feat: add type validation to params when tool is invoked

[twishabansal]

Adds type validation to inputs when a tool is invoked using pydantic.

[twishabansal]

I see base model also has a signiture -- can we reuse theirs instead of creating our own? https://docs.pydantic.dev/latest/api/base_model/#pydantic.BaseModel

There is also a .model_fields option -- maybe we use this to replace parameters to address my comment in the other PR about not using ToolSchema directly?

By default, Pydantic models treat all fields without defaults as keyword-only arguments in their generated init signature.

This means, if we use the pydantic model signature (self.__pydantic_model.__signature__),
tool calls like loaded_tool("some value") == "ok" will fail with errors: TypeError: too many positional arguments.
but tool calls like loaded_tool(param1="some value") == "ok" will pass.

I tried using kw_only=False (ref) as a constraint. But it still shows up as a keyword argument in the generated signature.