How praneyaarora/satquerryai works
SatQuery is an Earth-observation assistant: users upload GeoTIFFs and ask natural-language questions, then receive validated raster context, tool-driven geospatial analyses, and a summary through the React UI, REST API, or MCP interface. The graph emphasizes the documented upload and query workflows and preserves the major imagery, weather, analysis, vision, and workflow capabilities. Some orchestrator internals and route wiring are unsampled; uncertain connections are omitted rather than inferred.
Main entry points: executor.py Tool executor · satquery_server.py MCP tool server · graph.py Query orchestration · main.py FastAPI app
Components 29
User interfaces
- React chat UI
frontend/src/App.jsx - MCP tool server
satquery_server.py - Results and artifacts
API and ingestion
- FastAPI app
backend/main.py - Query routes
backend/api/routes/query.py - Raster upload
backend/api/routes/uploads.py - Raster preview routes
backend/api/routes/rasters.py - Raster ingestion
backend/orchestrator/ingest_graph.py
Planning and execution
- Query state
backend/orchestrator/state.py - Query orchestration
backend/orchestrator/graph.py - LLM planning
backend/orchestrator/llm.py - Tool registry
backend/orchestrator/registry.py - Tool executor
backend/tools/executor.py
Earth observation tools
- Mission workflows
satquery_workflows.py - Optical imagery
Tool_1_fetch_optical_imagery/fetch_optical_imagery.py - Multispectral imagery
Tool_2_fetch_multispectral_imagery/fetch_multispectral_imagery.py - SAR imagery
Tool_3_fetch_sar_imagery/fetch_sar_imagery.py - Weather data
Tool_4_fetch_weather_environment/fetch_weather_environment.py - Vegetation indices
Tool_5_compute_vegetation_indices/compute_vegetation_indices.py - GeoTIFF metadata
Tool_6_inspect_geotiff_metadata/inspect_geotiff_metadata.py - Temporal change
Tool_7_analyze_temporal_change/analyze_temporal_change.py - Landcover terrain
Tool_8_analyze_spatial_landcover_terrain/analyze_spatial_landcover_terrain.py - Web intelligence
Tool_9_fetch_web_intelligence/fetch_web_intelligence.py - Geocoding and POI
Tool_10_spatial_geocoding_poi/spatial_geocoding_poi.py - Affine markup
Tool_11_deterministic_affine_markup/deterministic_affine_markup.py
Vision and rendering
- Raster rendering
backend/rendering/raster_preview.py - Vision providers
backend/vision - Local model service
local_model_server/server.py
Other components
- User
Connections 45
- User → React chat UI: asks or uploads
- React chat UI → FastAPI app: HTTP requests
- FastAPI app → Query routes: mounts routes
backend/main.py - FastAPI app → Raster upload: mounts routes
backend/main.py - FastAPI app → Raster preview routes: mounts routes
backend/main.py - Query routes → Query orchestration: runs query
backend/api/routes/query.py - Raster upload → Raster ingestion: runs ingestion
backend/api/routes/uploads.py - Raster ingestion → Tool executor: executes metadata and geocode tools
backend/orchestrator/ingest_graph.py - Raster ingestion → GeoTIFF metadata: validates raster
backend/orchestrator/ingest_graph.py - Query routes → Query state: uses query state
backend/api/routes/query.py - Query orchestration → Query state: shares state
backend/orchestrator/graph.py - Query orchestration → LLM planning: plans requests
backend/orchestrator/graph.py - Query orchestration → Tool registry: loads tool registry
backend/orchestrator/graph.py - Tool executor → Tool registry: resolves tools
backend/tools/executor.py - Query orchestration → Tool executor: dispatches tools
- Tool executor → Optical imagery: calls
backend/tools/executor.py - Tool executor → Multispectral imagery: calls
backend/tools/executor.py - Tool executor → SAR imagery: calls
backend/tools/executor.py - Tool executor → Weather data: calls
backend/tools/executor.py - Tool executor → Vegetation indices: calls
backend/tools/executor.py - Tool executor → GeoTIFF metadata: calls
backend/tools/executor.py - Tool executor → Temporal change: calls
backend/tools/executor.py - Tool executor → Landcover terrain: calls
backend/tools/executor.py - Tool executor → Web intelligence: calls
backend/tools/executor.py - Tool executor → Geocoding and POI: calls
backend/tools/executor.py - Tool executor → Affine markup: calls
backend/tools/executor.py - Tool executor → Mission workflows: dispatches missions
backend/tools/executor.py - MCP tool server → Optical imagery: exposes tools
satquery_server.py - MCP tool server → Multispectral imagery: exposes tools
satquery_server.py - MCP tool server → SAR imagery: exposes tools
satquery_server.py - MCP tool server → Weather data: exposes tools
satquery_server.py - MCP tool server → Vegetation indices: exposes tools
satquery_server.py - MCP tool server → GeoTIFF metadata: exposes tools
satquery_server.py - MCP tool server → Temporal change: exposes tools
satquery_server.py - MCP tool server → Landcover terrain: exposes tools
satquery_server.py - MCP tool server → Web intelligence: exposes tools
satquery_server.py - MCP tool server → Geocoding and POI: exposes tools
satquery_server.py - MCP tool server → Affine markup: exposes tools
satquery_server.py - Raster preview routes → Raster rendering: renders preview
backend/api/routes/rasters.py - Tool executor → Raster rendering: renders vision input
backend/rendering/raster_preview.py - Raster rendering → Affine markup: uses markup engine
backend/rendering/raster_preview.py - Tool executor → Vision providers: invokes vision
backend/tools/executor.py - Vision providers → Local model service: optional local inference
- Query routes → Results and artifacts: returns query results
backend/api/routes/query.py - React chat UI → Results and artifacts: shows results