Add coverage tox target

[gnikonorov]

Add a tox target for generating coverage reports. I am doing this so that we can grow our test coverage. When the target is run, we get the following output after the pytest suite finishes:

-------------------------------------------------------------------------- generated html file: file:///home/gnikonorov/pytest-html/.tox/coverage/log/report.html --------------------------------------------------------------------------

----------- coverage: platform linux, python 3.6.9-final-0 -----------
Name                      Stmts   Miss Branch BrPart     Cover   Missing
------------------------------------------------------------------------
pytest_html/__init__.py       7      2      0      0    71.43%   7-9
pytest_html/extras.py        28     11      0      0    60.71%   14, 24, 28, 32, 36, 40, 44, 48, 52, 56, 60
pytest_html/hooks.py          5      0      0      0   100.00%
pytest_html/plugin.py       376    300    136      4    16.02%   33-38, 89->exit, 90->91, 91-92, 93->95, 95-96, 101->102, 102-103, 128-130, 134-135, 140-152, 156-196, 202-211, 216-234, 237-279, 282-320, 325-357, 360-364, 367-371, 374-384, 387-393, 396-406, 409-414, 418-419, 422-585, 588-612, 615-628, 631-683, 689, 692-693, 696, 699-701, 704
------------------------------------------------------------------------
TOTAL                       416    313    136      4    19.75%

The next steps assuming we're all on board would be:

• Set up codecov for this project
• Publish the results of this run to codecov
• Grow our coverage

I used pytest-cov's adhoc-layout example as a starting point

[BeyondEvil]

Nice! But given the discussion in #350, I think we should hold off on any work to up the coverage. Better to spend that time and energy on V3 imo.

That way we can set a coverage target from the start and block PR's that puts us under the target.

[gnikonorov]

Nice! But given the discussion in #350, I think we should hold off on any work to up the coverage. Better to spend that time and energy on V3 imo.

That way we can set a coverage target from the start and block PR's that puts us under the target.

I agree! This was just something small to get it going

[ssbarnea]

I do not hing that coverage is what we need at this stage for two reasons: upcoming major rework around dropping py and also use of pytest-cov is... controversial. This job does not run on CI so it does not provide much practical value.

Running coverage on CI is tricky as one also needs to collect results from each job and combine them. We should do it but I see it at the bottom of the priority list.

[gnikonorov]

I'll close for now then to not clutter the list of open PRs. I'll reopen when we feel ready to add coverage