feat: add support for complex number

[tmichela]

HED needs to save complex number for their next experiment.
h5py does support saving complex numbers, but netCDF does not yet, there's an option in xarrays's to_netcdf to save "invalid netCDF" files allowing these features, c.f. https://docs.xarray.dev/en/stable/user-guide/io.html#invalid-netcdf-files

I understand this PR in HDF5 would allow support for these data types in netCDF, but it's not clear when it will be available.

[codecov]

Codecov Report

Attention: Patch coverage is 84.61538% with 6 lines in your changes missing coverage. Please review.

Project coverage is 75.63%. Comparing base (8c90aed) to head (3eb25b9).
Report is 36 commits behind head on master.

Files with missing lines	Patch %	Lines
damnit/api.py	85.71%	3 Missing ⚠️
damnit/ctxsupport/damnit_ctx.py	0.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #374      +/-   ##
==========================================
+ Coverage   75.47%   75.63%   +0.16%     
==========================================
  Files          32       32              
  Lines        5137     5332     +195     
==========================================
+ Hits         3877     4033     +156     
- Misses       1260     1299      +39     

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

[takluyver]

Other than those things, this LGTM

[tmichela]

I just noticed that the run_variables table has a summary_type column that isn't used yet, but sounds like its purpose would be exactly handling this kind of object(?)
@takluyver you think it'd be better to use that column to define the summary type instead of adding a header to the blob data?

[tmichela]

I made the change anyway, let me know what you think

[takluyver]

Thanks, I think that makes sense. I had envisaged the summary_type for where there was a more specialised type based on an SQLite type that we also use, e.g. timestamps stored as floats, but I don't see any problems with using it for custom bytes types as well.

LGTM

[CammilleCC]

+1 to putting the complex data type to summary_type 🥺

Also, LGTM. Thanks for adding the tests!

[tmichela]

thanks both for the review!