NASC ingestion refactoring

[brandynlucca]

This draft PR includes refactored changes to the NASC ingestion, docstrings, and associated tests.

[leewujung]

@brandynlucca : testing structure looks great! thanks!

[leewujung]

Hey @brandynlucca : Thanks for the PR!

The flow you have under the main functions looks good so I didn't go through all the new functions carefully. I trusted that you have verified the details.

I only have a few suggestions/questions below:

• For the old tests, you can use pytestmark = pytest.mark.skip(reason="Temporarily disable this module") on top of a .py to disable all tests in lieu of commenting the code out.
• Why are setting column names to lowercase and impute_bad_coordinates needed in read_nasc_file? I was thinking df_nasc_all_ages (or df_nasc_no_age1) would be a "final" product so that people could just load and use those directly. Are the formatting and data in df_nasc_all_ages kept in a specific way to accommodate something else?
• Is filter_transect_intervals related to the mesh polygon creation downstream? Or is it just so that people have some control over what part of the geographical regions are included?
• I see many unit tests, but couldn't find integration tests. How about adding those for the functions that are exposed in feat_hake.py?
• Mirrorfeat_hake.py to have a notebook as a gateway for people to interactively try out the code?

[leewujung]

Oh also just noticed the merge conflict - seems just small things from my PRs added after your branched out this, so they ended up not in the commit history here.

[brandynlucca]

Why are setting column names to lowercase and impute_bad_coordinates needed in read_nasc_file? I was thinking df_nasc_all_ages (or df_nasc_no_age1) would be a "final" product so that people could just load and use those directly. Are the formatting and data in df_nasc_all_ages kept in a specific way to accommodate something else?

This mostly ensures backwards compatibility with previous FEAT survey years where the column name schemes are somewhat inconsistent. This is somewhat in anticipation of incorporating the validation step(s), which would incorporate any required formatting changes.

Is filter_transect_intervals related to the mesh polygon creation downstream? Or is it just so that people have some control over what part of the geographical regions are included?

This relates to removing off-effort transect intervals.