Add types for read_metadata()

[victorlin]

Description of proposed changes

For type checking.

I considered adding types for read_csv_with_index_col() but decided against it, as it relies on fragile imports from pandas._typing.¹

¹ https://github.com/pandas-dev/pandas-stubs/blob/64e978cbce28de05ff38105c169387283b6c95c2/pandas-stubs/io/parsers/readers.pyi#L39

Related issue(s)

Follow-up to #1917 (review)

Checklist

• Automated checks pass
• Check if you need to add a changelog message
• Check if you need to add tests
• Check if you need to update docs

[codecov]

Codecov Report

❌ Patch coverage is 80.00000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 74.45%. Comparing base (57c36c4) to head (9312413).
⚠️ Report is 7 commits behind head on master.

Files with missing lines	Patch %	Lines
augur/io/metadata.py	80.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1923      +/-   ##
==========================================
- Coverage   74.46%   74.45%   -0.02%     
==========================================
  Files          82       82              
  Lines        9098     9102       +4     
  Branches     1853     1855       +2     
==========================================
+ Hits         6775     6777       +2     
  Misses       2020     2020              
- Partials      303      305       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[jameshadfield]

The cost of overloaded signatures allows us to infer the return type based on the chunk_size type, yeah? I'm somewhat ambivalent about this as I find overload types a little confusing, but I presume this has been useful to you when writing scripts etc?

[victorlin]

Yeah, I added this to address the pyright error in #1961:

augur/augur/frequencies.py

Line 142 in 86e3991

metadata = read_metadata(

augur/augur/frequencies.py

Line 169 in 86e3991

dates, metadata[METADATA_DATE_COLUMN],

error: "__getitem__" method not defined on type "Generator[Unknown, None, None]" (reportIndexIssue)

[jameshadfield]

I considered adding types for read_csv_with_index_col() but decided against it, as it relies on fragile imports from pandas._typing.¹

I don't really know the state of python typing, how does this tend to work for non-core libraries? The TS community has been really proactive about publishing separate type libraries for most common JS-only packages, although ensuring the versions are in-sync has been a little painful. Pandas clearly has some types (pd.DataFrame), do they hide FilePath, ReadCsvBuffer etc away because the types aren't stable?

[victorlin]

pandas-stubs is basically the same thing (separate type library) and it's official, so sync is not much of an issue.

[victorlin]

I'm not sure why pandas._typing is hidden away, but it's explicitly noted in docs.

pandas.api.typing should be stable. v3 includes FilePath and ReadCsvBuffer, but we still target compatibility with v1.