[ROSETTA] RAGFlowTeamAPI.from_config ignores ragflow_http_timeout

[isolomatov-gd]

src/rosetta-mcp-server/rosetta_mcp/services/_ragflow_team_api.py, from_config (~line 47-57) constructs the client without passing timeout=, so it falls back to the hardcoded timeout: int = 30 default instead of the env-configurable config.ragflow_http_timeout used elsewhere in the codebase. Team-membership calls (used for read_policy=team authorization) always run with the hardcoded 30s value, inconsistent with the rest of the HTTP client configuration.

Fix: pass timeout=config.ragflow_http_timeout in from_config.

[github-actions]

Rosetta Triage

Classification: bug, good first issue — verified in code; single-line fix, well-scoped, low risk.
Priority assessment: Low. Only affects the timeout value used for read_policy=team membership checks (30s hardcoded default instead of the configured value, default 60s per constants.py); no functional breakage under normal conditions, just inconsistent/harder-to-tune timeout behavior.

Findings:

• Confirmed: RAGFlowTeamAPI.from_config (_ragflow_team_api.py:53-57) constructs the client without timeout=, so it always falls back to the hardcoded timeout: int = 30 default on __init__ (_ragflow_team_api.py:47), ignoring config.ragflow_http_timeout.
• Other RAGFlow HTTP call sites (e.g. tracing.py) correctly read the configurable timeout via config.ragflow_http_timeout / _get_ragflow_http_timeout(), confirming this is an inconsistency rather than intentional.

Suggestions:

• Fix as proposed: pass timeout=config.ragflow_http_timeout in from_config (_ragflow_team_api.py:56).

Description is clear, complete, and directly actionable — no further information needed.

Automated triage by Rosetta agent

[slegarraga]

I will take this one and open a focused fix.