dacapo.predict
==============

.. py:module:: dacapo.predict


Attributes
----------

.. autoapisummary::

   dacapo.predict.logger


Functions
---------

.. autoapisummary::

   dacapo.predict.predict


Module Contents
---------------

.. py:data:: logger

.. py:function:: predict(run_name: str | dacapo.experiments.Run, iteration: int | None, input_container: upath.UPath | str, input_dataset: str, output_path: dacapo.store.local_array_store.LocalArrayIdentifier | upath.UPath | str, output_roi: Optional[funlib.geometry.Roi | str] = None, num_workers: int = 1, output_dtype: numpy.dtype | str = np.uint8, overwrite: bool = True)

   Predict with a trained model.

   :param run_name: The name of the run to predict with or the Run object.
   :type run_name: str or Run
   :param iteration: The training iteration of the model to use for prediction.
   :type iteration: int or None
   :param input_container: The container of the input array.
   :type input_container: Path | str
   :param input_dataset: The dataset name of the input array.
   :type input_dataset: str
   :param output_path: The path where the prediction array will be stored, or a LocalArryIdentifier for the prediction array.
   :type output_path: LocalArrayIdentifier | str
   :param output_roi: The ROI of the output array. If None, the ROI of the input array will be used. Defaults to None.
   :type output_roi: Optional[Roi | str], optional
   :param num_workers: The number of workers to use for blockwise prediction. Defaults to 1 for local processing, otherwise 12.
   :type num_workers: int, optional
   :param output_dtype: The dtype of the output array. Defaults to np.uint8.
   :type output_dtype: np.dtype | str, optional
   :param overwrite: If True, the output array will be overwritten if it already exists. Defaults to True.
   :type overwrite: bool, optional

   :raises ValueError: If run_name is not found in config store

   .. rubric:: Examples

   >>> predict("run_name", 100, "input.zarr", "raw", "output.zarr", output_roi="[0:100,0:100,0:100]")