Encord Python SDK API Reference — Encord Python SDK documentation

Datasets / Data / Adding data / Adding data from a private cloud

Datasets / Managing a dataset

get_project(project_hash)[source]#

Get the Project class to access project fields and manipulate a project.

You will only have access to this project if you are one of the following

Project admin

Project team manager

Organisation admin of the project

Parameters: project_hash (str) – The Project ID
Return type: Project

Expand for references to encord.user_client.EncordUserClient.get_project

Authentication / User authentication / Authenticate with Encord - user authentication

Frequently asked questions (FAQ) / Why do I see an AuthenticationError when working with label hashes?

Bitmask annotations / 1. Downloading bitmask annotations from Encord

Bitmask annotations / 2. Uploading bitmask annotations to Encord

Projects / Project details / Collaborator session information

Projects / Project details

get_ontology(ontology_hash)[source]#

Return type: Ontology

Expand for references to encord.user_client.EncordUserClient.get_ontology

create_private_dataset(dataset_title, dataset_type, dataset_description=None)[source]#

DEPRECATED - please use create_dataset instead.

Return type: CreateDatasetResponse

create_dataset(dataset_title, dataset_type, dataset_description=None)[source]#

Parameters

dataset_title (str) – Title of dataset.
dataset_type (StorageLocation) – StorageLocation type where data will be stored.
dataset_description (Optional[str]) – Optional description of the dataset.

Return type

CreateDatasetResponse

Returns

CreateDatasetResponse

Expand for references to encord.user_client.EncordUserClient.create_dataset

Datasets / API keys / Creating a dataset API key with specific rights

create_dataset_api_key(dataset_hash, api_key_title, dataset_scopes)[source]#

Return type: DatasetAPIKey

Expand for references to encord.user_client.EncordUserClient.create_dataset_api_key

get_dataset_api_keys(dataset_hash)[source]#

Return type: List[DatasetAPIKey]

Expand for references to encord.user_client.EncordUserClient.get_dataset_api_keys

Datasets / API keys / Creating a master API key with full rights

get_or_create_dataset_api_key(dataset_hash)[source]#

Return type: DatasetAPIKey

Expand for references to encord.user_client.EncordUserClient.get_or_create_dataset_api_key

get_datasets(title_eq=None, title_like=None, desc_eq=None, desc_like=None, created_before=None, created_after=None, edited_before=None, edited_after=None)[source]#

List either all (if called with no arguments) or matching datasets the user has access to.

Parameters

title_eq (Optional[str]) – optional exact title filter
title_like (Optional[str]) – optional fuzzy title filter; SQL syntax
desc_eq (Optional[str]) – optional exact description filter
desc_like (Optional[str]) – optional fuzzy description filter; SQL syntax
created_before (Union[str, datetime, None]) – optional creation date filter, ‘less’
created_after (Union[str, datetime, None]) – optional creation date filter, ‘greater’
edited_before (Union[str, datetime, None]) – optional last modification date filter, ‘less’
edited_after (Union[str, datetime, None]) – optional last modification date filter, ‘greater’

Return type

List[Dict[str, Any]]

Returns

list of (role, dataset) pairs for datasets matching filter conditions.

Expand for references to encord.user_client.EncordUserClient.get_datasets

Datasets / Listing existing datasets

static create_with_ssh_private_key(ssh_private_key=None, password=None, requests_settings=RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3), ssh_private_key_path=None, **kwargs)[source]#

Creates an instance of EncordUserClient authenticated with private SSH key. Accepts the private key content, path to key file, that can be provided as method parameters or as following environment variables:

ENCORD_SSH_KEY: environment variable with the private key content
ENCORD_SSH_KEY_FILE: environment variable with the path to the key file

Parameters

ssh_private_key (Optional[str]) – the private key content
ssh_private_key_path (Optional[str | Path]) – the pah to the private key file
password (Optional[str]) – private key password

Return type

EncordUserClient

Expand for references to encord.user_client.EncordUserClient.create_with_ssh_private_key

Authentication / User authentication / Authenticate with Encord - user authentication

Frequently asked questions (FAQ) / Why do I see an AuthenticationError when working with label hashes?

Bitmask annotations / 1. Downloading bitmask annotations from Encord

Bitmask annotations / 2. Uploading bitmask annotations to Encord

Datasets / Data / Adding data / Adding data from a private cloud

Datasets / API keys / Creating a dataset API key with specific rights

Datasets / API keys / Creating a master API key with full rights

Datasets / Listing existing datasets

Datasets / Managing a dataset

Projects / Project details / Collaborator session information

Projects / API keys / Creating a project API key / Creating a master API key with full rights

Projects / Creating a Project

Projects / API keys / Creating a project API key / Creating a project API key with specific rights

Projects / Other Resources / CVAT Integration / Importing with the Encord SDK

Projects / Listing existing projects

Projects / Project details

get_projects(title_eq=None, title_like=None, desc_eq=None, desc_like=None, created_before=None, created_after=None, edited_before=None, edited_after=None)[source]#

List either all (if called with no arguments) or matching projects the user has access to.

Parameters

title_eq (Optional[str]) – optional exact title filter
title_like (Optional[str]) – optional fuzzy title filter; SQL syntax
desc_eq (Optional[str]) – optional exact description filter
desc_like (Optional[str]) – optional fuzzy description filter; SQL syntax
created_before (Union[str, datetime, None]) – optional creation date filter, ‘less’
created_after (Union[str, datetime, None]) – optional creation date filter, ‘greater’
edited_before (Union[str, datetime, None]) – optional last modification date filter, ‘less’
edited_after (Union[str, datetime, None]) – optional last modification date filter, ‘greater’

Return type

List[Dict]

Returns

list of (role, projects) pairs for project matching filter conditions.

Expand for references to encord.user_client.EncordUserClient.get_projects

Projects / Listing existing projects

create_project(project_title, dataset_hashes, project_description='', ontology_hash='', workflow_settings=<encord.orm.project.ManualReviewWorkflowSettings object>, workflow_template_hash=None)[source]#

Creates a new project and returns its uid (‘project_hash’)

Parameters

project_title (str) – the title of the project
dataset_hashes (List[str]) – a list of the dataset uids that the project will use
project_description (str) – the optional description of the project
ontology_hash (str) – the uid of an ontology to be used. If omitted, a new empty ontology will be created
workflow_settings (Union[ManualReviewWorkflowSettings, BenchmarkQaWorkflowSettings]) – selects and configures the type of the quality control workflow to use, See encord.orm.project.ProjectWorkflowSettings for details. If omitted, ManualReviewWorkflowSettings is used.
workflow_template_hash (Optional[str]) – project will be created using a workflow based on the template provided.

Return type

str

Returns

the uid of the project.

Expand for references to encord.user_client.EncordUserClient.create_project

Projects / Creating a Project

create_project_api_key(project_hash, api_key_title, scopes)[source]#

Return type: str
Returns: The created project API key.

Expand for references to encord.user_client.EncordUserClient.create_project_api_key

Projects / API keys / Creating a project API key / Creating a project API key with specific rights

get_project_api_keys(project_hash)[source]#

Return type: List[ProjectAPIKey]

Expand for references to encord.user_client.EncordUserClient.get_project_api_keys

Projects / API keys / Creating a project API key / Creating a master API key with full rights

get_or_create_project_api_key(project_hash)[source]#

Return type: str

Expand for references to encord.user_client.EncordUserClient.get_or_create_project_api_key

get_dataset_client(dataset_hash, dataset_access_settings=DatasetAccessSettings(fetch_client_metadata=False), **kwargs)[source]#

DEPRECATED - prefer using get_dataset() instead.

Return type: EncordClientDataset

Expand for references to encord.user_client.EncordUserClient.get_dataset_client

Projects / Other Resources / CVAT Integration / Importing with the Encord SDK

get_project_client(project_hash, **kwargs)[source]#

DEPRECATED - prefer using get_project() instead.

Return type: Union[EncordClientProject, EncordClientDataset]

create_project_from_cvat(import_method, dataset_name, review_mode=ReviewMode.LABELLED, max_workers=None, *, transform_bounding_boxes_to_polygons=False)[source]#

Export your CVAT project with the “CVAT for images 1.1” option and use this function to import: your images and annotations into encord. Ensure that during you have the “Save images” checkbox enabled when exporting from CVAT.

Parameters

import_method (LocalImport) – The chosen import method. See the ImportMethod class for details.
dataset_name (str) – The name of the dataset that will be created.
review_mode (ReviewMode) –

Set how much interaction is needed from the labeler and from the reviewer for the CVAT labels.
See the ReviewMode documentation for more details.
max_workers (Optional[int]) – DEPRECATED: This argument will be ignored
transform_bounding_boxes_to_polygons – All instances of CVAT bounding boxes will be converted to polygons in the final Encord project.

Returns

If the project was successfully imported. CvatImporterError: If the project could not be imported.

Return type

CvatImporterSuccess

Raises

ValueError – If the CVAT directory has an invalid format.

Expand for references to encord.user_client.EncordUserClient.create_project_from_cvat

get_cloud_integrations()[source]#

Return type: List[CloudIntegration]

Expand for references to encord.user_client.EncordUserClient.get_cloud_integrations

Datasets / Data / Adding data / Adding data from a private cloud

get_ontologies(title_eq=None, title_like=None, desc_eq=None, desc_like=None, created_before=None, created_after=None, edited_before=None, edited_after=None)[source]#

List either all (if called with no arguments) or matching ontologies the user has access to.

Parameters

title_eq (Optional[str]) – optional exact title filter
title_like (Optional[str]) – optional fuzzy title filter; SQL syntax
desc_eq (Optional[str]) – optional exact description filter
desc_like (Optional[str]) – optional fuzzy description filter; SQL syntax
created_before (Union[str, datetime, None]) – optional creation date filter, ‘less’
created_after (Union[str, datetime, None]) – optional creation date filter, ‘greater’
edited_before (Union[str, datetime, None]) – optional last modification date filter, ‘less’
edited_after (Union[str, datetime, None]) – optional last modification date filter, ‘greater’

Return type

List[Dict]

Returns

list of (role, projects) pairs for ontologies matching filter conditions.

create_ontology(title, description='', structure=None)[source]#

Return type: Ontology

Expand for references to encord.user_client.EncordUserClient.create_ontology

Authentication / User authentication / Authenticate with Encord - user authentication

deidentify_dicom_files(dicom_urls, integration_hash, redact_dicom_tags=True, redact_pixels_mode=DeidentifyRedactTextMode.REDACT_NO_TEXT, save_conditions=None, upload_dir=None)[source]#

Deidentify DICOM files in external storage. Given links to DICOM files pointing to AWS, GCP, AZURE or OTC, for example: [ “https://s3.region-code.amazonaws.com/bucket-name/dicom-file-input.dcm” ] Function executes deidentification on those files, it removes all DICOM tags (https://dicom.nema.org/medical/Dicom/2017e/output/chtml/part06/chapter_6.html) from metadata except for:

x00080018 SOPInstanceUID
x00100010 PatientName
x00180050 SliceThickness
x00180088 SpacingBetweenSlices
x0020000d StudyInstanceUID
x0020000e SeriesInstanceUID
x00200032 ImagePositionPatient
x00200037 ImageOrientationPatient
x00280008 NumberOfFrames
x00281050 WindowCenter
x00281051 WindowWidth
x00520014 ALinePixelSpacing

Parameters

self – Encord client object.
dicom_urls (List[str]) – a list of urls to DICOM files, e.g. [ “https://s3.region-code.amazonaws.com/bucket-name/dicom-file-input.dcm” ]
integration_hash (str) – integration_hash parameter of Encord platform external storage integration
redact_dicom_tags (bool) – Specifies if DICOM tags redaction should be enabled.
redact_pixels_mode (DeidentifyRedactTextMode) – Specifies which text redaction policy should be applied to pixel data.
save_conditions (Optional[List[Union[SaveDeidentifiedDicomConditionNotSubstr, SaveDeidentifiedDicomConditionIn]]]) – Specifies a list of conditions which all have to be true for DICOM deidentified file to be saved.
upload_dir (Optional[str]) – Specifies a directory that files will be uploaded to. By default, set to None, deidentified files will be uploaded to the same directory as source files.

Return type

List[str]

Returns

Function returns list of links pointing to deidentified DICOM files, those will be saved to the same bucket and the same directory as original files with prefix ( deid_{timestamp}_ ). Example output: [ “https://s3.region-code.amazonaws.com/bucket-name/deid_167294769118005312_dicom-file-input.dcm” ]

class encord.user_client.ListingFilter(value)[source]#

Available properties_filter keys for get_projects() and get_datasets().

The values for _before and _after should be datetime objects.

TITLE_EQ = 'title_eq'#

TITLE_LIKE = 'title_like'#

DESC_EQ = 'desc_eq'#

DESC_LIKE = 'desc_like'#

CREATED_BEFORE = 'created_before'#

CREATED_AFTER = 'created_after'#

EDITED_BEFORE = 'edited_before'#

EDITED_AFTER = 'edited_after'#

Project#

class encord.project.Project(client, project_instance, ontology, client_v2)[source]#

Access project related data and manipulate the project.

Expand for references to encord.project.Project

Frequently asked questions (FAQ) / Why do I see an AuthenticationError when working with label hashes?

Projects / Project details / Collaborator session information

Projects / Project details

property project_hash: str#

Get the project hash (i.e. the Project ID).

Return type: str

property title: str#

Get the title of the project.

Return type: str

property description: str#

Get the description of the project.

Return type: str

property created_at: datetime.datetime#

Get the time the project was created at.

Return type: datetime

property last_edited_at: datetime.datetime#

Get the time the project was last edited at.

Return type: datetime

property ontology: dict#

Get the ontology of the project.

DEPRECATED: Prefer using the encord.Project.ontology_structure() method.

Return type: dict

property ontology_hash: str#

Get the ontology hash of the project’s ontology.

Return type: str

property ontology_structure: encord.objects.ontology_structure.OntologyStructure#

Get the ontology structure of the project’s ontology.

Return type: OntologyStructure

property datasets: list#

Get the associated datasets.

Prefer using the encord.objects.project.ProjectDataset() class to work with the data.

from encord.objects.project import ProjectDataset

project = user_client.get_project("<project_hash>")

project_datasets = ProjectDataset.from_list(project.datasets)

Return type: list

property label_rows: dict#

Get the label rows. DEPRECATED: Prefer using list_label_row_v2() method and LabelRowV2() class to work with the data.

from encord.orm.label_row import LabelRowMetadata

project = user_client.get_project("<project_hash>")

label_rows = LabelRowMetadata.from_list(project.label_rows)

Return type: dict

Expand for references to encord.project.Project.label_rows

Frequently asked questions (FAQ) / Why do I see an AuthenticationError when working with label hashes?

refetch_data()[source]#

The Project class will only fetch its properties once. Use this function if you suspect the state of those properties to be dirty.

Return type: None

refetch_ontology()[source]#

Update the ontology for the project to reflect changes on the backend

Return type: None

get_project()[source]#

This function is exposed for convenience. You are encouraged to use the property accessors instead.

Return type: Project

list_label_rows_v2(data_hashes=None, label_hashes=None, edited_before=None, edited_after=None, label_statuses=None, shadow_data_state=None, data_title_eq=None, data_title_like=None, workflow_graph_node_title_eq=None, workflow_graph_node_title_like=None)[source]#

Parameters

data_hashes (Optional[List[str]]) – List of data hashes to filter by.
label_hashes (Optional[List[str]]) – List of label hashes to filter by.
edited_before (Union[str, datetime, None]) – Optionally filter to only rows last edited before the specified time
edited_after (Union[str, datetime, None]) – Optionally filter to only rows last edited after the specified time
label_statuses (Optional[List[AnnotationTaskStatus]]) – Optionally filter to only those label rows that have one of the specified :class:`~encord.orm.label_row.AnnotationTaskStatus`es
shadow_data_state (Optional[ShadowDataState]) – On Optionally filter by data type in Benchmark QA projects. See ShadowDataState
data_title_eq (Optional[str]) – Optionally filter by exact title match
data_title_like (Optional[str]) – Optionally filter by fuzzy title match; SQL syntax
workflow_graph_node_title_eq (Optional[str]) – Optionally filter by exact match with workflow node title
workflow_graph_node_title_like (Optional[str]) – Optionally filter by fuzzy match with workflow node title; SQL syntax

Return type

List[LabelRowV2]

Returns

A list of LabelRowV2 instances for all the matching label rows

Expand for references to encord.project.Project.list_label_rows_v2

Bitmask annotations / 1. Downloading bitmask annotations from Encord

Bitmask annotations / 2. Uploading bitmask annotations to Encord

add_users(user_emails, user_role)[source]#

Add users to project

Parameters

user_emails (List[str]) – list of user emails to be added
user_role (ProjectUserRole) – the user role to assign to all users

Return type

List[ProjectUser]

Returns

ProjectUser

Raises

AuthorisationError – If the project API key is invalid.
ResourceNotFoundError – If no project exists by the specified project EntityId.
UnknownError – If an error occurs while adding the users to the project

Expand for references to encord.project.Project.add_users

copy_project(copy_datasets=False, copy_collaborators=False, copy_models=False, *, copy_labels=None, new_title=None, new_description=None)[source]#

Copy the current project into a new one with copied contents including settings, datasets and users. Labels and models are optional.

Parameters

copy_datasets (Union[bool, CopyDatasetOptions]) – if True, the datasets of the existing project are copied over, and new tasks are created from those datasets
copy_collaborators – if True, all users of the existing project are copied over with their current roles. If label and/or annotator reviewer mapping is set, this will also be copied over
copy_models – currently if True, all models with their training information will be copied into the new project
copy_labels (Optional[CopyLabelsOptions]) – options for copying labels, defined in CopyLabelsOptions
new_title (Optional[str]) – when provided, will be used as the title for the new project.
new_description (Optional[str]) – when provided, will be used as the title for the new project.

Return type

str

Returns

the EntityId of the newly created project

Raises

AuthorisationError – If the project API key is invalid.
ResourceNotFoundError – If no project exists by the specified project EntityId.
UnknownError – If an error occurs while copying the project.

Expand for references to encord.project.Project.copy_project

Frequently asked questions (FAQ) / Why do I see an AuthenticationError when working with label hashes?

submit_label_row_for_review(uid)[source]#

Submit a label row for review.

Note: this method is not supported for the workflow-based projects. See the documentation about the workflows

Parameters

uid (str) – A label_hash (uid) string.

Returns

Bool.

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
UnknownError – If an error occurs while submitting for review.
OperationNotAllowed – If the write operation is not allowed by the API key.

Expand for references to encord.project.Project.submit_label_row_for_review

add_datasets(dataset_hashes)[source]#

Add a dataset to a project

Parameters

dataset_hashes (List[str]) – List of dataset hashes of the datasets to be added

Return type

bool

Returns

Bool.

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ResourceNotFoundError – If one or more datasets don’t exist by the specified dataset_hashes.
UnknownError – If an error occurs while adding the datasets to the project.
OperationNotAllowed – If the write operation is not allowed by the API key.

Expand for references to encord.project.Project.add_datasets

remove_datasets(dataset_hashes)[source]#

Remove datasets from project

Parameters

dataset_hashes (List[str]) – List of dataset hashes of the datasets to be removed

Return type

bool

Returns

Bool.

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ResourceNotFoundError – If no dataset exists by the specified dataset_hash (uid).
UnknownError – If an error occurs while removing the datasets from the project.
OperationNotAllowed – If the operation is not allowed by the API key.

Expand for references to encord.project.Project.remove_datasets

get_project_ontology()[source]#

DEPRECATED - prefer using the ontology property accessor instead.

Return type: Ontology

add_object(name, shape)[source]#

Add object to an ontology.

ATTENTION: this legacy method will affect all the projects sharing the same ontology

Parameters

name (str) – the name of the object
shape (ObjectShape) – the shape of the object. (BOUNDING_BOX, POLYGON, POLYLINE or KEY_POINT)

Return type

bool

Returns

True if the object was added successfully and False otherwise.

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
UnknownError – If an error occurs while add te object to the project ontology
OperationNotAllowed – If the operation is not allowed by the API key.
ValueError – If invalid arguments are supplied in the function call

add_classification(name, classification_type, required, options=None)[source]#

Add classification to an ontology.

ATTENTION: this legacy method will affect all the projects sharing the same ontology

Parameters

name (str) – the name of the classification
classification_type (ClassificationType) – the classification type (RADIO, TEXT or CHECKLIST)
required (whether this classification is required by the annotator) –
options (Optional[Iterable[str]]) – the list of options for the classification (to be set to None for texts)

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
UnknownError – If an error occurs while add te classification to the project ontology
OperationNotAllowed – If the operation is not allowed by the API key.
ValueError – If invalid arguments are supplied in the function call

list_models()[source]#

List all models that are associated with the project. Use the encord.project.Project.get_training_metadata() to get more metadata about each training instance.

from encord.utilities.project_utilities import get_all_model_iteration_uids

project = client_instance.get_project(<project_hash>)

model_configurations = project.list_models()
all_model_iteration_uids = get_all_model_iteration_uids(model_configurations)
training_metadata = project.get_training_metadata(
    all_model_iteration_uids,
    get_model_training_labels=True,
)

Return type: List[ModelConfiguration]

get_training_metadata(model_iteration_uids, get_created_at=False, get_training_final_loss=False, get_model_training_labels=False)[source]#

Given a list of model_iteration_uids, get some metadata around each model_iteration.

Parameters

model_iteration_uids (Iterable[str]) – The model iteration uids
get_created_at (bool) – Whether the created_at field should be retrieved.
get_training_final_loss (bool) – Whether the training_final_loss field should be retrieved.
get_model_training_labels (bool) – Whether the model_training_labels field should be retrieved.

Return type

List[TrainingMetadata]

create_model_row(title, description, features, model)[source]#

Create a model row.

Parameters

title (str) – Model title.
description (str) – Model description.
features (List[str]) – List of <feature_node_hashes> which is id’s of ontology objects or classifications to be included in the model.
model (Union[AutomationModels, str]) – the model type to be used. For backwards compatibility purposes, we continuously allow strings corresponding to the values of the AutomationModels Enum.

Return type

str

Returns

The uid of the added model row.

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ModelFeaturesInconsistentError – If a feature type is different than what is supported by the model (e.g. if creating a classification model using a bounding box).

model_delete(uid)[source]#

Delete a model created on the platform.

Parameters

uid (str) – A model_hash (uid) string.

Return type

bool

Returns

bool

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ResourceNotFoundError – If no model exists by the specified model_hash (uid).
UnknownError – If an error occurs during training.

model_inference(uid, file_paths=None, base64_strings=None, conf_thresh=0.6, iou_thresh=0.3, device=Device.CUDA, detection_frame_range=None, allocation_enabled=False, data_hashes=None, rdp_thresh=0.005)[source]#

Run inference with model trained on the platform.

The image(s)/video(s) can be provided either as local file paths, base64 strings, or as data hashes if the data is already uploaded on the Encord platform.

Parameters

uid (str) – A model_iteration_hash (uid) string.
file_paths (Optional[List[str]]) – List of local file paths to image(s) or video(s) - if running inference on files.
base64_strings (Optional[List[bytes]]) – List of base 64 strings of image(s) or video(s) - if running inference on base64 strings.
conf_thresh (float) – Confidence threshold (default 0.6).
iou_thresh (float) – Intersection over union threshold (default 0.3).
device (Device) – Device (CPU or CUDA, default is CUDA).
detection_frame_range (Optional[List[int]]) – Detection frame range (for videos).
allocation_enabled (bool) – Object UID allocation (tracking) enabled (disabled by default).
data_hashes (Optional[List[str]]) – list of hash of the videos/image_groups you’d like to run inference on.
rdp_thresh (float) – parameter specifying the polygon coarseness to be used while running inference. The higher the value, the less points in the segmented image

Returns

A dict of inference results.

Return type

Inference results

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ResourceNotFoundError – If no model exists by the specified model_iteration_hash (uid).
UnknownError – If an error occurs while running inference.
FileTypeNotSupportedError – If the file type is not supported for inference (has to be an image or video)
FileSizeNotSupportedError – If the file size is too big to be supported.
DetectionRangeInvalidError – If a detection range is invalid for video inference

model_train(uid, label_rows=None, epochs=None, batch_size=24, weights=None, device=Device.CUDA)[source]#

Train a model created on the platform.

Parameters

uid (str) – A model_hash (uid) string.
label_rows (Optional[List[str]]) – List of label row uid’s (hashes) for training.
epochs (Optional[int]) – Number of passes through training dataset.
batch_size (int) – Number of training examples utilized in one iteration.
weights (Optional[ModelTrainingWeights]) – Model weights.
device (Device) – Device (CPU or CUDA, default is CUDA).

Returns

A model iteration object.

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ModelWeightsInconsistentError – If the passed model weights are incompatible with the selected model.
ResourceNotFoundError – If no model exists by the specified model_hash (uid).
UnknownError – If an error occurs during training.

object_interpolation(key_frames, objects_to_interpolate)[source]#

Run object interpolation algorithm on project labels (requires an editor ontology and feature uid’s).

Interpolation is supported for bounding box, polygon, and keypoint.

Parameters

key_frames –

Labels for frames to be interpolated. Key frames are consumed in the form:

{
    "<frame_number>": {
        "objects": [
            {
                "objectHash": "<object_hash>",
                "featureHash": "<feature_hash>",
                "polygon": {
                    "0": { "x": x1, "y": y1, },
                    "1": { "x": x2, "y": y2, },
                    # ...,
                }
            },
            # ...
        ]
    },
    # ...,
}

objects_to_interpolate – List of object uid’s (hashes) of objects to interpolate.

Returns

Full set of filled frames including interpolated objects.

Return type

Interpolation results

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
UnknownError – If an error occurs while running interpolation.

fitted_bounding_boxes(frames, video)[source]#

Parameters

frames (dict) –

Labels for frames to be fitted. Frames are consumed in the form:

{
    "<frame_number>": {
        "objects": [
            {
                "objectHash": "<object_hash>",
                "featureHash": "<feature_hash>",
                "polygon": {
                    "0": { "x": x1, "y": y1, },
                    "1": { "x": x2, "y": y2, },
                    # ...,
                }
            },
            # ...
        ]
    },
    # ...,
}

video (dict) –
Metadata of the video for which bounding box fitting needs to be run:
```
{
    "width" : w,
    "height" : h,
}
```

Returns

Full set of filled frames including fitted objects.

Return type

Fitting results

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
UnknownError – If an error occurs while running interpolation.

get_data(data_hash, get_signed_url=False)[source]#

Retrieve information about a video or image group.

Parameters

data_hash (str) – The uid of the data object
get_signed_url (bool) – Optionally return signed URLs for timed public access to that resource (default False)

Return type

Tuple[Optional[Video], Optional[List[Image]]]

Returns

A consisting of the video (if it exists) and a list of individual images (if they exist)

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
UnknownError – If an error occurs while retrieving the object.

get_websocket_url()[source]#

Return type: str

get_label_logs(user_hash=None, data_hash=None, from_unix_seconds=None, to_unix_seconds=None, after=None, before=None, user_email=None)[source]#

Get label logs, which represent the actions taken in the UI to create labels.

All arguments can be left as None if no filtering should be applied.

Parameters

user_hash (Optional[str]) – Filter the label logs by the user.
data_hash (Optional[str]) – Filter the label logs by the data_hash.
from_unix_seconds (Optional[int]) – Filter the label logs to only include labels after this timestamp. Deprecated: use parameter after instead
to_unix_seconds (Optional[int]) – Filter the label logs to only include labels before this timestamp. Deprecated: use parameter before instead
after (Optional[datetime]) – Filter the label logs to only include labels after the specified time.
before (Optional[datetime]) – Filter the label logs to only include labels before the specified time.
user_email (Optional[str]) – Filter by the annotator email.

Return type

List[LabelLog]

get_cloud_integrations()[source]#

Return type: List[CloudIntegration]

list_label_rows(edited_before=None, edited_after=None, label_statuses=None, shadow_data_state=None, *, include_uninitialised_labels=False, label_hashes=None, data_hashes=None)[source]#

DEPRECATED - use list_label_rows_v2 to manage label rows instead.

Parameters

edited_before (Union[str, datetime, None]) – Optionally filter to only rows last edited before the specified time
edited_after (Union[str, datetime, None]) – Optionally filter to only rows last edited after the specified time
label_statuses (Optional[List[AnnotationTaskStatus]]) – Optionally filter to only those label rows that have one of the specified AnnotationTaskStatus
shadow_data_state (Optional[ShadowDataState]) – On Optionally filter by data type in Benchmark QA projects. See ShadowDataState
include_uninitialised_labels – Whether to return only label rows that are “created” and have a label_hash (default). If set to True, this will return all label rows, including those that do not have a label_hash.
data_hashes (Optional[List[str]]) – List of data hashes to filter by.
label_hashes (Optional[List[str]]) – List of label hashes to filter by.

Return type

List[LabelRowMetadata]

Returns

A list of LabelRowMetadata instances for all the matching label rows

Raises

UnknownError – If an error occurs while retrieving the data.

set_label_status(label_hash, label_status)[source]#

DEPRECATED - this function is currently not maintained.

Set the label status for a label row to a desired value.

Parameters

self – Encord client object.
label_hash (str) – unique identifier of the label row whose status is to be updated.
label_status (LabelStatus) – the new status that needs to be set.

Return type

bool

Returns

Bool.

Raises

AuthorisationError – If the label_hash provided is invalid or not a member of the project.
UnknownError – If an error occurs while updating the status.

get_label_row(uid, get_signed_url=True, *, include_object_feature_hashes=None, include_classification_feature_hashes=None, include_reviews=False)[source]#

DEPRECATED: Prefer using the list_label_rows_v2 function to interact with label rows.

Retrieve label row. If you need to retrieve multiple label rows, prefer using encord.project.Project.get_label_rows() instead.

A code example using the include_object_feature_hashes and include_classification_feature_hashes filters can be found in encord.project.Project.get_label_row().

Parameters

uid (str) – A label_hash (uid) string.
get_signed_url (bool) – Whether to generate signed urls to the data asset. Generating these should be disabled if the signed urls are not used to speed up the request.
include_object_feature_hashes (Optional[Set[str]]) – If None all the objects will be included. Otherwise, only objects labels will be included of which the feature_hash has been added.
include_classification_feature_hashes (Optional[Set[str]]) – If None all the classifications will be included. Otherwise, only classification labels will be included of which the feature_hash has been added.
include_reviews (bool) – Whether to request read only information about the reviews of the label row.

Returns

A label row instance.

Return type

LabelRow

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ResourceNotFoundError – If no label exists by the specified label_hash (uid).
UnknownError – If an error occurs while retrieving the label.
OperationNotAllowed – If the read operation is not allowed by the API key.

Expand for references to encord.project.Project.get_label_row

Projects / Project details / Collaborator session information

get_label_rows(uids, get_signed_url=True, *, include_object_feature_hashes=None, include_classification_feature_hashes=None, include_reviews=False)[source]#

DEPRECATED: Prefer using the list_label_rows_v2 function to interact with label rows.

Retrieve a list of label rows. Duplicates will be dropped. The result will come back in a random order.

This return is undefined behaviour if any of the uids are invalid (i.e. it may randomly fail or randomly succeed and should not be relied upon).

# Code example of using the object filters.
from encord.objects.common import Shape
from encord.objects.ontology_structure import OntologyStructure

project = ... # assuming you already have instantiated this Project object

# Get all feature hashes of the objects which are of type `Shape.BOUNDING_BOX`
ontology = OntologyStructure.from_dict(project.ontology)
only_bounding_box_feature_hashes = set()
for object_ in ontology.objects:
    if object_.shape == Shape.BOUNDING_BOX:
        only_bounding_box_feature_hashes.add(object_.feature_node_hash)

no_classification_feature_hashes = set()  # deliberately left empty

# Get all labels of tasks that have already been initiated.
# Include only labels of bounding boxes and exclude all
# classifications
label_hashes = []
for label_row in project.label_rows:
    # Trying to run `get_label_row` on a label_row without a `label_hash` would fail.
    if label_row["label_hash"] is not None:
        label_hashes.append(label_row["label_hash"])

all_labels = project.get_label_rows(
    label_hashes,
    include_object_feature_hashes=only_bounding_box_feature_hashes,
    include_classification_feature_hashes=no_classification_feature_hashes,
)

Parameters

uids (List[str]) – A list of label_hash (uid).
get_signed_url (bool) –

Whether to generate signed urls to the data asset. Generating these should be disabled
if the signed urls are not used to speed up the request.

include_object_feature_hashes:
If None all the objects will be included. Otherwise, only objects labels will be included of which the feature_hash has been added.

include_classification_feature_hashes:
If None all the classifications will be included. Otherwise, only classification labels will be included of which the feature_hash has been added.

include_reviews: Whether to request read only information about the reviews of the label row.

Raises

MultiLabelLimitError – If too many labels were requested. Check the error’s maximum_labels_allowed field to read the most up to date error limit.
AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ResourceNotFoundError – If no label exists by the specified label_hash (uid).
UnknownError – If an error occurs while retrieving the label.
OperationNotAllowed – If the read operation is not allowed by the API key.

Return type

List[LabelRow]

save_label_row(uid, label)[source]#

DEPRECATED: Prefer using the list_label_rows_v2 function to interact with label rows.

Save existing label row.

If you have a series of frame labels and have not updated answer dictionaries, call the construct_answer_dictionaries utilities function to do so prior to saving labels.

Parameters

uid – A label_hash (uid) string.
label – A label row instance.

Returns

Bool.

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
ResourceNotFoundError – If no label exists by the specified label_hash (uid).
UnknownError – If an error occurs while saving the label.
OperationNotAllowed – If the write operation is not allowed by the API key.
AnswerDictionaryError – If an object or classification instance is missing in answer dictionaries.
CorruptedLabelError – If a blurb is corrupted (e.g. if the frame labels have more frames than the video).

create_label_row(uid)[source]#

DEPRECATED: Prefer using the list_label_rows_v2 function to interact with label rows.

Create a label row (for data in a project not previously been labeled).

Parameters

uid (str) – the data_hash (uid) of the data unit being labeled. Available in client.get_project().get(‘label_rows’) where label_status is NOT_LABELLED.

Returns

A label row instance.

Return type

LabelRow

Raises

AuthenticationError – If the project API key is invalid.
AuthorisationError – If access to the specified resource is restricted.
UnknownError – If an error occurs while saving the label.
OperationNotAllowed – If the write operation is not allowed by the API key.
AnswerDictionaryError – If an object or classification instance is missing in answer dictionaries.
CorruptedLabelError – If a blurb is corrupted (e.g. if the frame labels have more frames than the video).
ResourceExistsError – If a label row already exists for this project data. Avoids overriding existing work.

create_bundle()[source]#

Initialises a bundle to reduce amount of network calls performed by the Encord SDK

See the encord.http.bundle.Bundle documentation for more details

Return type: Bundle

list_collaborator_timers(after, before=None, group_by_data_unit=True)[source]#

Provides information about time spent for each collaborator that has worked on the project within a specified range of dates.

Parameters

after (datetime) – the beginning of the period of interest.
before (Optional[datetime]) – the end of period of interest.
group_by_data_unit (bool) – if True, time spent by a collaborator for each data unit is provided separately, and if False, all time spent in the scope of the project is aggregated together.

Return type

Iterable[CollaboratorTimer]

Returns

Iterable[CollaboratorTimer]

Expand for references to encord.project.Project.list_collaborator_timers

Dataset#

class encord.dataset.Dataset(client, orm_dataset)[source]#

Access dataset related data and manipulate the dataset.

Expand for references to encord.dataset.Dataset

Authentication / User authentication / Authenticate with Encord - user authentication

Datasets / Data / Adding data / Adding data from a private cloud

Datasets / Managing a dataset

property dataset_hash: str#

Get the dataset hash (i.e. the Dataset ID).

Return type: str

property title: str#

Return type: str

property description: str#

Return type: str

property storage_location: encord.orm.dataset.StorageLocation#

Return type: StorageLocation

property data_rows: List[encord.orm.dataset.DataRow]#

Part of the response of this function can be configured by the encord.dataset.Dataset.set_access_settings() method.

dataset.set_access_settings(DatasetAccessSettings(fetch_client_metadata=True))
print(dataset.data_rows)

Return type: List[DataRow]

list_data_rows(title_eq=None, title_like=None, created_before=None, created_after=None, data_types=None)[source]#

Retrieve dataset rows (pointers to data, labels).

Parameters

title_eq (Optional[str]) – optional exact title row filter
title_like (Optional[str]) – optional fuzzy title row filter; SQL syntax
created_before (Union[str, datetime, None]) – optional datetime row filter
created_after (Union[str, datetime, None]) – optional datetime row filter
data_types (Optional[List[DataType]]) – optional data types row filter

Returns

A list of DataRows object that match the filter

Return type

List[DataRow]

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.

refetch_data()[source]#

The Dataset class will only fetch its properties once. Use this function if you suspect the state of those properties to be dirty.

Return type: None

get_dataset()[source]#

This function is exposed for convenience. You are encouraged to use the property accessors instead.

Return type: Dataset

set_access_settings(dataset_access_settings, *, refetch_data=True)[source]#

Parameters

dataset_access_settings (DatasetAccessSettings) – The access settings to use going forward
refetch_data (bool) – Whether a refetch_data() call should follow the update of the dataset access settings.

Return type

None

add_users(user_emails, user_role)[source]#

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.

Parameters

user_emails (List[str]) – list of user emails to be added
user_role (DatasetUserRole) – the user role to assign to all users

Return type

List[DatasetUser]

upload_video(file_path, cloud_upload_settings=CloudUploadSettings(max_retries=None, backoff_factor=None, allow_failures=False), title=None)[source]#

Upload video to Encord storage.

Parameters

file_path (str) – path to video e.g. ‘/home/user/data/video.mp4’
cloud_upload_settings (CloudUploadSettings) – Settings for uploading data into the cloud. Change this object to overwrite the default values.
title (Optional[str]) – 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)

create_image_group(file_paths, max_workers=None, cloud_upload_settings=CloudUploadSettings(max_retries=None, backoff_factor=None, allow_failures=False), title=None, *, create_video=True)[source]#

Create an image group in Encord storage. Choose this type of image upload for sequential images. Else, you can choose the Dataset.upload_image() function.

Parameters

file_paths (Iterable[str]) – a list of paths to images, e.g. [‘/home/user/data/img1.png’, ‘/home/user/data/img2.png’]
max_workers (Optional[int]) – DEPRECATED: This argument will be ignored
cloud_upload_settings (CloudUploadSettings) – Settings for uploading data into the cloud. Change this object to overwrite the default values.
title (Optional[str]) – 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 (bool) – 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)

create_dicom_series(file_paths, cloud_upload_settings=CloudUploadSettings(max_retries=None, backoff_factor=None, allow_failures=False), title=None)[source]#

Upload a DICOM series to Encord storage

Parameters

file_paths (List[str]) – a list of paths to DICOM files, e.g. [‘/home/user/data/DICOM_1.dcm’, ‘/home/user/data/DICOM_2.dcm’]
cloud_upload_settings (CloudUploadSettings) – Settings for uploading data into the cloud. Change this object to overwrite the default values.
title (Optional[str]) – 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)

upload_image(file_path, title=None, cloud_upload_settings=CloudUploadSettings(max_retries=None, backoff_factor=None, allow_failures=False))[source]#

Upload a single image to Encord storage. If your images are sequential we recommend creating an image group via the Dataset.create_image_group() function. For more information please compare https://docs.encord.com/docs/annotate-images and https://docs.encord.com/docs/annotate-videos

Parameters

file_path (Union[Path, str]) – The file path to the image
title (Optional[str]) – 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 (CloudUploadSettings) – Settings for uploading data into the cloud. Change this object to overwrite the default values.

Return type

Image

delete_image_group(data_hash)[source]#

Delete an image group in Encord storage.

Parameters: data_hash (str) – the hash of the image group to delete

delete_data(data_hashes)[source]#

Delete a video/image group from a dataset.

Parameters: data_hashes (List[str]) – list of hash of the videos/image_groups you’d like to delete, all should belong to the same dataset

add_private_data_to_dataset(integration_id, private_files, ignore_errors=False)[source]#

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 https://python.docs.encord.com/tutorials/datasets.html#adding-data-from-a-private-cloud

Parameters

integration_id (str) – The EntityId of the cloud integration you wish to use.
private_files (Union[str, Dict, Path, TextIO]) – A str path or Path object to a json file, json str or python dictionary of the files you wish to add
ignore_errors (bool) – When set to True, this will prevent individual errors from stopping the upload process.

Return type

AddPrivateDataResponse

Returns

add_private_data_response List of DatasetDataInfo objects containing data_hash and title

Expand for references to encord.dataset.Dataset.add_private_data_to_dataset

Datasets / Data / Adding data / Adding data from a private cloud

add_private_data_to_dataset_start(integration_id, private_files, ignore_errors=False)[source]#

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 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.

Parameters

integration_id (str) – The EntityId of the cloud integration you wish to use.
private_files (Union[str, Dict, Path, TextIO]) – A str path or Path object to a json file, json str or python dictionary of the files you wish to add
ignore_errors (bool) – When set to True, this will prevent individual errors from stopping the upload process.

Return type

str

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.

Expand for references to encord.dataset.Dataset.add_private_data_to_dataset_start

Datasets / Data / Adding data / Adding data from a private cloud

add_private_data_to_dataset_get_result(upload_job_id, timeout_seconds=604800)[source]#

Fetch data upload status, perform long polling process for timeout_seconds.

Parameters

upload_job_id (str) – UUID Identifier of upload job. This id enables the user to track the job progress via SDK, or web app.
timeout_seconds (int) – 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.

Return type

DatasetDataLongPolling

Returns

DatasetDataLongPolling: Response containing details about job status, errors and progress.

Expand for references to encord.dataset.Dataset.add_private_data_to_dataset_get_result

Datasets / Data / Adding data / Adding data from a private cloud

update_data_item(data_hash, new_title)[source]#

DEPRECATED: Use the individual setter properties of the respective encord.orm.dataset.DataRow instance instead. These can be retrieved via the Dataset.data_rows() function.

Update a data item

Parameters

data_hash (str) – str Data hash of the item being updated
new_title (str) – String containing the new title of the data item being updated

Return type

bool

Returns

Returns a boolean for whether the update was successful

re_encode_data(data_hashes)[source]#

Launches an async task that can re-encode a list of videos.

Parameters: data_hashes (List[str]) – 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.

re_encode_data_status(job_id)[source]#

Returns the status of an existing async task which is aimed at re-encoding videos.

Parameters

job_id (int) – id of the async task that was launched to re-encode the videos

Returns

Object containing the status of the task, along with info about the new encoded videos: in case the task has been completed

Return type

ReEncodeVideoTask

run_ocr(image_group_id)[source]#

Returns an optical character recognition result for a given image group :type image_group_id: str :param image_group_id: the id of the image group in this dataset to run OCR on

Return type: List[ImageGroupOCR]
Returns: Returns a list of ImageGroupOCR objects representing the text and corresponding coordinates found in each frame of the image group

get_cloud_integrations()[source]#

Return type: List[CloudIntegration]

Ontology#

class encord.ontology.Ontology(querier, config, instance)[source]#

Access ontology related data and manipulate the ontology. Instantiate this class via encord.user_client.EncordUserClient.get_ontology()

property ontology_hash: str#

Get the ontology hash (i.e. the Ontology ID).

Return type: str

property title: str#

Get the title of the ontology.

Return type: str

property description: str#

Get the description of the ontology.

Return type: str

property created_at: datetime.datetime#

Get the time the ontology was created at.

Return type: datetime

property last_edited_at: datetime.datetime#

Get the time the ontology was last edited at.

Return type: datetime

property structure: encord.objects.ontology_structure.OntologyStructure#

Get the structure of the ontology.

Return type: OntologyStructure

Expand for references to encord.ontology.Ontology.structure

Projects / API keys / Creating a project API key / Creating a project API key with specific rights

refetch_data()[source]#

The Ontology class will only fetch its properties once. Use this function if you suspect the state of those properties to be dirty.

Return type: None

save()[source]#

Sync local state to the server, if updates are made to structure, title or description fields

Return type: None

Ontology Structure#

class encord.objects.ontology_labels_impl.OntologyStructure(objects=<factory>, classifications=<factory>)[source]#

objects: List[encord.objects.ontology_object.Object]#

classifications: List[encord.objects.classification.Classification]#

get_child_by_hash(feature_node_hash, type_=None)[source]#

Returns the first child node of this ontology tree node with the matching feature node hash. If there is more than one child with the same feature node hash in the ontology tree node, then the ontology would be in an invalid state. Throws if nothing is found or if the type is not matched.

Parameters

feature_node_hash (str) – the feature_node_hash of the child node to search for in the ontology.
type – The expected type of the item. If the found child does not match the type, an error will be thrown.

Return type

~OntologyElementT

get_child_by_title(title, type_=None)[source]#

Returns a child node of this ontology tree node with the matching title and matching type if specified. If more than one child in this Object have the same title, then an error will be thrown. If no item is found, an error will be thrown as well.

Parameters

title (str) – The exact title of the child node to search for in the ontology.
type – The expected type of the child node. Only a node that matches this type will be returned.

Return type

~OntologyElementT

get_children_by_title(title, type_=None)[source]#

Returns all the child nodes of this ontology tree node with the matching title and matching type if specified. Title in ontologies do not need to be unique, however, we recommend unique titles when creating ontologies.

Parameters

title (str) – The exact title of the child node to search for in the ontology.
type – The expected type of the item. Only nodes that match this type will be returned.

Return type

List[~OntologyElementT]

classmethod from_dict(d)[source]#

Parameters: d (Dict[str, Any]) – a JSON blob of an “ontology structure” (e.g. from Encord web app)
Raises: KeyError – If the dict is missing a required field.
Return type: OntologyStructure

to_dict()[source]#

Return type: Dict[str, List[Dict[str, Any]]]
Returns: The dict equivalent to the ontology.
Raises: KeyError – If the dict is missing a required field.

add_object(name, shape, uid=None, color=None, feature_node_hash=None)[source]#

Adds an object class definition to the structure.

structure = ontology_structure.OntologyStructure()

eye = structure.add_object(
    name="Eye",
)
nose = structure.add_object(
    name="Nose",
)
nose_detail = nose.add_attribute(
    encord.objects.common.ChecklistAttribute,
)
nose_detail.add_option(feature_node_hash="2bc17c88", label="Is it a cute nose?")
nose_detail.add_option(feature_node_hash="86eaa4f2", label="Is it a wet nose? ")

Parameters

name (str) – the user-visible name of the object
shape (Shape) – the kind of object (bounding box, polygon, etc). See encord.objects.common.Shape enum for possible values
uid (Optional[int]) – integer identifier of the object. Normally auto-generated; omit this unless the aim is to create an exact clone of existing structure
color (Optional[str]) – the color of the object in the label editor. Normally auto-assigned, should be in ‘#1A2B3F’ syntax.
feature_node_hash (Optional[str]) – global identifier of the object. Normally auto-generated; omit this unless the aim is to create an exact clone of existing structure

Return type

Object

Returns

the created object class that can be further customised with attributes.

add_classification(uid=None, feature_node_hash=None)[source]#

Adds an classification definition to the ontology.

structure = ontology_structure.OntologyStructure()

cls = structure.add_classification(feature_node_hash="a39d81c0")
cat_standing = cls.add_attribute(
    encord.objects.common.RadioAttribute,
    feature_node_hash="a6136d14",
    name="Is the cat standing?",
    required=True,
)
cat_standing.add_option(feature_node_hash="a3aeb48d", label="Yes")
cat_standing.add_option(feature_node_hash="d0a4b373", label="No")

Parameters

uid (Optional[int]) – integer identifier of the object. Normally auto-generated; omit this unless the aim is to create an exact clone of existing structure
feature_node_hash (Optional[str]) – global identifier of the object. Normally auto-generated; omit this unless the aim is to create an exact clone of existing structure

Return type

Classification

Returns

the created classification node. Note that classification attribute should be further specified by calling its add_attribute() method.

class encord.objects.ontology_labels_impl.Object(feature_node_hash, uid, name, color, shape, attributes=<factory>)[source]#

uid: int#

name: str#

color: str#

shape: encord.objects.common.Shape#

feature_node_hash: str#

attributes: List[encord.objects.attributes.Attribute]#

property title: str#

Return type: str

property children: Sequence[encord.objects.ontology_element.OntologyElement]#

Return type: Sequence[OntologyElement]

create_instance()[source]#

Create a encord.objects.ObjectInstance to be used with a label row.

Return type: ObjectInstance

classmethod from_dict(d)[source]#

Return type: Object

to_dict()[source]#

Return type: Dict[str, Any]

T#: alias of TypeVar(‘T’, bound=encord.objects.attributes.Attribute)

add_attribute(cls, name, local_uid=None, feature_node_hash=None, required=False, dynamic=False)[source]#

Adds an attribute to the object.

Parameters

cls (Type[T]) – attribute type, one of RadioAttribute, ChecklistAttribute, TextAttribute
name (str) – the user-visible name of the attribute
local_uid (Optional[int]) – integer identifier of the attribute. Normally auto-generated; omit this unless the aim is to create an exact clone of existing ontology
feature_node_hash (Optional[str]) – global identifier of the attribute. Normally auto-generated; omit this unless the aim is to create an exact clone of existing ontology
required (bool) – whether the label editor would mark this attribute as ‘required’
dynamic (bool) – whether the attribute can have a different answer for the same object across different frames.

Return type

Returns

the created attribute that can be further specified with Options, where appropriate

Raises

ValueError – if specified local_uid or feature_node_hash violate uniqueness constraints

class encord.objects.ontology_labels_impl.Classification(feature_node_hash, uid, attributes)[source]#

Represents a whole-image classification as part of Ontology structure. Wraps a single Attribute that describes the image in general rather than an individual object.

uid: int#

feature_node_hash: str#

attributes: List[encord.objects.attributes.Attribute]#

property title: str#

Return type: str

property children: Sequence[encord.objects.ontology_element.OntologyElement]#

Return type: Sequence[OntologyElement]

create_instance()[source]#

Create a encord.objects.ClassificationInstance to be used with a label row.

Return type: ClassificationInstance

classmethod from_dict(d)[source]#

Return type: Classification

to_dict()[source]#

Return type: Dict[str, Any]

T#: alias of TypeVar(‘T’, bound=encord.objects.attributes.Attribute)

add_attribute(cls, name, local_uid=None, feature_node_hash=None, required=False)[source]#

Adds an attribute to the classification.

Parameters

cls (Type[T]) – attribute type, one of RadioAttribute, ChecklistAttribute, TextAttribute
name (str) – the user-visible name of the attribute
local_uid (Optional[int]) – integer identifier of the attribute. Normally auto-generated; omit this unless the aim is to create an exact clone of existing ontology
feature_node_hash (Optional[str]) – global identifier of the attribute. Normally auto-generated; omit this unless the aim is to create an exact clone of existing ontology
required (bool) – whether the label editor would mark this attribute as ‘required’

Return type

Returns

the created attribute that can be further specified with Options, where appropriate

Raises

ValueError – if the classification already has an attribute assigned

Encord Objects#

class encord.objects.common.PropertyType(value)[source]#

An enumeration.

RADIO = 'radio'#

TEXT = 'text'#

CHECKLIST = 'checklist'#

class encord.objects.common.Shape(value)[source]#

An enumeration.

BOUNDING_BOX = 'bounding_box'#

POLYGON = 'polygon'#

POINT = 'point'#

SKELETON = 'skeleton'#

POLYLINE = 'polyline'#

ROTATABLE_BOUNDING_BOX = 'rotatable_bounding_box'#

BITMASK = 'bitmask'#

class encord.objects.common.DeidentifyRedactTextMode(value)[source]#

An enumeration.

REDACT_ALL_TEXT = 'REDACT_ALL_TEXT'#

REDACT_NO_TEXT = 'REDACT_NO_TEXT'#

REDACT_SENSITIVE_TEXT = 'REDACT_SENSITIVE_TEXT'#

class encord.objects.common.SaveDeidentifiedDicomConditionType(value)[source]#

An enumeration.

NOT_SUBSTR = 'NOT_SUBSTR'#

IN = 'IN'#

class encord.objects.common.SaveDeidentifiedDicomConditionIn(value, dicom_tag, condition_type=SaveDeidentifiedDicomConditionType.IN)[source]#

value: List[str]#

dicom_tag: str#

condition_type: encord.objects.common.SaveDeidentifiedDicomConditionType = 'IN'#

to_dict()[source]#

Return type: Dict[str, Any]

class encord.objects.common.SaveDeidentifiedDicomConditionNotSubstr(value, dicom_tag, condition_type=SaveDeidentifiedDicomConditionType.NOT_SUBSTR)[source]#

value: str#

dicom_tag: str#

condition_type: encord.objects.common.SaveDeidentifiedDicomConditionType = 'NOT_SUBSTR'#

to_dict()[source]#

Return type: Dict[str, Any]

encord.objects.utils.short_uuid_str()[source]#

This is being used as a condensed uuid.

Return type: str

encord.objects.utils.check_type(obj, type_)[source]#

Return type: None

encord.objects.utils.does_type_match(obj, type_)[source]#

Return type: bool

encord.objects.utils.checked_cast(obj, type_)[source]#

Return type: ~T

encord.objects.utils.filter_by_type(objects, type_)[source]#

Return type: List[~T]

encord.objects.utils.is_valid_email(email)[source]#

Validate that an email is a valid one

Return type: bool

encord.objects.utils.check_email(email)[source]#

Return type: None

class encord.objects.project.ProjectDataset(dataset_hash, title, description)[source]#

dataset_hash: str#

title: str#

description: str#

static from_dict(d)[source]#

Return type: ProjectDataset

classmethod from_list(l)[source]#

Return type: List[ProjectDataset]

Label Row V2#

class encord.objects.LabelRowV2(label_row_metadata, project_client, ontology)[source]#

This class represents a single label row. It is corresponding to exactly one data row within a project. It holds all the labels for that data row.

You can access a many metadata fields with this class directly. If you want to read or write labels you will need to call initialise_labels() first. To upload your added labels call save().

property label_hash: Optional[str]#

Return type: Optional[str]

property data_hash: str#

Return type: str

property dataset_hash: str#

Return type: str

property dataset_title: str#

Return type: str

property data_title: str#

Return type: str

property data_type: encord.constants.enums.DataType#

Return type: DataType

property label_status: encord.orm.label_row.LabelStatus#

Returns the current labeling status for the label row.

Note: This method is not supported for workflow-based projects. Please see our workflow documentation for more details.

Return type: LabelStatus

property annotation_task_status: encord.orm.label_row.AnnotationTaskStatus#

Returns the current annotation task status for the label row.

Note: This method is not supported for workflow-based projects. Please see our workflow documentation for more details.

Return type: AnnotationTaskStatus

property workflow_graph_node: Optional[encord.orm.label_row.WorkflowGraphNode]#

Return type: Optional[WorkflowGraphNode]

property is_shadow_data: bool#

Return type: bool

property created_at: Optional[datetime.datetime]#

The creation date of the label row. None if the label row was not yet created.

Return type: Optional[datetime]

property last_edited_at: Optional[datetime.datetime]#

The time the label row was updated last as a whole. None if the label row was not yet created.

Return type: Optional[datetime]

property number_of_frames: int#

Return type: int

property duration: Optional[float]#

Only a value for Video data types.

Return type: Optional[float]

property fps: Optional[float]#

Only a value for Video data types.

Return type: Optional[float]

property data_link: Optional[str]#

The data link in either your cloud storage or the encord storage to the underlying object. This will be None for DICOM series or image groups that have been created without performance optimisations, as there is no single underlying file for these data types.

Return type: Optional[str]

property width: Optional[int]#

This is None for image groups without performance optimisation, as there is no single underlying width for this data type.

Return type: Optional[int]

property height: Optional[int]#

This is None for image groups without performance optimisation, as there is no single underlying width for this data type.

Return type: Optional[int]

property ontology_structure: encord.objects.ontology_structure.OntologyStructure#

Get the corresponding ontology structure

Return type: OntologyStructure

property is_labelling_initialised: bool#

Whether you can start labelling or not. If this is False, call the member initialise_labels() to read or write specific ObjectInstances or ClassificationInstances.

Return type: bool

initialise_labels(include_object_feature_hashes=None, include_classification_feature_hashes=None, include_reviews=False, overwrite=False, bundle=None)[source]#

Call this function to download or export labels stored on the Encord server, as well as to perform any other reading or writing operations. If you only want to inspect a subset of labels, you can filter them. Please note that if you filter the labels, and upload them later, you will effectively delete all the labels that were previously filtered.

If the label was not yet in progress, this will set the label status to LabelStatus.LABEL_IN_PROGRESS.

You can call this function at any point to overwrite the current labels stored in this class with the most up to date labels stored in the Encord servers. This would only matter if you manipulate the labels while someone else is working on the labels as well. You would need to supply the overwrite parameter to True

Parameters

include_object_feature_hashes (Optional[Set[str]]) – If None all the objects will be included. Otherwise, only objects labels will be included of which the feature_hash has been added. WARNING: it is only recommended to use this filter if you are reading (not writing) the labels. If you are requesting a subset of objects and later, save the label, you will effectively delete all the object instances that are stored in the Encord platform, which were not included in this filtered subset.
include_classification_feature_hashes (Optional[Set[str]]) – If None all the classifications will be included. Otherwise, only classification labels will be included of which the feature_hash has been added. WARNING: it is only recommended to use this filter if you are reading (not writing) the labels. If you are requesting a subset of classifications and later, save the label, you will effectively delete all the classification instances that are stored in the Encord platform, which were not included in this filtered subset.
include_reviews (bool) – Whether to request read only information about the reviews of the label row.
overwrite (bool) – If the label row was already initialised, you need to set this flag to True to overwrite the current labels with the labels stored in the Encord server. If this is False and the label row was already initialised, this function will throw an error.
bundle (Optional[Bundle]) – If not passed, initialisation is performed independently. If passed, it will be delayed and initialised along with other objects in the same bundle.

Return type

None

from_labels_dict(label_row_dict)[source]#

If you have a label row dictionary in the same format that the Encord servers produce, you can initialise the LabelRow from that directly. In most cases you should prefer using the initialise_labels method.

This function also initialises the label row.

Calling this function will reset all the labels that are currently stored within this class.

Parameters: label_row_dict (dict) – The dictionary of all labels as expected by the Encord format.
Return type: None

get_image_hash(frame_number)[source]#

Get the corresponding image hash of the frame number. Return None if the frame number is out of bounds. Raise an error if this function is used for non-image data types.

Return type: Optional[str]

get_frame_number(image_hash)[source]#

Get the corresponding image hash of the frame number. Return None if the image hash was not found with an associated frame number. Raise an error if this function is used for non-image data types.

Return type: Optional[int]

save(bundle=None)[source]#

Upload the created labels with the Encord server. This will overwrite any labels that someone has created in the platform in the meantime.

Parameters: bundle (Optional[Bundle]) – if not passed, save is executed immediately. If passed, it is executed as a part of the bundle
Return type: None

get_frame_view(frame=0)[source]#

Parameters: frame (Union[int, str]) – Either the frame number or the image hash if the data type is an image or image group. Defaults to the first frame.
Return type: FrameView

get_frame_views()[source]#

Return type: List[FrameView]
Returns: A list of frame views in order of available frames.

get_object_instances(filter_ontology_object=None, filter_frames=None)[source]#

Parameters

filter_ontology_object (Optional[Object]) – Optionally filter by a specific ontology object.
filter_frames (Union[int, List[int], Range, List[Range], None]) – Optionally filter by specific frames.

Return type

List[ObjectInstance]

Returns

All the `ObjectInstance`s that match the filter.

add_object_instance(object_instance, force=True)[source]#

Add an object instance to the label row. If the object instance already exists, it

Parameters

object_instance (ObjectInstance) – The object instance to add.
force (bool) – overwrites current objects, otherwise this will replace the current object.

Return type

None

add_classification_instance(classification_instance, force=False)[source]#

Add a classification instance to the label row.

Parameters

classification_instance (ClassificationInstance) – The object instance to add.
force (bool) – overwrites current objects, otherwise this will replace the current object.

Return type

None

remove_classification(classification_instance)[source]#: Remove a classification instance from a label row.

add_to_single_frame_to_hashes_map(label_item, frame)[source]#

This is an internal function, it is not meant to be called by the SDK user.

Return type: None

get_classification_instances(filter_ontology_classification=None, filter_frames=None)[source]#

Parameters

filter_ontology_classification (Optional[Classification]) – Optionally filter by a specific ontology classification.
filter_frames (Union[int, List[int], Range, List[Range], None]) – Optionally filter by specific frames.

Return type

List[ClassificationInstance]

Returns

All the `ObjectInstance`s that match the filter.

remove_object(object_instance)[source]#: Remove an object instance from a label row.

to_encord_dict()[source]#

This is an internal helper function. Likely this should not be used by a user. To upload labels use the save() function.

Return type: Dict[str, Any]

workflow_reopen()[source]#

A label row is returned to the first annotation stage for re-labeling. No data will be lost during this call.

This method is only relevant for the projects that use the Workflow feature, and will raise an error for pre-workflow projects.

Return type: None

workflow_complete()[source]#

A label row is moved to the final workflow node, marking it as ‘Complete’.

This method can be called only for labels for which initialise_labels() was called at least ance, and consequentially the “label_hash” field is not None.

Please note that labels need not be initialized every time the workflow_complete() method is called.

This method is only relevant for the projects that use the Workflow feature, and will raise an error for projects that don’t use Workflows.

Return type: None

class FrameView(label_row, label_row_read_only_data, frame)[source]#

This class can be used to inspect what object/classification instances are on a given frame or what metadata, such as a image file size, is on a given frame.

property image_hash: str#

Return type: str

property image_title: str#

Return type: str

property file_type: str#

Return type: str

property frame: int#

Return type: int

property width: int#

Return type: int

property height: int#

Return type: int

property data_link: Optional[str]#

Return type: Optional[str]

add_object_instance(object_instance, coordinates, *, overwrite=False, created_at=None, created_by=None, last_edited_at=None, last_edited_by=None, confidence=None, manual_annotation=None)[source]#

Return type: None

add_classification_instance(classification_instance, *, overwrite=False, created_at=None, created_by=None, confidence=1.0, manual_annotation=True, last_edited_at=None, last_edited_by=None)[source]#

Return type: None

get_object_instances(filter_ontology_object=None)[source]#

Parameters: filter_ontology_object (Optional[Object]) – Optionally filter by a specific ontology object.
Return type: List[ObjectInstance]
Returns: All the `ObjectInstance`s that match the filter.

get_classification_instances(filter_ontology_classification=None)[source]#

Parameters: filter_ontology_classification (Optional[Classification]) – Optionally filter by a specific ontology object.
Return type: List[ClassificationInstance]
Returns: All the `ObjectInstance`s that match the filter.

class FrameLevelImageGroupData(image_hash, image_title, file_type, frame_number, width, height, data_link=None)[source]#

This is an internal helper class. A user should not directly interact with it.

image_hash: str#

image_title: str#

file_type: str#

frame_number: int#

width: int#

height: int#

data_link: Optional[str] = None#

class LabelRowReadOnlyData(label_hash, created_at, last_edited_at, data_hash, data_type, label_status, annotation_task_status, workflow_graph_node, is_shadow_data, number_of_frames, duration, fps, dataset_hash, dataset_title, data_title, width, height, data_link, frame_level_data=<factory>, image_hash_to_frame=<factory>, frame_to_image_hash=<factory>)[source]#

This is an internal helper class. A user should not directly interact with it.

label_hash: Optional[str]#: This is None if the label row does not have any labels and was not initialised for labelling.

created_at: Optional[datetime.datetime]#: This is None if the label row does not have any labels and was not initialised for labelling.

last_edited_at: Optional[datetime.datetime]#: This is None if the label row does not have any labels and was not initialised for labelling.

data_hash: str#

data_type: encord.constants.enums.DataType#

label_status: encord.orm.label_row.LabelStatus#

annotation_task_status: Optional[encord.orm.label_row.AnnotationTaskStatus]#

workflow_graph_node: Optional[encord.orm.label_row.WorkflowGraphNode]#

is_shadow_data: bool#

number_of_frames: int#

duration: Optional[float]#

fps: Optional[float]#

dataset_hash: str#

dataset_title: str#

data_title: str#

width: Optional[int]#

height: Optional[int]#

data_link: Optional[str]#

frame_level_data: Dict[int, encord.objects.ontology_labels_impl.LabelRowV2.FrameLevelImageGroupData]#

image_hash_to_frame: Dict[str, int]#

frame_to_image_hash: Dict[int, str]#

Label ObjectInstance#

class encord.objects.ObjectInstance(ontology_object, *, object_hash=None)[source]#

An object instance is an object that has coordinates and can be places on one or multiple frames in a label row.

is_assigned_to_label_row()[source]#

Return type: Optional[LabelRowV2]

property object_hash: str#

A unique identifier for the object instance.

Return type: str

property ontology_item: encord.objects.ontology_object.Object#

Return type: Object

property feature_hash: str#

Feature node hash from the project ontology

Return type: str

property object_name: str#

Object name from the project ontology

Return type: str

get_answer(attribute, filter_answer=None, filter_frame=None, is_dynamic=None)[source]#

Get the answer set for a given ontology Attribute. Returns None if the attribute is not yet answered.

For the ChecklistAttribute, it returns None if and only if the attribute is nested and the parent is unselected. Otherwise, if not yet answered it will return an empty list.

Parameters

attribute (Attribute) – The ontology attribute to get the answer for.
filter_answer (Union[str, Option, Iterable[Option], None]) – A filter for a specific answer value. Only applies to dynamic attributes.
filter_frame (Optional[int]) – A filter for a specific frame. Only applies to dynamic attributes.
is_dynamic (Optional[bool]) – Optionally specify whether a dynamic answer is expected or not. This will throw if it is set incorrectly according to the attribute. Set this to narrow down the return type.

Return type

Union[str, Option, Iterable[Option], List[AnswerForFrames], None]

Returns

If the attribute is static, then the answer value is returned, assuming an answer value has already been set. If the attribute is dynamic, the AnswersForFrames object is returned.

set_answer(answer, attribute=None, frames=None, overwrite=False)[source]#

Set the answer for a given ontology Attribute. This is the equivalent of e.g. selecting a checkbox in the UI after drawing the ObjectInstance. There is only one answer per ObjectInstance per Attribute, unless the attribute is dynamic (check the args list for more instructions on how to set dynamic answers).

Parameters

answer (Union[str, Option, Sequence[Option]]) – The answer to set.
attribute (Optional[Attribute]) – The ontology attribute to set the answer for. If not set, this will be attempted to be inferred. For answers to encord.objects.common.RadioAttribute or encord.objects.common.ChecklistAttribute, this can be inferred automatically. For encord.objects.common.TextAttribute, this will only be inferred there is only one possible TextAttribute to set for the entire object instance. Otherwise, a encord.exceptionsLabelRowError will be thrown.
frames (Union[int, List[int], Range, List[Range], None]) – Only relevant for dynamic attributes. The frames to set the answer for. If None, the answer is set for all frames that this object currently has set coordinates for (also overwriting current answers). This will not automatically propagate the answer to new frames that are added in the future. If this is anything but None for non-dynamic attributes, this will throw a ValueError.
overwrite (bool) – If True, the answer will be overwritten if it already exists. If False, this will throw a LabelRowError if the answer already exists. This argument is ignored for dynamic attributes.

Return type

None

set_answer_from_list(answers_list)[source]#

This is a low level helper function and should usually not be used directly.

Sets the answer for the classification from a dictionary.

Parameters: answers_list (List[Dict[str, Any]]) – The list of dictionaries to set the answer from.
Return type: None

delete_answer(attribute, filter_answer=None, filter_frame=None)[source]#

This resets the answer of an attribute as if it was never set.

Parameters

attribute (Attribute) – The attribute to delete the answer for.
filter_answer (Union[str, Option, Iterable[Option], None]) – A filter for a specific answer value. Delete only answers with the provided value. Only applies to dynamic attributes.
filter_frame (Optional[int]) – A filter for a specific frame. Only applies to dynamic attributes.

Return type

None

check_within_range(frame)[source]#

Return type: None

set_for_frames(coordinates, frames=0, *, overwrite=False, created_at=None, created_by=None, last_edited_at=None, last_edited_by=None, confidence=None, manual_annotation=None, reviews=None, is_deleted=None)[source]#

Places the object onto the specified frame. If the object already exists on the frame and overwrite is set to True, the currently specified values will be overwritten.

Parameters

coordinates (Union[BoundingBoxCoordinates, RotatableBoundingBoxCoordinates, PointCoordinate, PolygonCoordinates, PolylineCoordinates, SkeletonCoordinates, BitmaskCoordinates]) – The coordinates of the object in the frame. This will throw an error if the type of the coordinates does not match the type of the attribute in the object instance.
frames (Union[int, List[int], Range, List[Range]]) – The frames to add the object instance to. Defaulting to the first frame for convenience.
overwrite (bool) – If True, overwrite existing data for the given frames. This will not reset all the non-specified values. If False and data already exists for the given frames, raises an error.
created_at (Optional[datetime]) – Optionally specify the creation time of the object instance on this frame. Defaults to datetime.now().
created_by (Optional[str]) – Optionally specify the creator of the object instance on this frame. Defaults to the current SDK user.
last_edited_at (Optional[datetime]) – Optionally specify the last edit time of the object instance on this frame. Defaults to datetime.now().
last_edited_by (Optional[str]) – Optionally specify the last editor of the object instance on this frame. Defaults to the current SDK user.
confidence (Optional[float]) – Optionally specify the confidence of the object instance on this frame. Defaults to 1.0.
manual_annotation (Optional[bool]) – Optionally specify whether the object instance on this frame was manually annotated. Defaults to True.
reviews (Optional[List[dict]]) – Should only be set by internal functions.
is_deleted (Optional[bool]) – Should only be set by internal functions.

Return type

None

get_annotation(frame=0)[source]#

Get the annotation for the object instance on the specified frame.

Parameters: frame (Union[int, str]) – Either the frame number or the image hash if the data type is an image or image group. Defaults to the first frame.
Return type: Annotation

copy()[source]#

Creates an exact copy of this ObjectInstance but with a new object hash and without being associated to any LabelRowV2. This is useful if you want to add the semantically same ObjectInstance to multiple `LabelRowV2`s.

Return type: ObjectInstance

get_annotations()[source]#

Get all annotations for the object instance on all frames it has been placed to.

Return type: List[Annotation]
Returns: A list of ObjectInstance.Annotation in order of available frames.

remove_from_frames(frames)[source]#: Ensure that it will be removed from all frames.

is_valid()[source]#

Check if is valid, could also return some human/computer messages.

Return type: None

are_dynamic_answers_valid()[source]#

Whether there are any dynamic answers on frames that have no coordinates.

Return type: None

class Annotation(object_instance, frame)[source]#

This class can be used to set or get data for a specific annotation (i.e. the ObjectInstance for a given frame number).

property frame: int#

Return type: int

property coordinates: Union[encord.objects.coordinates.BoundingBoxCoordinates, encord.objects.coordinates.RotatableBoundingBoxCoordinates, encord.objects.coordinates.PointCoordinate, encord.objects.coordinates.PolygonCoordinates, encord.objects.coordinates.PolylineCoordinates, encord.objects.coordinates.SkeletonCoordinates, encord.objects.bitmask.BitmaskCoordinates]#

Return type: Union[BoundingBoxCoordinates, RotatableBoundingBoxCoordinates, PointCoordinate, PolygonCoordinates, PolylineCoordinates, SkeletonCoordinates, BitmaskCoordinates]

property created_at: datetime.datetime#

Return type: datetime

property created_by: Optional[str]#

Return type: Optional[str]

property last_edited_at: datetime.datetime#

Return type: datetime

property last_edited_by: Optional[str]#

Return type: Optional[str]

property confidence: float#

Return type: float

property manual_annotation: bool#

Return type: bool

property reviews: Optional[List[Dict[str, Any]]]#

A read only property about the reviews that happened for this object on this frame.

Return type: Optional[List[Dict[str, Any]]]

property is_deleted: Optional[bool]#

This property is only relevant for internal use.

Return type: Optional[bool]

class FrameInfo(created_at=<factory>, created_by=None, last_edited_at=<factory>, last_edited_by=None, confidence=1.0, manual_annotation=True, reviews=None, is_deleted=None)[source]#

created_at: datetime.datetime#

created_by: Optional[str] = None#: None defaults to the user of the SDK once uploaded to the server.

last_edited_at: datetime.datetime#

last_edited_by: Optional[str] = None#: None defaults to the user of the SDK once uploaded to the server.

confidence: float = 1.0#

manual_annotation: bool = True#

reviews: Optional[List[dict]] = None#

is_deleted: Optional[bool] = None#

static from_dict(d)[source]#

update_from_optional_fields(created_at=None, created_by=None, last_edited_at=None, last_edited_by=None, confidence=None, manual_annotation=None, reviews=None, is_deleted=None)[source]#

Return a new instance with the specified fields updated.

Return type: None

class FrameData(coordinates, object_frame_instance_info)[source]#

coordinates: Union[encord.objects.coordinates.BoundingBoxCoordinates, encord.objects.coordinates.RotatableBoundingBoxCoordinates, encord.objects.coordinates.PointCoordinate, encord.objects.coordinates.PolygonCoordinates, encord.objects.coordinates.PolylineCoordinates, encord.objects.coordinates.SkeletonCoordinates, encord.objects.bitmask.BitmaskCoordinates]#

object_frame_instance_info: encord.objects.ontology_object_instance.ObjectInstance.FrameInfo#

Label ClassificationInstance#

class encord.objects.ClassificationInstance(ontology_classification, *, classification_hash=None)[source]#

property classification_hash: str#

A unique identifier for the classification instance.

Return type: str

property ontology_item: encord.objects.classification.Classification#

Return type: Classification

property classification_name: str#

Classification name from the project ontology

Return type: str

property feature_hash: str#

Feature node hash from the project ontology

Return type: str

is_assigned_to_label_row()[source]#

Return type: bool

set_for_frames(frames=0, *, overwrite=False, created_at=None, created_by=None, confidence=1.0, manual_annotation=True, last_edited_at=None, last_edited_by=None, reviews=None)[source]#

Places the classification onto the specified frame. If the classification already exists on the frame and overwrite is set to True, the currently specified values will be overwritten.

Parameters

frames (Union[int, List[int], Range, List[Range]]) – The frame to add the classification instance to. Defaulting to the first frame for convenience.
overwrite (bool) – If True, overwrite existing data for the given frames. This will not reset all the non-specified values. If False and data already exists for the given frames, raises an error.
created_at (Optional[datetime]) – Optionally specify the creation time of the classification instance on this frame. Defaults to datetime.now().
created_by (Optional[str]) – Optionally specify the creator of the classification instance on this frame. Defaults to the current SDK user.
last_edited_at (Optional[datetime]) – Optionally specify the last edit time of the classification instance on this frame. Defaults to datetime.now().
last_edited_by (Optional[str]) – Optionally specify the last editor of the classification instance on this frame. Defaults to the current SDK user.
confidence (float) – Optionally specify the confidence of the classification instance on this frame. Defaults to 1.0.
manual_annotation (bool) – Optionally specify whether the classification instance on this frame was manually annotated. Defaults to True.
reviews (Optional[List[dict]]) – Should only be set by internal functions.

Return type

None

get_annotation(frame=0)[source]#

Parameters: frame (Union[int, str]) – Either the frame number or the image hash if the data type is an image or image group. Defaults to the first frame.
Return type: Annotation

remove_from_frames(frames)[source]#

Return type: None

get_annotations()[source]#

Return type: List[Annotation]
Returns: A list of ClassificationInstance.Annotation in order of available frames.

is_valid()[source]#

Return type: None

set_answer(answer, attribute=None, overwrite=False)[source]#

Set the answer for a given ontology Attribute. This is the equivalent of e.g. selecting a checkbox in the UI after adding a ClassificationInstance. There is only one answer per ClassificationInstance per Attribute.

Parameters

answer (Union[str, Option, Sequence[Option]]) – The answer to set.
attribute (Optional[Attribute]) – The ontology attribute to set the answer for. If not set, this will be attempted to be inferred. For answers to encord.objects.common.RadioAttribute or encord.objects.common.ChecklistAttribute, this can be inferred automatically. For encord.objects.common.TextAttribute, this will only be inferred there is only one possible TextAttribute to set for the entire object instance. Otherwise, a encord.exceptionsLabelRowError will be thrown.
overwrite (bool) – If True, the answer will be overwritten if it already exists. If False, this will throw a LabelRowError if the answer already exists.

Return type

None

set_answer_from_list(answers_list)[source]#

This is a low level helper function and should not be used directly.

Sets the answer for the classification from a dictionary.

Parameters: answers_list (List[Dict[str, Any]]) – The list to set the answer from.
Return type: None

get_answer(attribute=None)[source]#

Get the answer set for a given ontology Attribute. Returns None if the attribute is not yet answered.

For the ChecklistAttribute, it returns None if and only if the attribute is nested and the parent is unselected. Otherwise, if not yet answered it will return an empty list.

Parameters: attribute (Optional[Attribute]) – The ontology attribute to get the answer for.
Return type: Union[str, Option, Iterable[Option], None]

delete_answer(attribute=None)[source]#

This resets the answer of an attribute as if it was never set.

Parameters: attribute (Optional[Attribute]) – The ontology attribute to delete the answer for. If not provided, the first level attribute is used.
Return type: None

copy()[source]#

Creates an exact copy of this ClassificationInstance but with a new classification hash and without being associated to any LabelRowV2. This is useful if you want to add the semantically same ClassificationInstance to multiple `LabelRowV2`s.

Return type: ClassificationInstance

get_all_static_answers()[source]#

A low level helper function.

Return type: List[Answer]

class Annotation(classification_instance, frame)[source]#

This class can be used to set or get data for a specific annotation (i.e. the ClassificationInstance for a given frame number).

property frame: int#

Return type: int

property created_at: datetime.datetime#

Return type: datetime

property created_by: Optional[str]#

Return type: Optional[str]

property last_edited_at: datetime.datetime#

Return type: datetime

property last_edited_by: Optional[str]#

Return type: Optional[str]

property confidence: float#

Return type: float

property manual_annotation: bool#

Return type: bool

property reviews: Optional[List[dict]]#

A read only property about the reviews that happened for this object on this frame.

Return type: Optional[List[dict]]

class FrameData(created_at=<factory>, created_by=None, confidence=1.0, manual_annotation=True, last_edited_at=<factory>, last_edited_by=None, reviews=None)[source]#

created_at: datetime.datetime#

created_by: Optional[str] = None#

confidence: float = 1.0#

manual_annotation: bool = True#

last_edited_at: datetime.datetime#

last_edited_by: Optional[str] = None#

reviews: Optional[List[dict]] = None#

static from_dict(d)[source]#

Return type: FrameData

update_from_optional_fields(created_at=None, created_by=None, confidence=None, manual_annotation=None, last_edited_at=None, last_edited_by=None, reviews=None)[source]#

Return type: None

class encord.objects.AnswerForFrames(answer, ranges)[source]#

answer: Union[str, encord.objects.options.Option, Iterable[encord.objects.options.Option]]#

ranges: List[encord.objects.frames.Range]#: The ranges are essentially a run length encoding of the frames where the unique answer is set. They are sorted in ascending order.

Label Frame Utilities#

encord.objects.frames.Frames#: alias of Union[int, List[int], encord.objects.frames.Range, List[encord.objects.frames.Range]]

encord.objects.frames.Ranges#: alias of List[encord.objects.frames.Range]

class encord.objects.frames.Range(start, end)[source]#

start: int#

end: int#

encord.objects.frames.frame_to_range(frame)[source]#

Convert a single frame to a Range.

Parameters: frame (int) – The single frame
Return type: Range

encord.objects.frames.frames_to_ranges(frames)[source]#

Create a sorted list (in ascending order) of run length encoded ranges of the frames. The Ranges will not be overlapping.

Parameters: frames (Collection[int]) – A collection of integers representing frames
Return type: List[Range]

encord.objects.frames.ranges_to_list(ranges)[source]#

Convert a list of Range to a list of lists (run length encoded) of integers.

Parameters: ranges (List[Range]) – A list of Range objects
Return type: List[List[int]]

encord.objects.frames.range_to_ranges(range_)[source]#

Convert a single Range to a list of Ranges.

Parameters: range – The single Range
Return type: List[Range]

encord.objects.frames.range_to_frames(range_)[source]#

Convert a single Range (run length encoded) to a list of integers. Useful to flatten out run length encoded values.

Parameters: range – The single Range
Return type: List[int]

encord.objects.frames.ranges_to_frames(range_list)[source]#

Convert a list of Ranges (run length encoded) to a list of integers. Useful to flatten out run length encoded values.

Parameters: range_list (List[Range]) – A list of Ranges
Return type: List[int]

encord.objects.frames.ranges_list_to_ranges(range_list)[source]#

Convert a list of lists (run length encoded) of integers to a list of Ranges.

Parameters: range_list (List[List[int]]) – A list of lists (run length encoded) of integers
Return type: List[Range]

encord.objects.frames.frames_class_to_frames_list(frames_class)[source]#

Convert a flexible Frames class to a list of integers.

Parameters: frames_class (Union[int, List[int], Range, List[Range]]) – A Frames class
Return type: List[int]

Utilities - Client#

Expand for references to encord.utilities.client_utilities

Projects / Other Resources / CVAT Integration / Importing with the Encord SDK

encord.utilities.client_utilities.pretty_print(data)[source]#

class encord.utilities.client_utilities.APIKeyScopes(value)[source]#

The APIKeyScope is used to provide specific access rights to a project through EncordUserClient.create_project_api_key(). The options are a follows:

LABEL_READ: access to tutorials/projects:Getting label rows

LABEL_WRITE: access to tutorials/projects:Saving label rows

MODEL_INFERENCE: access to Inference

MODEL_TRAIN: access to Creating a model row and Training

LABEL_LOGS_READ: access to Reviewing label logs

ALGO_LIBRARY: access to algorithms like Object interpolation

Expand for references to encord.utilities.client_utilities.APIKeyScopes

Projects / API keys / Creating a project API key / Creating a project API key with specific rights

LABEL_READ = 'label.read'#

LABEL_WRITE = 'label.write'#

MODEL_INFERENCE = 'model.inference'#

MODEL_TRAIN = 'model.train'#

LABEL_LOGS_READ = 'label_logs.read'#

ALGO_LIBRARY = 'algo.library'#

class encord.utilities.client_utilities.LocalImport(file_path)[source]#

file_path: Supply the path of the exported folder which contains the images and annotations.xml file. Make sure to select “Save images” when exporting your CVAT Task or Project.

Expand for references to encord.utilities.client_utilities.LocalImport

Projects / Other Resources / CVAT Integration / Importing with the Encord SDK

file_path: str#

encord.utilities.client_utilities.ImportMethod#: Using images/videos in cloud storage as an alternative import method will be supported in the future.

class encord.utilities.client_utilities.Issue(issue_type, instances)[source]#

For each issue_type there may be multiple occurrences which are documented in the instances. The instances list can provide additional information on how the issue was encountered. If there is no additional information available, the instances list will be empty.

issue_type: str#

instances: List[str]#

class encord.utilities.client_utilities.Issues(errors, warnings, infos)[source]#

Any issues that came up during importing a project. These usually come from incompatibilities between data saved on different platforms.

errors: List[encord.utilities.client_utilities.Issue]#

warnings: List[encord.utilities.client_utilities.Issue]#

infos: List[encord.utilities.client_utilities.Issue]#

static from_dict(d)[source]#

Return type: Issues

class encord.utilities.client_utilities.CvatImporterSuccess(project_hash, dataset_hash, issues)[source]#

Expand for references to encord.utilities.client_utilities.CvatImporterSuccess

Projects / Other Resources / CVAT Integration / Importing with the Encord SDK

project_hash: str#

dataset_hash: str#

issues: encord.utilities.client_utilities.Issues#

class encord.utilities.client_utilities.CvatImporterError(dataset_hash, issues)[source]#

dataset_hash: str#

issues: encord.utilities.client_utilities.Issues#

encord.utilities.client_utilities.parse_datetime(key, val)[source]#

Return type: Optional[str]

encord.utilities.client_utilities.optional_set_to_list(s)[source]#

Return type: Optional[List[~T]]

Utilities - Project User#

Expand for references to encord.utilities.project_user

class encord.utilities.project_user.ProjectUserRole(value)[source]#

An enumeration.

Expand for references to encord.utilities.project_user.ProjectUserRole

Datasets / Data / Adding data / Adding data from a private cloud

ADMIN = 0#

ANNOTATOR = 1#

REVIEWER = 2#

ANNOTATOR_REVIEWER = 3#

TEAM_MANAGER = 4#

class encord.utilities.project_user.ProjectUser(user_email, user_role, project_hash)[source]#

user_email: str#

user_role: encord.utilities.project_user.ProjectUserRole#

project_hash: str#

classmethod from_dict(json_dict)[source]#

Utilities - Label#

encord.utilities.label_utilities.construct_answer_dictionaries(label_row)[source]#

Adds answer object and classification answer dictionaries from a label row if they do not exist. Integrity checks are conducted upon saving of labels.

Parameters: label_row – A label row.
Returns: A label row instance with updated answer dictionaries
Return type: LabelRow

encord.utilities.label_utilities.add_answers_to_items(items, classification_answers, object_answers)[source]#: If object_hash (uid) or classification_hash (uid) are not in answer dictionaries, add key entry with empty classification list.

Utilities - ORM classes#

class encord.orm.api_key.ApiKeyMeta(dic)[source]#

ApiKeyMeta contains key information.

ORM:

title, resource_type

DB_FIELDS: collections.OrderedDict = {'resource_type': <class 'str'>, 'title': <class 'str'>}#

NON_UPDATABLE_FIELDS: set = {'resource_type'}#

class encord.orm.base_orm.BaseORM(dic)[source]#

Base ORM for all database objects.

DB_FIELDS: collections.OrderedDict = {}#

NON_UPDATABLE_FIELDS: set = {}#

static from_db_row(row, db_field)[source]#: Static method for conveniently converting db row to client object. :param row: :param db_field: :return:

to_dic(time_str=True)[source]#

Conveniently set client object as dict. Only considers the dict items, no other object attr will be counted

Parameters: time_str (bool) – if set to True, will convert datetime field to str with format %Y-%m-%d %H:%M:%S. If False, will keep the original datetime type. Default will be True.

updatable_fields()[source]#

class encord.orm.base_orm.BaseListORM(iter_)[source]#

A wrapper for a list of objects of a specific ORM.

BASE_ORM_TYPE#: alias of encord.orm.base_orm.BaseORM

Expand for references to encord.orm.cloud_integration

class encord.orm.cloud_integration.CloudIntegration(id, title)[source]#

Expand for references to encord.orm.cloud_integration.CloudIntegration

Datasets / Data / Adding data / Adding data from a private cloud

id: str#

title: str#

Expand for references to encord.orm.dataset

Datasets / Data / Adding data / Adding data from a private cloud

Datasets / Data / Re-encoding videos / Check the status of a re-encoding task

Datasets / API keys / Creating a dataset API key with specific rights

Datasets / API keys / Creating a master API key with full rights

Datasets / Listing existing datasets

class encord.orm.dataset.DatasetUserRole(value)[source]#

An enumeration.

ADMIN = 0#

USER = 1#

class encord.orm.dataset.DatasetUser(user_email, user_role, dataset_hash)[source]#

user_email: str#

user_role: encord.orm.dataset.DatasetUserRole#

dataset_hash: str#

classmethod from_dict(json_dict)[source]#

class encord.orm.dataset.DatasetUsers[source]#

class encord.orm.dataset.DataClientMetadata(payload)[source]#

payload: dict#

class encord.orm.dataset.ImageData(image_hash, title, file_link, file_type, file_size, storage_location, created_at, last_edited_at, width, signed_url, height)[source]#

Information about individual images within a single DataRow of type DataType.IMG_GROUP. Get this information via the DataRow.images property.

property image_hash: str#

Return type: str

property title: str#

Return type: str

property file_link: str#

Return type: str

property file_type: str#

The MIME type of the file.

Return type: str

property file_size: int#

The size of the file in bytes.

Return type: int

property storage_location: encord.orm.dataset.StorageLocation#

Return type: StorageLocation

property created_at: datetime.datetime#

Return type: datetime

property last_edited_at: datetime.datetime#

Return type: datetime

property height: int#

Return type: int

property width: int#

Return type: int

property signed_url: Optional[str]#

The signed URL if one was generated when this class was created.

Return type: Optional[str]

classmethod from_dict(json_dict)[source]#

Return type: ImageData

classmethod from_list(json_list)[source]#

Return type: List[ImageData]

class encord.orm.dataset.DataRow(uid, title, data_type, created_at, last_edited_at, width, height, file_link, file_size, file_type, storage_location, client_metadata, frames_per_second, duration, images_data, signed_url, is_optimised_image_group)[source]#

Each individual DataRow is one upload of a video, image group, single image, or DICOM series.

This class has dict-style accessors for backwards compatibility. Clients who are using this class for the first time are encouraged to use the property accessors and setters instead of the underlying dictionary. The mixed use of the dict style member functions and the property accessors and setters is discouraged.

WARNING: Do NOT use the .data member of this class. Its usage could corrupt the correctness of the datastructure.

property uid: str#

The unique identifier for this data row. Note that the setter does not update the data on the server.

Return type: str

property title: str#

The data title.

The setter updates the custom client metadata. This queues a request for the backend which will be executed on a call of DataRow.upload().

Return type: str

property data_type: encord.constants.enums.DataType#

Return type: DataType

property created_at: datetime.datetime#

Return type: datetime

property frames_per_second: Optional[int]#

If the data type is DataType.VIDEO this returns the actual number of frames per second for the video. Otherwise, it returns None as a frames_per_second field is not applicable.

Return type: Optional[int]

property duration: Optional[int]#

If the data type is DataType.VIDEO this returns the actual duration for the video. Otherwise, it returns None as a duration field is not applicable.

Return type: Optional[int]

property client_metadata: Optional[dict]#

The currently cached client metadata. To cache the client metadata, use the fetch() function.

The setter updates the custom client metadata. This queues a request for the backend which will be executed on a call of DataRow.upload().

Return type: Optional[dict]

property width: Optional[int]#

An actual width of the data asset. This is None for data types of DataType.IMG_GROUP where is_image_sequence is False, because each image in this group can have a different dimension. Inspect the images to get the height of individual images.

Return type: Optional[int]

property height: Optional[int]#

An actual height of the data asset. This is None for data types of DataType.IMG_GROUP where is_image_sequence is False, because each image in this group can have a different dimension. Inspect the images to get the height of individual images.

Return type: Optional[int]

property last_edited_at: datetime.datetime#

Return type: datetime

property file_link: Optional[str]#

A permanent file link of the given data asset. When stored in StorageLocation.CORD_STORAGE this will be the internal file path. In private bucket storage location this will be the full path to the file. If the data type is DataType.DICOM then this returns None as no single file is associated with the series.

Return type: Optional[str]

property signed_url: Optional[str]#

The cached signed url of the given data asset. To cache the signed url, use the fetch() function.

Return type: Optional[str]

property file_size: int#

The file size of the given data asset in bytes.

Return type: int

property file_type: str#

A MIME file type of the given data asset as a string

Return type: str

property storage_location: encord.orm.dataset.StorageLocation#

Return type: StorageLocation

property images_data: Optional[List[encord.orm.dataset.ImageData]]#

A list of the cached ImageData objects for the given data asset. Fetch the images with appropriate settings in the fetch() function. If the data type is not DataType.IMG_GROUP then this returns None.

Return type: Optional[List[ImageData]]

property is_optimised_image_group: Optional[bool]#

If the data type is an DataType.IMG_GROUP, returns whether this is a performance optimised image group. Returns None for other data types.

DEPRECATED: This method is deprecated and will be removed in uplocming library version. Please usie is_image_sequence() instead

Return type: Optional[bool]

property is_image_sequence: Optional[bool]#

If the data type is an DataType.IMG_GROUP, returns whether this is an image sequence. Returns None for other data types.

For more details refer to the documentation on image sequnces

Return type: Optional[bool]

refetch_data(*, signed_url=False, images_data_fetch_options=None, client_metadata=False)[source]#

Fetches all the most up-to-date data. If any of the parameters are falsy, the current values will not be updated.

Parameters

signed_url (bool) – If True, this will fetch a generated signed url of the data asset.
images_data_fetch_options (Optional[ImagesDataFetchOptions]) – If not None, this will fetch the image data of the data asset. You can additionally specify what to fetch with the ImagesDataFetchOptions class.
client_metadata (bool) – If True, this will fetch the client metadata of the data asset.

save()[source]#

Sync local state to the server, if updates are made. This is a blocking function.

The newest values from the Encord server will update the current DataRow object.

Return type: None

classmethod from_dict(json_dict)[source]#

Return type: DataRow

classmethod from_dict_list(json_list)[source]#

Return type: List[DataRow]

class encord.orm.dataset.DataRows(data_rows)[source]#

This is a helper class that forms request for filtered dataset rows Not intended to be used directly

classmethod from_dict(json_dict)[source]#

Return type: DataRow

class encord.orm.dataset.DatasetInfo(dataset_hash, user_hash, title, description, type, created_at, last_edited_at)[source]#

This class represents a dataset in the context of listing

Expand for references to encord.orm.dataset.DatasetInfo

dataset_hash: str#

user_hash: str#

title: str#

description: str#

type: int#

created_at: datetime.datetime#

last_edited_at: datetime.datetime#

class encord.orm.dataset.Dataset(title, storage_location, data_rows, dataset_hash, description=None)[source]#

property dataset_hash: str#

Return type: str

property title: str#

Return type: str

property description: str#

Return type: str

property storage_location: encord.orm.dataset.StorageLocation#

Return type: StorageLocation

property data_rows: List[encord.orm.dataset.DataRow]#

Return type: List[DataRow]

classmethod from_dict(json_dict)[source]#

Return type: Dataset

class encord.orm.dataset.DatasetDataInfo(data_hash, title)[source]#

data_hash: str#

title: str#

classmethod from_dict(json_dict)[source]#

Return type: DatasetDataInfo

class encord.orm.dataset.AddPrivateDataResponse(dataset_data_list)[source]#

Response of add_private_data_to_dataset

Expand for references to encord.orm.dataset.AddPrivateDataResponse

Datasets / Data / Adding data / Adding data from a private cloud

dataset_data_list: List[encord.orm.dataset.DatasetDataInfo]#

classmethod from_dict(json_dict)[source]#

Return type: AddPrivateDataResponse

class encord.orm.dataset.DatasetAPIKey(dataset_hash, api_key, title, key_hash, scopes)[source]#

Expand for references to encord.orm.dataset.DatasetAPIKey

Datasets / API keys / Creating a dataset API key with specific rights

Datasets / API keys / Creating a master API key with full rights

dataset_hash: str#

api_key: str#

title: str#

key_hash: str#

scopes: List[encord.orm.dataset.DatasetScope]#

classmethod from_dict(json_dict)[source]#

Return type: DatasetAPIKey

class encord.orm.dataset.CreateDatasetResponse(title, storage_location, dataset_hash, user_hash)[source]#

Expand for references to encord.orm.dataset.CreateDatasetResponse

property title: str#

Return type: str

property storage_location: encord.orm.dataset.StorageLocation#

Return type: StorageLocation

property dataset_hash: str#

Return type: str

property user_hash: str#

Return type: str

classmethod from_dict(json_dict)[source]#

Return type: CreateDatasetResponse

class encord.orm.dataset.StorageLocation(value)[source]#

An enumeration.

Expand for references to encord.orm.dataset.StorageLocation

Datasets / API keys / Creating a dataset API key with specific rights

CORD_STORAGE = 0#

AWS = 1#

GCP = 2#

AZURE = 3#

OTC = 4#

NEW_STORAGE = -99#: This is a placeholder for a new storage location that is not yet supported by your SDK version. Please update your SDK to the latest version.

static from_str(string_location)[source]#

Return type: StorageLocation

get_str()[source]#

Return type: str

encord.orm.dataset.DatasetType#: For backwards compatibility

class encord.orm.dataset.DatasetScope(value)[source]#

An enumeration.

Expand for references to encord.orm.dataset.DatasetScope

READ = 'dataset.read'#

WRITE = 'dataset.write'#

class encord.orm.dataset.DatasetData(dic)[source]#

Video base ORM.

DB_FIELDS: collections.OrderedDict = {'data_hash': <class 'str'>, 'images': <class 'list'>, 'video': <class 'dict'>}#

class encord.orm.dataset.SignedVideoURL(dic)[source]#

A signed URL object with supporting information.

DB_FIELDS: collections.OrderedDict = {'data_hash': <class 'str'>, 'file_link': <class 'str'>, 'signed_url': <class 'str'>, 'title': <class 'str'>}#

class encord.orm.dataset.SignedImageURL(dic)[source]#

A signed URL object with supporting information.

DB_FIELDS: collections.OrderedDict = {'data_hash': <class 'str'>, 'file_link': <class 'str'>, 'signed_url': <class 'str'>, 'title': <class 'str'>}#

class encord.orm.dataset.SignedImagesURL(iter_)[source]#

A signed URL object with supporting information.

BASE_ORM_TYPE#: alias of encord.orm.dataset.SignedImageURL

class encord.orm.dataset.SignedDicomURL(dic)[source]#

A signed URL object with supporting information.

DB_FIELDS: collections.OrderedDict = {'data_hash': <class 'str'>, 'file_link': <class 'str'>, 'signed_url': <class 'str'>, 'title': <class 'str'>}#

class encord.orm.dataset.SignedDicomsURL(iter_)[source]#

A signed URL object with supporting information.

BASE_ORM_TYPE#: alias of encord.orm.dataset.SignedDicomURL

class encord.orm.dataset.Video(dic)[source]#

A video object with supporting information.

DB_FIELDS: collections.OrderedDict = {'data_hash': <class 'str'>, 'file_link': <class 'str'>, 'title': <class 'str'>}#

NON_UPDATABLE_FIELDS: set = {'data_hash'}#

class encord.orm.dataset.ImageGroup(dic)[source]#

An image group object with supporting information.

DB_FIELDS: collections.OrderedDict = {'data_hash': <class 'str'>, 'file_link': <class 'str'>, 'title': <class 'str'>}#

NON_UPDATABLE_FIELDS: set = {'data_hash'}#

class encord.orm.dataset.Image(dic)[source]#

An image object with supporting information.

DB_FIELDS: collections.OrderedDict = {'data_hash': <class 'str'>, 'file_link': <class 'str'>, 'title': <class 'str'>}#

NON_UPDATABLE_FIELDS: set = {'data_hash'}#

class encord.orm.dataset.SingleImage(dic)[source]#

For native single image upload.

success: bool#

class encord.orm.dataset.Images(success)[source]#

Uploading multiple images in a batch mode.

success: bool#

class encord.orm.dataset.DicomSeries(data_hash, title)[source]#

data_hash: str#

title: str#

class encord.orm.dataset.DicomDeidentifyTask(dicom_urls, integration_hash)[source]#

dicom_urls: List[str]#

integration_hash: str#

class encord.orm.dataset.ImageGroupOCR(processed_texts)[source]#

processed_texts: Dict#

class encord.orm.dataset.ReEncodeVideoTaskResult(data_hash, signed_url, bucket_path)[source]#

data_hash: str#

signed_url: Optional[str]#

bucket_path: str#

class encord.orm.dataset.ReEncodeVideoTask(status, result=None)[source]#

A re encode video object with supporting information.

Expand for references to encord.orm.dataset.ReEncodeVideoTask

Datasets / Data / Re-encoding videos / Check the status of a re-encoding task

status: str#

result: Optional[List[encord.orm.dataset.ReEncodeVideoTaskResult]] = None#

classmethod from_dict(json_dict)[source]#

class encord.orm.dataset.DatasetAccessSettings(fetch_client_metadata)[source]#

Settings for using the dataset object.

fetch_client_metadata: bool#: Whether client metadata should be retrieved for each data_row.

class encord.orm.dataset.ImagesDataFetchOptions(fetch_signed_urls=False)[source]#

fetch_signed_urls: bool = False#: Whether to fetch signed urls for each individual image. Only set this to true if you need to download the images.

class encord.orm.dataset.LongPollingStatus(value)[source]#

An enumeration.

Expand for references to encord.orm.dataset.LongPollingStatus

Datasets / Data / Adding data / Adding data from a private cloud

PENDING = 'PENDING'#: Job will automatically start soon (waiting in queue) or already started processing.

DONE = 'DONE'#

Job has finished successfully (possibly with errors if ignore_errors=True) If ignore_errors=False was specified in encord.dataset.Dataset.add_private_data_to_dataset_start() , job will only have the status DONE if there were no errors. If ignore_errors=True was specified in encord.dataset.Dataset.add_private_data_to_dataset_start() , job will always show the status DONE once complete and will never show ERROR status if this flag was set to True. There could be errors that were ignored. Information about number of errors and stringified exceptions is available in the units_error_count: int and errors: List[str] attributes.

Expand for references to encord.orm.dataset.LongPollingStatus.DONE

Datasets / Data / Adding data / Adding data from a private cloud

ERROR = 'ERROR'#

Job has completed with errors. This can only happen if ignore_errors was set to False. Information about errors is available in the units_error_count: int and errors: List[str] attributes.

Expand for references to encord.orm.dataset.LongPollingStatus.ERROR

Datasets / Data / Adding data / Adding data from a private cloud

class encord.orm.dataset.DatasetDataLongPolling(status, data_hashes_with_titles, errors, units_pending_count, units_done_count, units_error_count)[source]#

Response of the upload job’s long polling request.

Note: An upload job consists of job units, where job unit could be either a video, image group, dicom series, or a single image.

status: encord.orm.dataset.LongPollingStatus#: Status of the upload job. Documented in detail in LongPollingStatus()

data_hashes_with_titles: List[encord.orm.dataset.DatasetDataInfo]#: Information about data which was added to the dataset.

errors: List[str]#: Stringified list of exceptions.

units_pending_count: int#: Number of upload job units that have pending status.

units_done_count: int#: Number of upload job units that have done status.

units_error_count: int#: Number of upload job units that have error status.

classmethod from_dict(json_dict)[source]#

Return type: DatasetDataLongPolling

class encord.orm.dataset_with_user_role.DatasetWithUserRole(user_role, dataset)[source]#

This is a helper class denoting the relationship between the current user and a project

user_role: int#

dataset: dict#

class encord.orm.formatter.Formatter[source]#

abstract classmethod from_dict(json_dict)[source]#

Return type: ~T

class encord.orm.label_log.LabelLog(log_hash, user_hash, user_email, annotation_hash, identifier, data_hash, feature_hash, action, label_name, time_taken, created_at, frame)[source]#

log_hash: str#

user_hash: str#

user_email: str#

annotation_hash: str#

identifier: str#

data_hash: str#

feature_hash: str#

action: encord.orm.label_log.Action#

label_name: str#

time_taken: int#

created_at: datetime.datetime#

frame: int#

class encord.orm.label_log.Action(value)[source]#

An enumeration.

ADD = 0#

EDIT = 1#

DELETE = 2#

START = 3#

END = 4#

MARK_AS_NOT_LABELLED = 5#

MARK_AS_IN_PROGRESS = 6#

MARK_AS_LABELLED = 7#

MARK_AS_REVIEW_REQUIRED = 8#

MARK_AS_REVIEWED = 9#

MARK_AS_REVIEWED_TWICE = 10#

SUBMIT_TASK = 11#

APPROVE_LABEL = 12#

REJECT_LABEL = 13#

CLICK_SAVE = 14#

CLICK_UNDO = 15#

CLICK_REDO = 16#

CLICK_BULK = 17#

CLICK_ZOOM = 19#

CLICK_BRIGHTNESS = 20#

CLICK_HOTKEYS = 21#

CLICK_SETTINGS = 22#

ADD_ATTRIBUTE = 23#

EDIT_ATTRIBUTE = 24#

DELETE_ATTRIBUTE = 25#

APPROVE_NESTED_ATTRIBUTE = 26#

REJECT_NESTED_ATTRIBUTE = 27#

SUBMIT_LABEL = 28#

SUBMIT_NESTED_ATTRIBUTE = 29#

BUFFERING_OVERLAY_SHOWN = 30#

BITRATE_WARNING_SHOWN = 31#

SEEKING_OVERLAY_SHOWN = 32#

class encord.orm.label_log.LabelLogParams(**data)[source]#

user_hash: Optional[str]#

data_hash: Optional[str]#

to_unix_seconds: Optional[int]#

from_unix_seconds: Optional[int]#

user_email: Optional[str]#

include_user_email_and_interface_key: bool#

class encord.orm.label_row.LabelRow(dic)[source]#

A label row contains a data unit or a collection of data units and associated labels, and is specific to a data asset with type video or img_group:

A label row with a data asset of type video contains a single data unit.
A label row with a data asset of type img_group contains any number of data units.

The label row ORM is as follows:

label_hash (uid) is the unique identifier of the label row
dataset_hash (uid) is the unique identifier of the dataset which contains the particular video or image group
dataset_title is the title of the dataset which contains the particular video or image group
data_title is the title of the video or image group
data_type either video or img_group depending on data type
data_units a dictionary with (key: data hash, value: data unit) pairs.
object_answers is a dictionary with (key: object hash, value: object answer) pairs.
classification_answers is a dictionary with (key: classification hash, value: classification answer) pairs.
object_actions is a dictionary with (key: <object_hash>, value: object action) pairs.
label_status is a string indicating label status. It can take the values enumerated in encord.orm.label_row.LabelStatus. Note that this does not reflect thes status shown in the Projects->Labels section on the web-app.

A data unit, mentioned for the dictionary entry data_units above, has in the form:

label_row = {  # The label row
    # ...
    "data_units": {
        "<data_hash>": {
            "data_hash": "<data_hash>",  # A data_hash (uid) string
            "data_title": "A data title",
            "data_link": "<data_link>",  # Signed URL that expiring after 7 days
            "data_type": "<data_type>",  # (video/mp4, image/jpeg, etc.)
            "data_fps": 24.95,           # For video, the frame rate
            "data_sequence": "0",        # Defines order of data units
            "width": 640,                # The width of the content
            "height": 610,               # The height of the content
            "labels": {
                # ...
            }
        },
        # ...,
    }
}

A data unit can have any number of vector labels (e.g. bounding box, polygon, keypoint) and classifications.

Objects and classifications

A data unit can have any number of vector labels (e.g., bounding boxes, polygons, polylines, keypoints) and classifications. Each frame-level object and classification has unique identifiers ‘objectHash’ and ‘classificationHash’. Each frame-level entity has a unique feature identifier ‘featureHash’, defined in the editor ontology.

The object and classification answers are contained separately from the individual data units to preserve space for video, sequential images, DICOM, etc.

The objects and classifications answer dictionaries contain classification ‘answers’ (i.e. attributes that describe the object or classification). This is to avoid storing the information at every frame in the blurb, of particular importance for videos.

A labels dictionary for video is in the form:

label_row["data_units"]["<data_hash>"]["labels"] = {
    "<frame_number>": {
        "objects": [
            # { object 1 },
            # { object 2 },
            # ...
        ],
        "classifications": [
            # { classification 1 },
            # { classification 2 },
            # ...
        ],
    }
}

A labels dictionary for an img_group data unit is in the form:

label_row["data_units"]["<data_hash>"]["labels"] = {
    "objects": [
        # { object 1 },
        # { object 2 },
        # ...
    ],
    "classifications": [
        # { classification 1 },
        # { classification 2 },
        # ...
    ],
}

The object answers dictionary is in the form:

label_row["object_answers"] = {
    "<object_hash>": {
        "objectHash": "<object_hash>",
        "classifications": [
            # {answer 1},
            # {answer 2},
            # ...
        ]
    },
    # ...
}

The classification answers dictionary is in the form:

label_row["classification_answers"] = {
    "<classification_hash>": {
        "classificationHash": "<classification_hash>",
        "classifications": [
            # {answer 1},
            # {answer 2},
            # ...
        ],
    },
    # ...
}

The object actions dictionary is in the form:

label_row["object_actions"] = {
    "<object_hash>": {
        "objectHash": "<object_hash>",
        "actions": [
            # {answer 1},
            # {answer 2},
            # ...
        ],
    },
    # ...
}

DB_FIELDS: collections.OrderedDict = {'annotation_task_status': <class 'str'>, 'classification_answers': <class 'dict'>, 'created_at': <class 'str'>, 'data_hash': <class 'str'>, 'data_title': <class 'str'>, 'data_type': <class 'str'>, 'data_units': <class 'dict'>, 'dataset_hash': <class 'str'>, 'dataset_title': <class 'str'>, 'label_hash': <class 'str'>, 'label_status': <class 'str'>, 'last_edited_at': <class 'str'>, 'object_actions': <class 'dict'>, 'object_answers': <class 'dict'>}#

NON_UPDATABLE_FIELDS: set = {'label_hash'}#

class encord.orm.label_row.Review[source]#

class encord.orm.label_row.AnnotationTaskStatus(value)[source]#

An enumeration.

QUEUED = 'QUEUED'#

ASSIGNED = 'ASSIGNED'#

IN_REVIEW = 'IN_REVIEW'#

RETURNED = 'RETURNED'#

COMPLETED = 'COMPLETED'#

class encord.orm.label_row.ShadowDataState(value)[source]#

Specifies the kind of data to fetch when working with a BenchmarkQa project

ALL_DATA = 'ALL_DATA'#: Fetch all the label rows

SHADOW_DATA = 'SHADOW_DATA'#

the annotator’s view of the benchmark

Type: Only fetch the label rows that were submitted against “shadow data”

NOT_SHADOW_DATA = 'NOT_SHADOW_DATA'#: Only fetch the label rows for “production” data

class encord.orm.label_row.LabelStatus(value)[source]#

An enumeration.

NOT_LABELLED = 'NOT_LABELLED'#

LABEL_IN_PROGRESS = 'LABEL_IN_PROGRESS'#

LABELLED = 'LABELLED'#

REVIEW_IN_PROGRESS = 'REVIEW_IN_PROGRESS'#

REVIEWED = 'REVIEWED'#

REVIEWED_TWICE = 'REVIEWED_TWICE'#

MISSING_LABEL_STATUS = '_MISSING_LABEL_STATUS_'#: This value will be displayed if the Encord platform has a new label status and your SDK version does not understand it yet. Please update your SDK to the latest version.

class encord.orm.label_row.WorkflowGraphNode(uuid, title)[source]#

uuid: str#

title: str#

classmethod from_optional_dict(json_dict)[source]#

Return type: Optional[WorkflowGraphNode]

class encord.orm.label_row.LabelRowMetadata(label_hash, created_at, last_edited_at, data_hash, dataset_hash, dataset_title, data_title, data_type, data_link, label_status, annotation_task_status, workflow_graph_node, is_shadow_data, number_of_frames, duration, frames_per_second, height, width)[source]#

Contains helpful information about a label row.

label_hash: Optional[str]#: Only present if the label row is initiated

created_at: Optional[datetime.datetime]#: Only present if the label row is initiated

last_edited_at: Optional[datetime.datetime]#: Only present if the label row is initiated

data_hash: str#

dataset_hash: str#

dataset_title: str#

data_title: str#

data_type: str#

data_link: Optional[str]#: Can be None for label rows of image groups or DICOM series.

label_status: encord.orm.label_row.LabelStatus#: Can be None for TMS2 projects

annotation_task_status: Optional[encord.orm.label_row.AnnotationTaskStatus]#: Only available for TMS2 project

workflow_graph_node: Optional[encord.orm.label_row.WorkflowGraphNode]#

is_shadow_data: bool#

number_of_frames: int#

duration: Optional[float]#: Only available for the VIDEO data_type

frames_per_second: Optional[int]#: Only available for the VIDEO data_type

height: Optional[int]#

width: Optional[int]#

classmethod from_dict(json_dict)[source]#

Return type: LabelRowMetadata

classmethod from_list(json_list)[source]#

Return type: List[LabelRowMetadata]

to_dict()[source]#

Return type: dict
Returns: The dict equivalent of LabelRowMetadata.

class encord.orm.labeling_algorithm.LabelingAlgorithm(dic)[source]#

Labeling algorithm base ORM.

ORM:

algorithm_name, algorithm_params

DB_FIELDS: collections.OrderedDict = {'algorithm_name': <class 'str'>, 'algorithm_parameters': <class 'dict'>}#

class encord.orm.labeling_algorithm.ObjectInterpolationParams(dic)[source]#

Labeling algorithm parameters for interpolation algorithm

ORM:

key_frames, objects_to_interpolate

DB_FIELDS: collections.OrderedDict = {'key_frames': <class 'dict'>, 'objects_to_interpolate': <class 'list'>}#

class encord.orm.labeling_algorithm.BoundingBoxFittingParams(dic)[source]#

Labeling algorithm parameters for bounding box fitting algorithm

ORM:

labels, video

DB_FIELDS: collections.OrderedDict = {'labels': <class 'dict'>, 'video': <class 'dict'>}#

class encord.orm.model.ModelOperations(value)[source]#

An enumeration.

INFERENCE = 0#

TRAIN = 1#

CREATE = 2#

class encord.orm.model.Model(dic)[source]#

Model base ORM.

ORM:

model_operation, model_parameters,

DB_FIELDS: collections.OrderedDict = {'model_operation': <class 'int'>, 'model_parameters': <class 'dict'>}#

class encord.orm.model.ModelConfiguration(model_uid, title, description, feature_node_hashes, model, model_iteration_uids)[source]#

model_uid: str#

title: str#

description: str#

feature_node_hashes: List[str]#: The corresponding feature node hashes of the ontology object

model: encord.constants.model.AutomationModels#

model_iteration_uids: List[str]#: All the UIDs of individual model training instances

classmethod from_dict(json_dict)[source]#

class encord.orm.model.ModelTrainingLabelMetadata(label_uid, data_uid, data_link)[source]#

label_uid: str#

data_uid: str#

data_link: str#

static from_dict(json_dict)[source]#

Return type: ModelTrainingLabelMetadata

classmethod from_list(json_list)[source]#

Return type: List[ModelTrainingLabelMetadata]

class encord.orm.model.ModelTrainingLabel(label_metadata_list, feature_uids)[source]#

label_metadata_list: List[encord.orm.model.ModelTrainingLabelMetadata]#

feature_uids: List[str]#

static from_dict(json_dict)[source]#

Return type: ModelTrainingLabel

class encord.orm.model.TrainingMetadata(model_iteration_uid, created_at=None, training_final_loss=None, model_training_labels=None)[source]#

model_iteration_uid: str#

created_at: Optional[datetime.datetime] = None#

training_final_loss: Optional[float] = None#

model_training_labels: Optional[encord.orm.model.ModelTrainingLabel] = None#

classmethod from_dict(json_dict)[source]#

static get_created_at(json_dict)[source]#

Return type: Optional[datetime]

static get_model_training_labels(json_dict)[source]#

Return type: Optional[ModelTrainingLabel]

class encord.orm.model.ModelRow(dic)[source]#

A model row contains a set of features and a model (resnet18, resnet34, resnet50, resnet101, resnet152, vgg16, vgg19, faster_rcnn, mask_rcnn).

ORM:

model_hash (uid), title, description, features, model,

DB_FIELDS: collections.OrderedDict = {'description': <class 'str'>, 'features': <class 'list'>, 'model': <class 'str'>, 'model_hash': <class 'str'>, 'title': <class 'str'>}#

class encord.orm.model.ModelInferenceParams(dic)[source]#

Model inference parameters for running models trained via the platform.

ORM:

local_file_path, conf_thresh, iou_thresh, device detection_frame_range (optional)

DB_FIELDS: collections.OrderedDict = {'allocation_enabled': <class 'bool'>, 'conf_thresh': <class 'float'>, 'data_hashes': <class 'list'>, 'detection_frame_range': <class 'list'>, 'device': <class 'str'>, 'files': <class 'list'>, 'iou_thresh': <class 'float'>, 'rdp_thresh': <class 'float'>}#

class encord.orm.model.ModelTrainingWeights(dic)[source]#

Model training weights.

ORM:

training_config_link, training_weights_link,

DB_FIELDS: collections.OrderedDict = {'model': <class 'str'>, 'training_config_link': <class 'str'>, 'training_weights_link': <class 'str'>}#

class encord.orm.model.ModelTrainingParams(dic)[source]#

Model training parameters.

ORM:

model_hash, label_rows, epochs, batch_size, weights, device

DB_FIELDS: collections.OrderedDict = {'batch_size': <class 'int'>, 'device': <class 'str'>, 'epochs': <class 'int'>, 'label_rows': <class 'list'>, 'model_hash': <class 'str'>, 'weights': <class 'encord.orm.model.ModelTrainingWeights'>}#

Expand for references to encord.orm.project

class encord.orm.project.Project(dic)[source]#

DEPRECATED - prefer using the encord.project.Project class instead.

A project defines a label ontology and is a collection of datasets and label rows.

ORM:

title,
description,
editor_ontology,
ontology_hash,

datasets:

[
   {
        dataset_hash (uid),
        title,
        description,
        dataset_type (internal vs. AWS/GCP/Azure),
   },
   ...
],

label_rows:

[
    {
        label_hash (uid),
        data_hash (uid),
        dataset_hash (uid),
        dataset_title,
        data_title,
        data_type,
        label_status
    },
    ...
]

annotation_task_status

DB_FIELDS: collections.OrderedDict = {'created_at': <class 'datetime.datetime'>, 'datasets': (<class 'list'>, <class 'str'>), 'description': <class 'str'>, 'editor_ontology': (<class 'dict'>, <class 'str'>), 'label_rows': (<class 'list'>, <class 'str'>), 'last_edited_at': <class 'datetime.datetime'>, 'ontology_hash': <class 'str'>, 'project_hash': <class 'str'>, 'source_projects': <class 'list'>, 'title': <class 'str'>}#

NON_UPDATABLE_FIELDS: set = {'datasets', 'editor_ontology', 'label_rows'}#

get_labels_list()[source]#

Returns a list of all optional label row IDs (label_hash uid) in a project. If no label_hash is found, a None value is appended. This can be useful for working with fetching additional label row data via encord.project.Project.get_label_rows() for example.

project = client_instance.get_project(<project_hash>)
project_orm = project.get_project()

labels_list = project_orm.get_labels_list()
created_labels_list = []
for label in labels_list:
    if label is not None:  # None values will fail the operation
        created_labels_list.append(label)

label_rows = project.get_label_rows(created_labels_list, get_signed_url=False)

Return type: List[Optional[str]]

property project_hash: str#

Return type: str

property title: str#

Return type: str

property description: str#

Return type: str

property editor_ontology#

property datasets#

property label_rows#

property ontology_hash: str#

Return type: str

property source_projects#

property created_at: datetime.datetime#

Return type: datetime

property last_edited_at: datetime.datetime#

Return type: datetime

class encord.orm.project.ProjectCopy[source]#

class encord.orm.project.ProjectUsers[source]#

class encord.orm.project.ProjectDataset[source]#

class encord.orm.project.ProjectCopyOptions(value)[source]#

An enumeration.

COLLABORATORS = 'collaborators'#

DATASETS = 'datasets'#

MODELS = 'models'#

LABELS = 'labels'#

class encord.orm.project.ReviewApprovalState(value)[source]#

An enumeration.

Expand for references to encord.orm.project.ReviewApprovalState

APPROVED = 'APPROVED'#

PENDING = 'PENDING'#

REJECTED = 'REJECTED'#

DELETED = 'DELETED'#

NOT_SELECTED_FOR_REVIEW = 'NOT_SELECTED_FOR_REVIEW'#

class encord.orm.project.CopyDatasetAction(value)[source]#

An enumeration.

Expand for references to encord.orm.project.CopyDatasetAction

ATTACH = 'ATTACH'#: Attach the datasets associated with the original project to the copy project.

CLONE = 'CLONE'#

Clone the data units from the associated datasets into a new dataset an attach it to the copy project.

Expand for references to encord.orm.project.CopyDatasetAction.CLONE

class encord.orm.project.CopyDatasetOptions(action=CopyDatasetAction.ATTACH, dataset_title=None, dataset_description=None, datasets_to_data_hashes_map=<factory>)[source]#

Options for copying the datasets associated with a project.

Expand for references to encord.orm.project.CopyDatasetOptions

action: encord.orm.project.CopyDatasetAction = 'ATTACH'#: One of CopyDatasetAction.ATTACH or CopyDatasetAction.CLONE. (defaults to ATTACH)

dataset_title: Optional[str] = None#

dataset_description: Optional[str] = None#

datasets_to_data_hashes_map: Dict[str, List[str]]#

List[<data_unit_hash>]}`. When provided with a CLONE action this will filter the copied data units. When combined with CopyLabelsOptions, only labels from specific data units will be copied.

Type: A dictionary of `{ <dataset_hash>

class encord.orm.project.CopyLabelsOptions(accepted_label_hashes=None, accepted_label_statuses=None)[source]#

Options for copying the labels associated with a project.

Expand for references to encord.orm.project.CopyLabelsOptions

accepted_label_hashes: Optional[List[str]] = None#: A list of label hashes that will be copied to the new project

accepted_label_statuses: Optional[List[encord.orm.project.ReviewApprovalState]] = None#: A list of label statuses to filter the labels copied to the new project, defined in ReviewApprovalState

class encord.orm.project.CopyProjectPayload(copy_project_options=<factory>, copy_labels_options=None, project_copy_metadata=None)[source]#

WARN: do not use, this is only for internal purpose.

copy_project_options: List[ProjectCopyOptions]#

copy_labels_options: Optional[_CopyLabelsOptions] = None#

project_copy_metadata: Optional[_ProjectCopyMetadata] = None#

class encord.orm.project.ProjectWorkflowType(value)[source]#

An enumeration.

MANUAL_QA = 'manual'#

BENCHMARK_QA = 'benchmark'#

class encord.orm.project.ManualReviewWorkflowSettings[source]#: Sets the project QA workflow to “manual reviews”. There are currently no available settings for this workflow

class encord.orm.project.BenchmarkQaWorkflowSettings(source_projects=<factory>)[source]#

Sets the project QA workflow to “Automatic”, with benchmark data being presented to all the annotators

source_projects: List[str]#: For Benchmark QA projects, a list of project ids (project_hash-es) that contain the benchmark source data

encord.orm.project.ProjectWorkflowSettings#

A variant type that allows you to select the type of quality assurance workflow for the project, and configure it.

Currently one of:: ManualReviewWorkflowSettings: a workflow with optional manual reviews BenchmarkQaWorkflowSettings: annotators are presented with “benchmark” or “honeypot” data

alias of Union[encord.orm.project.ManualReviewWorkflowSettings, encord.orm.project.BenchmarkQaWorkflowSettings]

class encord.orm.project.ReviewMode(value)[source]#

UNLABELLED:

The labels are added to the images. However, the one person must still go over: all of the labels before submitting them for review.

LABELLED:

The labels are added to the images and are marked as labelled. A reviewer will: still need to approve those.

REVIEWED:

The labels are added to the images and considered reviewed. No more action is: required from the labeler or reviewer.

UNLABELLED = 'unlabelled'#

LABELLED = 'labelled'#

REVIEWED = 'reviewed'#

class encord.orm.project.ProjectImporter(dic)[source]#

DB_FIELDS: collections.OrderedDict = {'errors': <class 'list'>, 'project_hash': typing.Union[str, NoneType]}#

class encord.orm.project.CvatExportType(value)[source]#

An enumeration.

PROJECT = 'project'#

TASK = 'task'#

class encord.orm.project.ProjectImporterCvatInfo(dic)[source]#

DB_FIELDS: collections.OrderedDict = {'export_type': <enum 'CvatExportType'>}#

Expand for references to encord.orm.project_api_key

class encord.orm.project_api_key.ProjectAPIKey(api_key, title, scopes)[source]#

Expand for references to encord.orm.project_api_key.ProjectAPIKey

api_key: str#

title: str#

scopes: List[encord.utilities.client_utilities.APIKeyScopes]#

classmethod from_dict(json_dict)[source]#

class encord.orm.project_with_user_role.ProjectWithUserRole(user_role, project)[source]#

This is a helper class denoting the relationship between the current user an a project

user_role: int#

project: dict#

Utilities - Other#

class encord.http.utils.CloudUploadSettings(max_retries=None, backoff_factor=None, allow_failures=False)[source]#

The settings for uploading data into the GCP cloud storage. These apply for each individual upload. These settings will overwrite the encord.http.constants.RequestsSettings() which is set during encord.EncordUserClient creation.

max_retries: Optional[int] = None#: Number of allowed retries when uploading

backoff_factor: Optional[float] = None#: With each retry, there will be a sleep of backoff_factor * (2 ** (retry_number - 1) )

allow_failures: bool = False#: If failures are allowed, the upload will continue even if some items were not successfully uploaded even after retries. For example, upon creation of a large image group, you might want to create the image group even if a few images were not successfully uploaded. The unsuccessfully uploaded images will then be logged.

Expand for references to encord.http.constants

class encord.http.constants.RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3)[source]#

The settings for all outgoing network requests. These apply for each individual request.

Expand for references to encord.http.constants.RequestsSettings

Authentication / Resource authentication / Authenticate with Encord - resource authentication

max_retries: int = 3#: Number of allowed retries when a request is sent. It only affects idempotent retryable requests.

backoff_factor: float = 1.5#: With each retry, there will be a sleep of backoff_factor * (2 ** (retry_number - 1) )

connection_retries: int = 3#: Number of allowed retries to establish TCP connection when a request is sent.

encord.utilities.project_utilities.get_all_model_iteration_uids(model_configurations)[source]#

A convenience function that works well with encord.project.Project.list_models()

Return type: Set[str]

class encord.constants.enums.DataType(value)[source]#

An enumeration.

VIDEO = 'video'#

IMG_GROUP = 'img_group'#

DICOM = 'dicom'#

IMAGE = 'image'#

DICOM_STUDY = 'dicom_study'#

MISSING_DATA_TYPE = '_MISSING_DATA_TYPE_'#

static from_upper_case_string(string)[source]#

Return type: DataType

to_upper_case_string()[source]#

Return type: str

Configs#

class encord.configs.BaseConfig(endpoint, requests_settings=RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3))[source]#

abstract define_headers(data)[source]#

Return type: Dict

class encord.configs.UserConfig(private_key, domain='https://api.encord.com', requests_settings=RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3))[source]#

define_headers(data)[source]#

Return type: Dict

static from_ssh_private_key(ssh_private_key, password='', requests_settings=RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3), **kwargs)[source]#

Instantiate a UserConfig object by the content of a private ssh key.

Parameters

ssh_private_key (str) – The content of a private key file.
password (Optional[str]) – The password for the private key file.
requests_settings (RequestsSettings) – The requests settings for all outgoing network requests.

Returns

UserConfig.

Raises

ValueError – If the provided key content is not of the correct format.

class encord.configs.Config(resource_id=None, web_file_path='/public', domain=None, websocket_endpoint='wss://message-api.cord.tech/websocket', requests_settings=RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3))[source]#

Config defining endpoint, project id, API key, and timeouts.

abstract get_websocket_url()[source]#

encord.configs.get_env_resource_id()[source]#

Return type: str

encord.configs.get_env_api_key()[source]#

Return type: str

encord.configs.get_env_ssh_key()[source]#

Returns the raw ssh key by looking up the ENCORD_SSH_KEY_FILE and ENCORD_SSH_KEY environment variables in the mentioned order and returns the first successfully identified key.

Return type: str

class encord.configs.ApiKeyConfig(resource_id=None, api_key=None, domain=None, requests_settings=RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3))[source]#

define_headers(data)[source]#

Return type: Dict

get_websocket_url()[source]#

encord.configs.EncordConfig#: alias of encord.configs.ApiKeyConfig

encord.configs.CordConfig#: alias of encord.configs.ApiKeyConfig

class encord.configs.SshConfig(user_config, resource_type, resource_id=None)[source]#

define_headers(data)[source]#

Return type: Dict

get_websocket_url()[source]#

encord.configs.ENCORD_DOMAIN = 'https://api.encord.com'

str(object=’’) -> str str(bytes_or_buffer[, encoding[, errors]]) -> str

Create a new string object from the given object. If encoding or errors is specified, then the object must expose a data buffer that will be decoded using the given encoding and error handler. Otherwise, returns the result of object.__str__() (if defined) or repr(object). encoding defaults to sys.getdefaultencoding(). errors defaults to ‘strict’.

Exceptions#

class encord.exceptions.ExceptionContext(timestamp=<factory>)[source]#

timestamp: str#

exception encord.exceptions.EncordException(message, context=None)[source]#: Base class for all exceptions.

encord.exceptions.CordException#: alias of encord.exceptions.EncordException

exception encord.exceptions.InitialisationError(message, context=None)[source]#: Exception thrown when API key fails to initialise.

exception encord.exceptions.AuthenticationError(message, context=None)[source]#: Exception thrown when API key fails authentication.

exception encord.exceptions.AuthorisationError(message, context=None)[source]#: Exception thrown when access is unauthorised. (E.g. access to a data asset or method).

exception encord.exceptions.ResourceNotFoundError(message, context=None)[source]#: Exception thrown when a requested resource is not found. (E.g. label, data asset).

exception encord.exceptions.TimeOutError(message, context=None)[source]#: Exception thrown when a request times out.

exception encord.exceptions.RequestException(message, context=None)[source]#: Ambiguous exception while handling request.

exception encord.exceptions.UnknownException(message, context=None)[source]#: Unknown error.

exception encord.exceptions.InvalidDateFormatError(message, context=None)[source]#: Invalid date format error

exception encord.exceptions.MethodNotAllowedError(message, context=None)[source]#: Exception thrown when HTTP method is not allowed.

exception encord.exceptions.OperationNotAllowed(message, context=None)[source]#: Exception thrown when a read/write operation is not allowed. The API key blocks the operation.

exception encord.exceptions.AnswerDictionaryError(message, context=None)[source]#: Exception thrown when answer dictionaries are incomplete. Occurs when an object or classification is missing.

exception encord.exceptions.CorruptedLabelError(message, context=None)[source]#: Exception thrown when a label is corrupted. (E.g. the frame labels have more frames than the video).

exception encord.exceptions.FileTypeNotSupportedError(message, context=None)[source]#: Exception thrown when a file type is not supported. Supported file types are: image/jpeg, image/png, video/webm, video/mp4.

exception encord.exceptions.FileSizeNotSupportedError(message, context=None)[source]#: Exception thrown when the combined size of the input files is larger than the supported limit.

exception encord.exceptions.FeatureDoesNotExistError(message, context=None)[source]#: If a feature uid does not exist in a given project ontology.

exception encord.exceptions.ModelWeightsInconsistentError(message, context=None)[source]#: Exception thrown when an attempted model training iteration has a different type of weights than what is recorded (i.e. if type of model_hash (uid) is faster_rcnn, but is attempted trained with different model weights).

exception encord.exceptions.ModelFeaturesInconsistentError(message, context=None)[source]#: If a feature type is different than what is supported by the model (e.g. if creating a classification model using a bounding box).

exception encord.exceptions.UploadOperationNotSupportedError(message, context=None)[source]#: Exception thrown when trying to upload a video/image group to non-Encord storage dataset

exception encord.exceptions.DetectionRangeInvalidError(message, context=None)[source]#: Exception thrown when a detection range is invalid. (E.g. negative or higher than num frames in video).

exception encord.exceptions.InvalidAlgorithmError(message, context=None)[source]#: Exception thrown when invalid labeling algorithm name is sent.

exception encord.exceptions.ResourceExistsError(message, context=None)[source]#: Exception thrown when trying to re-create a resource. Avoids overriding existing work.

exception encord.exceptions.DuplicateSshKeyError(message, context=None)[source]#: Exception thrown when using an SSH key that was added twice to the platform.

exception encord.exceptions.SshKeyNotFound(message, context=None)[source]#: Exception thrown when using an SSH key that was not added to the platform.

exception encord.exceptions.InvalidArgumentsError(message, context=None)[source]#: Exception thrown when the arguments are invalid.

exception encord.exceptions.GenericServerError(message, context=None)[source]#: The server has reported an error which is not recognised by this SDK version. Try upgrading the SDK version to see the precise error that is reported.

exception encord.exceptions.CloudUploadError(message, context=None)[source]#: The upload to the cloud was not successful

exception encord.exceptions.MultiLabelLimitError(message, maximum_labels_allowed, context=None)[source]#: Too many labels were requested

exception encord.exceptions.LabelRowError(message, context=None)[source]#: An error thrown when the construction of a LabelRow class is invalid.

exception encord.exceptions.OntologyError(message, context=None)[source]#: An error thrown when using the ontology class with an error.

exception encord.exceptions.WrongProjectTypeError(message, context=None)[source]#: An error thrown when project type does not match the operation E.g. when TMS2 specific operations are attempted on non-TMS2 project

Helpers#

class encord.http.bundle.Bundle[source]#

This class allows to perform operations in bundles to improve performance by reducing number of network calls.

It is not supposed to be instantiated directly by a user. Use encord.project.Project.create_bundle() method to initiate bundled operations.

To execute batch you can either call execute() directly, or use a Context Manager.

# Code example of performing batched label initialisation
project = ... # assuming you already have instantiated this Project object
label_rows = project.list_label_rows_v2()
bundle = project.create_bundle()
for label_row in label_rows:
    label_row.initialise_labels(bundle=bundle)
    # no real network operations happened at this point

# now, trigger the actual network interaction
bundle.execute()

# all labels are initialised at this point

And this is the same flow with the Context Manager approach:

# Code example of performing batched label initialisation
project = ... # assuming you already have instantiated this Project object
label_rows = project.list_label_rows_v2()
with project.create_bundle() as bundle:
    for label_row in label_rows:
        label_row.initialise_labels(bundle=bundle)
        # no real network operations happened at this point

# At this point all labels will be initialised

add(operation, request_reducer, result_mapper, payload, limit)[source]#

This is an internal method and normally is not supposed to be used externally.

Adds an operation to a bundle for delayed execution.

Return type: None

execute()[source]#

Executes all scheduled operations in bundles and populates results

Return type: None

DEPRECATED Client#

Using this client is deprecated. You are encouraged to use the Project and Dataset instead.

encord.client provides a simple Python client that allows you to query project resources through the Encord API.

Here is a simple example for instantiating the client for a project and obtaining project info:

Expand for references to encord.client

Authentication / Resource authentication / Authenticate with Encord - resource authentication

class encord.client.EncordClient(querier, config)[source]#

Encord client. Allows you to query db items associated with a project (e.g. label rows, datasets).

Expand for references to encord.client.EncordClient

static initialise(resource_id=None, api_key=None, domain='https://api.encord.com', requests_settings=RequestsSettings(max_retries=3, backoff_factor=1.5, connection_retries=3))[source]#

Create and initialize a Encord client from a resource EntityId and API key.

Parameters

resource_id (Optional[str]) –
either of the following
- A <project_hash>. If None, uses the ENCORD_PROJECT_ID environment variable. The CORD_PROJECT_ID environment variable is supported for backwards compatibility.
- A <dataset_hash>. If None, uses the ENCORD_DATASET_ID environment variable. The CORD_DATASET_ID environment variable is supported for backwards compatibility.
api_key (Optional[str]) – An API key. If None, uses the ENCORD_API_KEY environment variable. The CORD_API_KEY environment variable is supported for backwards compatibility.
domain (str) – The encord api-server domain. If None, the encord.configs.ENCORD_DOMAIN is used
requests_settings (RequestsSettings) – The RequestsSettings from this config

Returns

A Encord client instance.

Return type

EncordClient

Expand for references to encord.client.EncordClient.initialise

Authentication / Resource authentication / Authenticate with Encord - resource authentication

static initialise_with_config(config)[source]#

Create and initialize a Encord client from a Encord config instance.

Parameters: config (ApiKeyConfig) – A Encord config instance.
Returns: A Encord client instance.
Return type: EncordClient

get_cloud_integrations()[source]#

Return type: List[CloudIntegration]

class encord.client.EncordClientDataset(querier, config, dataset_access_settings=DatasetAccessSettings(fetch_client_metadata=False))[source]#

DEPRECATED - prefer using the encord.dataset.Dataset instead

Expand for references to encord.client.EncordClientDataset

Authentication / Resource authentication / Authenticate with Encord - resource authentication