These are chat archives for dereneaton/ipyrad

16th
Feb 2016
Deren Eaton
@dereneaton
Feb 16 2016 00:27 UTC
Yeah, it seemed to make more sense.
And you, for sure. Users should be able to do everything with the API
if they want
My though is that the API provides a nicer framework for people who are really into creating reproducible workflows since you can record every step from start to finish, whereas the CLI has the hidden step of people filling in params files by hand.
But whether the assembly is done in the CLI or API, the API should be the go-to place for downstream analyses if users want to do them. Since it's way easier for us to add-on analysis tools to the API rather than build more and more options into the CLI which would make it feel messy eventually.
also, I have a pretty hard deadline on some stuff early this week, so I'm not gonna get much done the next few days.
Isaac Overcast
@isaacovercast
Feb 16 2016 00:33 UTC
cool. I have some stuff tomorrow and wednesday, should be able to make some progress.
Ive been having some issues getting logging info to output to the log file, did something change?
Deren Eaton
@dereneaton
Feb 16 2016 01:02 UTC
I think I fixed it recently. I added a line to __init__ that resets the log file when it gets too big, and versioner was fucking with its debug setting.
but I"ll double check that its not still broken
Deren Eaton
@dereneaton
Feb 16 2016 04:19 UTC
I'm noticing that the API seems broken now that I'm trying it again after a while. Maybe something I did with launch2(). ugh.
Isaac Overcast
@isaacovercast
Feb 16 2016 16:58 UTC
working on CLI docs, i'm working on a "full tutorial" styled after the one for pyrad. I imagine there will be this, and there will be an advanced_CLI.rst with examples of reference mapping, branching, MPI, etc.
Deren Eaton
@dereneaton
Feb 16 2016 17:07 UTC
Sweeeeet
Isaac Overcast
@isaacovercast
Feb 16 2016 22:33 UTC
Do you think a -i flag would be useful for printing paramsinfo at the command line?
i.e: ipyrad -i 10
    (10) phred_Qscore_offset ---------------------------------------------
    Examples:
    ----------------------------------------------------------------------
    data.set_params(10) = 33
    data.set_params("phred_Qscore_offset") = 33
    ----------------------------------------------------------------------
Deren Eaton
@dereneaton
Feb 16 2016 22:35 UTC
if it's easy enough to do, I would say go for it.
How's the docs coming?
"sim_rad_test" worked fine for me when I ran it using the API, I'll have to test with the CLI still.
Isaac Overcast
@isaacovercast
Feb 16 2016 22:36 UTC
Getting there. I'm going through the pyrad docs and copy/pasting, and then basically rewriting everything so it make sense for the new system. Still getting used to the weird markup.
I would expect a significant chunk to be done today, got the rest of the night blocked out.
Deren Eaton
@dereneaton
Feb 16 2016 22:56 UTC
awesome.
Isaac Overcast
@isaacovercast
Feb 16 2016 22:57 UTC
-i flag was easy.
Isaac Overcast
@isaacovercast
Feb 16 2016 23:05 UTC
I can't imagine how CLI and API would be different for step 7 for "simrad_test_R1.fastq.gz", i just did a git pull, rebuilt, reran from step1 and i still see the assert. o_O
at CLI, natch.
Isaac Overcast
@isaacovercast
Feb 16 2016 23:11 UTC
Do you have a better workflow for working on docs than write->git commit->push->wait->wait->wait->look at rtd? That sucks.
Deren Eaton
@dereneaton
Feb 16 2016 23:16 UTC
I don't. but there is one. You can host the docs locally.
its on the rtd docs somewhere
Ok, I'll give it a try with the CLI soon.
well, later tonight.
duuude, I made a native implementation of svdquartets today.
for ipyrad.analysis
and it works!
it'll be a while before I get a chance to speed test it against the original. I could imagine it might be a little slower for each test, but it will be easy to use and infinitely parallized, so that's a big improvement.
Isaac Overcast
@isaacovercast
Feb 16 2016 23:19 UTC
Nice!
That rules. I like that method.
Deren Eaton
@dereneaton
Feb 16 2016 23:20 UTC
oh, and it will have checkpointing, so you can start, stop, restart, etc. Sample all those quartets.
Isaac Overcast
@isaacovercast
Feb 16 2016 23:22 UTC
nice move. People are going to love that!
Deren Eaton
@dereneaton
Feb 16 2016 23:22 UTC
it was actually super easy. Literally took half a day. You should check out the code when I push it.
Isaac Overcast
@isaacovercast
Feb 16 2016 23:23 UTC
will do.
I got invited to Tulane to do a bioinformatics wkshop for the derryberry lab in May. Be another good chance to spread the ipyrad word, by May this shit should be tight!
Isaac Overcast
@isaacovercast
Feb 16 2016 23:39 UTC
How would you feel if i added the parameter names to the params.txt file like this:
Deren Eaton
@dereneaton
Feb 16 2016 23:39 UTC
Thats awesome!
Isaac Overcast
@isaacovercast
Feb 16 2016 23:40 UTC
wat                            ## [0] [assembly_name]: Assembly name. Used to name output directories for assembly steps
./                             ## [1] [project_dir]: Project dir (made in curdir if not present)
                               ## [2] [raw_fastq_path]: Location of raw non-demultiplexed fastq files
                               ## [3] [barcodes_path]: Location of barcodes file
Is that dumb?
REason being the -i flag w/ no arguments prints a list of available params to get descriptions for, i think this would make it more explicit which parameter is which.
Deren Eaton
@dereneaton
Feb 16 2016 23:44 UTC
No, I was thinking the same thing.
Isaac Overcast
@isaacovercast
Feb 16 2016 23:45 UTC
oh good, cuz i did it already ;p
Deren Eaton
@dereneaton
Feb 16 2016 23:45 UTC
Its good for the publication too. Users can say they set max_Ns
Isaac Overcast
@isaacovercast
Feb 16 2016 23:46 UTC
Right! def