These are chat archives for dereneaton/ipyrad

13th
Feb 2016
Isaac Overcast
@isaacovercast
Feb 13 2016 01:48
bleh, that's annoying.
would be good to have json for the assemblies tho, i agree it would make them more flexible.
Deren Eaton
@dereneaton
Feb 13 2016 02:09
uuugh, this shit is so annoying
been working on the json for hours
Isaac Overcast
@isaacovercast
Feb 13 2016 02:13
oh no dude!
housekeeping is always more annoying that it should be.
Isaac Overcast
@isaacovercast
Feb 13 2016 02:41
Fixed so now main creates -n wat params as params-wat.txt
Deren Eaton
@dereneaton
Feb 13 2016 02:55
when I push this all old objects are gonna be broken
I should have left the old load_assembly in tact
I guess I'll try to put it back in place
I've got the json working now though. Runs the exact same as before, but now the object is human-readable, and reading in different version objects should be easier for us to fix.
Isaac Overcast
@isaacovercast
Feb 13 2016 03:06
Cool. go for it.
Deren Eaton
@dereneaton
Feb 13 2016 03:10
ok, both formats can be loaded. If there's a json and assembly file it takes the json file. It saves in json. So .assembly should be phased out pretty quick.
Isaac Overcast
@isaacovercast
Feb 13 2016 21:27
In set_param for raw_fastq_path it's testing for a directory + wildcard. Right now if you just have one raw file and you specify the absolute path it fails to read it. I think it's a good idea to support the ability to just pass in the exact file name if you only have one fq.gz file. Any thoughts?
Deren Eaton
@dereneaton
Feb 13 2016 21:31
yeah, for sure, sometimes people have all their data in one file.
so we should have it check for both
I just spent a few more hours on the JSON encoding. Now it conserves tuples correctly.
Isaac Overcast
@isaacovercast
Feb 13 2016 21:33
nvmind im an idiot. it does work.
it was an auto-troll. I gunzip'd the raw file, but was trying to step1 on the .gz. derp
Deren Eaton
@dereneaton
Feb 13 2016 21:38
ugh, don't download 0.1.44, I broke a bunch of shit
working on fixing it.
Isaac Overcast
@isaacovercast
Feb 13 2016 21:39
ok.
Deren Eaton
@dereneaton
Feb 13 2016 21:48
ok, I think it's good.
Deren Eaton
@dereneaton
Feb 13 2016 22:08
only ~one week left until our hopeful launch date!
Isaac Overcast
@isaacovercast
Feb 13 2016 22:10
i know dude, so psyched.
Question, in step 3, if a sample is already in step 2.5 should we skip mapreads/clustall and just pass it forward to muscle_align?
My vote is "yes"
Deren Eaton
@dereneaton
Feb 13 2016 22:27
I think so, that was the idea. Unless they use --force
there's a lot of work left in steps 5-7
Isaac Overcast
@isaacovercast
Feb 13 2016 22:59
i know, it's gonna be a ffun week