These are chat archives for dereneaton/ipyrad

31st
Mar 2016
Deren Eaton
@dereneaton
Mar 31 2016 19:04
yeah, I just ran into the same problem... we must be doing something dumb.
I know that multi_muscle_align is creating way too many tmp files for aligning, because the estimate of optim doesn't work, cuz it looks for clusters_total even tho that value is still 0, so it defaults to optim=100.
That should be fixed, but it still shouldn't be causing a memory error.
Deren Eaton
@dereneaton
Mar 31 2016 19:09
unless the shutil.rmtree arg is crashing cuz we're handing it a dir with too many files in it
that could actually be the problem...
or a problem... still doesn't seem like it should build 10GB of RAM
And it's not saving their state in 2.5 when it crashes, ugh.
Deren Eaton
@dereneaton
Mar 31 2016 19:39
when samples are in state=2.5 force flag does not restart them from state 2, currently, it seems...
Deren Eaton
@dereneaton
Mar 31 2016 21:46
step 5 does not create a s5_consens_stats.txt file yet.
Deren Eaton
@dereneaton
Mar 31 2016 22:17
we're hitting an error in reference alignment in the tutorial
 --------------------------------------------------
  ipyrad [v.0.1.84]
  Interactive assembly and analysis of RADseq data
 --------------------------------------------------
  loading Assembly: data2
  saved location: ~/iptutorial/data2.json
  ipyparallel setup: Local connection to 4 Engines

  Step3: Clustering/Mapping reads
    **************************************************************
    Indexing reference sequence. This only needs to be done once.
    This should take less than 10 minutes.
    **************************************************************
    Done indexing reference sequence
[Engine Exception]
---------------------------------------------------------------------------IndexError                                Traceback (most recent call last)<string> in <module>()
/Users/student/anaconda/lib/python2.7/site-packages/ipyrad/assemble/cluster_within.pyc in muscle_align(args)
    149         if len(names) <= 1:
    150             if names:
--> 151                 stack = [names[0]+"\n"+seqs[0]]
    152         else:
    153             ## split seqs if paired end seqs
IndexError: list index out of range

None
    Saving Assembly.
Deren Eaton
@dereneaton
Mar 31 2016 22:22
we gotta fix this ASAP
Deren Eaton
@dereneaton
Mar 31 2016 22:46
we should set another deadline for ourselves to make sure we stay on track getting this out.
shit, we're hitting 10GB on the example pedicularis tutorial too!
if you exit out before it crashes you DO get some samples in state3 and some in state 2.5.
weird