feat: expose the wrapped callable on FunctionTool (closes #3381)

[fede-kamel]

Summary

Adds a public FunctionTool.function attribute that holds the original Python callable wrapped by @function_tool.

Today the wrapped function is only reachable by walking the closure of on_invoke_tool (...on_invoke_tool._invoke_tool_impl.__closure__), which is brittle and got an extra hop in 0.16 with the failure-handling indirection. This gives code outside the Agents runtime a stable hook to introspect, directly test, or re-run a tool's body without constructing a ToolContext and doing a JSON encode/decode cycle.

• New field is None for tools not backed by a plain function (agent-as-tool, hosted-tool wrappers), so the attribute is always present and safe to check.
• Added as a keyword-only dataclass field, so it does not change the positional constructor order of FunctionTool.
• Documented via the field docstring (renders in the API reference).

This revives the approach from #2146 (which was closed only over an unrelated pytest fixture-collection regression, not the concept). Happy to rename to func or also wire __wrapped__ for inspect.unwrap() if preferred.

Test plan

• New tests/test_function_tool_exposes_function.py: sync + async tools expose a runnable .function; directly-constructed tools expose None.
• tests/test_function_tool.py (49 tests) still pass — no regression.
• make format, make lint, make typecheck (mypy + pyright) clean.

Issue number

Closes #3381

Checks

• I've added new tests (if relevant)
• I've added/updated the relevant documentation
• I've run make lint and make format
• I've made sure tests pass

[fede-kamel]

Following up on the duplicate label — agreed, this overlaps with the earlier #3396, which exposes the same original callable (there as FunctionTool.func; here as FunctionTool.function). The implementations are close to equivalent, so I'm happy to close this one in favor of #3396 if that's the preferred candidate — or to fold anything useful from this PR's tests into it. Whichever keeps the review load smallest.

[fede-kamel]

Closing in favor of the earlier #3396, which covers the same ground. Hope the extra test angles here are useful as a reference if needed.