WIP: Custom HTML Rendering for Nested Columns

[dougbrn]

Change Description

Resolves #41 .

• My PR includes a link to the issue that I am addressing

Solution Description

This is a WIP attempt at #41. We're able to modify the representation via the Pandas Styler object, however in my current implementation the styler has some tradeoffs with using the default dataframe representation. Particularly the behavior around how tables are truncated is different (large tables are not truncated down to the first and last 5 entries, instead always truncating to the first max_rows rows. In addition, the truncated df does not preview the shape of the dataframe which feels like a big loss. Overall I'm not happy with where this is currently, but I will keep looking to see if the styler can be brought closer in line with the default pandas behavior. It's possible there are other implementations that might be more appropriate than this approach (can we modify the series styling itself?)

[github-actions]

Before [6152580]	After [7de2f7e]	Ratio	Benchmark (Parameter)
34.9±3ms	36.2±3ms	1.04	benchmarks.AssignSingleDfToNestedSeries.time_run
1.08±0.01ms	1.13±0.05ms	1.04	benchmarks.NestedFrameReduce.time_run
6.43±0.1ms	6.57±0.5ms	1.02	benchmarks.NestedFrameQuery.time_run
84M	84M	1	benchmarks.NestedFrameAddNested.peakmem_run
88.1M	88.1M	1	benchmarks.NestedFrameQuery.peakmem_run
87.3M	87.2M	1	benchmarks.NestedFrameReduce.peakmem_run
70.7±8ms	71.0±2ms	1	benchmarks.ReassignHalfOfNestedSeries.time_run
276M	272M	0.99	benchmarks.ReassignHalfOfNestedSeries.peakmem_run
259M	255M	0.98	benchmarks.AssignSingleDfToNestedSeries.peakmem_run
10.1±0.9ms	9.91±0.6ms	0.98	benchmarks.NestedFrameAddNested.time_run

Click here to view all benchmarks.

[codecov]

Codecov Report

Attention: Patch coverage is 14.28571% with 6 lines in your changes missing coverage. Please review.

Project coverage is 98.11%. Comparing base (0bb042a) to head (098a0f2).
Report is 2 commits behind head on main.

Files	Patch %	Lines
src/nested_pandas/nestedframe/core.py	14.28%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #103      +/-   ##
==========================================
- Coverage   98.77%   98.11%   -0.66%     
==========================================
  Files          15       15              
  Lines         895      902       +7     
==========================================
+ Hits          884      885       +1     
- Misses         11       17       +6     

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

[dougbrn]

I've revisted this briefly, this is what this PR can implement. I'm not happy with the loss of information as described above, and am beginning to believe the pandas styler object just does not give us the tools to do something about this.