Support for catboost inference #299
Conversation
openeo/processes.py
Outdated
| :return: The predicted value. | ||
| """ | ||
| return predict_catboost(data=self, model=model) | ||
|
|
There was a problem hiding this comment.
openeo/processes.py is a generated file
openeo-python-client/openeo/processes.py
Lines 2 to 3 in 6f0a865
so it's nor recommended to manually add processes here (they will be removed with the next update)
There was a problem hiding this comment.
as long as predict_catboost is not an official process, we can not put it here
openeo/rest/datacube.py
Outdated
| a job id (``str``) or a :py:class:`RESTJob` instance of a batch job that saved a single model. | ||
| :param dimension: dimension along which to apply the ``reduce_dimension`` process. | ||
|
|
||
| .. versionadded:: 0.10.0 |
openeo/rest/datacube.py
Outdated
|
|
||
| :param model: a :py:class:`MlModel` (e.g. loaded from :py:meth:`~openeo.Connection.load_ml_model`) | ||
| or a reference to such a model: a URL (``str``) to a STAC Item (implementing the `ml-model` extension) | ||
| a job id (``str``) or a :py:class:`RESTJob` instance of a batch job that saved a single model. |
There was a problem hiding this comment.
see latest doc of predict_random_forest for a cleaner parameter description:
openeo-python-client/openeo/rest/datacube.py
Lines 1870 to 1876 in 6f0a865
openeo/rest/datacube.py
Outdated
| if not isinstance(model, MlModel): | ||
| model = MlModel.load_ml_model(connection=self.connection, id=model) | ||
| from openeo.processes import predict_catboost | ||
| reducer = lambda data, context: predict_catboost(data=data, model=context) |
There was a problem hiding this comment.
instead use
from openeo.processes import process
reducer = lambda data, context: process('predict_catboost', data=data, model=model)
|
It's probably better to avoid model-specific predict processes like |
No description provided.