- 2.4.8 was broken for > Python 3.6
- include new
Project.enable_model_assisted_labeling
method for turning on model-assisted labeling
- fix failing
next
call Labelbox#74
Ontology
schema for interacting with ontologies and their schema nodes
- fix failing
create_metadata
calls
- retry capabilities for common flaky API failures
- protection against improper types passed into
Project.upload_anntations
- pass thru API error messages when possible
BulkImportRequest
data typeProject.upload_annotation
supports uploading via a local ndjson file, url, or a iterable of annotations
Client.upload_data
will now pass the correctcontent-length
when uploading data.
Dataset.create_data_row
andDataset.create_data_rows
will now upload with content type to ensure the Labelbox editor can show videos.
Prediction
andPredictionModel
data types.
Client.execute
now automatically extracts the 'data' value from the returneddict
. This breaks existing code that directly uses theClient.execute
method.- Major code reorganization, naming and test improvements.
Label.seconds_to_label
field value is now optional when creating aLabel
. Default value is 0.0.
Project.upsert_review_queue
method.Project.extend_reservations
method.Label.created_by
relationship (To-One User).- Changelog.
Dataset.create_data_row
upload of local file data.
Changelog not maintained before version 2.2.