How mrvenom17/sihp-intelrisk works
Project type and purpose
• Full-stack, real-time disaster intelligence platform.
• Front-end: React + TypeScript SPA for interactive mapping, dashboards, fact-checking.
• Back-end: Node.js + Express API (JWT auth, rate limiting, validation) plus a PostgreSQL database.
• AI Processing Pipeline: Python modules (extractor, detecter, analyser, checker) orchestrated together.
File structure
• Top level:
• Front-End/ – React app (components, hooks, services).
• Back-End/ – Express API (config, middleware, routes, services).
• src/ – Python "agents" (analyser, checker, detecter, extractor), core, schema, orchestrator.
• server/ – Python API mock server (for local microservice testing).
• docs/ – ADRs detailing design of each Python agent and orchestrator.
• benchmarks/ – performance tests for each Python agent.
• Patterns:
• Front vs Back separation
• Node API → Python microservices pipeline
• MVC-style on Express (routes → services → database)
• Microservice/pipeline pattern for AI stages
README insights
• Tech stack summary (React, Express, PostgreSQL, JWT, Joi, Tailwind, Leaflet).
• Feature list (hotspots, fact checker, export).
• Environment variables (.env).
• API endpoints under /api/v1 for auth, hotspots, fact-check, export.
• Quick start scripts (start-dev.sh, docker-compose.yml).
Detail
• Inside the AI Pipeline cluster, detail each stage (extractor, detecter, analyser, checker) and the orchestrator flow.
• Data formats flowing between components (JSON over HTTP, CSV/JSON export).
• External dependencies (OpenStreetMap, third-party AI models if any).