feat: pin netcdf to >1.6 and add explicit encoding [all tests ci]

[lsetiawan]

Overview

This PR unpins the netCDF4 version and set it to >1.6 so that we are getting the latest. Also, the set_netcdf_encodings utils has been updated so that we set zlib=False for strings... it looks like the older netcdf4 is already doing this under the hood.

[codecov-commenter]

Codecov Report

Merging #1112 (90f4a45) into dev (4b3c70e) will increase coverage by 0.03%.
Report is 3 commits behind head on dev.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##              dev    #1112      +/-   ##
==========================================
+ Coverage   78.14%   78.17%   +0.03%     
==========================================
  Files          65       65              
  Lines        6232     6241       +9     
==========================================
+ Hits         4870     4879       +9     
  Misses       1362     1362              

Flag	Coverage Δ
unittests	78.17% <100.00%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
echopype/utils/coding.py	90.42% <100.00%> (+0.77%)	⬆️

... and 4 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[emiliom]

I've reviewed the PR, and it looks good. Next, I'll run it locally.

[emiliom]

Everything seems to be working great!

Here's the encoding on Beam_group1.channel, with a conda environment built with netcdf4 1.6:

In [7]: ds16 = xr.open_dataset(
   'with_netcdf4v16/Summer2017-D20170615-T190214.nc', group='Sonar/Beam_group1', decode_cf=False
)

In [8]: ds16.channel.encoding
Out[8]:
{'zlib': False,
 'shuffle': False,
 'complevel': 0,
 'fletcher32': False,
 'contiguous': True,
 'chunksizes': None,
 'source': '/home/mayorga/.echopype/temp_output/with_netcdf4v16/Summer2017-D20170615-T190214.nc',
 'original_shape': (3,),
 'dtype': str}

With netcdf4 1.5, the encoding is identical.

File sizes are trivially different, with the netcdf4 1.6 nc file being 10 bytes smaller out of 1.4 MB.

[leewujung]

Thanks @lsetiawan @emiliom !

To close the loop: the changes boiled down to set string type encoding with "zlib" : False. It seems this PR addresses everything in #988. Can you confirm?

[emiliom]

To close the loop: the changes boiled down to set string type encoding with "zlib" : False.

Yup.

It seems this PR addresses everything in #988. Can you confirm?

Ultimately, yes. The core motivation for that issue has been fully addressed. But on re-reading, there are some subtleties that I'll comment on there.