Camelia, the Perl 6 bug

IRC log for #bioperl, 2009-09-25

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

All times shown according to UTC.

Time Nick Message
09:32 ptl ...evolution is a fraud, I'll sequence every genome up to its last telomeres and prove that!!!
12:10 deafferret I'm listening. How will that prove your hypothesis?  :)
13:04 ptl That's simple, we'll run an information detector on it and find God^WThe Designer's signature. Hence Intelligent Design is true and Evolution is false.
13:39 deafferret looking forward to your results!
14:34 kyanardag joined #bioperl
14:45 ptl joined #bioperl
17:18 ptl_ joined #bioperl
17:53 denisem joined #bioperl
17:53 denisem Morning all
17:54 rbuels denisem: morning
17:54 rbuels you must be on the west coast
17:55 rbuels ah yes, seattle
17:55 * rbuels is in portland
17:55 denisem Ineed :)
17:56 denisem I have a question, but I'm trying to frame it in a manner that's understandable by someone else ;)
17:56 denisem so it'll probably take me a few mins
18:04 ptl_ or maybe your question is just incognoscible (I love that world... hope it really exists)
18:11 denisem that could be true
18:19 denisem ok
18:20 denisem so the error that I'm having is that I have a Bio::SeqFeature::Generic that has a 'peptide_shift' entry in the object (I dunno if it's in the right spot though) that then the command $self->feature->get_tag_values('peptide_shift'); doesn't work to get the peptide shift value
18:24 rbuels denisem: make sure you're calling get_tag_values in list context
18:24 rbuels if you're sure there's just one you could do my ($pep_shift) = $feature->get_tag_values('peptide_shift')
18:25 denisem hmm
18:25 denisem ok, will try that
18:25 rbuels cause it returns a *list* of scalars
18:25 denisem actually
18:25 denisem here's the code that actually gets it:
18:25 denisem my @alleles = $self->feature->get_tag_values('peptide_shift');   @alleles    = split /\//,$alleles[0] unless @alleles >= 2;   my $size = 2 + 10 * ($#alleles +1);
18:26 denisem The error that it gives: MSG: asking for tag value that does not exist peptide_shift
18:26 denisem Here's the object if I print it out :  http://gist.github.com/193750
18:26 rbuels oh.  that's different
18:27 denisem sorry :x
18:27 * rbuels rubs deafferret's chin
18:27 rbuels hmmmm
18:27 denisem so I'm using Bio::Graphics::Panel to use Bio::Graphics::Glyph::gm_allele_tower
18:28 denisem could it be a bioperl install error or something?
18:28 denisem because the very confounding thing is that this works on our dev box... but not on our production server...
18:28 denisem -_-
18:30 rbuels are they different versions of bioperl?
18:30 rbuels or different versions of whatever is producing these feature objects?
18:30 rbuels does the dump of the object look different on the dev box?
18:31 denisem just making that dump now
18:31 rbuels probably does.
18:31 denisem nod
18:31 denisem the bioperl is the same
18:31 denisem but the production box is running perl 5.8.8 and the dev box has perl 5.8.5
18:32 denisem cuz our IT people updated our production server -_-
18:32 rbuels that won't make a diff..
18:32 denisem k
18:32 rbuels the proximate problem is that the stuff you're looking for is in a subfeature of your obj
18:32 rbuels but if this same code works on the dev box.....its feature must be structured differently
18:33 denisem nod
18:33 rbuels somebody has done $feature->add_SeqFeature( ) to it
18:33 rbuels when they probably wanted to do the add_SeqFeature on something else
18:33 rbuels would be my guess
18:34 rbuels my @alleles = map $_->get_tag_values('peptide_shift'​),$self->feature->get_SeqFeatures;   @alleles    = split /\//,$alleles[0] unless @alleles >= 2;   my $size = 2 + 10 * ($#alleles +1);
18:34 rbuels would get our peptide shifts in this case
18:34 rbuels but that woudn't work on the dev box, if its feature is structured differents
18:35 rbuels if this is coming out of some pipeline, probably the pipeline code is a different version or something on the dev box
18:36 * rbuels does not really know enough about your devel setup to speculate too much
18:36 denisem hrmm... so my dev box doesn't even try to print out that particular feature
18:36 denisem it feels like this is a feature that should be skipped and my dev box does, but my production box doesn't
18:36 rbuels well the code is certainly different somewhere.
18:37 rbuels you can diff the two bioperl installations .....
18:37 rbuels and the two application code installations ..
18:37 denisem how can I do that across two machines?
18:37 denisem (sorry, newb question)
18:37 rbuels yes, that's the trick, lol
18:38 rbuels find your_bioperl_dir -name '*.pm' -exec cat {} ';' > /tmp/bioperl_1.txt
18:38 rbuels on other machine: find your_bioperl_dir -name '*.pm' -exec cat {} ';' > /tmp/bioperl_2.txt
18:39 denisem but yes, the production machine seems to be creating those subFeature things and the dev box just has Bio::SeqFeature::Generic entries with no subfeatures
18:39 rbuels scp /tmp/bioperl_1.txt other_machine:/tmp
18:39 rbuels diff /tmp/bioperl_1.txt /tmp/bioperl_2.txt
18:39 rbuels where are these features coming from?
18:39 rbuels a FeatureIO object?
18:39 denisem sec, gotta trace that
18:42 denisem http://gist.github.com/193756
18:42 denisem $panel is a Bio::Graphics::Panel object
18:43 denisem collection_start and collection_stop are just chr locations
18:43 rbuels that's where these features are going *to*, not coming from, right?
18:43 rbuels or are you getting them *out* of the panel object, which is coming from somewhere?
18:44 denisem that's where they're going to I guess
18:44 denisem oi
18:45 rbuels i've gotta go to lunch.....
18:45 denisem np :)
18:45 denisem I'll try to figure out where the info is coming
18:46 denisem it's coming from a database
18:46 rbuels back in a few ......
18:46 denisem but I dunno how it's getting in
18:46 denisem ;)
18:53 denisem hrmm... I tried doing that diff and for some reason it listed the modules in a different order (ie: my dev box list has Bio::Biblio first and my production machine has Bio::Align::AlignI first) even though they have the same basic directory list (both have Align and AlignIO as the first directories, etc)
19:18 rbuels find your_bioperl_dir -name '*.pm' | sort | xargs cat > /tmp/bioperl_1.txt
19:19 rbuels denisem: but it could very well be differences in how the prod and dev dbs were loaded, if you're getting these from dbs
19:19 rbuels (and if you have separate prod and dev dbs)
20:23 deafferret wow
20:57 denisem joined #bioperl
21:12 rbuels ;-)
21:17 rbuels deafferret: have you seen this yet?  http://www.youtube.com/watch?v=zSgiXGELjbc
21:18 rbuels autotuning 'Cosmos'
21:18 rbuels carl sagan++
22:33 deafferret rbuels: will check at home  :)
22:35 rbuels deafferret: have a good weekend
22:38 deafferret still wurkn
22:38 deafferret in another dudes office so cat youtube

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