nurse Ratched
20:50 dave_messina Medication time!
20:50 dave_messina Then you'll definitely have "no problem"
20:50 * perl_splut runs to the padded room *
20:52 * perl_splut goes back to staring at Java code *
20:53 rbuels java is amusing.  so many lines to do so little.
20:53 rbuels it's very good for limiting the damage a bad programmer can do, though.
20:54 rbuels in The Enterpise (R)
20:54 philsf left #bioperl
20:54 perl_splut maybe... or can amplify their inability to write good code behind lots of bad objects...
20:55 perl_splut or in my case, trying to decipher code in a language I've never used before, ever...
20:59 deafferret 2M lines of code == job security! woot!
21:00 perl_splut there are 9 million bicycles in Beijing...
21:00 dave_messina left #bioperl
21:08 sheenams rbuels: i have a few questions about prove, got a minute?
21:09 rbuels sheenams: yep
21:09 mmlevitt joined #bioperl
21:09 mmlevitt hello, #bioper
21:09 sheenams rbuels: so i get the -I and the -l, but i'm not sure i understand what -r is doing.
21:10 sheenams rbuels: perldoc give two definitions, recurse and reverse
21:10 rbuels sheenams: -r is for 'recursive', meaning 'run all tests in the directory and any subdirectories thereof i am pointing you at'
21:10 rbuels sheenams: no, reverse is only --reverse
21:10 rbuels sheenams: it doesn't have a one-letter version.
21:11 rbuels sheenams: whereas recurse can be --recurse, or just -r
21:11 rbuels sheenams: make sense?
21:11 sheenams rbuels: yep.
21:11 * rbuels nods sagely
21:11 * rbuels strokes his fu manchu moustache
21:11 sheenams rbuels: so by saying: -lr t/ its saying to run all the tests I have in the t/ directory
21:11 sheenams which is just basic.t at this point
21:12 rbuels sheenams: yep, exactly
21:12 rbuels sheenams: 90% of the time I run prove, in all the projects I work on, it's prove -lr t/
21:12 rbuels actually the ending / is not really necessary
21:12 sheenams rbuels: k. now i just have to figure out why it doesn't work.
21:12 rbuels sheenams: the exception is bioperl-live, which for hysterical reasons does not have its modules in lib/
21:14 sheenams rbuels: i got my directories all worked out. apparently 2 am is not a good time to try learning anything.
21:14 rbuels haha
21:15 rbuels sheenams: if you're ever up at 2am again, i will most likely not be,  but some people in #perl on this same server probably will be
21:15 rbuels sheenams: type /join #perl to see that channel
21:16 sheenams rbuels: that's not a normal working hour for me.
21:16 rbuels :-)
21:17 * rbuels wonders where py<tab> has gone
21:17 rbuels sheenams: fun fact:  most of the time you can tab-complete people's names in IRC
21:17 * rbuels thinks she probably already knew this
21:18 sheenams i didn't, so thanks. i have never used IRC before this bioperl adventure
21:18 kyanardag left #bioperl
21:19 sheenams rbuels: and we're still calling it an 'adventure' as opposed to anything more hellish
21:19 rbuels sheenams: wow really?  i should dial up the difficulty then
21:19 * rbuels chuckles
21:19 sheenams rbuels: ha...i am only pretending that i'm waving and not drowning...
21:20 rbuels there we go, that's the spirit
21:20 rbuels ;-)
21:21 sheenams still have to write my gsoc application, as well as learn this 'assignment' (in addition to my real assignments)
21:23 mmlevitt gsoc will be a great experience!
21:23 sheenams rbuels: the prove line is telling me it can't locate the cxgn/tools in cxgn-corelibs/lib/cxgn path.
21:24 sheenams mmlevitt: just trying to get up to par for the gsoc is turning into a great experience. I'm learning a lot!
21:25 rbuels sheenams: are you doing something like:  cd Bio-AGP; prove -lr -I ../cxgn-corelibs/lib t/
21:26 sheenams rbuels: prove -I /sheena-warrior/cxgn-corelibs/lib/CXGN -lr t/  while in the Bio-AGP dir
21:27 rbuels sheenams: that's an absolute path.
21:27 rbuels (/sheena-warrior/cxgn-corelibs/lib/CXGN)
21:27 rbuels sheenams: that's probably not right
21:28 sheenams rbuels: using the cd Bio-AGP line from the sheena-warrior dir gives the same error
21:28 rbuels sheenams: so.  you should be in the Bio-AGP directory.
21:28 rbuels sheenams: and the path you should give to -I is the absolute or relative path to the 'lib' dir in cxgn-corelibs
21:29 rbuels sheenams: you can tell if it's right by running something like 'ls <that path>'
21:29 rbuels sheenams: which, if that path is right, should show you CXGN/ and Bio/ and the other things in the cxgn-corelibs/lib dir
21:30 kyanardag joined #bioperl
21:37 mmlevitt I was not able to get GMOD's GBrowse, postgresql, or a chado installation replicated, but I was able to get ace and contig files from AMOS.
21:37 mmlevitt which Linux OS is recommended as base system.  I tried Ubuntu, but now I will try Fedora.
21:39 rbuels mmlevitt: i like ubuntu, mostly
21:39 mmlevitt rbuels: me, too.
21:40 rbuels mmlevitt: that's what i use for my development
21:41 mmlevitt deb packages werent configured automatically, so I will try again with Fedora.
21:46 mmlevitt but, yes, I do most computing with Ubuntu!
21:59 sheenams rbuels: does the prove -l command work the same as the use lib 'file' command?
21:59 rbuels sheenams: -l is equivalent to -I lib is equivalent to 'use lib "lib"' in code
22:02 deafferret Debian++
22:02 sheenams rbuels: ok. can i email you my error print out for some help?
22:03 rbuels sheenams: use http://gist.github.com
22:03 rbuels sheenams: that, or similar things like pastebin.com, are the best way to exchange error outputs over irc
22:03 rbuels there are million paste services
22:06 rbuels sheenams: just paste the url here after you make the paste
22:06 sheenams rbuels: git@gist.github.com:55f1abc9c497c5ac9465.git
22:07 rbuels sheenams: eh, probably you wanted http://gist.github.com/55f1abc9c497c5ac9465
22:07 * rbuels looks
22:07 rbuels i don't want to clone it :-)
22:07 rbuels sheenams: look at the printout of @INC there.
22:08 sheenams rbuels: im new, i just copied the link it gave me :)
22:08 rbuels sheenams: you can see what it thinks you mean by 'sheena-warrior/cxgn-corelibs/lib'
22:08 rbuels sheenams: read this: http://linux.about.com/od/i​tl_guide/a/gdeitl29t01.htm
22:08 sheenams rbuels: yes, it thinks it in the Bio-AGP dir
22:08 rbuels sheenams: so, you need the path to where you actually have it
22:09 rbuels sheenams: you're in Bio-AGP, you gave it a relative path sheena-warrior/cxgn-corelibs/lib, so it thinks you mean /home/sheena-warrior/Bio-AGP/sh​eena-warrior/cxgn-corelibs/lib
22:10 sheenams so it wants /home/sheena-warrior/cxgn....
22:11 rbuels sheenams: so you would probably want to say either -I ../cxgn-corelibs/lib or -I /home/sheena-warrior/cxgn-corelibs/lib
22:11 rbuels sheenams: those are the two ways to specify that path:  first one is relative, second one is absolute
22:11 rbuels sheenams: not that those two dots, '..' are not just for show.
22:12 rbuels sheenams: er, note that
22:12 sheenams rbuels: noted
22:13 sheenams rbuels: well its giving me a new error, so I'm at least getting somewhere
22:13 rbuels sheenams: you can reuse and edit the same gist if you want
22:13 rbuels sheenams: just tell me when to refresh
22:15 sheenams rbuels: refresh
22:15 rbuels sheenams: ah, so Tie::URLEncoder, you can see, is being used by something that Bio::AGP::LowLevel is using
22:15 rbuels sheenams: see how the error report is a back-trace?
22:16 rbuels sheenams: starts at the failure, then moves through what called what?
22:16 rbuels sheenams: giving file names and line numbers
22:16 rbuels sheenams: so in this case, you don't have Tie::URLEncoder installed, and that's used by CXGN::Tools::Identifiers, which is used by Bio::AGP::LowLevel
22:17 sheenams rbuels: which is why i was completely lost trying to figure out what Tie::URLEncoder was...
22:17 rbuels sheenams: probably the easiest thing to do would be to either a.) install that module, or b.) progress onward to removing the uses of CXGN::Tools::Identifiers from the Bio::AGP::LowLevel module
22:18 rbuels sheenams: http://lmctfy.org/Tie::URLEncoder/
22:18 rbuels sheenams: so that tells you what Tie::UrlEncoder is :-)
22:20 sheenams rbuels: removing the CXGN::Tools::Identifiers line lets the test run finally
22:21 sheenams rbuels: it fails, but it runs at least
22:21 * rbuels nods
22:23 kyanardag left #bioperl
22:24 sheenams rbuels: its looking for the file 'chr04.v3.agp' listed in basic.t, but can't find it (its looking in a dir i don't have : Bio-AGP/t/data/chr04.v3.agp)
22:24 rbuels sheenams: ah, forgot about that.
22:25 rbuels sheenams: grab it out of the cxgn-corelibs/t/data dir
22:27 sheenams rbuels: success!!!
22:27 rbuels sheenams++
22:28 rbuels sheenams: now you need to work up to the point where the tests pass without the -I that includes the module's old home, and put the dist in your github account
22:29 rbuels sheenams: use git to consolidate your gains.
22:29 rbuels sheenams: like, you should commit now
22:29 sheenams rbuels: gees, let me enjoy my success for a second before i have to struggle more :)
22:30 rbuels sheenams: well, i think it might be beer thirty
22:30 sheenams rbuels: yeah, i hear a spire hard cider calling my name...
22:32 sheenams rbuels: so each time i commit, do i do all three steps you listed (init, add, commit)?
22:32 rbuels sheenams: no, init makes a repo.
22:32 rbuels sheenams: only have to do that the first time
22:33 rbuels sheenams: the add and commit steps can be compressed together once you learn more git foo
22:33 sheenams rbuels: so just add (to add my new/changed files) and then commit with -m message
22:34 rbuels sheenams: yep
22:34 rbuels sheenams: learn the other git commands too
22:34 rbuels sheenams: git show, git log, git reset, git push, git pull
22:35 rbuels sheenams: git help <command>
22:35 rbuels sheenams: git diff
22:35 rbuels sheenams: git diff --cached
22:39 sheenams rbuels: question about the lowlevel.pm file.
22:40 sheenams rbuels: in the synopsis it still lists some of the subroutines that I deleted. Do i worry about that at this point?
22:40 rbuels sheenams: eh, should delete them at your convenience
22:45 mmlevitt postgresql was accessed successfull on Fedora 14, and BioPerl will be installed.
22:46 rbuels mmlevitt: nice, congrats
22:52 philsf joined #bioperl
22:57 mmlevitt left #bioperl
23:16 sheenams left #bioperl
23:26 perl_splut left #bioperl
23:44 mmlevitt joined #bioperl
23:44 mmlevitt left #bioperl
23:49 mmlevitt joined #bioperl
23:56 philsf left #bioperl
23:57 sheenams joined #bioperl

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