fix(agno): normalize user_id and session_id with diagnostic logging (#3404)

[dlowzzxx]

Description

Fixes #3404.

In openinference-instrumentation-agno, user_id and session_id can be provided as integers (or other non-string types) by users or framework run arguments. Currently:

1. Per the OpenInference semantic conventions, user.id and session.id are defined as strings. When non-string values like 123 were passed, they could be exported as non-string attribute values, causing downstream UIs (such as Phoenix or Arize) that expect string IDs to surface them as blank or unindexed.
2. In truthiness checks like if user_id: and if session_id:, numeric zero (0) was silently dropped without any warning or attribute emission.
3. No diagnostic feedback was logged when non-string types were received, making type mismatches difficult to diagnose.

Proposed Changes

• Introduced _normalize_id(value: Any, name: str) -> Optional[str] helper in openinference.instrumentation.agno.utils:
  • Returns None if value is None or an empty/whitespace-only string.
  • Strips strings.
  • If a non-string value is encountered, logs a diagnostic warning (Expected <name> to be a string, but received <value> (type: <type>). Coercing to string.) and coerces to string (str(value)).
  • Correctly preserves numeric 0 as "0".
• Applied _normalize_id in:
  • _runs_wrapper.py: _run_arguments() (normalizing both user_id and session_id, checking is not None instead of truthiness) and _agent_run_attributes() (for Agent and Team runs).
  • _workflow_wrapper.py: _workflow_run_arguments(), _WorkflowWrapper.run(), run_stream(), arun(), and _WorkflowExecuteWrapper.execute() / aexecute().
• Added comprehensive unit tests in tests/test_instrumentor.py:
  • test_run_arguments_user_and_session_id_coercion: validates string coercion, diagnostic logging via caplog, and preservation of 0.
  • test_agent_run_attributes_session_and_user_id_coercion: validates agent instance attribute coercion and 0 preservation.
  • test_workflow_run_arguments_coercion: validates workflow run arguments coercion and diagnostic warnings.
  • test_normalize_id_unit: unit tests for _normalize_id.

Verification

• Ran uv run --extra test pytest across all 159 tests in openinference-instrumentation-agno (100% pass).
• Verified uv run --with ruff ruff check and uv run --with ruff ruff format --check (0 issues, 100% clean).
• Verified strict mypy type checking (Success: no issues found in 7 source files).