Accept pandas Series/ExtensionArray for Data; lift pandas<3 cap

[rly]

Summary

• Accept pd.Series and pandas.api.extensions.ExtensionArray (incl. StringArray/ArrowStringArray) as data in Data and its subclasses, normalizing to numpy at the Data.__init__/Data.extend boundary so every subclass (VectorData, VectorIndex, ScratchData, ElementIdentifiers, …) picks up the fix without per-class changes.
• Reject pd.NA/NaN and pandas nullable numeric/boolean dtypes (IntegerArray, BooleanArray, FloatingArray) with informative TypeErrors — the former would crash at HDF5 vlen-string write time, the latter would silently widen on .to_numpy().
• Lift the pandas<3 cap from pyproject.toml.

Fixes #1384.

Why

Pandas 3.0 makes PyArrow-backed strings the default for all DataFrame string columns. df['col'].values is now ArrowStringArray, so VectorData(name=..., data=df['col'].values) (and any other typical user pattern that hands HDMF a string column) now fails docval type validation. Centralizing the fix at the Data construction boundary means VectorData, add_unit, add_electrode, from_dataframe, etc. all keep working with no further changes.

Behavior

• ArrowStringArray, StringArray, pd.Series (any backing dtype), pd.Categorical → converted to np.ndarray silently.
• pandas input containing pd.NA or NaN → TypeError pointing at the missing-values cause and asking the user to fill with a sentinel.
• IntegerArray/BooleanArray/FloatingArray → TypeError asking the user to cast explicitly (.astype('int64').to_numpy() or .to_numpy(dtype=...)), since defaulting .to_numpy() would silently change the dtype.
• Non-pandas inputs are pass-through; no behavior change for existing callers.

Verification

• Reproducer from Pandas 3.0 String Type Compatibility Breaking HDMF Data Ingestion #1384 now succeeds.
• HDF5 roundtrip on DynamicTable.from_dataframe(df=...) with pandas 3.0.2 default string columns works end-to-end.
• Full unit suite on a pandas 3.0.2 environment: 1801 passed, 111 skipped, 1 xfailed, 0 failed.

Test plan

• Unit tests added for coerce_pandas_data covering StringArray, ArrowStringArray, plain numeric Series, Categorical, NA-bearing inputs, and nullable int/bool.
• End-to-end test through VectorData for both Series and df.values paths.
• Manual HDF5 roundtrip with pandas 3.0.
• CI passes on Python 3.10–3.13 with pandas 1.4 (lower bound), pandas 2.x, and pandas 3.x.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.20%. Comparing base (0d61982) to head (2788270).

Additional details and impacted files

@@            Coverage Diff             @@
##              dev    #1469      +/-   ##
==========================================
+ Coverage   93.18%   93.20%   +0.01%     
==========================================
  Files          41       41              
  Lines       10176    10195      +19     
  Branches     2103     2108       +5     
==========================================
+ Hits         9483     9502      +19     
  Misses        415      415              
  Partials      278      278              

☔ 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.