Raise a clear error for unknown TypedDict fields in extract_fields

[breken-ai]

When @extract_fields gets a list of field names and the function returns a TypedDict, a name that is not in the TypedDict is recorded in errors. It is then looked up anyway on the next line. Users get a bare KeyError at graph-build time instead of the InvalidDecoratorException that lists the invalid fields.

class MyDict(TypedDict):
    test: int

@extract_fields(["test", "tset"])
def d() -> MyDict: ...

# on main: KeyError: 'tset'
# expected: InvalidDecoratorException: ... The following fields were not valid:
#           ["tset is not a field in the `TypedDict` ..."]

The dict form (extract_fields({"tset": int})) already reports this properly, because that branch continues after recording the error.

Changes

• hamilton/function_modifiers/expanders.py: in _determine_fields_to_extract, continue after recording an unknown field, matching the dict branch just below it.
• tests/function_modifiers/test_expanders.py: one more case in test_extract_fields_invalid_annotations_for_inferred_types, ("MyDict", (["test", "not_a_field"])).

How I tested this

• New case on unmodified main (661aa888): fails with KeyError: 'not_a_field'.
• With the fix, it passes (the expected InvalidDecoratorException is raised).
• pytest tests/function_modifiers tests/test_function_modifiers.py tests/test_end_to_end.py tests/test_hamilton_driver.py tests/test_async_driver.py: 528 passed.
• ruff check and ruff format --check (0.15.14) are clean. git diff --check is clean.

Notes

This was introduced with the list-of-names support for TypedDict in d601a92 (2025-07).

Checklist

• PR has an informative and human-readable title (this will be pulled into the release notes)
• Changes are limited to a single goal (no scope creep)
• Code passed the pre-commit check & code is left cleaner/nicer than when first encountered.
• Any change in functionality is tested
• New functions are documented (with a description, list of inputs, and expected output) — no new functions
• Placeholder code is flagged / future TODOs are captured in comments — none added
• Project documentation has been updated if adding/changing functionality — no documented behavior changes; this makes the code match the existing docs

---

AI disclosure (per the ASF Generative Tooling guidance): this change was written with an AI coding tool (Claude Code, Claude Opus 5.5) working through the breken-ai account. The tool found the bug, wrote the fix and the test, and wrote this description. The commit carries a Generated-by: trailer. The diff is a small original change to existing Hamilton code and includes no third-party material. The red/green runs above are real and can be re-run from the diff. If you would rather not take AI-assisted contributions here, say so and I will close this.