These are chat archives for dereneaton/ipyrad

5th
Nov 2016
Deren Eaton
@dereneaton
Nov 05 2016 16:01
@edgardomortiz OK, give the latest change to Master branch a try. I think I got it this time. There was confusion about measuring the array dimensions using mindepth_statistical for some parts versusmindepth_majrule in others that was leading to the shape errors when the two were not equal.
Edgardo M. Ortiz
@edgardomortiz
Nov 05 2016 17:15
great, I will test it now
Edgardo M. Ortiz
@edgardomortiz
Nov 05 2016 17:31
just to clarify the previous tests: when I didn't change any parameters step 4 failed and when I changed both in a branch step 4 succeeded
Deren Eaton
@dereneaton
Nov 05 2016 19:47
@edgardomortiz argh, I found a bug that needs fixing, so no rush on testing it.
Edgardo M. Ortiz
@edgardomortiz
Nov 05 2016 20:13

@dereneaton yeah, is this one?


 -------------------------------------------------------------
  ipyrad [v.0.4.8]
  Interactive assembly and analysis of RAD-seq data
 -------------------------------------------------------------
  loading Assembly: test-s4
  from saved path: /scratch/02728/emo347/ddRAD/astereae/e_read-lengths_ipyrad/test-s4.json
  host compute node: [48 cores] on nid00019

  Step 4: Joint estimation of error rate and heterozygosity
  [####################] 100%  inferring [H, E]      | 0:34:04
ERROR:ipyrad.assemble.util:  Sample PIO-Dcine failed with error OSError([Errno 12] Cannot allocate memory)

  Encountered an error, see ./ipyrad_log.txt.
    Sample PIO-Dcine failed step 4

I am reducing the numbers of cores to 12 just to check

Deren Eaton
@dereneaton
Nov 05 2016 20:13
yeah, it's writing really big files
maybe better to just wait til it's fixed, it shouldn't be writing these big tmp files at all.
I'm working on it now
Edgardo M. Ortiz
@edgardomortiz
Nov 05 2016 20:33
Also, what is your opinion for those clusters made up of merged reads and pairs linked by nnnn ? I tried this:
dereneaton/ipyrad#194
but I am not sure how those clusters should be treated...
Deren Eaton
@dereneaton
Nov 05 2016 21:32
@edgardomortiz OK bug fix pushed.
v.0.4.9 now up with a bug fix for a memory issue introduced into 0.4.8.
Edgardo M. Ortiz
@edgardomortiz
Nov 05 2016 21:34
nice! testing...
Deren Eaton
@dereneaton
Nov 05 2016 21:34
I like your idea for removing the separator when some of the reads merge, makes sense.
Edgardo M. Ortiz
@edgardomortiz
Nov 05 2016 21:48
Oh no, this happened again, do you think is better to reun step 3 with the new version?

 -------------------------------------------------------------
  ipyrad [v.0.4.9]
  Interactive assembly and analysis of RAD-seq data
 -------------------------------------------------------------
  loading Assembly: test-s4
  from saved path: /scratch/02728/emo347/ddRAD/astereae/e_read-lengths_ipyrad/test-s4.json
  host compute node: [48 cores] on nid00022

  Step 4: Joint estimation of error rate and heterozygosity
  [####################] 100%  inferring [H, E]      | 0:05:47
ERROR:ipyrad.assemble.util:  Sample PIO-Dcine failed with error IndexError(index 125 is out of bounds for axis 0 with size 125)

  Encountered an error, see ./ipyrad_log.txt.
    Sample PIO-Dcine failed step 4
Deren Eaton
@dereneaton
Nov 05 2016 21:49
I guess so. I tried to make it so that you can start step4 from an old assembly, but maybe best to do it from step3.
Edgardo M. Ortiz
@edgardomortiz
Nov 05 2016 21:50
Also because when the parameters change and step 4 recalculates the dimensions everything went well
Deren Eaton
@dereneaton
Nov 05 2016 21:50
that's what I'm hoping for