These are chat archives for dereneaton/ipyrad

14th
Dec 2017
arminf
@arminf82
Dec 14 2017 17:48

Hi @isaacovercast and @dereneaton,
I have the same problem some people have with the ABBA test.
The notebook works fine with the example data.
My own dataset includes 52-97 samples and around 6000 final loci with minCov around 60%.
When I start the notebook with my data and generate_tests_from_tree depending on
the contrains and samples I assign to Px I either get 0 or e.g. 74 or 5000 tests generated from tree. It will crash after a short period of time “0% calculating D-stats | 0:00:08 |” with the error: 0: ValueError(cannot assign slice from input of different size). I think sometimes I also get the error: 0: IndexError(list index out of range). What is wrong with my locifile?

This is how I setup the dicts:
bb.generate_tests_from_tree(
constraint_dict={
"p4": ["Cal_1522"],
"p4": ["MRT_1425"],
})

Isaac Overcast
@isaacovercast
Dec 14 2017 17:58
@JStarrett There's not a straightforward way to extract locus information from phylip/nexus files, it's not really a use-case we considered. The easiest way probably is to just parse the .loci file and construct a partitioned file.
@jlblancopastor_twitter Can you show me the results of ls -l on the _fastqs directory? Can you rerun ipyrad with the -d flag and show me the last 10-20 lines of the ipyrad_log.txt file once it stops doing anything?
@DylanHCohen_twitter The bpp build looks like it works fine. Not sure why anaconda not being present would matter, but did you try running this and then rerunning the build conda install anaconda-client
Dylan
@DylanHCohen_twitter
Dec 14 2017 18:30
@jlblancopastor_twitter I tried these commands again last night
git clone https://github.com/dereneaton/ipyrad.git conda-build ipyrad/conda.recipe/bpp/ conda install --use-local bpp
and appears bpp is working fine now. I did try installing anaconda client when I was having problems, but that did not seem to work.