Accept pandas Series/ExtensionArray in add_scratch; lift pandas<3 cap

[rly]

Motivation

Pandas 3.0 makes ArrowStringArray the default backing for every DataFrame string column, so df['col'].values is now an ExtensionArray rather than an ndarray. NWBFile.add_scratch rejected such values at the PyNWB docval layer before HDMF's coercion could run, and PyNWB capped pandas<3.

The HDMF side was fixed in hdmf-dev/hdmf#1469 (shipped in HDMF 6.1.0), which centralizes coercion of pd.Series/ExtensionArray to numpy at the Data.__init__/Data.extend boundary. Because PyNWB never overrides the array_data macro, that fix propagates automatically to TimeSeries subclasses, add_unit, add_electrode, DynamicTable.from_dataframe, etc. The only PyNWB-specific surface needing a change is add_scratch, which uses its own hand-rolled type tuple.

Context: hdmf-dev/hdmf#1384, hdmf-dev/hdmf#1469.

How to test the behavior

import pandas as pd
from pynwb import NWBFile
from datetime import datetime, timezone

nwb = NWBFile(session_description='x', identifier='y',
              session_start_time=datetime.now(timezone.utc))
df = pd.DataFrame({'animal': ['cat', 'dog']})
nwb.add_scratch(df['animal'].values, name='a', description='d')  # ArrowStringArray on pandas 3

Changes

• NWBFile.add_scratch: accept pd.Series and pandas.api.extensions.ExtensionArray in the data docval type tuple and the array-branch isinstance check; updated docstring and error messages. Coercion to numpy happens in HDMF's Data.__init__, not in PyNWB.
• pyproject.toml: lifted the pandas<3 cap and bumped the minimum HDMF dependency to >=6.1.0.
• tests/unit/test_scratch.py: added tests passing pd.Series (numeric and string), StringArray, and (pyarrow-guarded) ArrowStringArray; updated the widened error-message assertions.

Verified the full unit suite passes under pandas 3.0.2 + pyarrow + HDMF with #1469.

Checklist

• Did you update CHANGELOG.md with your changes?
• Have you checked our Contributing document?
• Have you ensured the PR clearly describes the problem and the solution?
• Is your contribution compliant with our coding style? This can be checked running ruff from the source directory.
• Have you checked to ensure that there aren't other open or previously closed Pull Requests for the same change?
• Have you included the relevant issue number using "Fix #XXX" notation where XXX is the issue number? By including "Fix #XXX" you allow GitHub to close issue #XXX when the PR is merged.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.29%. Comparing base (dc3ff5e) to head (ac81125).

Additional details and impacted files

@@           Coverage Diff           @@
##              dev    #2208   +/-   ##
=======================================
  Coverage   95.29%   95.29%           
=======================================
  Files          30       30           
  Lines        3038     3039    +1     
  Branches      450      450           
=======================================
+ Hits         2895     2896    +1     
  Misses         87       87           
  Partials       56       56           

Flag	Coverage Δ
integration	73.14% <100.00%> (+<0.01%)	⬆️
unit	85.98% <100.00%> (+<0.01%)	⬆️

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

☔ View full report in Codecov by Harness.
📢 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.

[Copilot]

Pull request overview

This PR updates PyNWB to be compatible with pandas 3.x by allowing NWBFile.add_scratch to accept pandas.Series and pandas ExtensionArray inputs (including pandas 3’s PyArrow-backed string arrays), relying on HDMF’s newer coercion behavior. It also lifts the pandas<3 dependency cap and raises the minimum HDMF version accordingly.

Changes:

• Expanded NWBFile.add_scratch input validation to accept pd.Series and pandas.api.extensions.ExtensionArray, and updated related docs/error messages.
• Updated dependencies to require hdmf>=6.1.0 and removed the pandas<3 cap.
• Added unit tests for add_scratch with pd.Series, StringArray, and (optionally) PyArrow-backed string arrays; updated changelog.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
tests/unit/test_scratch.py	Adds coverage for add_scratch with Series and extension arrays, plus a pyarrow-guarded Arrow string array case.
src/pynwb/file.py	Extends add_scratch docval/type checks to accept Series and ExtensionArray inputs and updates messaging/docs.
pyproject.toml	Bumps minimum HDMF to 6.1.0 and removes the pandas<3 constraint.
CHANGELOG.md	Documents pandas 3 compatibility and the HDMF minimum version bump rationale.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[oruebel]

@rly overall this looks good to me, but copilot suggested a couple of changes that look important. Ping me when I should re-review to approve.