[prassu017]

Link to Issue or Description of Change

1. Link to an existing issue (if applicable):

• Closes: ApigeeLlm (Chat Completions) ignores FunctionCallingConfig.allowed_function_names and the VALIDATED mode #8100

Problem:
CompletionsHTTPClient._map_tools maps only FunctionCallingConfig.mode. With mode=ANY and allowed_function_names=["lookup_order"], the Chat Completions request carries "tool_choice": "required", so the model may call any declared tool.

Solution:
When mode=ANY has exactly one allowed name, send the named choice {"type": "function", "function": {"name": ...}}, the same rule AnthropicLlm._build_tool_choice uses. Other cases are unchanged ("required" for zero or several names). This PR does not cover VALIDATED or a multi-name subset (allowed_tools / filtering tools), which the issue also mentions; that needs a design choice from maintainers.

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

New test_construct_payload_any_mode_with_allowed_function_names (one name, two names, none). On main (59b0dd6) the one-name case fails with assert 'required' == {'function': {'name': 'lookup_order'}, 'type': 'function'}; with the fix:

pytest tests/unittests/models/test_apigee_llm.py
81 passed

Only that file ran (Python 3.13); the full suite and tox were not run. pyink --check and isort --check-only pass on both files.

Manual End-to-End (E2E) Tests:

Not run (no Apigee proxy). The unit test asserts the payload sent to /chat/completions.

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.

Additional context

AI-assisted; tests run locally.