Reading from csv files

[williamjameshandley]

Description

Following conversations with @yallup, this PR implements reading csv files which have been written with the .to_csv method. Getting this right for all variations on labels/weights being dropped is a little fiddly, and the current implementation prioritises robustness over speed (reading multiple times in some cases), but will function as a low-latency start point.

For consistency this also moves WeightedLabelledPandas into its own file anesthetic.weighted_labelled_pandas.py rather than anesthetic.samples, which reduces clutter for users just looking at the latter file.

Checklist:

• I have performed a self-review of my own code
• My code is PEP8 compliant (flake8 anesthetic tests)
• My code contains compliant docstrings (pydocstyle --convention=numpy anesthetic)
• New and existing unit tests pass locally with my changes (python -m pytest)
• I have added tests that prove my fix is effective or that my feature works
• I have appropriately incremented the semantic version number in both README.rst and anesthetic/_version.py

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (3ea992c) to head (0fec842).

Additional details and impacted files

@@            Coverage Diff            @@
##            master      #364   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           34        36    +2     
  Lines         2979      3043   +64     
=========================================
+ Hits          2979      3043   +64     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[williamjameshandley]

@yallup does this perform as expected?

[yallup]

Just tested and the functionality seems good. Including read_csv into the read_chains logic chain may be useful and then plugs into the gui automatically to boot?

[williamjameshandley]

I agree this is a good idea, but leaves a small choice:

If we want it to be consistent with read_chains (read_polychord etc), then this should really take a root rather than a filename (root.csv), which differs from the pandas default read_csv.

Thoughts/preferences?

[williamjameshandley]

(could always accept both)

[yallup]

I agree this is a good idea, but leaves a small choice:

If we want it to be consistent with read_chains (read_polychord etc), then this should really take a root rather than a filename (root.csv), which differs from the pandas default read_csv.

Thoughts/preferences?

I was going to remark (before checking that it is actually consistent with pandas as is), that root as a filename arg would be preferable. Although we override pandas I think there is enough local precedent to make this conform to the "root" style chains file reading.

[williamjameshandley]

OK, I've implemented both, so you can pass root or root.csv to anesthetic.read.csv.read_csv. The gui and read_chains should now both work with csv files.

[yallup]

Does everything I was looking for, am not testing any of the intricacies around labels but looks good to me(rge)