Usage¶
The two workflows¶
Initial calibration fits a new model to a historical window, once. Global optimizer, train/test split, sensitivity analysis. See Calibration.
Online recalibration keeps a running model on track. Local optimizer, capped step size, variance trigger. See Calibration.
Package layout¶
| Module | Contents |
|---|---|
pyadm1ode_calibration.calibration |
Calibrator, InitialCalibrator, OnlineCalibrator, CalibrationResult, bounds |
pyadm1ode_calibration.calibration.optimization |
optimizers, objective functions, constraints |
pyadm1ode_calibration.calibration.analysis |
sensitivity and identifiability |
pyadm1ode_calibration.io.loaders |
MeasurementData, CSVHandler, PlantSchema, MeasurementBuilder |
pyadm1ode_calibration.io.persistence |
Database, ORM models, repositories |
pyadm1ode_calibration.io.validation |
DataValidator, OutlierDetector |
pyadm1ode_calibration.plants |
plant topology builders |
The most-used names are re-exported at the top level, so
from pyadm1ode_calibration import Calibrator, MeasurementData works.
Where to look next¶
- Tutorials for the notebook series.
- Examples for runnable scripts.
- API Reference for signatures.