dacapo.utils.view

Classes

`BestScore`	Represents the best score achieved during a run.
`NeuroglancerRunViewer`	A class for viewing neuroglancer runs.

Functions

`get_viewer`(→ neuroglancer.Viewer \| IPython.display.IFrame)	Creates a neuroglancer viewer to visualize arrays.
`add_seg_layer`(state, name, data, voxel_size[, meshes])	Add a segmentation layer to the Neuroglancer state.
`add_scalar_layer`(state, name, data, voxel_size)	Add a scalar layer to the state.

Module Contents

dacapo.utils.view.get_viewer(arrays: dict, width: int = 1500, height: int = 600, headless: bool = True, bind_address: str = '0.0.0.0', bind_port: int = 0) → neuroglancer.Viewer | IPython.display.IFrame

Creates a neuroglancer viewer to visualize arrays.

Parameters:

arrays (dict) – A dictionary containing arrays to be visualized.
width (int, optional) – The width of the viewer window in pixels. Defaults to 1500.
height (int, optional) – The height of the viewer window in pixels. Defaults to 600.
headless (bool, optional) – If True, returns the viewer object. If False, returns an IFrame object embedding the viewer. Defaults to True.
bind_address (str, optional) – Bind address for Neuroglancer webserver.
bind_port (int, optional) – Bind port for Neuroglancer webserver.

Returns:

The neuroglancer viewer object or an IFrame object embedding the viewer.

Return type:

neuroglancer.Viewer | IFrame

Raises:

ValueError – If the array is not a numpy array or a neuroglancer.LocalVolume object.

Examples

>>> import numpy as np
>>> import neuroglancer
>>> from dacapo.utils.view import get_viewer
>>> arrays = {
...     "raw": {
...         "array": np.random.rand(100, 100, 100)
...     },
...     "seg": {
...         "array": np.random.randint(0, 10, (100, 100, 100)),
...         "is_seg": True
...     }
... }
>>> viewer = get_viewer(arrays)
>>> viewer

dacapo.utils.view.add_seg_layer(state, name, data, voxel_size, meshes=False)

Add a segmentation layer to the Neuroglancer state.

Parameters:

state (neuroglancer.ViewerState) – The Neuroglancer viewer state.
name (str) – The name of the segmentation layer.
data (ndarray) – The segmentation data.
voxel_size (list) – The voxel size in nm.
meshes (bool, optional) – Whether to generate meshes for the segments. Defaults to False.

Raises:

ValueError – If the data is not a numpy array.

Examples

>>> import numpy as np
>>> import neuroglancer
>>> from dacapo.utils.view import add_seg_layer
>>> state = neuroglancer.ViewerState()
>>> data = np.random.randint(0, 10, (100, 100, 100))
>>> voxel_size = [1, 1, 1]
>>> add_seg_layer(state, "seg", data, voxel_size)

dacapo.utils.view.add_scalar_layer(state, name, data, voxel_size)

Add a scalar layer to the state.

Parameters:

state (neuroglancer.ViewerState) – The viewer state to add the layer to.
name (str) – The name of the layer.
data (ndarray) – The scalar data to be displayed.
voxel_size (list) – The voxel size in nm.

Raises:

ValueError – If the data is not a numpy array.

Examples

>>> import numpy as np
>>> import neuroglancer
>>> from dacapo.utils.view import add_scalar_layer
>>> state = neuroglancer.ViewerState()
>>> data = np.random.rand(100, 100, 100)
>>> voxel_size = [1, 1, 1]
>>> add_scalar_layer(state, "raw", data, voxel_size)

class dacapo.utils.view.BestScore(run: dacapo.experiments.run.Run)

Represents the best score achieved during a run.

run

The run object associated with the best score.

Type:: Run

score

The best score achieved.

Type:: float

iteration

The iteration number at which the best score was achieved.

Type:: int

parameter

The parameter associated with the best score.

Type:: Optional[str]

validation_parameters: The validation parameters used during the run.

array_store: The array store object used to store prediction arrays.

stats_store: The stats store object used to store iteration scores.

ds: The dataset object associated with the best score.

get_ds(iteration, validation_dataset): Retrieves the dataset object associated with the best score.

does_new_best_exist(): Checks if a new best score exists.

run: dacapo.experiments.run.Run

score: float

iteration: int = 0

parameter: str | None = None

validation_parameters

array_store

stats_store

get_ds(iteration, validation_dataset)

Retrieves the dataset object associated with the best score.

Parameters:

iteration (int) – The iteration number.
validation_dataset – The validation dataset object.

Raises:

FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import BestScore
>>> run = Run()
>>> best_score = BestScore(run)
>>> iteration = 0
>>> validation_dataset = run.datasplit.validate[0]
>>> best_score.get_ds(iteration, validation_dataset)

does_new_best_exist()

Checks if a new best score exists.

Returns:: True if a new best score exists, False otherwise.
Return type:: bool
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import BestScore
>>> run = Run()
>>> best_score = BestScore(run)
>>> new_best_exists = best_score.does_new_best_exist()

class dacapo.utils.view.NeuroglancerRunViewer(run: dacapo.experiments.run.Run, embedded=False)

A class for viewing neuroglancer runs.

run

The run object.

Type:: Run

best_score

The best score object.

Type:: BestScore

embedded

Whether the viewer is embedded.

Type:: bool

viewer: The neuroglancer viewer.

raw: The raw dataset.

gt: The ground truth dataset.

segmentation: The segmentation dataset.

most_recent_iteration: The most recent iteration.

run_thread: The run thread.

array_store: The array store object.

updated_neuroglancer_layer(layer_name, ds): Update the neuroglancer layer with the given name and data source.

deprecated_start_neuroglancer(): Deprecated method to start the neuroglancer viewer.

start_neuroglancer(): Start the neuroglancer viewer.

start(): Start the viewer.

open_from_array_identitifier(array_identifier): Open the array from the given identifier.

get_datasets(): Get the datasets for validation.

update_best_info(): Update the best info.

update_neuroglancer(): Update the neuroglancer viewer.

update_best_layer(): Update the best layer.

new_validation_checker(): Start a new validation checker thread.

update_with_new_validation_if_possible(): Update with new validation if possible.

stop(): Stop the viewer.

run: dacapo.experiments.run.Run

best_score

embedded

updated_neuroglancer_layer(layer_name, ds)

Update the neuroglancer layer with the given name and data source.

Parameters:

layer_name (str) – The name of the layer.
ds – The data source.

Raises:

FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> layer_name = "prediction"
>>> ds = viewer.run.datasplit.validate[0].raw._source_array
>>> viewer.updated_neuroglancer_layer(layer_name, ds)

deprecated_start_neuroglancer()

Deprecated method to start the neuroglancer viewer.

Returns:: The embedded viewer.
Return type:: IFrame
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.deprecated_start_neuroglancer()

start_neuroglancer(bind_address='0.0.0.0', bind_port=None)

Start the neuroglancer viewer.

Parameters:

bind_address (str, optional) – Bind address for Neuroglancer webserver.
bind_port (int, optional) – Bind port for Neuroglancer webserver.

Returns:

The embedded viewer.

Return type:

IFrame

Raises:

FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.start_neuroglancer()

start()

Start the viewer.

Returns:: The embedded viewer.
Return type:: IFrame
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.start()

open_from_array_identitifier(array_identifier)

Open the array from the given identifier.

Parameters:: array_identifier – The array identifier.
Returns:: The opened dataset or None if it doesn’t exist.
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> array_identifier = viewer.run.datasplit.validate[0].raw._source_array
>>> ds = viewer.open_from_array_identitifier(array_identifier)

get_datasets()

Get the datasets for validation.

Parameters:: run (Run) – The run object.
Returns:: The raw and ground truth datasets for validation.
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.get_datasets()

update_best_info()

Update the best info.

Parameters:: run (Run) – The run object.
Returns:: The embedded viewer.
Return type:: IFrame
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.update_best_info()

update_neuroglancer()

Update the neuroglancer viewer.

Parameters:: run (Run) – The run object.
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.update_neuroglancer()

update_best_layer()

Update the best layer.

Parameters:: run (Run) – The run object.
Returns:: The embedded viewer.
Return type:: IFrame
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.update_best_layer()

new_validation_checker()

Start a new validation checker thread.

Parameters:: run (Run) – The run object.
Returns:: The embedded viewer.
Return type:: IFrame
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.new_validation_checker()

update_with_new_validation_if_possible()

Update with new validation if possible.

Parameters:: run (Run) – The run object.
Returns:: The embedded viewer.
Return type:: IFrame
Raises:: FileNotFoundError – If the dataset object does not exist.

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.update_with_new_validation_if_possible()

stop()

Stop the viewer.

Parameters:: run (Run) – The run object.
Raises:: FileNotFoundError – If the dataset object does not exist.
Returns:: The embedded viewer.
Return type:: IFrame

Examples

>>> from dacapo.experiments.run import Run
>>> from dacapo.utils.view import NeuroglancerRunViewer
>>> run = Run()
>>> viewer = NeuroglancerRunViewer(run)
>>> viewer.stop()