Refactor bcftbx library to reorganise modules into subdirectories

[pjbriggs]

The bcftbx library is currently a flat set of modules, some of which have grown to be quite unwieldy (in particular IlluminaData). Also the capitalization of some modules (e.g. IlluminaData, SolidData etc) does not follow accepted module naming conventions.

This issue proposes reorganising the library modules into subdirectories and in some cases maybe splitting them, to make it easier to navigate in future.

There are various issues, the most significant being that this has the potential to introduce new bugs and will almost definitely break external applications which depend on these libraries (for example auto_process_ngs, GFFUtils etc).

Possible rearranged scheme:

• SolidData, Experiment -> solid/
• IlluminaData -> illumina/
• FASTQFile, TabFile -> files/
• JobRunner, Pipeline -> jobs/
• SpreadSheet, simple_xls -> xls/
• utils, Md5sum -> utils/
• htmlpagewriter -> html/

Thoughts:

• Might it be possible to use symlinks for the first round of refactoring, to try and manage breakages?
• Could pull in GFFFile and GTFFile modules from GFFUtils?
• Could move code from Python programs into new libraries? (e.g.

[pjbriggs]

Consider using the warnings module to indicate deprecation: https://python.readthedocs.io/en/stable/library/warnings.html

See also https://pymotw.com/2/warnings/

[pjbriggs]

PR #221 is a first round of refactoring the SOLiD code from bcftbx.SolidData into a new module platforms.solid.data.