Add plotly support to pyani graphics, or remove the option from the parser

[baileythegreen]

Summary:

plotly is listed as one of the three plotting options when running pyani plot -h, but there is no code currently supporting this.

Reproducible Steps:

Assuming an existing database with the default name (.pyani/pyanidb), with at least one run in it:

pyani plot -h
pyani plot outdir 1 --method plotly

If it's not possible to reproduce the issue, please include a description of how you discovered the issue.

Current Output:

Traceback (most recent call last):
  File "/Users/baileyharrington/miniconda3/bin/pyani", line 33, in <module>
    sys.exit(load_entry_point('pyani', 'console_scripts', 'pyani')())
  File "/Users/baileyharrington/Software/pyani/pyani/scripts/pyani_script.py", line 117, in run_main
    returnval = args.func(args)
  File "/Users/baileyharrington/Software/pyani/pyani/scripts/subcommands/subcmd_plot.py", line 94, in subcmd_plot
    write_run_heatmaps(run_id, session, outfmts, args)
  File "/Users/baileyharrington/Software/pyani/pyani/scripts/subcommands/subcmd_plot.py", line 134, in write_run_heatmaps
    write_heatmap(
  File "/Users/baileyharrington/Software/pyani/pyani/scripts/subcommands/subcmd_plot.py", line 202, in write_heatmap
    GMETHODS[args.method](
KeyError: 'plotly'

Expected Output:

Plots.

pyani Version:

0.3.0-alpha

[widdowquinn]

We've discussed this at meetings. Adding plotly support is a goal, in part intended for the proposed Flask support, so that we can have interactive, rather than static, graphics in that interface.

[widdowquinn]

It's OK for the documentation to be slightly out of sync during development.

[baileythegreen]

I don't remember discussion plotly, though I know there is a discussion about Flask. I just noticed that option wasn't implemented when I was adding scatterplot code.

[widdowquinn]

I remember discussion of plotly in the context of matplotlib vs seaborn (vs…) and in the context of a candidate Flaskinterface. In any case, the point about synchronisation of code and docs is the important one. Writing docs ahead of implementation (including inline docs) can set a target to code to, just as writing tests ahead of implementation (aka test-driven development) gives a target to code to.

My view is that the appropriate course of action here is to create an issue like #124 for plotly as an output option and include it in the milestones for a release version in the Project section.