feat: Stage 5 of nox implementation - adding coverage targets

[mf2199]

As suggested in #466 , this represents item 5 of the recommended list aimed to implement nox testing automation in multiple steps.

Change list:

1. Added cover session to noxfile.py + configuration parameters;

The code, at it current state, shows only 41% test coverage:

nox > Running session cover
nox > Creating virtual environment (virtualenv) using python.exe in .nox\cover
nox > pip install coverage pytest-cov
nox > coverage report --show-missing --fail-under=99
Name Stmts Miss Branch BrPart Cover Missing
-------------------------------------------------------------------------------------
django_spanner\__init__.py 41 41 12 0 0% 7-73
django_spanner\base.py 49 49 2 0 0% 7-146
django_spanner\client.py 3 3 0 0 0% 7-11
django_spanner\compiler.py 48 48 28 0 0% 7-90
django_spanner\creation.py 46 46 14 0 0% 7-75
django_spanner\expressions.py 10 10 4 0 0% 7-22
django_spanner\features.py 27 27 4 0 0% 7-342
django_spanner\functions.py 52 52 4 0 0% 7-104
django_spanner\introspection.py 65 65 28 0 0% 7-217
django_spanner\lookups.py 90 90 36 0 0% 7-144
django_spanner\operations.py 158 158 68 0 0% 7-275
django_spanner\schema.py 127 127 72 0 0% 7-262
django_spanner\utils.py 7 7 2 0 0% 1-14
django_spanner\validation.py 10 10 2 0 0% 1-21
spanner_dbapi\__init__.py 20 2 4 2 83% 85->86, 86, 89->90, 90
spanner_dbapi\connection.py 62 37 6 0 37% 24-26, 32-33, 44-46, 49-50, 53-54, 57-58, 61-69, 72, 84-88, 91-110, 113-115, 118-120, 123, 128, 131-132
spanner_dbapi\cursor.py 173 130 54 0 19% 43-52, 64-93, 96, 102-110, 113-139, 145-151, 155-158, 161-184, 187, 190, 193, 197-213, 217, 223-224, 227-228, 231-235, 238-240, 243-245, 248-253, 256-258,
273-285, 289, 292, 295, 298, 301, 304, 307, 313-319, 325-338, 341-351
spanner_dbapi\exceptions.py 20 0 0 0 100%
spanner_dbapi\parse_utils.py 156 8 89 8 93% 39->40, 40, 43->49, 49, 141->142, 142, 156->157, 157, 249->250, 250, 273->274, 274, 321->322, 322, 338->339, 339
spanner_dbapi\parser.py 142 18 78 10 85% 49->50, 50, 51->52, 52, 53->54, 54, 55->56, 56, 81, 87->88, 88, 91->92, 92, 96->97, 97, 115->119, 119, 126->127, 127, 139, 196->233, 233-242, 246-247
spanner_dbapi\types.py 31 1 0 0 97% 43
spanner_dbapi\utils.py 39 0 10 0 100%
spanner_dbapi\version.py 7 0 0 0 100%
tests\spanner_dbapi\__init__.py 0 0 0 0 100%
tests\spanner_dbapi\test_connect.py 46 6 0 0 87% 57, 70, 80-82, 95-97
tests\spanner_dbapi\test_globals.py 7 0 0 0 100%
tests\spanner_dbapi\test_parse_utils.py 74 0 26 0 100%
tests\spanner_dbapi\test_parser.py 52 0 24 0 100%
tests\spanner_dbapi\test_types.py 43 0 2 0 100%
tests\spanner_dbapi\test_utils.py 19 0 0 0 100%
tests\spanner_dbapi\test_version.py 14 0 0 0 100%
-------------------------------------------------------------------------------------
TOTAL 1638 935 569 20 41%
Coverage failure: total of 41 is less than fail-under=99
nox > Command coverage report --show-missing --fail-under=99 failed with exit code 2
nox > Session cover failed.

Towards #474.

[c24t]

The coverage target itself looks good, but some questions about the config:

Bacause of .coveragerc, running coverage alone fails under 100, but running nox -e coverage fails under 99, and running nox generates a coverage report but never fails. Why the three different limits? And what's the point of generating the coverage report on each build if we never fail because of coverage?

I see this is how python-spanner does it too (https://github.com/googleapis/python-spanner/blob/master/noxfile.py, https://github.com/googleapis/python-spanner/blob/master/.coveragerc), I just don't understand why they're doing it this way.

[mf2199]

I see this is how python-spanner does it too (https://github.com/googleapis/python-spanner/blob/master/noxfile.py, https://github.com/googleapis/python-spanner/blob/master/.coveragerc), I just don't understand why they're doing it this way.

This is a common way among other API as well, e.g. Storage, Firestore, Bigtable etc.