Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-08-03

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

All times shown according to UTC.

Time Nick Message
02:18 CIA-93 bioperl-live: Chris Fields master * r6f12850 / t/Align/Graphics.t : [bug 2725] some updates to tests so they actually test accurately (they still pass!) - http://bit.ly/cJuQZF
02:18 CIA-93 bioperl-live: Chris Fields master * rb88c67c / Build.PL : move YAML to recommends - http://bit.ly/cgOSE2
02:24 deejoe joined #bioperl
04:28 CIA-93 bioperl-live: Chris Fields master * r0abbbde / Changes : updates - http://bit.ly/9zXK7C
04:28 CIA-93 bioperl-live: Chris Fields master * r6f662c6 / scripts/Bio-SeqFeature-Sto​re/bp_seqfeature_gff3.PLS : make sure the returned format is GFF3 - http://bit.ly/aStU1y
05:49 bag_ joined #bioperl
07:52 brandi joined #bioperl
07:52 brandi left #bioperl
10:07 brandi1 joined #bioperl
10:23 brandi1 left #bioperl
14:57 newtobioperl joined #bioperl
14:57 newtobioperl hello everyone.
14:57 newtobioperl I have a question: is there anyway that you can find a extron, intron of a sequence using bioperl????
15:00 newtobioperl for example: >GIQPZ7U01CI6YV-W1_10_21 TCCAGTGTACACCCTATTATACAGTACCTTCTCCCCTTTGATTGACAGGC​CTTGTAATCCCTGTGGGCCAGTCGGTCCTTCAGAACCTTTTTCACCCTGA​AAGTACAAGGGAGGGATATCAATGCACCAACCAAAATAAGAATTCCAGGA
15:01 kyanardag_ joined #bioperl
15:14 deafferret doh!
15:20 pyrimidine joined #bioperl
15:32 pyrimidine http://www.lyricsreg.com/lyrics/c​racker/Turn+On+Tune+In+Drop+Out/
15:32 pyrimidine :)
15:40 deafferret way ahead of ya   :)
15:41 pyrimidine catch the perl ref in the song?
15:41 pyrimidine just happened to be listening to it in the middle of coding perl... spooky
15:42 Jun joined #bioperl
15:43 deafferret most things that happen in my life happen "in the middle of coding perl"  :)
15:47 deafferret http://blip.fm/profile/deafferret/blip/51007383
15:48 pyrimidine :)
15:55 spekki0 Silly question but i using  Bio::Seq::SeqBuilder and only want GI numbers from the genbank file im parsing through. What do i add to add_wanted_slot('???') for gi numbers?
15:56 pyrimidine any reason not to use grep?
15:56 deafferret spekki0: SeqBuilder might be "all features" or none
15:57 spekki0 hmm i guess not lol
15:57 deafferret db_xref gi  isn't special I don't think
15:57 deafferret (Genbank)
15:58 pyrimidine spekki0: I'm wondering why not use grep/ack in this case, or a simple perl script.
15:58 pyrimidine bioperl seems like overkill for something simple like this
15:58 pyrimidine (and I'm saying that as a core dev :)
15:59 spekki0 well it's just a small part of a bigger bioperl program i have going. so i was just going with the rest of it
15:59 pyrimidine ok
16:00 spekki0 if it was by itself i think your right that grep would be quicker /easier
16:00 pyrimidine I think, if you want the GI, I added a method for that in Bio::Seq::RichSeq: ncbi_gi()
16:02 pyrimidine let me double-check...
16:02 spekki0 kk, tx
16:04 pyrimidine sequence GI ID is primary_id(), sequence accession is accession() (strangely enough :)
16:04 pyrimidine ncbi_gi() was in Bio::Search, so my bad
16:04 pyrimidine so yuo should get the GI from $seq->primary_id
16:04 pyrimidine *you
16:05 spekki0 ok cool
16:05 spekki0 lemme go try it out
16:05 * pyrimidine 's fingers are crossed, hoping NCBI didn't bork this
16:06 spekki0 cool it worked
16:06 spekki0 thank pyrimidine
16:07 pyrimidine np
16:07 spekki0 thanks*
16:12 deafferret so with SeqBuilder you *can* get the GI out 10x faster than parsing all features?  http://www.bioperl.org/wiki/HOWTO:SeqIO#Speed.2C_Bio::Seq::SeqBuilder
16:12 deafferret primary_id()  ?
16:12 pyrimidine You can leave out the features, I'm guessing, which would significantly speed things up
16:13 pyrimidine still think this design is backwards
16:14 pyrimidine should be passing data into handlers that do the work
16:14 pyrimidine ah, if I only had time...
16:17 rbuels pyrimidine: this is an instance of paying later for absence of code review up front
16:17 deafferret code review? haha
16:17 pyrimidine well, the problem is a bit deeper than that
16:18 * rbuels grasp of the problem is admittedly shallow
16:18 pyrimidine the parsers were designed to simply push data into objects for processing, w/o forethought on downstream implications
16:19 pyrimidine such as the problem spekki0 had above
16:19 deafferret rbuels: you meant rbuels'   . Your IRC client needs improved emote logic
16:19 * pyrimidine 's sad for rbuels :(
16:20 rbuels lol
16:20 pyrimidine I have to add a space in mine
16:20 pyrimidine otherwise I get 'unknown command'
16:20 rbuels pyrimidine: so i guess this argues for keeping a layer of separation between the parser and the data model
16:20 rbuels pyrimidine: is that what you're saying?
16:20 pyrimidine yes
16:21 * rbuels nods
16:21 pyrimidine the parser should be a generic middle layer
16:21 rbuels probably shouldn't even have anything to do with bioperl, really
16:21 pyrimidine exactly
16:21 rbuels (where 'bioperl' is mostly data model)
16:22 pyrimidine There should be a generic set of seq format parsers
16:22 pyrimidine the problem is, how would one standardize this?
16:23 rbuels not sure you can.
16:23 rbuels you might not really need to.
16:23 rbuels the big difference is going to be, some will be pull, some will be push
16:24 pyrimidine but, if the parser is abstracted out, that becomes moot
16:24 rbuels well...yeah i see what you're saying though.  if each parser has an idiosyncratic api
16:24 rbuels it's a pain.
16:24 pyrimidine Well, it may not be too bad
16:25 pyrimidine there are commonalities with most seq formats
16:25 pyrimidine so, maybe look at how some of the most popular formats cluster data
16:26 pyrimidine seq, features, annotation => sequence record
16:26 pyrimidine etc etc
16:26 rbuels yes, maybe
16:27 rbuels it's delicate, because the more uniform structure you impose on the api of each parser, the closer you are getting to just defining another data model
16:27 pyrimidine right
16:28 pyrimidine but we definitely don't want a SAX-style event parser
16:28 rbuels well it's a good thing this computer code stuff is trivial and unimportant anyway
16:28 pyrimidine :)
16:29 genehack even if you have the parser abstracted out, you're going to have to settle on what the internal/"middle" representation of the data is, aren't you?
16:29 rbuels genehack: yeah, if you want standardization
16:29 * genehack parachutes into the middle of the discussion...
16:30 pyrimidine If you look at the design of the Seq stuff in bp, it's pretty much hammered together based on the formats
16:31 * rbuels wanders off to fix bugs, SGN is almost migrated to catalyst
16:31 pyrimidine ok
16:32 genehack ok -- i haven't gotten into the guts of that stuff yet.
16:32 pyrimidine that's why there is a RichSeq
16:32 pyrimidine (should just be merged into Bio::Seq, and rename Bio::Seq as Bio::SeqRecord or somesuch)
16:32 pyrimidine it's a bit hairy
16:33 genehack I think it says that on the outside of the tin. 8^)
16:34 pyrimidine ;-)
17:00 flu rbuels, speaking of catalyst...  Do you have any experience instantiating models from a conf file?
17:01 flu I can do it but only when a model is a top level config option.  Nesting it in a config section breaks it.
17:02 rbuels flu: which model?  Catalyst::Model::DBIC?
17:03 flu No, it is a custom model that uses Catalyst::Model::Adaptor to instantiate a non catalyst moose class
17:04 flu I'll do some more digging and then head on over to #catalyst if I can't find anything.
17:05 rbuels flu: i've never used it, no real insight here.
17:05 rbuels flu: #catalyst is your best bet
17:20 CIA-93 bioperl-live: Chris Fields master * r10dcb4d / (Bio/Tools/Est2Genome.pm Changes): [bug 2982] score<->percID mixup, patch courtesy Alexei Papanicolaou - http://bit.ly/dflS43
17:20 CIA-93 bioperl-live: Chris Fields master * rc4efe6a / Bio/LocatableSeq.pm : strand should be defined in a comparison - http://bit.ly/cBe8VX
18:44 bag_ joined #bioperl
19:16 brandi joined #bioperl
19:17 brandi left #bioperl
19:51 pyrimidine Ah, close one bug, open another...
19:55 CIA-93 bioperl-live: Chris Fields master * r563e929 / (4 files in 4 dirs): [bug 2984] end was calculated based on arbitrary symbols; punt and let LocatableSeq do this, for consistency - http://bit.ly/cdA3bD
20:22 spekki0 im running into a bit of an odd problem, here is what im attempting to do, I have a genbank file with a couple thousand records in it and I want to go through each of the records and test if it has been replaced in genbank via getting the summary and checking the status field. Here is the code im using http://codepad.org/5CDIzUg5 I was having issues with posting the whole set of records initially so i post in chunks of 20. But here is the p
20:22 spekki0 roblem some posts and retreivals of summarys spit out this error, http://codepad.org/nNazlWpb and I have a feeling when this happens no summary info is returned for the current set of 20 records that get posted. Reason I have this feeling is that the code basically splits the records from the original genbank file into 2 other files, records that are suppressed, and everything else. And I find that the addition of the 2 record counts from t
20:22 spekki0 wo files doesnt add up to the original amount of records from the genbank file. So im jsut wondering am i diong something wrong using epost/esummary or is it something else? Sry for the very long post :(
20:30 pyrimidine spekki0: are you using UIDs (not accessions)?
20:31 spekki0 yep, uid's. GI numbers specifically
20:37 pyrimidine not seeing anything out of the ordinary.  You don't really need '-keep_histories' (that was a queuing thing, not sure if that will stay in the API long-term)
20:37 pyrimidine spekki0: have a few example sequences to test this on?
20:38 pyrimidine can email them
20:38 spekki0 k lemme find some of the ones that make it break
20:38 pyrimidine brb
21:16 spekki0 this is dumb i removed that -keep_histories, and i was trying to get a set of records that kept causing the issue but now i can't get it to come back...
21:25 pyrimidine could be a server side issue
21:25 pyrimidine email me some seqs to play with. I can try working it out.
21:26 spekki0 kk
21:28 spekki0 k its sent
21:29 spekki0 the only change i made was to remove -keep_histories
21:29 spekki0 after that change i couldnt recreate the error, so i suggest trying it with -keep_histories in...
21:42 pyrimidine spekki0: caught up with a few things, but will look at this tonight.  Have to go
21:43 pyrimidine left #bioperl
21:43 spekki0 no worries tx for looking at it
22:36 CIA-93 bioperl-live: Chris Fields master * r503c736 / Bio/Tools/EUtilities/EUtilParameters.pm : partial revert, some small code changes snuck in - http://bit.ly/910FTo
23:15 nuba joined #bioperl

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