Trick extension array formatting

[hombit]

This PR fixes #50 with a temporal hack. We should open a pandas issue and work with pandas maintainers to contribute a proper solution to this problem.

Examples of formatting with this PR

>>> from nested_pandas.datasets import generate_data

>>> generate_data(3, 1)
          a         b                                        nested
0  0.263827  0.453190   [{t: 5.850555, flux: 84.944634, band: 'g'}]
1  0.561989  0.788521   [{t: 16.96538, flux: 81.045302, band: 'g'}]
2  0.752709  1.946682  [{t: 11.489896, flux: 52.094699, band: 'r'}]

>>> generate_data(3, 3)
          a         b                                             nested
0  0.037709  1.962112  [{t: 16.999902, flux: 65.742084, band: 'g'}; …...
1  0.278047  1.636276  [{t: 0.81424, flux: 11.695323, band: 'g'}; …] ...
2  0.460416  0.109109  [{t: 13.668602, flux: 15.618081, band: 'r'}; …...

>>> nf = generate_data(3, 3)
>>> nf.nested.nest['t'] = nf.nested.nest['t'].round(1)
>>> nf.nested.nest['flux'] = nf.nested.nest['flux'].round(1)
>>> nf
          a         b                                          nested
0  0.789335  0.542048   [{t: 5.8, flux: 92.2, band: 'g'}; …] (3 rows)
1  0.631588  0.953179  [{t: 13.6, flux: 59.3, band: 'r'}; …] (3 rows)
2  0.215239  0.840790   [{t: 7.6, flux: 25.3, band: 'g'}; …] (3 rows)

[codecov]

Codecov Report

Attention: Patch coverage is 76.59574% with 11 lines in your changes missing coverage. Please review.

Project coverage is 98.47%. Comparing base (bbe1866) to head (54dac6c).
Report is 173 commits behind head on main.

Files with missing lines	Patch %	Lines
src/nested_pandas/series/ext_array.py	76.59%	11 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #195      +/-   ##
==========================================
- Coverage   99.44%   98.47%   -0.97%     
==========================================
  Files          13       13              
  Lines        1074     1118      +44     
==========================================
+ Hits         1068     1101      +33     
- Misses          6       17      +11     

☔ 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.

[github-actions]

Before [bbe1866] <v0.3.4>	After [80f4ab0]	Ratio	Benchmark (Parameter)
32.8±2ms	46.5±8ms	~1.42	benchmarks.ReassignHalfOfNestedSeries.time_run
10.1±0.06ms	10.6±0.3ms	1.06	benchmarks.NestedFrameAddNested.time_run
24.9±0.3ms	25.1±0.7ms	1.01	benchmarks.AssignSingleDfToNestedSeries.time_run
269M	269M	1.00	benchmarks.AssignSingleDfToNestedSeries.peakmem_run
91.6M	91.6M	1.00	benchmarks.NestedFrameAddNested.peakmem_run
96.1M	96M	1.00	benchmarks.NestedFrameQuery.peakmem_run
10.7±0.2ms	10.8±0.03ms	1.00	benchmarks.NestedFrameQuery.time_run
95.3M	95.2M	1.00	benchmarks.NestedFrameReduce.peakmem_run
289M	288M	1.00	benchmarks.ReassignHalfOfNestedSeries.peakmem_run
1.25±0.02ms	1.23±0.2ms	0.99	benchmarks.NestedFrameReduce.time_run

Click here to view all benchmarks.

[dougbrn]

Thanks a ton for providing a detailed comment in the code for this. I don't have any real reservations about using this trick and think the output here is really useful.