Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-06-01

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

All times shown according to UTC.

Time Nick Message
04:55 CIA-95 joined #bioperl
06:16 bag_ joined #bioperl
07:27 dnewkirk joined #bioperl
11:31 gabytex joined #bioperl
11:48 brandi joined #bioperl
11:48 brandi left #bioperl
13:24 driley joined #bioperl
16:04 brandi joined #bioperl
16:04 brandi left #bioperl
16:05 spekki01 when using the bioperl hmmer3 parser here is the code that I use to basically go through all my hits and grab some of there info: http://codepad.org/P42y2ZB1 Now for some reason it seems to not go through all the results returned it stops when it hits this line: http://codepad.org/9o5Z8G7w Basically there are 3492 targets reported over the threshold but then there is this other inclusion threshold which is only 2436 records. How do I have
16:05 spekki01 bioperl parse the remaining records?
16:35 brandi joined #bioperl
16:35 brandi left #bioperl
17:23 deafferret spekki01: Line 1: error: expected identifier or '(' before numeric constant
17:23 deafferret ?
17:23 deafferret is that actually happening, or a codepad.org artifact?
17:24 spekki01 codepad.org problem
17:24 spekki01 it never runs my code properly, i forget to uncheck that box at the bottom too
17:24 deafferret spekki01: so do you have a runnable case so I can reproduce?  (on github.com or somewhere?)
17:25 spekki01 but i figured it out i guess, i just went and removed that -----inclusion threshold----- line and it went through all my records not just the ones above the line
17:25 spekki01 um i could set it up so you can test it if you like
17:25 deafferret ???  '-----inclusion threshold-----'  was literally in the file?
17:26 spekki01 yes, it's even explained in the hmmer3 documentation as to what it is
17:26 * deafferret laughs nervously
17:26 spekki01 but it's kinda weird cause it kills the file parsing
17:26 spekki01 i couldnt find anyway to keep going past it other than removing the line
17:26 deafferret this is why runnable cases are so important -- I never would have guessed in a million years that that was actually sitting in the file  :)
17:27 deafferret spekki01: ya, so are you using trunk code or new stuff ?
17:27 spekki01 errr sry what does trunk code mean?
17:27 deafferret sorry - branch 'master' from github.com/bioperl/bioperl-live
17:27 * deafferret beats himself with the "not in SVN any more" stick
17:28 deafferret several people (I thought it was you?) were working on hmmer3 over the last 30 days. dunno where all that code is
17:28 spekki01 well for this i used that stand alone thing, well kinda standalone, http://github.com/kblin/bioperl-hmmer3
17:29 spekki01 i dl'ed it 2 yesterday
17:29 spekki01 ops scratch the 2
17:29 deafferret ah, ok. So kblin's new stuff needs to be enhanced to gracefully skip over the '-----inclusion threshold-----' line
17:29 deafferret so you can branch his branch, fix it, and send him a pull request
17:30 deafferret err... fork his repo, fix it, and send him a pull request
17:30 spekki01 yeah, i kinda brought it up yesterday, when i was talking to him, but i dont think i explained it very well the codepad posts would have made it much more clear
17:30 deafferret or have him add you as a co-maintainer, then you can commit into his repo directly
17:30 spekki01 well, I can try that, i've not done much on github so it'll take a bit of reading and figuring it out but id be glad to help with it
17:31 deafferret spekki01: the SUPER AWESOME thing to do is to write inclusion_threshold.t which fails when not handled correctly
17:31 deafferret then also fix it, and the .t starts passing  :)
17:31 deafferret TDD++
17:32 deafferret kblin: PING! can haz commit bits?
17:32 spekki01 well I'll try that lol, i've never done that kinda stuff but it cant be to hard once i get my head wrapped around it
17:33 deafferret programming is easy!  :D
17:33 spekki01 lol i know its just always a little daunting when its something new
17:33 deafferret :)
17:34 deafferret kblin: spekki01 and I want to add bugs to http://github.com/kblin/bioperl-hmmer3 directly  :)
17:34 deafferret kblin: seems a touch silly to fork a fork of bioperl-live?  :)
17:35 spekki01 i really have to read up on github and what all the terminology means for it lol
17:35 deafferret spekki01: ya, there's a lot to learn  :)   Facebook for super nerds
17:35 spekki01 lol
17:36 brandi joined #bioperl
17:36 spekki01 can i have my own profile page ;P
17:36 deafferret spekki01: on github.com ?
17:36 brandi left #bioperl
17:36 spekki01 oh lol do they actually have those, i was making a joke lol
17:37 deafferret http://github.com/jhannah
17:37 deafferret http://github.com/kblin
17:37 spekki01 lol thats so cool
17:37 deafferret that's about it for "profile page" I guess
17:38 deafferret if you don't have an account yet, create one  :)
17:38 spekki01 kk, brb
17:55 spekki01 ok well i made a github and gravatar accounts and am good to go, but now im hungry so im going to eat lunch :)
17:57 deafferret :)
18:27 brandi joined #bioperl
18:27 brandi left #bioperl
18:42 spekki01 ok so how do i get started on helping out with enhancing the bioperl-hmmer3 parsing?
18:42 spekki01 kblin: can i lend a hand in fixing it to work around that inclusion threshold stuff?
18:58 spekki01 this may seem dumb, but when doing the hmmsearch result parsing what does Hsp stand for? I keep seing it alonside Hit all the time.
19:01 deafferret High Scoring Pair
19:02 deafferret Query | Hit | HSP
19:02 spekki01 ah ok
19:05 deafferret spekki01: perhaps not idea, but you can always fork his repo and start enhancing it, and he can pull your changes into his
19:05 deafferret perhaps not ideal
19:06 spekki01 works for me, that'll give me a chance to get understanding the code better anyway
19:08 deafferret ya, it's definately a great system... I only hesitate because you're forking a fork  :)
19:08 spekki01 what did he fork off of?
19:08 deafferret theoretically it shouldn't matter, but I've never done that in reality so don't know if that gets messy
19:08 deafferret his repo is a fork off of github.com/bioperl
19:08 deafferret it says so right below the name of his repo
19:09 spekki01 gooing to look
19:10 spekki01 ah ok so theres the bioperl-hmmer3 then kblins
19:10 spekki01 I'll fork off of him anyway and give it a go, whats the worst that could happen lol
19:11 deafferret ya, the worst case is that someone somewhere has to apply changes manually, which should be no big deal
19:11 spekki01 questions, who head the main bioperl-hmmer3 module?
19:11 deafferret in the best case yours and his history merge seamlessly back into the bioperl-live repo someday soonish  :)
19:12 deafferret um, looking
19:12 deafferret "sharpton"
19:13 spekki01 kk
19:13 deafferret if you search the bioperl-l archive you should find him talking about his hmmer3 code in March
19:13 spekki01 yeah, i saw that yesterday when i was reading up on hmmer3 parsing
19:14 spekki01 well im gonna figure out my ssh publik key thing for github
19:14 deafferret interesting.... sharpton, kblin, and cjfields have all been in here   http://github.com/bioperl/bi​operl-hmmer3/commits/master/
19:14 deafferret spekki01: ya, you'll need keys no matter where your code ends up  :)
19:15 deafferret and they're github-wide, so the one-time setup is pretty painless for all the benefit you get forever across all of github.com
19:51 dave_messina joined #bioperl
19:54 spekki01 hmmm ok this seems strange, im trying to fork kblin/bioperl-hmmer3 and i get this message saying this tree seems to be invalid :(
19:55 deafferret spekki01: wut? really? lemme try
19:55 deafferret spekki01: whoah! cool  :)
19:55 spekki01 yeah im not sure, im doing it on the website, all i did was go to his page and click the fork button
19:56 deafferret ya.... that....  should work  :)
19:56 spekki01 so you got the error too?
19:56 deafferret yessir
19:56 spekki01 hmmm :(
19:56 deafferret http://github.com/jhannah/bioperl-hmmer3
19:56 deafferret theoretically that is not possible  :D
19:56 deafferret hmm
19:57 spekki01 yeah im really new to github so im not even sure what might be up
19:58 spekki01 oh well contacting support...
19:58 deafferret wow. weird.
19:59 deafferret so mine magically starts working after I push a whitespace change back to github
19:59 spekki01 um can you walk me through that i've never pushed anything before
19:59 deafferret so try ignoring the website for now. Clone your fork then...
20:00 deafferret sure
20:00 deafferret what is your github username?
20:00 spekki01 pboutet
20:00 deafferret So here: http://github.com/pboutet/bioperl-hmmer3
20:01 spekki01 yep thats my broken fork
20:01 deafferret so now you want to clone your fork down to your workstation. do you code on a linux box?
20:01 spekki01 yep
20:02 deafferret so see the SSH command right above "
20:02 deafferret SearchIO::hmmer3 Save the full..."
20:02 spekki01 so git-clone http://github.com/pboutet/bioperl-hmmer3?
20:02 deafferret you want the ssh version
20:02 deafferret see the SSH | HTTP | Git Read-Only   gizmo?
20:02 deafferret you want the SSH version. the box on the right copies that into your clipboard
20:03 deafferret then you can paste that
20:03 deafferret for me its   git clone git@github.com:jhannah/bioperl-hmmer3.git
20:03 spekki01 ah ok got it lemme put that command in
20:03 deafferret for you it's that only pboutet instead of jhannah
20:04 spekki01 ok so im my directory i made a folder called github, and inside that i now have bioperl-hmmer3 downloaded from that git clone command
20:05 deafferret cool. now you can add a blank line to README
20:05 deafferret then   git add README; git commit; git push
20:05 deafferret and your github.com website is mysteriously no longer broken
20:05 deafferret I've never seen that before. some deep fork bug somewhere
20:06 deafferret I'm going to click the support button and report it
20:07 spekki01 hmm ok i dont think i did that commit and push properly
20:07 spekki01 lemme try again
20:08 deafferret whoah.... http://github.com/blog/644​-subversion-write-support
20:12 deafferret I reported the problem. My report is "awaiting moderation".
20:13 deafferret spekki01: looks like yours is still broke. having problems?  :)
20:13 spekki01 k so when i do git-commit it opens up vim and wants me to "please enter the commit message for your changes ....?
20:13 deafferret ya. that's where you describe what you changed
20:14 deafferret "Whitespace change to work around github fork bug."
20:14 deafferret or whatever
20:15 spekki01 does it matter where i put that, does it have to have a # before it or not?
20:15 deafferret it needs to NOT have # in front of it. those are comments that will not be part of your commit message
20:16 spekki01 ok there we go now were rolling
20:16 spekki01 sry wasnt sure what to do with that commit file, noob at this.
20:17 spekki01 looks like its working now on the website http://github.com/pboutet/bioperl-hmmer3
20:20 brandi joined #bioperl
20:20 spekki01 when i follow kblin's hmmer3 repo where does it announce changes and stuff he makes?
20:20 brandi left #bioperl
20:21 deafferret on your "homepage"
20:21 spekki01 kk
20:21 deafferret http://github.com/pboutet on the right
20:21 deafferret I think you can RSS subscribe to that too
20:21 spekki01 k i see it now
20:26 spekki01 yeah this will definetly take some time to get used to using, on a side note the history stuff is pretty cool from my 5 min of playing with it.
20:26 deafferret ya, there's lots of neat toys in git* land
20:26 deafferret GitX on my Mac is the most recent one I've been playing with
20:27 spekki01 when did github come about?
20:27 deafferret 2009?
20:28 deafferret git's been around since ~2002 I think...?
20:50 bag_ joined #bioperl
20:54 spekki01 in github do you follow repos or user or both?
20:54 spekki01 and is that different from waching?
21:03 spekki01 k well i think i figured it out... sigh new stuff lol
21:05 deafferret i never actually watch any of it
21:06 deafferret I'm buried in a gajillion things to do, let alone social networking spam let alone that attrocity people call "email"  :)
21:11 spekki01 lol sounds rough
21:13 deafferret sooo lazy...
21:19 kblin spekki01: to be honest I didn't get around to work on the table parser so far, busy day at work
21:19 spekki01 oh no worries
21:20 spekki01 I can always give it a go, i got started on github and made a fork of your repo to play around with it
21:21 spekki01 kblin: can i ask a quick question about getting scores our of hmmsearch result files
21:21 kblin sure
21:22 spekki01 my result file has 2 types of scores one for "full sequence" the other for "best 1 domain" how do i get the domain score, cause i know my $score = $hit->raw_score(); give me "full sequence" score.
21:22 spekki01 but not how to get the other.
21:24 deafferret kblin: is that code from sharpton a good jump-start?
21:24 kblin when you iterate over the hits, you get the domain's score from that
21:25 spekki01 oh so ->raw_score is domain score?
21:25 kblin deafferret: yeah, I had to change a few minor things, I mostly added some test cases
21:25 deafferret yay!  :)
21:26 kblin spekki01: no, result->raw_score is the "full sequence" score
21:26 kblin and every result has an array of hits you can iterate with result->next_hit
21:26 kblin and the hit has a score as well
21:26 kblin have a look at the test case, that has an example call :)
21:27 spekki01 the test_hmmparser.pl ?
21:27 kblin oh, never tried that
21:28 kblin t/SearchIO/hmmer3.t
21:28 spekki01 k
21:28 kblin if you copy that into the bioperl-live sources, along with the real parser code of course
21:30 kblin er, hang on, I asked you to look at it, not run it...
21:30 kblin It's getting late, obviously :)
21:31 deafferret what does this huge red button do?
21:31 * deafferret can't help himself, pushes it
21:32 kblin you could have waited for the paint to dry, at least
21:47 spekki01 kblin: im not sure what im missing here but here is the basic code im working with http://codepad.org/TlwqYfWp and when i do ->raw_score that is the score for the "full sequence" so where do i pull out the "domain" score, the outer loop or inner? I know you said it comes from result but with the outer loop it only shows me one result.
21:48 kblin spekki01: ah, my bad, I mixed up hit and hsp
21:48 kblin as I said, it's getting late
21:48 kblin sorry :/
21:48 spekki01 np your being helpful so im not complaining lol
21:49 spekki01 k so it should be in hsp?
21:57 kblin yeah, so if you do.. my $hsp = $hit->next_domain, $hsp->score is the domain socore
21:57 kblin *score
21:58 kblin that seemed like the most logical way to get people access to all the scores
22:00 spekki01 k lemme try that out
22:06 spekki01 hmmm ok i see how it works now, it'll take a bit of tweaking to get the best 1 domain score from hit's that have multiple domains but at least im getting the scores now :)
22:09 kblin as far as I understand the SearchIO interface, there's no way to give back two scores on the "hit" level
22:10 spekki01 ah ok, i've never really looked atthe searchIO interface and how it parses stuff
22:11 kblin anyway, need to get some sleep now, night folks
22:11 spekki01 bye :)
22:11 deafferret o7

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