Skip to content

Latest commit

 

History

History
73 lines (45 loc) · 2.58 KB

File metadata and controls

73 lines (45 loc) · 2.58 KB

🕹️ Control Room Operator Guide (v2.0.0)

Welcome to the Command Center of the Semantic Memory Engine. Version 2.0.0 transitions SME from a collection of CLI scripts into a unified, containerized forensic laboratory.


🚀 Quick Launch

The v2.0.0 stack is designed to run in Docker to bypass environment conflicts and optimize your GPU (1660 Ti) automatically.

# Pull and start the laboratory
docker-compose up -d

Access Points:


🔌 Navigation: The Three Pillars

1. Connections Manager

Monitor the pulse of your infrastructure.

  • Service Status: Real-time health checks for the Operator, Sidecar, and Database.
  • AI Strategy: Switch between AI providers (Langflow, Ollama, Mock) on the fly without restarting the containers.
  • Hardware Pulse: Monitor VRAM usage to stay within the 6GB limit of the 1660 Ti.

2. The Harvester

The entry point for all semantic data.

  • Semantic Crawl: Input a URL to convert it into atomic facts.
  • Deep Domain Mapping: Enable Deep Crawl to recursively map an entire domain.
  • JS Rendering: Toggle JS Render for modern SPAs (Next.js/React).

3. Tool Lab

Access specialized forensic utilities in one place.

  • Run audits, sniffs, and visual bridges directly from the dashboard.

📂 Forensic Pathing (Container vs. Local)

Important

In v2.0.0, the "Source of Truth" for data is within the container volumes.

Local Path (Host) Container Path Purpose
d:/SME/data/ /app/data/ Knowledge Graph & Centrifuge DB
d:/SME/storage/ /app/storage/ Raw Harvester captures & Logs
d:/SME/extensions/ /app/extensions/ Custom logic & plugin manifests

Note on Manual Scripts: If you run a local script like python src/utils/auditor.py, ensure your environment variables (like SME_DB_PATH) point to the relative paths to match the container's expectations.


🛠️ Troubleshooting

  • Sidecar Offline: Ensure the sme-sidecar container is running (docker logs sme-sidecar).
  • Database Connection: Check if Postgres is healthy at localhost:5432.
  • UI Build Error: If running manually outside Docker, run npm install in the frontend folder.

Next Steps: See ADVANCED_QUICKSTART.md for deeper probe configurations.