add pylintrc file to each indicator

[sgsmob]

This should make the linter complain less about:

• Short variable names
• Pytest naming/argument conventions

[krivard]

It looks like this is using an identical .pylintrc for all indicators; the following modules had existing .pylintrc that have been overwritten. Was that intentional?

• _delphi_utils_python
• cds (can probably ignore this; cds is no longer being pursued)
• google_health
• jhu
• nchs_mortality
• quidel_covidtest
• safegraph
• safegraph_patterns
• usafacts

[chinandrew]

this isn't scoped just to the test dir right? so adding these lines to have tests lint correctly potentially allows these into the actual module code?

secondarily, how strongly do we feel we need to lint unit tests?

[sgsmob]

Yes, they would allow it into the module code, but they seem like narrow enough and innocuous enough problems that they could be caught in review.

If one uses an integrated linter in their IDE, it will read from these files and complain about the tests. FWIW we enforce style the same across implementation and test files.

[chinandrew]

I'm not a huge fan of relying on code review to catch problems that could have been linted, but agree they're pretty innocuous. One option is to add an additional pylintrc to the tests module, so any weird test things can get captured without worrying about affecting the module code. Not sure how clean that is though.

[sgsmob]

I thought about doing this but thought it created a lot of lint files (2 per indicator rather than the necessary 1). Can do if you think it is worthwhile.

[chinandrew]

Yea that's kind of annoying...I think they're pretty harmless for now, lets just leave as is.

[sgsmob]

It looks like this is using an identical .pylintrc for all indicators; the following modules had existing .pylintrc that have been overwritten. Was that intentional?

• _delphi_utils_python
• cds (can probably ignore this; cds is no longer being pursued)
• google_health
• jhu
• nchs_mortality
• quidel_covidtest
• safegraph
• safegraph_patterns
• usafacts

Yes, I felt that we should have a consistent style across all indicators.

[krivard]

Yes, I felt that we should have a consistent style across all indicators.

Sounds good! In that case, would you do a pass through the ones below and make sure they lint successfully with the new configuration? Otherwise the next PR on their deploy-* branch will have surprise failures unrelated to the feature being added.

• google_health
• jhu
• quidel_covidtest
• safegraph
• usafacts

[jingjtang]

A quick check for google_health and quidel_covidtest, both of them run into linter errors. Most of the errors are due to C0103 (invalid-name), a few of them are due to E1101, W0212 etc.

Specially, W0212 results from protected-access to to member _load_crosswalk in GeoMapper. This should be discussed with @dshemetov to make _load_crosswalk public (or not).

[sgsmob]

Yes, I felt that we should have a consistent style across all indicators.

Sounds good! In that case, would you do a pass through the ones below and make sure they lint successfully with the new configuration? Otherwise the next PR on their deploy-* branch will have surprise failures unrelated to the feature being added.

• google_health
• jhu
• quidel_covidtest
• safegraph
• usafacts

@krivard should I include the lint compliance changes in this PR?

[krivard]

lgtm