from pathlib import Path
from datetime import datetime
from typing import Dict, Iterable, List, Optional, TextIO, Union
from encord.client import EncordClientDataset
from encord.http.utils import CloudUploadSettings
from encord.orm.cloud_integration import CloudIntegration
from encord.orm.dataset import AddPrivateDataResponse, DataRow
from encord.orm.dataset import Dataset as OrmDataset
from encord.orm.dataset import (
DatasetDataLongPolling,
DatasetAccessSettings,
DatasetUser,
DatasetUserRole,
Image,
ImageGroupOCR,
StorageLocation,
)
from encord.constants.enums import DataType
[docs]class Dataset:
"""
Access dataset related data and manipulate the dataset.
"""
def __init__(self, client: EncordClientDataset):
self._client = client
self._dataset_instance: Optional[OrmDataset] = None
@property
def dataset_hash(self) -> str:
"""
Get the dataset hash (i.e. the Dataset ID).
"""
dataset_instance = self._get_dataset_instance()
return dataset_instance.dataset_hash
@property
def title(self) -> str:
dataset_instance = self._get_dataset_instance()
return dataset_instance.title
@property
def description(self) -> str:
dataset_instance = self._get_dataset_instance()
return dataset_instance.description
@property
def storage_location(self) -> StorageLocation:
dataset_instance = self._get_dataset_instance()
return dataset_instance.storage_location
@property
def data_rows(self) -> List[DataRow]:
"""
Part of the response of this function can be configured by the
:meth:`encord.dataset.Dataset.set_access_settings` method.
.. code::
dataset.set_access_settings(DatasetAccessSettings(fetch_client_metadata=True))
print(dataset.data_rows)
"""
dataset_instance = self._get_dataset_instance()
return dataset_instance.data_rows
[docs] def list_data_rows(
self,
title_eq: Optional[str] = None,
title_like: Optional[str] = None,
created_before: Optional[Union[str, datetime]] = None,
created_after: Optional[Union[str, datetime]] = None,
data_types: Optional[List[DataType]] = None,
) -> List[DataRow]:
"""
Retrieve dataset rows (pointers to data, labels).
Args:
title_eq: optional exact title row filter
title_like: optional fuzzy title row filter; SQL syntax
created_before: optional datetime row filter
created_after: optional datetime row filter
data_types: optional data types row filter
Returns:
List[DataRow]: A list of DataRows object that match the filter
Raises:
AuthorisationError: If the dataset API key is invalid.
ResourceNotFoundError: If no dataset exists by the specified dataset EntityId.
UnknownError: If an error occurs while retrieving the dataset.
"""
return self._client.list_data_rows(title_eq, title_like, created_before, created_after, data_types)
[docs] def refetch_data(self) -> None:
"""
The Dataset class will only fetch its properties once. Use this function if you suspect the state of those
properties to be dirty.
"""
self._dataset_instance = self.get_dataset()
[docs] def get_dataset(self) -> OrmDataset:
"""
This function is exposed for convenience. You are encouraged to use the property accessors instead.
"""
return self._client.get_dataset()
[docs] def set_access_settings(self, dataset_access_settings: DatasetAccessSettings, *, refetch_data: bool = True) -> None:
"""
Args:
dataset_access_settings: The access settings to use going forward
refetch_data: Whether a `refetch_data()` call should follow the update of the dataset access settings.
"""
self._client.set_access_settings(dataset_access_settings)
if refetch_data:
self.refetch_data()
[docs] def add_users(self, user_emails: List[str], user_role: DatasetUserRole) -> List[DatasetUser]:
"""
Add users to dataset. If the user was already added, this operation will succeed but the `user_role` will be
unchanged. The existing `user_role` will be reflected in the `DatasetUser` instance.
Args:
user_emails: list of user emails to be added
user_role: the user role to assign to all users
"""
return self._client.add_users(user_emails, user_role)
[docs] def upload_video(
self,
file_path: str,
cloud_upload_settings: CloudUploadSettings = CloudUploadSettings(),
title: Optional[str] = None,
):
"""
Upload video to Encord storage.
Args:
file_path: path to video e.g. '/home/user/data/video.mp4'
cloud_upload_settings:
Settings for uploading data into the cloud. Change this object to overwrite the default values.
title:
The video title. If unspecified, this will be the file name. This title should include an extension.
For example "encord_video.mp4".
Returns:
Bool.
Raises:
UploadOperationNotSupportedError: If trying to upload to external
datasets (e.g. S3/GPC/Azure)
"""
return self._client.upload_video(file_path, cloud_upload_settings=cloud_upload_settings, title=title)
[docs] def create_image_group(
self,
file_paths: Iterable[str],
max_workers: Optional[int] = None,
cloud_upload_settings: CloudUploadSettings = CloudUploadSettings(),
title: Optional[str] = None,
*,
create_video: bool = True,
):
"""
Create an image group in Encord storage. Choose this type of image upload for sequential images. Else, you can
choose the :meth:`.Dataset.upload_image` function.
Args:
file_paths: a list of paths to images, e.g.
['/home/user/data/img1.png', '/home/user/data/img2.png']
max_workers:
DEPRECATED: This argument will be ignored
cloud_upload_settings:
Settings for uploading data into the cloud. Change this object to overwrite the default values.
title:
The title of the image group. If unspecified this will be randomly generated for you. This title should
NOT include an extension. For example "encord_image_group".
create_video:
A flag specifying how image groups are stored. If `True`, a compressed video will be created from
the image groups. `True` was the previous default support. If `False`, the images
are saved as a sequence of images.
Returns:
Bool.
Raises:
UploadOperationNotSupportedError: If trying to upload to external
datasets (e.g. S3/GPC/Azure)
"""
return self._client.create_image_group(
file_paths,
cloud_upload_settings=cloud_upload_settings,
title=title,
create_video=create_video,
)
[docs] def create_dicom_series(
self,
file_paths: List[str],
cloud_upload_settings: CloudUploadSettings = CloudUploadSettings(),
title: Optional[str] = None,
):
"""
Upload a DICOM series to Encord storage
Args:
file_paths: a list of paths to DICOM files, e.g.
['/home/user/data/DICOM_1.dcm', '/home/user/data/DICOM_2.dcm']
cloud_upload_settings:
Settings for uploading data into the cloud. Change this object to overwrite the default values.
title:
The title of the DICOM series. If unspecified this will be randomly generated for you. This title should
NOT include an extension. For example "encord_image_group".
Returns:
Bool.
Raises:
UploadOperationNotSupportedError: If trying to upload to external
datasets (e.g. S3/GPC/Azure)
"""
return self._client.create_dicom_series(file_paths, cloud_upload_settings=cloud_upload_settings, title=title)
[docs] def upload_image(
self,
file_path: Union[Path, str],
title: Optional[str] = None,
cloud_upload_settings: CloudUploadSettings = CloudUploadSettings(),
) -> Image:
"""
Upload a single image to Encord storage. If your images are sequential we recommend creating an image group via
the :meth:`.Dataset.create_image_group` function. For more information please compare
https://docs.encord.com/docs/annotate/editor/images and https://docs.encord.com/docs/annotate/editor/videos
Args:
file_path: The file path to the image
title: The image title. If unspecified, this will be the file name. This title should include an extension.
For example "encord_image.png".
cloud_upload_settings:
Settings for uploading data into the cloud. Change this object to overwrite the default values.
"""
return self._client.upload_image(file_path, title, cloud_upload_settings)
[docs] def delete_image_group(self, data_hash: str):
"""
Delete an image group in Encord storage.
Args:
data_hash: the hash of the image group to delete
"""
return self._client.delete_image_group(data_hash)
[docs] def delete_data(self, data_hashes: List[str]):
"""
Delete a video/image group from a dataset.
Args:
data_hashes: list of hash of the videos/image_groups you'd like to delete, all should belong to the same
dataset
"""
return self._client.delete_data(data_hashes)
[docs] def add_private_data_to_dataset(
self,
integration_id: str,
private_files: Union[str, Dict, Path, TextIO],
ignore_errors: bool = False,
) -> AddPrivateDataResponse:
"""
Append data hosted on a private cloud to an existing dataset.
For a more complete example of safe uploads, please follow the guide found in our docs under
:ref:`https://python.docs.encord.com/tutorials/datasets.html#adding-data-from-a-private-cloud
<tutorials/datasets:Adding data from a private cloud>`
Args:
integration_id:
The `EntityId` of the cloud integration you wish to use.
private_files:
A `str` path or `Path` object to a json file, json str or python dictionary of the files you wish to add
ignore_errors:
When set to `True`, this will prevent individual errors from stopping the upload process.
Returns:
add_private_data_response List of DatasetDataInfo objects containing data_hash and title
"""
return self._client.add_private_data_to_dataset(integration_id, private_files, ignore_errors)
[docs] def add_private_data_to_dataset_start(
self,
integration_id: str,
private_files: Union[str, Dict, Path, TextIO],
ignore_errors: bool = False,
) -> str:
"""
Append data hosted on a private cloud to an existing dataset.
This method inititalizes the upload in Encord's backend.
Once the upload id has been returned, you can exit the terminal
while the job continues uninterrupted.
You can check upload job status at any point using
the :meth:`add_private_data_to_dataset_get_result` method.
This can be done in a separate python session to the one
where the upload was initialized.
Args:
integration_id:
The `EntityId` of the cloud integration you wish to use.
private_files:
A `str` path or `Path` object to a json file, json str or python dictionary of the files you wish to add
ignore_errors:
When set to `True`, this will prevent individual errors from stopping the upload process.
Returns:
str
`upload_job_id` - UUID Identifier of upload job.
This id enables the user to track the job progress via SDK, or web app.
"""
return self._client.add_private_data_to_dataset_start(integration_id, private_files, ignore_errors)
[docs] def add_private_data_to_dataset_get_result(
self,
upload_job_id: str,
timeout_seconds: int = 7 * 24 * 60 * 60, # 7 days
) -> DatasetDataLongPolling:
"""
Fetch data upload status, perform long polling process for `timeout_seconds`.
Args:
upload_job_id:
UUID Identifier of upload job. This id enables the user to track the job progress via SDK, or web app.
timeout_seconds:
Number of seconds the method will wait while waiting for a response.
If `timeout_seconds == 0`, only a single checking request is performed.
Response will be immediately returned.
Returns:
DatasetDataLongPolling
Response containing details about job status, errors and progress.
"""
return self._client.add_private_data_to_dataset_get_result(upload_job_id, timeout_seconds)
[docs] def update_data_item(self, data_hash: str, new_title: str) -> bool:
"""
DEPRECATED: Use the individual setter properties of the respective :class:`encord.orm.dataset.DataRow`
instance instead. These can be retrieved via the :meth:`.Dataset.data_rows` function.
Update a data item
Args:
data_hash: str
Data hash of the item being updated
new_title:
String containing the new title of the data item being updated
Returns:
Returns a boolean for whether the update was successful
"""
return self._client.update_data_item(data_hash, new_title)
[docs] def re_encode_data(self, data_hashes: List[str]):
"""
Launches an async task that can re-encode a list of videos.
Args:
data_hashes: list of hash of the videos you'd like to re_encode, all should belong to the same
dataset
Returns:
EntityId(integer) of the async task launched.
"""
return self._client.re_encode_data(data_hashes)
[docs] def re_encode_data_status(self, job_id: int):
"""
Returns the status of an existing async task which is aimed at re-encoding videos.
Args:
job_id: id of the async task that was launched to re-encode the videos
Returns:
ReEncodeVideoTask: Object containing the status of the task, along with info about the new encoded videos
in case the task has been completed
"""
return self._client.re_encode_data_status(job_id)
[docs] def run_ocr(self, image_group_id: str) -> List[ImageGroupOCR]:
"""
Returns an optical character recognition result for a given image group
Args:
image_group_id: the id of the image group in this dataset to run OCR on
Returns:
Returns a list of ImageGroupOCR objects representing the text and corresponding coordinates
found in each frame of the image group
"""
return self._client.run_ocr(image_group_id)
[docs] def get_cloud_integrations(self) -> List[CloudIntegration]:
return self._client.get_cloud_integrations()
def _get_dataset_instance(self):
if self._dataset_instance is None:
self._dataset_instance = self.get_dataset()
return self._dataset_instance