google-gemini
Support python `Literal` types when defining JSON or FunctionCalling schemas.
Description of the feature request:
String literal types should create Enums in the generated schema.
What problem are you trying to solve with this feature?
No response
Any other information you'd like to share?
No response
@MarkDaoust you mention "generated" - is there a private method that generates protos.Schema somewhere ?
I was about to write a feature request for a from_object() like in node.js to convert a pydantic object to Schema.
Hi @yan-hic,
The code is here:
https://github.com/google-gemini/generative-ai-python/blob/8f77cc6ac99937cd3a81299ecf79608b91b06bbb/google/generativeai/types/content_types.py#L337-L342
It uses pydantic to do the conversion already, but I haven't tested actually passing pydantic objects.
Similarly, I haven't tried this with types like Literal['a', 'b'].
These seem to work, as long as they're nested inside a class, just not at the top level.