Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-05-26

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

All times shown according to UTC.

Time Nick Message
06:04 bag_ joined #bioperl
09:20 vinnana joined #bioperl
12:22 slackaholic joined #bioperl
12:28 slackaholic joined #bioperl
12:31 slackaholic joined #bioperl
13:40 deafferret .....,...g..   .
14:03 ffoertter joined #bioperl
14:07 ffoertter_ joined #bioperl
14:07 ffoertter_ joined #bioperl
15:44 spekki01 morning
15:49 * deafferret waves
16:21 spekki01 deafferret: have any suggestions for me to figure out why my hmmsearch just keeps grinding away as if its not doing anything heres my top output with it running since yesterday http://codepad.org/k70HebP0
16:21 spekki01 im tempted to try this on some different hardeware, think that might make a difference
16:28 rbuels spekki01: it probably would not make a difference.  i don't know anything about hmmer in particular, but based on my experience with other tools i would bet that either 1.) this is just how long it takes, or 2.) you should have run it with different options
16:29 spekki01 fair enough, I'll look through the manual again and see what i can tweak.
16:34 kai hm, I think I just found a bug in the SeqIO embl generator, let's see if I can reproduce
16:34 kai spekki01: is that hmmer3 or hmmer2?
16:38 spekki01 sry was afk its hmmer3
16:39 kai hm, hmmer3 is pretty fast for me, are you running it standalone or from bioperl?
16:39 spekki01 standalone, how do you do it from bioperl?
16:40 spekki01 i do "hmmsearch profile.hmm mydb.fasta > profile.out
16:42 spekki01 does it load each sequence one at a time or try to put it all in memory?
16:45 kai hm, no idea, I've only been using hmmscan myself
16:45 kai however..
16:46 spekki01 brb gotta talk to my prof
16:46 * kai thumbs through the hmmer userguide
16:49 kai ah, hang on, I confused this with the steps necessary to build databases for  hmmscan, which you should hmmpress to make search faster
16:49 kai but for hmmsearch that makes no sense
16:58 spekki01 hmmscan is where you have multiple profiles and you make them into a database press them and then search a sequence against that database right?
16:59 spekki01 well atm im going to rule out if its a problem with the database file, yesterday i took the first 1000 sequences from the original db and made a small version, which i ran hmmsearch against and it worked out fine
17:00 spekki01 i got this smaller db from using bio::seqio and reading in each sequence and then writing it back out in the same fasta format
17:01 spekki01 so im gonna do this on the whole big db, read in each sequence and write it out to a new db file
17:01 rbuels (i would think hmmsearch would be mature enough that malformed input would not cause it to go nuts)
17:01 spekki01 i thought so to but im just trying to rule out anything at this point :(
17:02 rbuels yeah.
17:03 spekki01 kai: in your uses of hmmsearch how big or a database file have you searched through ?
17:03 spekki01 or = of*
17:27 spekki01 joined #bioperl
17:57 deafferret spekki01: did you try emailing eddy about some sort of verbose output so you'd see *something* ?
17:58 deafferret spekki01: http://selab.janelia.org/people/eddys/blog/?p=289
17:59 spekki01 no but thats a really good idea, ill write to him and see what he says
18:21 kblin spekki01: haven't used hmmsearch yet
18:22 kblin which reminds me, I need to go chase that embl writer bug
18:28 dave_messina joined #bioperl
18:28 kblin whoa
18:28 kblin hmmscan sure is fast compared to hmmpfam
18:28 bag_ joined #bioperl
18:29 bag__ joined #bioperl
18:50 dave_messina Rob Buels, you there?
18:52 deafferret rbuels: ^
18:52 rbuels eh wha
18:52 rbuels oh
18:52 deafferret rbuels: ^!!!
18:52 rbuels dave_messina: um, yes
18:52 deafferret dave_messina: he's lying
18:52 dave_messina Ah goody. :)
18:52 rbuels well, kind of
18:52 rbuels lol
18:53 dave_messina Rob, are there any summer of code projects that didn't get funded and you might be looking for someone to work on?
18:53 rbuels dave_messina: when you type my nick it flashes my irc client
18:53 perl_splut joined #bioperl
18:53 rbuels dave_messina: try typing rb<tab>
18:53 dave_messina Ah okay, rbuels.
18:53 dave_messina rbuels
18:53 dave_messina :)
18:53 rbuels dave_messina: umm, yes there are ......
18:53 deafferret rbuels: anywhere or just as the start of a message?
18:53 * rbuels looks
18:53 rbuels deafferret: anywhere in the message
18:53 * deafferret lools (rbuels)
18:54 rbuels um yes
18:54 dave_messina This guy Kanna posted to the list today and is looking for something to do.
18:54 deafferret rblools
18:54 dave_messina So I thought you might be able to suggest a project for him.
18:54 rbuels dave_messina: here's a list: http://bioperl.org/wiki/Google_Summer_of_Code
18:54 rbuels dave_messina: the only one that got funded was the align subsystem
18:55 dave_messina rbuels: thanks. I'll put that in my response to his post.
18:55 rbuels dave_messina: but, not all of these are necessarily suitable for new contributors ......
18:55 dave_messina Aha...okay.
18:55 rbuels dave_messina: without mentoring, anywa
18:55 dave_messina right. Wanna mentor then? :)
18:55 dave_messina kidding.
18:55 rbuels dave_messina: they are kind of designed to have an active person mentoring them ......
18:55 rbuels dave_messina: and to have the person do them full-time over a summer
18:55 * deafferret pushes everyone in the deep end
18:56 dave_messina hmm, okay, then maybe GSOC projects aren' the best place for him to start.
18:56 * rbuels looks at Kanna's email
18:56 rbuels dave_messina: sounds like his itch is in ontology handling ...
18:56 dave_messina I'm pretty sure the OBO thing is done — it says so right on the Project priority list.
18:56 dave_messina so I don't know why he suggested that one.
18:56 dave_messina But I don't know how up-to-date with reality that is.
18:56 rbuels dave_messina: so 1.) see if there are any bugs in bugzilla related to ontologies ...
18:57 dave_messina only two bugs.
18:57 rbuels dave_messina: well that's a start, ask him to work on those
18:57 dave_messina one of which he could maybe handle.
18:57 dave_messina Yep, sounds good.
18:57 rbuels that would at least see if he's savvy
18:57 rbuels probably just give him one
18:57 rbuels the easiest one
18:58 dave_messina Right, seems like it'd be better to start with something easy and see how it goes.
18:58 rbuels and ask him to have a look at it and submit a patch, or something
18:58 dave_messina OK, will do.
18:58 rbuels yes, you never know how savvy or not-savvy somebody is right off
18:58 rbuels best to start them on something easy
18:59 * dave_messina agreed
18:59 rbuels ok, meeting time for me ...
18:59 * rbuels joins a teleconference
18:59 rbuels ttyl
18:59 rbuels dave_messina: thanks for doing this.
18:59 dave_messina thank YOU rbuels. :)
18:59 dave_messina cya
19:06 deafferret "submit a patch"?  ew.  better to "fork on github.com, commit your changes, and send us a pull request"
19:07 dave_messina Yeah, yeah, you youngsters, with yer mp3s and yer GPS and yer fork on github.
19:07 dave_messina Back in my day, we'd buy albums, get lost, and submit patches.
19:08 deafferret welcome ... to the FUTURE!!  :D
19:09 dave_messina Seriously, though, deafferret, would you respond to my post on-list with a few Git pointers? I agree that we should promote that wherever possible.
19:09 deafferret on-list? email? ew!
19:09 deafferret ;)
19:10 dave_messina heh :)
19:10 deafferret email makes me feel so swamped, all the time.  IRC lets me live IN THE NOW
19:10 deafferret and wander off when things get hard or worky
19:10 dave_messina That's right. IRC=zen
19:10 deafferret I tend to think I suck at git so I don't know that I should be dolling out "best practice" advice
19:11 deafferret every time I try to get fancy I end up frustrated and confused for long blocks of time
19:11 deafferret so the best advice I've seen so far are the people who have posted extremely quick summaries of work flow
19:11 dave_messina I hear ya. I don't know what I'm doing with Git either.
19:12 deafferret (ignore the million bells and whistles and go like this: A, B, C, D, E
19:12 dave_messina But hell, man, I ain't askin' for a just throw up a link to "Using Git"
19:12 dave_messina for a treatise*
19:12 * deafferret clocks out, looks for dave's email
19:12 dave_messina Yep, recipes are key for me when trying to learn something.
19:13 dave_messina later.
19:13 deafferret dave_messina: which post are you asking me to respond to? What's the subject/date?
19:14 dave_messina the one just now. wasn't that what your "submit a patch" ew related to?
19:14 deafferret ah, yessir. will do
19:14 dave_messina cool, thanks.
19:15 deafferret whenever my copy arrives... I seem to be on a 30m email delay from that Mailman. It's hated me for years.
19:15 * deafferret coos to his IRC mistress
19:16 dave_messina Ah, so THAT's why you prefer IRC. You feel like a leper or something, ignored by the list. :)
19:17 deafferret ya, that's part of it. instant gratification addict I suppose
19:17 deafferret emails feel like "Hi, here's 3 months of work. What do you think?"
19:17 dave_messina yeah, now that I'm a big-britches core dev, I should ask for mailman privs and see if I can get that running a bit more smoothly.
19:18 dave_messina Hehe, 3 months? man, I thought I hated writing long emails.
19:18 deafferret dave_messina: ya, too bad rbuels diluted your cool.  :/
19:18 dave_messina :)
19:18 deafferret "core dev" used to be so respectable
19:19 dave_messina heck, he deserves it for managing GSOC alone. That's a crapload of I don't want.
19:19 deafferret ya  :)
19:19 deafferret he's good people. but don't tell him I said so.
19:19 dave_messina definitely not.
19:19 dave_messina notice how I waited for him to leave to say something nice. :)
19:19 deafferret :)   /me clocks back in, keeps one eye on the dreaded Inbox
19:19 dave_messina cya
19:20 deafferret :)
19:44 dnewkirk joined #bioperl
19:51 kyanardag_ joined #bioperl
19:52 CIA-95 bioperl-live: Chris Fields branch-1-6 * ra850b2b / Bio/DB/Taxonomy/flatfile.pm : return unless a taxid is given - http://bit.ly/dnXnMY
19:53 CIA-95 bioperl-live: Chris Fields master * r95025c0 / Bio/DB/Taxonomy/flatfile.pm : return unless taxid is given - http://bit.ly/b6GzDW
20:07 * kblin kills some time while waiting for his bugzilla account email to pass through the graylisting
20:08 deafferret email-- # pergatory
20:08 deafferret purgatory?
20:09 deafferret OOo! I get to use another zero-width negative look-behind assertion!
20:10 kblin hm, actually I should go find out if this is an error in the parser or the writer
20:13 spekki01 omfg hmmer2 is already in this stupid system, this whole time i've thought ive been using my install of hmmer3 no wonder its taking so long /cry
20:34 kblin ouch
20:35 spekki01 :(
20:35 rbuels oook.
20:35 rbuels which hmmsearch
20:35 * rbuels grimaces
20:35 deafferret /wrong/one/hmmsearch
20:36 deafferret damn, can't use this zero-width negative look-behind assertion
20:36 spekki01 i was searching with hmmsearch version 2 which is really slow apparently
20:36 deafferret variable width not implemented   :p
20:38 kblin ah, so this is an error in the EMBL parser. interesting
20:44 CIA-95 bioperl-live: Chris Fields branch-1-6 * r5f45dee / Bio/DB/Taxonomy/flatfile.pm :
20:44 CIA-95 bioperl-live: Revert "return unless a taxid is given"
20:44 CIA-95 bioperl-live: This reverts commit a850b2b4f79a7d40b66df5a10355b6cbb09f3da5. - http://bit.ly/bPWinH
22:22 dave_messina joined #bioperl
22:56 kyanardag_ joined #bioperl
23:33 ffoertter joined #bioperl

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