Camelia, the Perl 6 bug

IRC log for #bioperl, 2011-04-28

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

All times shown according to UTC.

Time Nick Message
00:18 carandraug joined #bioperl
01:14 carandraug left #bioperl
05:16 philsf joined #bioperl
05:16 philsf left #bioperl
05:16 philsf joined #bioperl
05:33 philsf left #bioperl
05:43 zenman joined #bioperl
05:58 zenman left #bioperl
06:39 koviel joined #bioperl
06:40 koviel hi everyone
06:40 koviel can i ask you a question&
06:41 koviel im a bioperl beginner and i really need help :)
06:44 koviel left #bioperl
07:01 koviel joined #bioperl
07:05 koviel hm.. im a very beginner, so ill try to ask, maybe someone will answer
07:05 koviel I wanted to make some changes in sequences from embl database. So i found that it is easy to make changes in cds: use Bio::SeqIO; my @cds_features = grep { $_->primary_tag eq 'CDS' } Bio::SeqIO->new(-file => $embl_file)->next_seq->get_SeqFeatures; my %gene_sequences = map {$_->get_tag_values('gene'), $_->spliced_seq->seq } @cds_features; foreach( keys %gene_sequences ) {   $gene_sequences{$_} =~ s/A/G/g; }  but how can i replace cds
07:05 deafferret ah, you're back  :)
07:06 deafferret replace? with what?
07:07 koviel with cds sequences
07:07 koviel i removed all A bases in cds stored in a hash
07:08 deafferret and you want to write a new embl_file ?
07:08 koviel now i want to put all canged cds back
07:08 philsf joined #bioperl
07:08 deafferret you'll need to write a new file
07:08 koviel not new
07:08 koviel ok
07:08 koviel how can i write a new one?
07:08 deafferret you can't replace inline. so write a new file then rename it at the end if you want  :)
07:09 deafferret my $out = Bio::SeqIO->new(-file => '>out.embl');   ?
07:09 koviel um...
07:10 koviel and this file will contan the old seuence with changes&
07:10 koviel ?
07:10 koviel with my changes*
07:10 deafferret http://www.bioperl.org/wiki/HOWTO:SeqIO
07:10 deafferret halfway down --  "# create one SeqIO object to read in,and another to write out"
07:10 koviel ive red it :(
07:11 koviel the problem is im a russian student :)
07:11 koviel cant get it
07:12 deafferret what do you need -- a whole new embl file? or just a list of CDSs that you've changed?
07:12 koviel the whole new embl file
07:12 deafferret did you find the example I refered you to? see how it reads a file and writes another one? doesn that make sense? run it and make sure it works for you
07:14 deafferret you'll do a foreach across every sequence, and then CHANGE the seq with your new seq, and write to the output file
07:15 apeiron joined #bioperl
07:15 apeiron left #bioperl
07:18 koviel hm... look
07:19 koviel $seqio_object = Bio::SeqIO->new(-file => 'test1.txt', -format => 'EMBL', -alphabet => 'dna');  while ($seq_object = $seqio_object->next_seq){  @cds_features = grep { $_->primary_tag eq 'CDS' } $seq_object ->get_SeqFeatures;  %gene_sequences = map {$_->get_tag_values('gene'), $_->spliced_seq->seq } @cds_features;  foreach( keys %gene_sequences ) {   $gene_sequences{$_} =~ s/A/G/g; } }  $seqio_obj = Bio::SeqIO->new(-file => '>
07:19 koviel i tried to do it lke that
07:19 deafferret use gist.github.com to show your code
07:20 koviel https://gist.github.com/945944
07:22 deafferret you need to indent inside { blocks to make the code readable
07:25 koviel https://gist.github.com/945949
07:25 deafferret much better  :)   you can update your gists if you want instead of posting new ones
07:26 koviel :)
07:26 koviel ok, but still i have mistakes
07:26 deafferret yes.   :)
07:26 koviel or i misunderstood smth
07:27 koviel i dunno what :)
07:27 deafferret call the first $seqio_in and the second $seqio_out
07:27 deafferret and $seqio_out needs to be at the top so you can write seqs to it as you go
07:28 deafferret then you'll need to loop $seq_object->get_SeqFeatures explicitly so you can manipulate each one
07:28 koviel ok
07:29 koviel im happy i found help here :)
07:29 koviel will you be here later?
07:30 koviel i need to go now, but ill correct my code
07:30 koviel and ask you more questions :)
07:30 deafferret heres a loop example  https://clabsvn.ist.unomaha.edu/anonsv​n/CLAB/perl_examples/parse_genbank.pl
07:30 koviel thanks
07:35 koviel left #bioperl
07:36 philsf left #bioperl
08:29 dbolser I got local::lib installed... how do I get perl 5.10?
10:01 dbolser I figured it would take me less time to rewrite my script than to install 5.10.0
10:01 dbolser I'd appreciate any feedback on the script, which seems to be working nicely
10:02 dbolser it's working so well, its showing me errors in SSAHA2 alignmetns
10:02 dbolser https://gist.github.com/0dbebac2441aa9eee4b1
10:03 dbolser note $mapper is just a Bio::Coordinate::Collection
12:09 takadonet joined #bioperl
12:26 genehack dbolser: use perlbrew
12:26 genehack http://search.cpan.org/~gugod/App-​perlbrew-0.19/lib/App/perlbrew.pm
12:26 genehack Look at the INSTALLATION section
12:27 genehack and given that 5.14 is gonna be out right around the first of May, I'd skip 5.10 and do /at least/ 5.12.3, if not 5.14
12:27 genehack of course, with perlbrew, the only thing standing between you and _all_ those versions is time and disk space. 8^)
12:38 carandraug joined #bioperl
13:05 dbolser hmm
13:05 * pyrimidine plans on testing on 5.14-rc1 soon-ish
13:05 pyrimidine testing bioperl, that is
13:06 dbolser pyrimidine: we need to stop syncing like that
13:06 pyrimidine :)
13:06 * pyrimidine loves the name of the latest Dancer release: "Silence of the ambs"
13:16 koviel joined #bioperl
13:28 genehack pyrimidine: i'm pretty sure it installs cleanly; further than that I haven't poked it.
13:31 genehack pyrimidine: yeah, just confirmed on another box -- 'cpanm Bio::Perl' on 5.14-RC1 Just Works(tm)
13:41 pyrimidine genehack: good!  at least I won't have to worry about the problems we had going from 5.8 -> 5.10, or 5.10 -> 5.12
13:41 * pyrimidine -> meeting
14:00 rziman joined #bioperl
14:19 rziman left #bioperl
14:22 gawbul joined #bioperl
14:24 gawbul left #bioperl
14:25 gawbul joined #bioperl
14:46 philsf joined #bioperl
14:46 philsf left #bioperl
14:46 philsf joined #bioperl
15:33 dbolser rbuels: does this look reasonable for a 'lift over' script?
15:33 dbolser https://gist.github.com/0dbebac2441aa9eee4b​1/f8e17e7b730ae0db00a02ae877ace282f0653af2
15:33 dbolser $mapper is just a Bio::Coordinate::Collection built from GFF or AGP
15:34 dbolser pyrimidine: too if you're back
15:35 rbuels dbolser: what's a split report?
15:35 dbolser its a meta assembly files
15:36 dbolser rbuels: it ends up being a Bio::Coordinate::Collection
15:36 dbolser it may just as well be an AGP
15:36 dbolser think of it as an AGP
15:36 dbolser (the code is essentially identical in that case)
15:37 dbolser note, I implement AGP as Bio::Coordinate::Collection, just in case you're wondering what I'm talking about when I say "it's an AGP"
15:37 * rbuels wonders how the test coverage is for these
15:38 dbolser ahhh...
15:38 rbuels (is talking about Bio::Coordinate::Collection)
15:38 dbolser I submitted a bunch of weird test failures for that
15:39 rbuels in redmine?
15:39 dbolser I expected some feedback from heki
15:39 dbolser no, I think just on the mailing list
15:39 * rbuels looks
15:39 dbolser there was behaviour I didn't understand, that was undocumented
15:39 rbuels how surprising
15:39 dbolser so I set it to fail...
15:40 dbolser wait, I'm either insulted or feeling normal...
15:40 dbolser rbuels: I think I may have been using it wrong back then, and getting ignored according to Warnock's dilemma
15:41 rbuels hehehee
15:41 rbuels i'm talking about weird undocumented behavior and bugs in bioperl
15:41 rbuels (to clarify)
15:41 rbuels :-)
15:42 dbolser ok, I'm normal
15:42 dbolser ;-)
15:42 dbolser [GitHub] dbolser bio coordinate pair tests [bioperl/bioperl-live GH-2]
15:42 dbolser (in my email archive)
15:42 dbolser can't find details on bp list
15:42 rbuels this toolkit was developed for a long time in a manner similar to the steep hillside behind a slum that lacks utility services
15:43 dbolser where did you read that?
15:43 rbuels i made it up
15:43 dbolser it's actually working quite nicely, once you get used to the weird gap/math results
15:43 dbolser s/math/match/
15:44 dbolser thought it was in the docs
15:44 dbolser "Question about Bio::Coordinate::Pair" <- bp mailing list topic
15:44 rbuels wow that was a long time ago
15:45 ank joined #bioperl
15:45 dbolser back when I was full of hope
15:46 rbuels the lack of response to your request should be seen as nominating you as the new community expert on this topic
15:46 rbuels hehe
15:46 dbolser but I'm not sure what teh problem was back then that I was trying to fix
15:46 dbolser lol
15:47 dbolser I think it comes down to weather you pass seq_id's to the locations or not, because bcp is somehow seq_id aware, although I see no reason why it should be
15:47 dbolser s/bcp/B:Coord:Pair/
15:47 dbolser Bio:Coord:Collection is seq_id aware, and that is good
15:48 rbuels has dbolser_bio_coordinate_pair_tests been looked at my anybody?
15:48 dbolser because it tells you the chromosome your feature is on when you map it via a superscaffold in an agp
15:48 dbolser do I count?
15:48 dbolser then no
15:49 rbuels dbolser: ok, point 1: you need to configure your editor to not emit tab characters
15:49 rbuels (or maybe you have done this since then)
15:50 rbuels dbolser: see how messed up the indentation is in your patch?  https://github.com/dbolser/bioperl-live/compare/m​aster...dbolser_bio_coordinate_pair_tests#L0R308
15:51 genehack rbuels: that steep hillside comment should go on the wiki somewhere... 8^)
15:51 * rbuels chuckles
15:51 * dbolser has already tweeted it against a picture of rbuels face
15:51 dbolser (only kidding)
15:52 dbolser rbuels: I don't follow about the tab
15:52 rbuels dbolser: you know what a tab character is?
15:52 dbolser yes
15:52 rbuels dbolser: you know what indentation is?
15:53 dbolser ... yeeees
15:53 * dbolser feels like hes walking into a trap
15:53 dbolser I don't see the problem with teh above path
15:53 rbuels dbolser: and do you see how, at line 308 in that diff, the stuff inside the if statement is protruding to the left of the if statement definition?
15:53 dbolser ahhhh
15:53 rbuels (in Bio::Coordinate::Pair)
15:53 rbuels dbolser: that's cause there's a tab character there
15:54 dbolser can't we all just agree a tab is 8 spaces?
15:54 rbuels dbolser: we could.
15:54 dbolser but point taken
15:54 rbuels dbolser: you must be the change you wish to see in the world
15:54 * rbuels chuckles
15:54 dbolser lol
15:55 dbolser in the real world, I don't like to complain about things I'm unwilling to try to fix
15:55 dbolser however, in cyberspace, everybody can hear you screem
15:55 dbolser so I tend to rant a lot more
15:55 rbuels change is just a find Bio -type f -and -name '*.pm' -exec perl -i -pE 's/\t/        /g' {} ';' away
15:55 gawbul left #bioperl
15:56 dbolser OK
15:56 rbuels and then, http://search.cpan.org/~bobtfish/T​est-NoTabs-1.0/lib/Test/NoTabs.pm
15:56 rbuels to seal the deal
15:56 dbolser btw, do you use emacs or eclipse or something else?
15:56 rbuels dbolser: i use emacs
15:56 dbolser cperl-mode?
15:57 rbuels dbolser: yes
15:57 dbolser OK, just to check there is precedent
15:57 dbolser anyway, lets not worry about those mysterious bugs
15:58 rbuels dbolser: if i see a tab character outta you again, soldier, you'll be scrubbing the latrine with your toothbrush
15:58 * rbuels chuckles
15:59 dbolser hmm...
15:59 rbuels anyway.
15:59 dbolser :-)
15:59 dbolser I hadn't considered actually doing anything about those tabs till you said that
15:59 * rbuels continues reading dbolser's patch
15:59 dbolser its just a bunch of tests
16:02 pyrimidine I think we need to make that the new BioPerl slogan:  "A toolkit developed for a long time in a manner similar to the steep hillside behind a slum that lacks utility services"
16:02 rbuels hahhah
16:02 rbuels small characters, italics, below the bioperl logo
16:03 pyrimidine but I think the words 'duck tape' and 'pixie sticks' need to be inserted in there somewhere
16:03 dbolser rbuels: what is your "Cperl Indent Level"?
16:04 rbuels dbolser: i use 4 chars
16:04 pyrimidine same here
16:04 rbuels that's relatively standard
16:04 dbolser me has 2
16:04 pyrimidine ick
16:04 pyrimidine :)
16:04 * dbolser has nothing to prove ;-)
16:06 genehack haha.
16:06 rbuels dbolser: sure you do.  t/Coordinate/CoordinateMapper.t
16:06 * genehack is also in the 2 camp unless forced by circumstances or work
16:06 rbuels oh, man, now you've done it.
16:06 dbolser heh
16:06 * rbuels reaches for his scimitars
16:06 * dbolser has set 4 so as not to look like a fool
16:07 genehack really i'm in the "use whatever the existing code uses" camp
16:07 genehack just so happens all my code uses 2. 8^)
16:07 dbolser I don't see tabs mentioned in Cperl Indentation Details
16:07 pyrimidine genehack: pretty much the same here.  I'm more bothered about use of tabs instead of spaces
16:07 dbolser genehack: I'm in the 'how did this author cope without my indentation standards, so I'll fix it' camp
16:08 dbolser is it cperl or just emacs?
16:08 dbolser C Insert Tab Function
16:10 genehack pyrimidine: All Tabs Must Die
16:10 genehack and i got my eye on that trailing whitespace nonsense too.
16:10 dbolser genehack: how many chars in a tab ;-)
16:11 genehack <sfx: Sam Jackson voice, ala Pulp Fiction>IT DOESN'T MATTER, because I'd never use one of the dirty things.
16:11 dbolser http://www.jwz.org/doc/tabs-vs-spaces.html
16:12 pyrimidine rbuels: do you think my response on the list to sheenams was comprehensive enough? ;)
16:12 genehack dbolser: i think there is some way to tell cperl-mode to use tabs.
16:12 dbolser (setq indent-tabs-mode nil)
16:13 genehack right.
16:13 * pyrimidine is just awaiting a tl;dr response...
16:13 rbuels what is tl;dr?
16:13 rbuels i've been meaning to ak
16:13 dbolser genehack: you mean _not_ to use tabs
16:13 pyrimidine "too long; didn't read"
16:13 genehack dbolser: yeah
16:13 rbuels oh, right
16:13 dbolser Parent groups: Fill - found it eventually
16:14 genehack pyrimidine: i thought it was a good response, although i was a little tempted to send a direct message to her saying "don't get sucked into the bikeshed, JFDI!"
16:14 pyrimidine genehack: you should do that :)
16:14 dbolser JFDI?
16:14 dbolser just do it?
16:15 pyrimidine Just Fscking Do It
16:15 genehack just fscking do it
16:15 pyrimidine heh, same censor I see
16:15 rbuels wow.  they both used fsck
16:15 rbuels o_O
16:16 pyrimidine habit.  Gets around the censor-bots patrolling the Matrix
16:19 dbolser rbuels: your example tab fix, -E should be -e right?
16:21 rbuels dbolser: i always use -E, which is like -e but turns on all the non-backcompat features whatever version of perl you have supports
16:22 rbuels dbolser: like 'say'
16:22 rbuels which is really good to have for one-liners
16:22 dbolser Unrecognized switch: -E
16:22 dbolser nm
16:22 rbuels dbolser: but i think it's only 5.10 and above that support it
16:22 dbolser cool
16:22 rbuels dbolser: change it to -e, should work on 5.8
16:23 dbolser yup
16:23 pyrimidine dbolser: get perlbrew installed yet? :)
16:23 dbolser no
16:23 dbolser ;-)
16:23 dbolser I'll look at it, but I had that script to write today
16:23 dbolser hence the question, would it be better with FeatureIO?
16:23 dbolser (implied question)
16:24 dbolser lastlog -pE
16:24 rbuels pyrimidine: did you see the stuff dbolser and i were talking about, https://github.com/dbolser/bioperl-live/compa​re/master...dbolser_bio_coordinate_pair_tests and "what is the proper strand of a gap?"
16:26 pyrimidine rbuels: no, hadn't seen that.  I would think gaps don't have a strand (or it is set to 0).
16:27 rbuels pyrimidine: it's related to http://bioperl.org/pipermail/bio​perl-l/2010-October/034100.html
16:27 pyrimidine actually, is this a gap in terms of assembly (not an alignment gap)?
16:27 dbolser it qas poping up undef iirc
16:27 rbuels pyrimidine: which was warnocked
16:28 dbolser assembly gap, like when building a gene from a genbank thingamy
16:29 dbolser http://www.bioperl.org/wiki/​Module:Bio::Coordinate::Pair
16:29 dbolser Bio::Coordinate::Result
16:29 rbuels pyrimidine: so, he changes the way the strandedness of gaps is, and it makes some of the tests fail that specifically test that strandedness
16:29 dbolser its the Bio::Coordinate::Result that contains gaps
16:29 rbuels pyrimidine: but it is lost in the mists of time whether those tests are, themselves correct
16:30 dbolser rbuels: but setting the strand has other weird consequences iirc
16:30 rbuels pyrimidine: i.e. maybe they are just testing the current behavior of the object (which might be incorrcet)
16:30 pyrimidine I would think this type of Gap is strandless
16:30 pyrimidine or, it doesn't make a difference
16:30 rbuels dbolser: what weird consequences?
16:31 dbolser would you like me to run those tests and let you know why I thougth the behaviour was strange?
16:31 dbolser I forgot, didn't I discuss it in the test?
16:32 pyrimidine a strand popping up undef to me is a bug
16:32 rbuels pyrimidine, dbolser: i just checked out that branch, and merged master into it, and here is the only test failure: https://gist.github.com/946690
16:32 pyrimidine IIRC a Bio::Range, as defined, has a strand of -1, 0, or 1
16:34 dbolser look at line 254 for example (in the test)
16:35 pyrimidine I tend to think the tests are bad.  I have seen a few cases of that, actually (in particular, with SearchIO)
16:36 dbolser I'm just trying to co that branch
16:37 pyrimidine just a warning, re: locations; there is some inherent funkiness (or maybe that should be fsckiness) with Bio::Location and strand (I think in particular Bio::Location::Split)
16:37 pyrimidine can't recall what it was specifically, but some nasty hackiness afoot IIRC
16:39 dbolser Try commenting out the three -seq_id lines below to observe strange
16:39 dbolser ## interactions!
16:39 dbolser should that be considered a bug?
16:39 pyrimidine to paraphrase Nicholson's Joker 'this code could use an enema'
16:41 dbolser https://github.com/dbolser/bioperl-live/blo​b/2d3f4f336da8215b945d15777c18e509725a430f/​t/Coordinate/CoordinateBoundaryTest.t#L344
16:41 dbolser if I comment that out, I get a failure
16:42 dbolser Failed test at t/Coordinate/CoordinateBoundaryTest.t line 515. got: '1125' expected: '6001'
16:43 dbolser its funny because if I comment out the seq id on line 334 alone, there is no error
16:43 rbuels so this means that this code is depending on the seq_id's to do the remapping??
16:43 dbolser but if I comment out the seq_id on line 334 and 344, there is a different error
16:43 dbolser yeah, but it doesn't fail if it's missing, I just does something else
16:44 rbuels oh, i guess it would have to, since the location is determined partly with the seq id
16:44 dbolser and there is no usefull way to use the seq_id here I think
16:44 dbolser but when you're dealing with a pair of sequences, there is only one in and one out sequence
16:44 dbolser and there is no matching to the seq-id done here
16:45 dbolser i.e. it can be an arbitrary random string
16:45 dbolser and not all tests fail, just some...
16:45 pyrimidine I think this should be filed as a bug
16:45 pyrimidine seems like one to me, at least
16:46 rbuels pyrimidine: you mean the unset seq_id thing?
16:46 rbuels pyrimidine: causing the mapping to fail?
16:46 dbolser the mapping doesn't fail, it just changes
16:46 pyrimidine dbolser: okay, but is it correct?
16:47 dbolser thats what I wanted to know too ;-)
16:47 dbolser I need to look further, because this is after changing Bio/Coordinate/Pair.pm
16:47 dbolser how do I pull the master Bio/Coordinate/Pair.pm into my current branch?
16:48 pyrimidine have you set upstream?
16:49 dbolser I don't know, so I guess no
16:49 pyrimidine something like: git remote add upstream git://github.com/bioperl/bioperl-live.git
16:50 dbolser pyrimidine: I think I have that as 'origin', my fork is 'dbolser'
16:53 pyrimidine oh, you pulled in your fork as a branch.  As long as origin is up to date, I think it's just a 'git merge origin/master' when you are in your current branch
16:57 dbolser Auto-merging Bio/Coordinate/Pair.pm
17:00 dbolser now I want to temporarily change the Bio/Coordinate/Pair.pm in this branch with the one in master
17:04 rbuels dbolser: use git checkout master Bio/Coordinate/Pair.pm maybe
17:04 rbuels dbolser: which will overwrite your bio/coordinate/pair in your working copy with the one from master
17:04 rbuels dbolser: (which you can then revert after you're done poking it)
17:04 rbuels dbolser: (using git reset)
17:04 rbuels dbolser: (or git checkout again)
17:10 dbolser ty
17:11 dbolser now I'm seeing the original issue
17:11 dbolser not ok 162
17:11 dbolser #   Failed test at t/Coordinate/CoordinateBoundaryTest.t line 494.
17:11 dbolser #          got: undef
17:11 dbolser #     expected: '-1'
17:11 dbolser (this is before any of the 'comment out the seq_id' hilarity)
17:12 dbolser things start to come back to me... the strand of the gap is the same as the match in some cases, but undef in others
17:12 dbolser I have no idea if this is desireable, expected, or buggy
17:12 dbolser how did we get on this topic?
17:12 rbuels dbolser: you started talking about it
17:13 rbuels dbolser: maybe you should email heikki directly
17:13 dbolser can I get those hours back?
17:13 * rbuels just wants to bring closure
17:13 dbolser yup
17:13 rbuels dbolser: since he's not paying attention to the mailing list
17:13 dbolser ic
17:18 dbolser 12:32 <@rbuels> pyrimidine, dbolser: i just checked out that branch, and merged master into it, and here is the only test failure: https://gist.github.com/946690
17:18 dbolser rbuels: sorry I missed that
17:18 pyrimidine dbolser: heikki should be responsive (I think he uses Bio::Coordinate quite a bit in his work)
17:19 dbolser rbuels: was that the master tests on my patch of Pair.pm?
17:19 pyrimidine also, this is a really good opportunity to use what you are doing as the incentive to writing up docs (HOWTO or whatever) that help explain how Bio::Coordinate works
17:19 rbuels dbolser: that was checking out your branch, and merging msater.
17:19 rbuels (into it)
17:19 rbuels (to bring things up to date)
17:20 dbolser my branch changed the tests and pair.pm together
17:20 dbolser when I run the test on that branch using the Pair.pm from master, I get these results:
17:20 pyrimidine I would say, though, if you are getting strand not being set (or set to undef) somewhere that is a bug
17:21 dbolser I'll email heikki
17:21 dbolser https://gist.github.com/cccdd2f4c8f8cc3f4a7f
17:21 dbolser ty pyrimidine. I guess I confused the issue by trying to fix that bug and then asking if my fix was reasonable
17:22 dbolser rather than keeping it simple and just logging a bug
17:22 dbolser thanks for help, sorry for distraction
17:24 pyrimidine yeah, strand should default to '0', but I think there were problems setting it up to do that. That's one nice thing about Moose/MooseX::Types, you can limit the value allowed and have a default
17:24 dbolser cool
17:28 dbolser In the bug fix email I talk about 'setting strand consistently', I think I just set the strand of the Bio::Coordinate::Result::Gap to match the Bio::Coordinate::Result::Match within the Bio::Coordinate::Result, which then returns a non-undef strand in response to a non-specific -> strand method call
17:39 dbolser pyrimidine: I'm not sure this is a bug afterall, if a Bio::Coordinate::Result has a gap and a match feature with different strands, it can't be expected to have an overall strand
17:40 pyrimidine dbolser: right, so the strand should be 0
17:40 dbolser I think the way the strands (of the match and gap features) are being set here is correct
17:40 dbolser pyrimidine: ok, I didn't realise that was the standard...
17:40 pyrimidine no, it's not.  undef != 0
17:41 dbolser I think the issue is that it has a valid strand function in some contexts, but an invalid strand function in other contextds
17:41 dbolser you think it should return 0 when it doesn't really make sense for it to have a strand?
17:42 pyrimidine a strand of 0 means it doesn't have a strand
17:42 dbolser ok
17:42 pyrimidine or that it is unknown
17:43 dbolser its defo confusing, because the pair has two locations, and is then used to map a third location
17:43 dbolser each with it's own seq_id
17:43 pyrimidine this is one of the problems with Bio::Location::Spluit, actually (I think it returns undef if the contained locations don't have the same strand)
17:43 dbolser yup
17:44 pyrimidine *Split
17:48 dbolser other than that, all the behaviour seems standard
18:03 rbuels you mean behavior?
18:03 * rbuels ducks
20:14 sl33v3 left #bioperl
20:15 sl33v3 joined #bioperl
21:03 zenman_ joined #bioperl
21:25 zenman_ left #bioperl
21:40 philsf left #bioperl

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