How homssalomssa/geo-rag-platform works
Geo-RAG Platform accepts satellite images and analysis settings through a React UI, submits asynchronous jobs to FastAPI, and returns progress and completed reports. Backend workers extract visual features, reserve a spatial-enrichment stage, retrieve urban-geography knowledge, and generate an answer; the API exposes job, health, cache, and spatial status functions. A separate orchestrator coordinates the same analysis services. The README documents the main stages; some worker links are supported by the source index, while unsampled implementation details are not inferred.
Main entry points: job_queue.py Job Queue · App.jsx App Flow · routes.py HTTP Routes · retriever.py Retrieval Worker
Components 20
Analysis UI
- App Flow
frontend/src/App.jsx - Upload Controls
frontend/src/pages/UploadPage.jsx - Results View
frontend/src/pages/ResultsPage.jsx - System Status
frontend/src/pages/SystemStatusPage.jsx - API Client
frontend/src/utils/api.js
API and Jobs
- HTTP Routes
app/api/routes.py - API Schemas
app/api/schemas.py - Job Queue
app/services/job_queue.py
Analysis Pipeline
- Vision Worker
app/workers/vision_extractor.py - Retrieval Worker
app/workers/retriever.py - Answer Worker
app/workers/answer_generator.py - Analysis Orchestrator
app/core/orchestrator.py
Knowledge and Context
- Result Cache
app/utils/cache_manager.py - Spatial Enrichment
app/utils/spatial_enricher.py - Knowledge Retrieval
app/services/rag_service.py
Local Inference
- Prompt Builder
app/core/prompt_builder.py - Vision Service
app/services/vision_service.py - LLM Service
app/services/llm_service.py
Other components
- Analyst
- Ollama
Connections 30
- Analyst → App Flow: sets analysis
- App Flow → Upload Controls: renders
frontend/src/App.jsx - App Flow → API Client: submits and polls
frontend/src/App.jsx - App Flow → Results View: renders results
frontend/src/App.jsx - App Flow → System Status: renders status
frontend/src/App.jsx - API Client → HTTP Routes: calls API
- HTTP Routes → API Schemas: uses schemas
app/api/routes.py - HTTP Routes → Job Queue: tracks jobs
app/api/routes.py - HTTP Routes → Result Cache: manages cache
app/api/routes.py - HTTP Routes → Spatial Enrichment: checks status
app/api/routes.py - Job Queue → Vision Worker: dispatches stages
app/workers/vision_extractor.py - Job Queue → Retrieval Worker: dispatches stages
app/workers/retriever.py - Job Queue → Answer Worker: dispatches stages
app/workers/answer_generator.py - Vision Worker → Vision Service: extracts features
app/workers/vision_extractor.py - Vision Worker → Result Cache: caches features
app/workers/vision_extractor.py - Vision Worker → Job Queue: updates progress
app/workers/vision_extractor.py - Retrieval Worker → Knowledge Retrieval: retrieves knowledge
app/workers/retriever.py - Retrieval Worker → Spatial Enrichment: enriches context
app/workers/retriever.py - Retrieval Worker → Result Cache: caches retrieval
app/workers/retriever.py - Retrieval Worker → Job Queue: updates progress
app/workers/retriever.py - Answer Worker → Prompt Builder: builds prompt
app/workers/answer_generator.py - Answer Worker → LLM Service: generates answer
app/workers/answer_generator.py - Answer Worker → Result Cache: reads and stores
app/workers/answer_generator.py - Answer Worker → Job Queue: updates progress
app/workers/answer_generator.py - Analysis Orchestrator → Vision Service: extracts features
app/core/orchestrator.py - Analysis Orchestrator → Knowledge Retrieval: retrieves context
app/core/orchestrator.py - Analysis Orchestrator → Prompt Builder: builds prompt
app/core/orchestrator.py - Analysis Orchestrator → LLM Service: generates answer
app/core/orchestrator.py - Vision Service → Ollama: requests inference
- LLM Service → Ollama: requests generation
app/services/llm_service.py