Updating quickmodel with more options and to use parser function.

[sonyahanson]

Updating quickmodel with a few more options and so that it now uses parser.py, which I've been working on for Greg and Mehtap to be usable outside of quickmodel.

Options to add:

• [ ] Methap's plate format.
• [ ] Trace png's
• [ ] Plot two datasets to compare (a la 'constph-grant' or 'GK v. WT' requests)

[sonyahanson]

Okay, so it turns out that the things that I thought would be easy to get done before the more advanced changes above took a bit longer than expected. What HAS been done here:

• Removed parsing to a function, which actually has allowed us to remove the need to separate between spectra and single wavelength assays because computers are smart (the parsing function looks for 'wavelength' being defined in the inputs file:
• Because of this new development this now requires input files have the 'file_set' defined in them, as spectra input files have always had. Changed example input files accordingly.
• Added some outputs to the outputs json that were previously missing, most notably the actual data points fit in the model, the equilibration time, and the credibility intervals.

The other things will have to wait for now, but were added to #72, see discussion on #91 for more about 'adding Mehtap's plate format'.

Mergeable if tests are passing.

[jchodera]

Looks like there is a py3 issue with the JSON serialization: https://travis-ci.org/choderalab/assaytools/jobs/228916368#L671-L688

The failure seems to be caused by the json.dump of the metadata:

Traceback (most recent call last):
  File "/home/travis/miniconda/envs/test/bin/quickmodel", line 11, in <module>
    load_entry_point('assaytools==0.2.0', 'console_scripts', 'quickmodel')()
  File "/home/travis/miniconda/envs/test/lib/python3.4/site-packages/assaytools/scripts/quickmodel.py", line 241, in entry_point
    quick_model(inputs, nsamples=args.nsamples, nthin=args.nthin)
  File "/home/travis/miniconda/envs/test/lib/python3.4/site-packages/assaytools/scripts/quickmodel.py", line 211, in quick_model
    json.dump(metadata, outfile, sort_keys = True, indent = 4, ensure_ascii=False)
  File "/home/travis/miniconda/envs/test/lib/python3.4/json/__init__.py", line 178, in dump
    for chunk in iterable:
  File "/home/travis/miniconda/envs/test/lib/python3.4/json/encoder.py", line 422, in _iterencode
    yield from _iterencode_dict(o, _current_indent_level)
  File "/home/travis/miniconda/envs/test/lib/python3.4/json/encoder.py", line 396, in _iterencode_dict
    yield from chunks
  File "/home/travis/miniconda/envs/test/lib/python3.4/json/encoder.py", line 429, in _iterencode
    o = _default(o)
  File "/home/travis/miniconda/envs/test/lib/python3.4/json/encoder.py", line 173, in default
    raise TypeError(repr(o) + " is not JSON serializable")
TypeError: 0 is not JSON serializable

[jchodera]

My best guess is that the error we see:

is due to a 0 that is not a normal Python int but a np.int64 that isn't automatically being converted to Python int. For example, in python 3.5:

mski1776:benchmark choderaj$ python
Python 3.5.3 |Continuum Analytics, Inc.| (default, Mar  6 2017, 12:15:08) 
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import json
>>> import numpy as np
>>> outfile = open('test.json', 'w')
>>> # This works fine:
>>> json.dump(0, outfile, sort_keys = True, indent = 4, ensure_ascii=False)
>>> # This breaks:
>>> json.dump(np.int64(0), outfile, sort_keys = True, indent = 4, ensure_ascii=False)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/choderaj/miniconda3/lib/python3.5/json/__init__.py", line 178, in dump
    for chunk in iterable:
  File "/Users/choderaj/miniconda3/lib/python3.5/json/encoder.py", line 436, in _iterencode
    o = _default(o)
  File "/Users/choderaj/miniconda3/lib/python3.5/json/encoder.py", line 179, in default
    raise TypeError(repr(o) + " is not JSON serializable")
TypeError: 0 is not JSON serializable

[jchodera]

I can't tell if it's inputs or outputs that contains the unserializable thing. You'll have to double-check to make sure nothing in there is still a numpy object before JSON serializing it.

[sonyahanson]

I think this should fix it, but wasn't able to reproduce the error in a python3 environment, so not sure. Will just have to wait!

[sonyahanson]

Checks are passing. Merging.