Rename ext_array._pa_array to _chunked_array

[hombit]

pandas' ArrowExtensionArray uses ._pa_array for the underlying array data, but it is a bit confusing because it is actually a chunked array. This PR changes this name in our implementation and also introduces a public chunked_array property.

It adds a data model description to NestedExtensionArray docs.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.76%. Comparing base (0ed3746) to head (d193826).
Report is 66 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #94   +/-   ##
=======================================
  Coverage   98.76%   98.76%           
=======================================
  Files          15       15           
  Lines         890      893    +3     
=======================================
+ Hits          879      882    +3     
  Misses         11       11           

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

[github-actions]

Before [0ed3746]	After [c7fd4f3]	Ratio	Benchmark (Parameter)
54.9±2ms	57.0±2ms	1.04	benchmarks.ReassignHalfOfNestedSeries.time_run
5.13±0.05ms	5.22±0.3ms	1.02	benchmarks.NestedFrameReduce.time_run
257M	260M	1.01	benchmarks.AssignSingleDfToNestedSeries.peakmem_run
270M	273M	1.01	benchmarks.ReassignHalfOfNestedSeries.peakmem_run
86.5M	86.5M	1	benchmarks.NestedFrameAddNested.peakmem_run
8.73±0.08ms	8.74±0.08ms	1	benchmarks.NestedFrameAddNested.time_run
90.1M	89.7M	1	benchmarks.NestedFrameQuery.peakmem_run
6.13±0.03ms	6.12±0.03ms	1	benchmarks.NestedFrameQuery.time_run
89.4M	89.7M	1	benchmarks.NestedFrameReduce.peakmem_run
32.1±1ms	31.1±1ms	0.97	benchmarks.AssignSingleDfToNestedSeries.time_run

Click here to view all benchmarks.

[dougbrn]

Nice, looks good