[zhouh (OUAO-FRANK)]

Why are these changes needed?

MemoryContent.mime_type accepts MemoryMimeType | str, and a field_serializer writes an enum member as its string value. There was no matching validator on the way back in, so after a dump_component() / load_component() roundtrip (e.g. saving and restoring a ListMemory) the enum degenerated into a plain str. Any code that dispatches on content.mime_type is MemoryMimeType.X silently changed behaviour after a save/load.

This adds a field_validator(mode="before") that coerces a string matching an enum value back to MemoryMimeType, while leaving arbitrary custom MIME-type strings as str.

Related issue number

Closes #8293

Checks

• I've included any doc changes needed for https://microsoft.github.io/autogen/. See https://github.com/microsoft/autogen/blob/main/CONTRIBUTING.md to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

Testing

Added test_memory_mime_type_roundtrip and test_memory_mime_type_custom_string_stays_str to python/packages/autogen-core/tests/test_memory.py. The roundtrip test fails on the previous code ('text/markdown' vs the enum) and passes now; the custom-string test guards against over-coercing unknown MIME types.

$ pytest tests/test_memory.py -q
12 passed

ruff check and ruff format --check (v0.4.8) are clean on the two changed files.

[zhouh (OUAO-FRANK)]

Closing this myself: I opened it before noticing the project is archived and no longer maintained (README points to microsoft/agent-framework as the successor). No point adding to an unmaintained repo. Sorry for the noise.