Add open_files function to read GPM files from list of file paths

[ghiggi]

Prework

• I understand and agree to this repository's code of conduct.
• I understand and agree to this repository's contributing guidelines.
• I have already submitted an issue or discussion thread to discuss my idea with the maintainers.

What kind of change does this PR introduce? (check at least one)

• Bugfix
• Feature
• Documentation
• Tutorial
• Code style update
• Refactor
• Build-related changes
• Other, please describe:

Does this PR introduce a breaking change? (check one)

• Yes
• No

If yes, please describe the impact and communicate accordingly:

The PR fulfills these requirements:

• It's submitted to the branch named as follow:
  • Fix a bug: bugfix-<some_key>-<word>
  • Improve the doc: doc-<some_key>-<word>
  • Improve a tutorial tutorial-<some_key>-<word>
  • Add a new feature: feature-<some_key>-<word>
  • Refactor some code: refactor-<some_key>-<word>
  • Optimize some code: optimize-<some_key>-<word>
• When resolving a specific issue, it's referenced in the PR's title (e.g. fix #xxx[,#xxx], where "xxx" is the issue number)
• Don't forget to link PR to issue if you are solving one.
• All tests are passing.
• New/updated tests are included

Summary

This PR adds the gpm.open_files function which allows to read a list of GPM files given the specified filepaths.
This PR address #80.

[ghiggi]

HI @kmuehlbauer ! I add this in mind the entire week so I spent 2 hours this morning to implement it.

Can you try if it works well for you use case and report possible improvements?
Especially maybe try it out with parallel=True argument to see if you experience some troubles.

To avoid netCDF locking, I typically run it by initializing a dask client as follow:

import os
os.environ["HDF5_USE_FILE_LOCKING"] = "FALSE"
from dask.distributed import Client, LocalCluster
cluster = LocalCluster(
        n_workers=20,
        threads_per_worker=1, # important to set to 1 to avoid netcdf locking ! 
        processes=True,
  )
 client = Client(cluster)

FYI: The PR tests fails for a minor problem related to an update of the polars library, but this affect a specific functionality of the software which should not concern you. I will fix it as soon as I have time.

[codecov]

Codecov Report

❌ Patch coverage is 89.15663% with 9 lines in your changes missing coverage. Please review.
✅ Project coverage is 91.29%. Comparing base (c5d3d25) to head (61a3bbb).

Files with missing lines	Patch %	Lines
gpm/dataset/coords.py	63.63%	4 Missing ⚠️
gpm/dataset/dataset.py	86.95%	3 Missing ⚠️
gpm/dataset/granule.py	81.81%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #81      +/-   ##
==========================================
+ Coverage   91.18%   91.29%   +0.10%     
==========================================
  Files         135      135              
  Lines       17214    17270      +56     
==========================================
+ Hits        15696    15766      +70     
+ Misses       1518     1504      -14     

☔ 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.

[coveralls]

coverage: 91.291% (+0.1%) from 91.182%
when pulling 61a3bbb on add-open_mfdataset
into c5d3d25 on main.