Helium-Starlink Coupling

Train Models

Trigger HSC model training with latest data.

Retrains the three HSC models (anomaly detector, precursor classifier, RF corrector) using the most recent Starlink + Helium observations centred on the given location.

Training uses RandomizedSearchCV for hyperparameter tuning, 5-fold cross-validation, and CalibratedClassifierCV for probability calibration. Results are registered in the model registry (common/model_registry.py).

This endpoint is intended for scheduled retraining or manual triggering after significant new data ingestion.

Pricing: Included in subscription tiers (legacy list price: GET /api/v1/billing/pricing).

post/api/v1/cross-physics/helium-starlink/train

Request

  • Base URL: https://www.deepmapai.com
  • URL: https://www.deepmapai.com/api/v1/cross-physics/helium-starlink/train
  • Auth: none declared

Query parameters

x_api_keystring nullable

Request body

latnumber required

Latitude in decimal degrees (WGS84)

lonnumber required

Longitude in decimal degrees (WGS84)

radius_kmnumber

Analysis radius in kilometres (default 50)

include_trainingboolean

If True, trigger model retraining with latest data

Response

Successful Response

statusstring required

Training status

models_trainedinteger required

Number of models trained

resultsobject

Per-model training results (accuracy, F1, etc.)

Changes

No recorded changes to this operation across all 1 revision of this API.