Phase 1: Fluent API with .with_detail_response() pattern

[tpellissier-msft]

Summary

This PR implements Phase 1 of the SDK restructure plan, introducing the fluent .with_detail_response() API pattern for accessing telemetry while maintaining full backward compatibility.

• PR 1.1: Add core result types (RequestMetadata, DataverseResponse, FluentResult)
• PR 1.2: Add internal metadata capture to _odata.py and _http.py
• PR 1.3: Integrate FluentResult into client.py with backward compatibility

Backward Compatibility

Existing code continues to work unchanged:

# Old code still works
ids = client.create("account", {"name": "Contoso"})
print(ids[0])  # Works via __getitem__
for id in ids:  # Works via __iter__
    print(id)

New Telemetry Access

# New code can access telemetry
response = client.create("account", {"name": "A"}).with_detail_response()
print(response.telemetry['timing_ms'])      # Request duration
print(response.telemetry['client_request_id'])  # For tracing
print(response.telemetry['batch_info'])     # For bulk operations

Key Changes

New Types in core/results.py:

• RequestMetadata - HTTP request/response metadata dataclass
• DataverseResponse[T] - Generic response with result + telemetry dict
• FluentResult[T] - Wrapper enabling .with_detail_response() pattern

Updated Methods in client.py:

• create() → FluentResult[List[str]]
• update() → FluentResult[None]
• delete() → FluentResult[Optional[str]]
• get() (single record) → FluentResult[Dict]

Test plan

• All 81 existing unit tests pass
• New tests for RequestMetadata, DataverseResponse, FluentResult (44 tests)
• Updated test_client.py tests for new _with_metadata methods
• Backward compatibility tests for iteration, indexing, equality

🤖 Generated with Claude Code