fix(type-inference): map Any to pa.null() for empty container handling (ENG-389)

[kurodo3]

Summary

• list[Any] and dict[Any, Any] (inferred from empty containers) no longer raise ValueError: Unsupported Python type: typing.Any when passed to python_type_to_arrow_type
• Added Any: pa.null() to _PYTHON_TO_ARROW_MAP in UniversalTypeConverter (forward path)
• Added explicit pa.types.is_null → Any check in _convert_arrow_to_python (reverse path), suppressing a spurious fallback warning
• Removed now-unreachable Any-specific hint from the error branch in _convert_python_to_arrow
• Added 8 converter/round-trip tests and 2 inference documentation tests
• Logged resolved entry UC1 in DESIGN_ISSUES.md

Test plan

• test_any_to_arrow_type — python_type_to_arrow_type(Any) == pa.null()
• test_list_any_to_arrow_type — list[Any] maps to pa.large_list(pa.null())
• test_dict_any_any_to_arrow_type — dict[Any, Any] maps to expected struct
• test_null_arrow_to_any_python_type — pa.null() maps back to Any
• test_list_any_round_trip and test_dict_any_any_round_trip — full round trips
• test_empty_container_inference_to_arrow_no_error — end-to-end, no exception
• test_pyarrow_empty_list_with_null_type — PyArrow accepts empty lists for null-typed columns
• test_infer_empty_list_schema and test_infer_empty_dict_schema — inference already correct
• Full test suite: 3495 passed, 56 skipped, 0 failures

Closes ENG-389

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[Copilot]

Pull request overview

This PR fixes ENG-389 by making typing.Any (commonly produced by empty-container schema inference like list[Any] / dict[Any, Any]) convertible to/from PyArrow types in UniversalTypeConverter, using pa.null() as the Arrow representation for “unknown element type”.

Changes:

• Map typing.Any → pa.null() in the Python→Arrow type map to prevent Unsupported Python type: typing.Any for inferred empty containers.
• Map pa.null() → typing.Any in the Arrow→Python conversion path to ensure clean round-trips and avoid spurious fallback warnings.
• Add targeted tests covering conversion, round-trips, and empty-container inference behavior; document the resolved design issue.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
src/orcapod/semantic_types/universal_converter.py	Adds Any ↔ pa.null() handling and removes now-unreachable Any-specific error hint.
tests/test_semantic_types/test_universal_converter.py	Adds converter + round-trip tests ensuring Any/empty-container inference no longer raises.
tests/test_semantic_types/test_pydata_utils.py	Adds tests documenting that empty containers infer as list[Any] / dict[Any, Any].
DESIGN_ISSUES.md	Records ENG-389 (UC1) as resolved with the implemented approach.
superpowers/specs/2026-06-08-eng-389-empty-container-type-inference.md	Adds an approved spec describing the problem and chosen solution.
superpowers/plans/2026-06-08-eng-389-empty-container-type-inference.md	Adds an implementation plan reflecting the changes and verification steps.

---

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

[kurodo3]

Review round response

Copilot reviewed all 6 changed files and generated no comments — no actionable issues were raised. No code changes are required for this review round.

The implementation is complete as submitted:

• Any: pa.null() mapping added to _PYTHON_TO_ARROW_MAP (forward path)
• pa.types.is_null → Any check added to _convert_arrow_to_python (reverse path)
• Dead Any-specific error hint removed from _convert_python_to_arrow
• 10 new tests (8 converter/round-trip + 2 inference documentation), all passing
• Full suite: 3495 passed, 56 skipped, 0 failures