fix: support TypedDict schema generation

[markphelps]

Fixes: #2973

Summary

• fix legacy runtime schema generation so TypedDict inputs are treated like dicts without accepting arbitrary dict subclasses
• add static schema support for TypedDict inputs and outputs, including qualified imports, inheritance, and Required / NotRequired key semantics
• add regression coverage in Python, Go parser tests, and integration txtar scripts for static and legacy TypedDict paths

Notes

• added integration txtar coverage for typeddict_input, typeddict_input_legacy, and typeddict_output

[ask-bonk]

I'm Bonk, and I've done a quick review of your PR.

This PR adds comprehensive TypedDict support for both static schema generation (Go parser) and legacy runtime schema generation (Python ADT). The implementation correctly handles:

• TypedDict inputs treated as opaque dicts (ANY type)
• TypedDict outputs with proper field extraction
• Required/NotRequired field qualifiers
• TypedDict inheritance (parent classes)
• Cross-file TypedDict imports
• total=False semantics

All Go tests pass. The integration test coverage is excellent with three txtar scripts covering static input, legacy input, and output paths.

One minor issue found:

[ask-bonk]

I've posted a review on PR #2978 identifying one minor issue where qualifiedEntry is assigned but never used in the error handling code at pkg/schema/types.go:320. The review is now visible on the PR.

github run

[ask-bonk]

LGTM

github run

[ask-bonk]

Looking at PR #2978 which adds TypedDict support for both static schema generation (Go parser) and legacy runtime schema generation (Python ADT), I can see the implementation is comprehensive and well-tested.

From the context provided, I can see that:

1. A previous review already identified and addressed the qualifiedEntry issue in pkg/schema/types.go
2. The author acknowledged the parameter count concern for some functions

The implementation correctly handles:

• TypedDict inputs treated as opaque dicts (ANY type)
• TypedDict outputs with proper field extraction
• Required/NotRequired field qualifiers
• TypedDict inheritance (parent classes)
• Cross-file TypedDict imports
• total=False semantics

All Go tests pass and the integration test coverage includes three txtar scripts. The Python ADT changes properly check for TypedDict using typing.is_typeddict.

LGTM

github run

[ask-bonk]

LGTM

github run

[ask-bonk]

LGTM

github run

[michaeldwan]

nice fix