Docs: README is stale — shows synchronous ML, omits RQ/SSE/MinIO/monitoring

[Alimedhat000]

Summary

README.md is stale and no longer describes the real system. Its sequence diagram shows synchronous ML inference (ML-->Tests: Analysis results) and there is zero mention of RQ workers, Redis pub/sub + SSE, MinIO/S3 storage, or the Prometheus/Grafana/Loki monitoring stack.

Evidence

• README.md sequence diagram (lines ~114-149): requests get analysis results directly in the response.
• Real flow: routes enqueue RQ jobs (routes/upload_routes.py:33-35 get_ml_queue, queue ml), an RQ worker runs ml/tasks.run_inference and writes scores to the DB asynchronously.
• ESP32 push events use Redis pub/sub + SSE (utils/esp32_connection_manager.py, routes/esp32_routes.py:95 /api/esp32/stream) — undocumented in README.
• Storage is local uploads/ or MinIO/S3 behind STORAGE_BACKEND — undocumented.
• Deploy is Oracle Cloud via GHCR + watchtower — undocumented.
• README also documents /api/test/... routes that are actually /api/tests/... (both routes/test_routes.py and routes/upload_routes.py use /api/tests).

The accurate docs already exist: API_ROUTES.md (~50KB) and docker/README.md.

Suggested fix

Rewrite README.md to match the current architecture: async inference via RQ, event-delivery pipeline (Redis pub/sub + SSE connection manager), hybrid storage, and the deployment topology. Keep the utility-script section (it's current). Update the stale /api/test references to /api/tests.