DOC: Improve the documentation for running integration tests locally

[tswast]

I'd love to see a TESTING.md file that showed how to run the tests locally before sending off a PR.

I can run pytest but is that running the integration tests, too? It seemed to run a bit fast for that to be the case.

[jreback]

http://pandas-gbq.readthedocs.io/en/latest/contributing.html

is pretty comprehensive

[jreback]

of course if u think it needs updates then pls send a PR

[parthea]

I set the following when running integration tests locally:

PROJECT_ID = None
PRIVATE_KEY_JSON_PATH = None
PRIVATE_KEY_JSON_CONTENTS = None

I scanned through the contributing docs and it appears that this step is missing (it wasn't in the pandas contributing docs either). I'll re-open this issue so we can update the docs.

https://github.com/pydata/pandas-gbq/blob/master/pandas_gbq/tests/test_gbq.py#L22

[parthea]

@tswast The integration tests are skipped if you do not provide credentials/project id . I'm open to improving the local integration testing process as well. Some users may not want to use their default gcloud credentials to run pandas-gbq integration tests but another alternative could be to use environment variables, or have a separate file for PROJECT_ID, PRIVATE_KEY_JSON_PATH and PRIVATE_KEY_JSON_CONTENTS . Thoughts?

[tswast]

That makes sense to skip, even if default gcloud credentials are available. I'll try running through the current workflow this morning to see how it goes.

[tswast]

Question: Does it make any sense to populate PRIVATE_KEY_JSON_PATH but not PRIVATE_KEY_JSON_CONTENTS?

I'm thinking of always loading the contents of the key file in _get_private_key_contents.

[parthea]

@tswast You're right. It seems redundant to have PRIVATE_KEY_JSON_CONTENTS in test_gbq.py since we could read Json from the file in PRIVATE_KEY_JSON_PATH.