fix(oci): preserve all parallel tool results in OCI GenAI messages

[bigfish-emily]

Link to Issue or Description of Change

Closes: #6589

Problem:

When an agent backed by OCIGenAILlm issues two or more parallel tool calls in
a single turn, only the first tool result is forwarded to the OCI model API on
the next request; all other results are silently dropped. Root cause:
_content_to_oci_message() collects every function_response part into
tool_results but builds only a single ToolMessage from
tool_results[0], and _build_chat_details() maps each Content to exactly
one message without flattening.

Solution:

Mirror the LiteLLM path (_content_to_message_param in
models/lite_llm.py): when a Content carries multiple function_response
parts, return one ToolMessage per result, and flatten the returned messages
in _build_chat_details() so parallel tool results are all preserved.

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
  • test_content_to_oci_message_multiple_function_responses: two
    function_response parts yield two ToolMessages with the correct tool
    call ids.
  • test_build_chat_details_flattens_multiple_tool_messages: a single
    Content with two parallel tool results produces two flattened messages
    in the outgoing chat request.
• All unit tests pass locally (the new tests are stub-verified against the
  real module in this sandbox; the full suite runs in CI).

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.