Align accepted filetypes with docstring description

[maxrjones]

The docs state the filetype Can be one of {‘netCDF3’, ‘netCDF4’, ‘HDF’, ‘TIFF’, ‘GRIB’, ‘FITS’, ‘dmrpp’, ‘zarr_v3’, ‘kerchunk’}" but in reality few of these arguments actually worked due to case sensitivity.

[TomNicholas]

Thanks! Is the string type tested explicitly anywhere?

In general this function signature is a little messy, especially as we have file types and backends. I think once we generalise this into an entry point we should clean it up.

[TomNicholas]

Also TIFF and GRIB don't currently work

[maxrjones]

Also TIFF and GRIB don't currently work

TIFF doesn't even work for Kerchunk-backed virtualization?

Edit, nevermind forgot about #291 😓

[maxrjones]

Thanks! Is the string type tested explicitly anywhere?

String type is tested, but a non-string and non-FileType type isn't. I'm working on getting to 100% code coverage for this section now (apologies for not thinking to put this as a draft first)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 78.90%. Comparing base (7f1232f) to head (3a69e66).
Report is 2 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff             @@
##             main     #361       +/-   ##
===========================================
+ Coverage   58.80%   78.90%   +20.10%     
===========================================
  Files          32       32               
  Lines        1903     1906        +3     
===========================================
+ Hits         1119     1504      +385     
+ Misses        784      402      -382     

Files with missing lines	Coverage Δ
virtualizarr/backend.py	97.10% <100.00%> (+7.70%)	⬆️

... and 16 files with indirect coverage changes

[TomNicholas]

TIFF doesn't even work for Kerchunk-backed virtualization?

Not right now, see #291

(apologies for not thinking to put this as a draft first)

No worries at all - thank you for taking initiative to improve this!

[maxrjones]

I'm confident the test failures are unrelated to this PR but won't have time to investigate more until Thursday or Friday

[TomNicholas]

(failure is #371)