Camelia, the Perl 6 bug

IRC log for #bioperl, 2011-03-09

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

All times shown according to UTC.

Time Nick Message
00:09 bag left #bioperl
00:10 mzgrideng left #bioperl
02:01 ank_ joined #bioperl
02:01 ank_ left #bioperl
02:02 ank_ joined #bioperl
02:03 ank left #bioperl
02:03 ank_ is now known as ank
03:47 CIA-57 bioperl6: takadonet master * re026953 / lib/Bio/Tools/CodonTable.pm : should have been a elsif instead of a if statement. Passing one more test - http://bit.ly/icwPP5
04:11 CIA-57 bioperl6: takadonet master * r6a62ff2 / t/Tools/IUPAC.t : added IUPAC.t which is p6 version of the one that's found in bioperl-live - http://bit.ly/fAP0pY
04:24 CIA-57 bioperl6: takadonet master * r26a6e8b / (lib/Bio/Tools/IUPAC.pm t/Tools/IUPAC.t): Commented out documentation so easier to work with in emacs. Now have class scoped variables. Commented out most methods for now. Added more tests - http://bit.ly/icqHrk
04:30 CIA-57 bioperl6: takadonet master * re120f1c / lib/Bio/Tools/CodonTable.pm : Removed hardcoded hash table and loading instead from Bio::Tools::IUPAC - http://bit.ly/dL1pd8
06:14 ank left #bioperl
06:15 ank joined #bioperl
06:27 bag joined #bioperl
07:53 bag left #bioperl
09:47 cassj joined #bioperl
10:44 komal_ joined #bioperl
11:04 cassj left #bioperl
11:14 komal_ left #bioperl
11:30 cassj joined #bioperl
11:42 cassj left #bioperl
12:13 dbolser is bioperl depreciated in favour of biomoose?
12:22 Lynx_ left #bioperl
13:11 takadonet dbolser: biomoose is still a work in progress
13:52 dbolser oh, is bioperl finished? ;-)
13:52 dbolser I need to read more on biomoose
13:54 pyrimidine biomoose?
13:55 dbolser It's my new name for bioperl use moose
13:55 dbolser wher's your repo?
13:55 dbolser and are you using complex subtypes?
13:56 pyrimidine dbolser: It's Biome: https://github.com/cjfields/biome
13:56 pyrimidine and one can use those, yes
13:56 dbolser to what extent do you think it should be done in 'biomoose'?
13:57 pyrimidine the biomoose name was nixed, per a past discussion on the mail list
13:57 dbolser nixed?
13:57 pyrimidine 'not used'
13:57 dbolser can I at least call it biomoo?
13:57 pyrimidine sure :)
13:57 dbolser :D
13:57 * dbolser emails
13:58 pyrimidine it's more a test-bed right now, but I have found some nice optimizations.  Haven't checked memory usage yet,
13:58 pyrimidine but feel free to hack on it.
13:59 dbolser I just attended a short talk about moose, and was marveling at the richness with which one can model ... stuff
13:59 dbolser like subtypes for example
13:59 pyrimidine dbolser: you should check out perl6
13:59 dbolser and type ontologies
13:59 dbolser that wasn't what the speaker said...
14:00 pyrimidine what did the speaker say?  Moose's types, roles, and classes are based on perl 6
14:00 dbolser oh, look at perl 6 in the sense of understanding moose... ok
14:00 dbolser he said 'perl 6 is a bag of shite'
14:00 pyrimidine who was the speaker?  mst
14:00 pyrimidine ?
14:01 dbolser our local perl guru, a guy called tom walsh
14:01 pyrimidine eh, he doesn't know from whence he speaks
14:01 dbolser http://twitter.com/#!/walshtp
14:01 dbolser I dare you!
14:01 takadonet flamewar--
14:01 dbolser but then I don't know from whence I speaks
14:02 dbolser takadonet: sorry, all in jest
14:02 takadonet dbolser: np
14:02 pyrimidine well, perl6, as it now stands, is slow (Rakudo), but it has also been gaining speed gradually
14:03 pyrimidine lots of devs seem to forget how slow Java used to be
14:03 dbolser he said its depreciated
14:03 dbolser then he said, its banned
14:03 pyrimidine I'm not sure I follow his reasoning
14:04 pyrimidine 'depreciated' means it's reduced in value
14:04 dbolser he didn't use anything so prosaic
14:04 dbolser I think he was joking
14:04 dbolser but the message was, forget 6
14:05 pyrimidine well, there are a lot of killer features that it has that other langs don't
14:05 pyrimidine so, when it does spread, I think it will see wider adoption
14:05 takadonet i noticed a huge speed up in the last year
14:06 pyrimidine takadonet: yes, and I think it will get faster with a few things on the horizon re: parrot optimizations
14:06 takadonet and once a major branch gets put into the trunk in the next few weeks..... it would be sweet
14:06 takadonet jnthn branch as we;;
14:06 takadonet well*
14:07 pyrimidine ah yes, the nqp-rx reworking for his metamodel stuff, IIRC
14:07 takadonet also should fix some of my workarounds that I have now
14:07 pyrimidine yep. One staller is pmichaud may be out for a bit
14:09 dbolser is metadata one word?
14:10 takadonet pyrimidine: niecza is also picking up speed
14:11 pyrimidine dbolser: guess it depends; metagenomics is written as one word
14:11 pyrimidine takadonet: yes, been watching that.  Any idea how many tests it's passing now?
14:11 pyrimidine *spectests
14:12 takadonet pyrimidine: no idea but he just started running the spectest a few weeks ago
14:12 takadonet some code right now run the same speed or faster then p5
14:12 pyrimidine I think his blocker right now is nqp fixes, but that's on the horizon
14:13 takadonet https://github.com/sorear/niec​za/blob/master/t/spectest.data
14:13 takadonet ya
14:14 pyrimidine dbolser: one of the problems I've been having with biome/bioperl6 is which one to focus my (ever-decreasing) energies on
14:14 dbolser pyrimidine: right
14:14 takadonet bioperl6
14:14 pyrimidine :)
14:14 dbolser I'm thinking that I should getinto fixing bioperl, but perhps I'll focus on biomoo
14:14 dbolser when will perl6 be released?
14:14 pyrimidine dbolser: good idea
14:15 takadonet I need someone to help make sense of the contradicting tests files!
14:15 pyrimidine dbolser: christmas
14:15 dbolser pyrimidine: which year?
14:15 dbolser ;-)
14:15 pyrimidine exactly
14:15 dbolser exactly why you should focus on biome
14:16 takadonet bioperl6 is the future man
14:16 pyrimidine well, right now I'm focusing on trying to get the next bioperl release out, then it'll probably be porting more to biome, with the side effect that optimizations would carry over to bioperl6 (and vice versa)
14:16 dbolser git push -u origin master # error: unknown switch `u'
14:17 dbolser pyrimidine: re: the richness of modeling in moose...
14:17 dbolser don't we need to sit down and decide how to model bp in moose?
14:17 dbolser seems there are prolly many ways to do it
14:17 pyrimidine by 'we', do you mean the current devs?
14:17 pyrimidine we talked about it at BOSC this last summer
14:18 takadonet pyrimidine: All I need from you is  *.t files  :)
14:18 dbolser the subset of bp devs who know moose...
14:18 pyrimidine takadonet: I can do that
14:18 dbolser pyrimidine: working document?
14:18 takadonet pyrimidine: I will do my best to deal with everything else
14:18 * dbolser wonders what -u is supposed to do on a decent version of git
14:18 pyrimidine dbolser: that subset is pretty small
14:18 dbolser heh
14:19 dbolser seems to crate a remote traking branch...
14:19 dbolser http://www.kernel.org/pub/softw​are/scm/git/docs/git-push.html
14:19 * dbolser wonders if it matters
14:20 pyrimidine I have it, git v 1.7.0.4
14:20 pyrimidine I haven't used it
14:20 dbolser centos seems to hate git
14:20 dbolser its used here: oh sorry, project specific git help page
14:21 pyrimidine It's probably set up as a convenient workaround
14:21 dbolser but not here: http://help.github.com/create-a-repo/
14:22 pyrimidine dbolser: they have to deal with multiple versions of git floating around out there, so they are picking the version that works for all of them
14:22 pyrimidine if you run 'git push
14:23 pyrimidine then git complains
14:24 pyrimidine should say, git initially complains
14:26 dbolser btw, https://github.com/twalsh/bio-6
14:26 dbolser same Tom "don't use perl 6" Walsh as I mentioned
14:28 takadonet heh
14:29 pyrimidine funny how that works :)
14:29 pyrimidine his gripe is probably the speed
14:29 pyrimidine that wouldn'
14:29 pyrimidine t surprise me
14:29 dbolser that is one of the things he mentioned
14:29 * pyrimidine has a fat pinky this morning
14:30 dbolser he said it'll be wonderfull in 10 years
14:30 dbolser well.. I should go back to $job
14:30 pyrimidine well, yes.  Java had the same curve
14:31 dbolser one final question,
14:31 pyrimidine the thing that most perl 5 users don't realize is, perl 6 is pretty much a different language
14:31 dbolser is MooseX::Descriptive > Java?
14:31 dbolser pyrimidine: I think that is part of what he didn't like ;-)
14:31 pyrimidine he should go back and read up exactly why perl 6 came about
14:31 pyrimidine there were very good reasons for it
14:32 pyrimidine just a few: lack of types, no method sigs, regex inconsistencies, etc
14:32 pyrimidine to fix them was essentially a backwards-incompatible fix, hence perl6
14:33 pyrimidine and, I have to say, when it does get to a point where it is fast, it will be very hard to compete with
14:34 pyrimidine ruby and python both have already started implementing code that mimics some aspects of perl 6 (not to mention perl5->Moose)
14:35 pyrimidine it's still very perl-ish (TIMTOWTDI), but has a very nice class system in place, as opposed to the hammered-on one in perl 5
14:36 pyrimidine which is very bare-bones (it essentially just associates, or 'blesses', a ref with a specific bag of behavior (the class).
14:37 pyrimidine that's okay, but Larry mentioned that he wanted it to be easy to do hard things, and hard to do impossible ones :)
14:39 pyrimidine and perl 5 essentially got in the way of that.
14:39 pyrimidine though, it's much better now, thanks to perl6 and Modern Perl
14:40 pyrimidine I think the two perls will end up co-existing for a while anyway
14:40 pyrimidine dbolser: couldn't find MooseX::Descriptive
14:41 pyrimidine unless you mean: http://search.cpan.org/dis​t/MooseX-MetaDescription/
14:57 mzgrideng joined #bioperl
14:57 mzgrideng left #bioperl
15:11 zenman joined #bioperl
15:21 zenman left #bioperl
15:33 dbolser pyrimidine: not sure, lets ask rbuels
15:34 dbolser MooseX::Declare
15:34 dbolser (sorry)
15:34 rbuels ask me what?
15:34 rbuels i have been unable to extract a question from the backscroll.
15:35 rbuels dbolser: ^^^^
15:36 dbolser hi rbuels
15:36 dbolser just the name of MooseX::Declare
15:36 dbolser and weather its > java
15:36 rbuels is moosex::declare better than java?
15:37 rbuels that's like asking if oranges are better than kiwis.
15:37 takadonet well they are
15:38 pyrimidine MooseX::Declare is actually even closer to perl6 than Moose alone
15:38 * rbuels shrugs
15:39 pyrimidine I actually stay away from MooseX::Declare, not b/c it isn't useful, but b/c there is a significant performance hit for using it
15:39 * rbuels doesn't use it either
15:40 pyrimidine we already suffer a performance hit with just vanilla bioperl, I want to reduce that not add to it :)
15:40 pyrimidine Moose, though, seems to help somewhat
15:42 pyrimidine rbuels: know of any ActiveResource (ruby) -like REST tools (client-end)?  The perl ActiveResource implementation is almost unusable (to the point I'm about to reimplement it in Moose)
15:42 pyrimidine :P
15:43 * rbuels doesn't know about activeresource, reads about it
15:44 pyrimidine it's very simple; I may actually just use the Ruby version
15:45 pyrimidine the perl version is too clever by half
15:46 pyrimidine though it is v 0.01 :)
15:52 pyrimidine rbuels: not sure if you have the link: http://api.rubyonrails.org/files​/activeresource/README_rdoc.html
15:56 rbuels so, basically serialize DBIC to and from XML
15:56 pyrimidine that's sort of what I'm thinking
15:56 rbuels or ... not XML, that's dumb.  JSON or YAML or something
15:57 rbuels probably json
15:57 pyrimidine actually, can be anything, but it is usually XML, JSON, or YAML
15:57 pyrimidine so, there would be a serialization component
15:58 pyrimidine *serialization/deserialization
16:00 rbuels i bet there is already something like this in cpan
16:00 * rbuels looks around
16:02 rbuels hmm, i guess it's app-framework-specific though
16:02 pyrimidine right
16:03 pyrimidine this would be DBIx::Class-like (have defined classes with attributes, etc, describing collections and elements)
16:03 rbuels it would be good if it was activeresource-compatible
16:03 pyrimidine there might be a way of doing this via DBIx::Class::Storage
16:04 pyrimidine if the storage was a RESTful service instead
16:04 rbuels pyrimidine: we should probably be discussing this in #catalyst
16:04 pyrimidine moving there now
16:05 mzgrideng joined #bioperl
16:05 mzgrideng left #bioperl
16:12 pyrimidine rbuels, this should do it: http://search.cpan.org/perldoc?Cat​alyst::Controller::DBIC::API::REST
16:12 pyrimidine thx
16:13 rbuels looks pretty nice
16:32 pyrimidine might be a bit overkill for what I want, but at least I don't have to rewrite anything
17:02 kyanardag_ joined #bioperl
17:11 dbolser pyrimidine: is it bio-me or bi-ome?
17:12 pyrimidine bi-ome
17:12 dbolser Tom has vetoed BioMoo because of this: http://search.cpan.org/~mstr​out/Moo-0.009007/lib/Moo.pm
17:12 pyrimidine right
17:12 pyrimidine Biome seems to fit, and it has biological context: http://en.wikipedia.org/wiki/Biome
17:13 dbolser what's on the front of the bioperl book?
17:13 dbolser I'm guessing it should be an archaea actually
17:13 pyrimidine what bioperl book? :)
17:14 dbolser since the discovery of the archaea is officially bioinformatics's greatest contribution to biology
17:14 dbolser the one we're writing
17:14 dbolser ;-)
17:14 dbolser you remember? I put you down for 3 chapters :-D
17:14 * dbolser is joking, but we should do it
17:14 pyrimidine my office is exactly one floor down from this man: http://mcb.illinois.edu/faculty/profile/1204
17:15 pyrimidine :)
17:15 dbolser can you link me to the thread wehre biomoose was rulled out?
17:15 pyrimidine so, there is a bit of archaea history here, one could say
17:15 dbolser uber!
17:16 dbolser context: http://www.kaggle.com/blog/2010/06/17/%EF%BB​%BFwhat-has-bioinformatics-ever-done-for-us/
17:17 dbolser pyrimidine: will you go punch wose in the face for me?
17:19 dbolser http://www.kaggle.com/blog/2010/06/17/​%EF%BB%BFwhat-has-bioinformatics-ever-​done-for-us/comment-page-1/#comments That Dan B seems to know what he's talking about!
17:19 pyrimidine :)
17:20 pyrimidine BTW, he's neck-deep in bioinformatics research himself
17:20 dbolser good
17:24 pyrimidine any reason why you have a violent reaction to dear Prof. Woese?
17:32 pyrimidine dbolser: ^^^
17:32 rbuels that dear man.
17:58 dbolser oh, I read somewhere that you should kill your idols
18:12 * pyrimidine just happened to be listening to a song called 'Bones of an Idol', is now perplexed
18:15 dbolser About 461,000 results
18:15 dbolser so, what is the standard way to encode key value pairs in GFF field 9
18:16 dbolser I'm guessing 'x=y;z=t' isn't recommended because x and z aren't keywords?
18:16 * dbolser chekc
18:16 dbolser s
18:16 pyrimidine dbolser: it's free-form if the key is lc
18:16 dbolser These tags have predefined meanings:
18:16 dbolser cool
18:16 pyrimidine upper-case keywords are reserved
18:16 * dbolser lc's his keys
18:17 pyrimidine http://www.sequenceontology.org/gff3.shtml
18:17 dbolser yup
18:18 pyrimidine 'All attributes that begin with an uppercase letter are reserved for later use.  Attributes that begin with a lowercase letter can be used freely by applications.'
18:18 dbolser tys
18:18 pyrimidine np, ttfn
18:18 dbolser l8r
18:18 * pyrimidine now feels 8 yrs old
18:19 dbolser wut?
18:19 dbolser ur !gr8
19:26 andrei_ joined #bioperl
19:26 andrei_ hello
19:27 andrei_ i have a question about genbank files
19:27 andrei_ I would like to modify the values of some tags from a genbank file
19:27 andrei_ what should I do after I loaded the file in a seqio object
19:28 andrei_ I can search for which tag_value has the value that I want
19:29 andrei_ but how can I change it and then save it in the genbank file?
19:30 andrei_ is there something like wright_tag_value?
19:30 andrei_ write_tag_value sorry
19:30 pyrimidine andrei_: have you read the relevant HOWTO documentation?
19:31 pyrimidine http://www.bioperl.org/wik​i/HOWTO:Feature-Annotation
19:31 pyrimidine basically, grab the feature of interest
19:31 andrei_ Yes I read HOWTO:Feature-Annotation
19:31 andrei_ but I still cannot figure it out
19:33 pyrimidine once you have the feature(s), use something like:
19:33 pyrimidine if ($sf->has_tag('foo')) { $sf->add_tag_value('foo', $new); }
19:33 andrei_ ok so it's add_tag_value
19:33 pyrimidine if you need to replace the tag value
19:33 pyrimidine you will need to remove the tag, then add back the new value to the same tag
19:33 andrei_ but what if I want to modify only half of that tag value?
19:34 pyrimidine store the tags in a temp array prior to removing them
19:34 pyrimidine not the best API, to be sure
19:34 pyrimidine but it works
19:35 pyrimidine do you want something like:
19:35 pyrimidine if ($sf->has_tag('foo')) { $sf->replace_tag_value('foo', 1, $new); }
19:35 pyrimidine ?
19:35 pyrimidine where '1' is the index?
19:35 andrei_ what I did is my $val = $feat_obj ->get_tag_values('gene');
19:35 andrei_ and then
19:36 pyrimidine you need:
19:36 pyrimidine my ($val) = $feat_obj ->get_tag_values('gene');
19:36 pyrimidine note the parens are $val
19:36 andrei_ $val =~ s/something/otherthing/;
19:37 pyrimidine you can have multiple values for a tag; last I recall get_tag_values always returned an array
19:38 pyrimidine so:
19:38 andrei_ yes I made a mistake
19:38 andrei_ my @val = $feat_obj ->get_tag_values('gene');
19:38 andrei_ this is what I meant
19:38 pyrimidine if ($sf->has_tag('foo')) { my $old = $sf->remove_tag('foo'
19:38 pyrimidine );
19:39 pyrimidine hold on...
19:39 andrei_ ok
19:41 pyrimidine https://gist.github.com/862829
19:41 andrei_ so basically I would have to get the tag value into an array, then modify that tag value; Remove the old tag and replace it with the modified one; Save everything to a new genbank file
19:41 pyrimidine something like that
19:42 pyrimidine if you are guaranteed to only have one value, you can use the idiom I have above
19:42 pyrimidine my ($val) = $feat_obj ->get_tag_values('gene');
19:43 andrei_ ok i'll give it a shot and see what results I get
19:43 pyrimidine ok
20:09 andrei_ i was able to do it by getting the tag value in an array, then removing the tag, modifying the array and add_tag_value the array
20:09 andrei_ this does exactly what i wanted
20:09 andrei_ but for some reason the replace_tag_value did not worked
20:10 andrei_ it kept giving me an error that it could not be found
20:12 andrei_ this is the small script:  git://gist.github.com/862897.git
20:14 andrei_ https://gist.github.com/862897
20:14 andrei_ sorry
20:54 andrei_ left #bioperl
20:54 cassj joined #bioperl
21:06 cassj left #bioperl
22:13 CIA-57 bioperl6: Takadonet master * r5604a74 / t/Tools/CodonTable.t : Changed test to check to see the array contents instead of just # of elems in the array - http://bit.ly/e3loH4
22:40 bag_ joined #bioperl
23:17 bag_ left #bioperl

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