How spandandey961-max/regrow works
ReGrow is a browser-based land-observation dashboard for Amaravati. Users choose an observation date and study sector, inspect mapped cover and paired changes over comparable usable cells, review vegetation and water histories, and explore a conditional restoration scenario. The repository also includes a reproducible HLS/WorldCover reference-model workflow and optional NISAR catalogue inspection; radar measurements are not used in dashboard classifications. The graph separates documented product behavior from sampled runtime wiring; unsampled connections are omitted rather than inferred.
Main entry points: inference-worker.js Browser inference · select_data.py HLS selection · prepare_train.py Reference model training · app.js Land dashboard
Components 18
Dashboard experience
- Land dashboard
web/app.js - Sector analysis
web/area-analysis.js - Planning outlook
web/outlook.js - Model evidence view
web/design.js
Inference runtime
- Hosting API adapter
web/hosting-api.js - Browser inference
web/inference-worker.js
Observation and model workflow
- Observation snapshots
data/dashboard.json - Fitted forest
data/forest.json - HLS selection
scripts/select_data.py - Source discovery
scripts/discover.py - Reference model training
scripts/prepare_train.py - Training evaluation
data/training-report.json
Radar exploration
- NISAR discovery
scripts/discover_sar.py - NISAR subset inspection
scripts/nisar_subset.py - NISAR catalogue records
data/sar/selected-nisar.json
Other components
- Dashboard visitor
- HLS catalogue service
- WorldCover data
Connections 12
- Dashboard visitor → Land dashboard: explores
README.md - Land dashboard → Sector analysis: opens sector report
README.md - Sector analysis → Planning outlook: renders scenario
web/area-analysis.js - Hosting API adapter → Browser inference: dispatches prediction
web/hosting-api.js - Browser inference → Observation snapshots: loads scenes
web/inference-worker.js - Browser inference → Fitted forest: loads model
web/inference-worker.js - HLS selection → Source discovery: imports search helper
scripts/select_data.py - HLS selection → HLS catalogue service: searches catalogue
scripts/select_data.py - Reference model training → HLS catalogue service: reads HLS assets
scripts/prepare_train.py - Reference model training → WorldCover data: reads reference labels
scripts/prepare_train.py - NISAR discovery → NISAR catalogue records: records catalogue
- NISAR subset inspection → NISAR catalogue records: selects products