def subscribe_ml_model(ml_model_dto: MLModelDTO) -> int:
"""
Subscribes an ML Model by saving it in the database
:params ml_model_dto
:raises DataError
:returns ID of the ml model
"""
new_ml_model = MLModel()
new_ml_model.create(ml_model_dto)
current_app.logger.info(new_ml_model)
return new_ml_model.id
def delete_ml_model(model_id: int):
"""
Deletes ML model and associated predictions
:params model_id
"""
ml_model = MLModel.get(model_id)
if ml_model:
ml_model.delete()
else:
raise NotFound('Model does not exist')
def get_ml_model_by_id(model_id: int):
"""
Get an ML Model for a given ID
:params model_id
:raises NotFound
:returns ML Model
"""
ml_model = MLModel.get(model_id)
if ml_model:
return ml_model.as_dto()
else:
raise NotFound('Model does not exist')
def tilejson(model_id, prediction_id):
"""
Get the TileJSON of the prediction id given
:params model_id
:params prediction_id
:returns dict
"""
tiles = PredictionTile.count(prediction_id)
if tiles.count == 0:
raise PredictionsNotFound('No Prediction Tiles exist')
ml_model = MLModel.get(model_id)
prediction = Prediction.get(prediction_id)
tilejson = {
"tilejson":
"2.1.0",
"name":
ml_model.name,
"description":
ml_model.project_url,
"inferences":
PredictionTile.inferences(prediction_id),
"token":
CONFIG.EnvironmentConfig.MAPBOX_TOKEN,
"attribution":
ml_model.source,
"version":
prediction.version,
"scheme":
"xyz",
"type":
"vector",
"tiles": [
"/v1/model/{0}/prediction/{1}/tiles/{{z}}/{{x}}/{{y}}.mvt".
format(model_id, prediction_id)
],
"minzoom":
0,
"maxzoom":
prediction.tile_zoom,
"bounds":
PredictionTile.bbox(prediction_id)
}
return tilejson
def update_ml_model(updated_ml_model_dto: MLModelDTO) -> int:
"""
Update an existing ML Model
:params model_id
:raises NotFound
:returns model_id
"""
ml_model = MLModel.get(updated_ml_model_dto.model_id)
if (ml_model):
ml_model.update(updated_ml_model_dto)
return updated_ml_model_dto.model_id
else:
raise NotFound('Model does not exist')
def get_all(model_filter: str, model_archived: bool):
"""
Get all ML Models
:raises NotFound
:returns array of ML Models
"""
ml_models = MLModel.get_all(model_filter, model_archived)
if (ml_models):
model_collection = []
for model in ml_models:
model_collection.append(model.as_dto().to_primitive())
return model_collection
else:
raise NotFound('No models exist')