How qingltc/pyrosync works
PyroSync is a fire-activity harmonization pipeline and Streamlit calendar. Its main workflow turns NASA FIRMS archives into filtered detections, comparable cell-day counts, calibrated MODIS-equivalent estimates, and processed tables that users explore by area and time. The map preserves both ratio and machine-learning approaches, their validation, and calendar analytics. Relationships are limited to calls and data transfers visible in excerpts or documented in the README; some app interactions and unsampled implementation details are therefore represented only at subsystem level.
Main entry points: build_dataset.py Dataset pipeline · app.py Streamlit calendar app · grid.py Fire cell-days · series.py Area time series
Components 15
Ingestion and gridding
- Pipeline settings
pyrosync/config.py - Read and filter detections
pyrosync/io.py - Fire cell-days
pyrosync/grid.py - Monthly coarse grid
pyrosync/grid.py - Dataset pipeline
scripts/build_dataset.py
Calibration and validation
- Layer A ratio
pyrosync/calibrate.py - NOAA-20 chain
pyrosync/calibrate.py - Layer B model
pyrosync/calibrate.py - Hold-out validation
pyrosync/calibrate.py
Processed data
- Processed tables
data/processed
Calendar experience
- Streamlit calendar app
app.py - Area time series
pyrosync/series.py - Calendar and anomalies
pyrosync/series.py
Other components
- Area and period user
- FIRMS archives
Connections 15
- FIRMS archives → Read and filter detections: reads archives
README.md - Dataset pipeline → Read and filter detections: loads detections
scripts/build_dataset.py - Dataset pipeline → Fire cell-days: counts cell-days
scripts/build_dataset.py - Fire cell-days → Monthly coarse grid: aggregates monthly
pyrosync/grid.py - Dataset pipeline → Monthly coarse grid: builds grid
scripts/build_dataset.py - Dataset pipeline → Layer A ratio: fits factors
scripts/build_dataset.py - Dataset pipeline → NOAA-20 chain: applies chain
scripts/build_dataset.py - Dataset pipeline → Layer B model: fits and predicts
scripts/build_dataset.py - Dataset pipeline → Hold-out validation: validates methods
scripts/build_dataset.py - Dataset pipeline → Processed tables: writes outputs
scripts/build_dataset.py - Area and period user → Streamlit calendar app: selects area
README.md - Streamlit calendar app → Processed tables: loads tables
app.py - Streamlit calendar app → Area time series: queries series
README.md - Streamlit calendar app → Calendar and anomalies: shows season story
app.py - Area time series → Calendar and anomalies: derives calendar
pyrosync/series.py