v3.1.0

Some new feature:

• Add a new project type, Question Answer Conversational. Enable user to input prompt, response and reference links. Also easy for review and download function.
  
  
• For project type Question Answer, enable user to select one question column with JsonArray data type, and also let user edit question and answer in annotation page.
  

Some fixed issue:

• Enhancement the pagination query and friendly index the annotated data.

v2.1.0

• Release note:

  • Version v2.1.0, which still keeps our old UI style, code saved at DAML-v2.1.0
  • The old UI e2e code saved in branch DAML-v2.1.0-e2e.

• Some new feature

  • Add one new annotation project type, Question Answer.

  • In project detail preview page, the latest data section, under the annotation tab, keep all annotators’ and reviewers’ annotate records, and separate these info into different datagrid.

  • APIs, add unique error code and msg.

  • Question Answer project support create, annotation, review, preview, append, generate and download datasets, annotation just like this
    

• Some fixed bugs

  • In project create page, once the slack assignment list is changed then update the email assignment form validation
  • During file generation file will download directly if target file size is less than 50MB

v3.0.0

Release note:

• DAML upgrade to v3.0.0 that mainly raise UI/UX to a higher standard
• The latest code saved in the branch master.
• The latest e2e code saved in branch e2e-test.

1. Some main points of UI/UX upgrade

• The header nav changed to left side nav. Dataset's datagrid adds labeling task column. Click dataset name can enter data detail page to create labeling task.
  
• Upload dataset feature has its own create page
  
• Labeling task datagrid enable hide/show column
  
• Click labeling task name enter task detail page can do annotate, review latest annotate data, check d3 charts, append dataset and view original dataset
  
• Create labeling task use wizard page which more clear and instructive for user
  
• DAML now support two roles power user and user. Power user can manage all users' datasets and labeling tasks.
  

2. All the feature enabled in DAML v2.1.0 are still active in DAML v3.0.0

v2.0.0

• Update angular from v8 to v11
• Update clarity from v2 to v5
• Some fixed bugs
  • When open wizard through click review data btn, should remove selected label column from the text checkbox columns
  • The MLfeedback project should open project edit modal without err
  • Add validation to label parse, creation and edit, should trim() and not allow comma
  • Fix log project can't preview normally some time
  • Project preview latest data datagrid cell info mislocation issue

v1.0.0

• Support for common annotation tasks:
  • Text classification
  • Named entity recognition. Support intersecting and overlapping text labeling as well as a secondary labeling.
  • Tabular classification and regression
  • Images recognition with bounding boxes and polygons
  • Log labeling
• Support label type
  • Text
  • Numeric
  • Upload hierarchical taxonomy file
• Active learning with uncertainly sampling to query unlabeled data. Allow users to choose active learning query strategy and select Spacy models (for different languages and model sizes)
• Project tracking with real time data aggregation and review process
  • Keep track of the progress of every annotator. Annotators are informed if they are not starting the task and not making progress before the entire labelling project is completed.
• User management panel with role-based access control
• Data management
  • Import in common data formats
  • Export in ML friendly formats
  • Data sharing through community datasets
  • Review the labelling quality for Text, Tabular, NER, and Image projects. Project owners can review the annotation, make modification, and pass review. Review assignment logic support most_uncertain, random and sequential.
• Swagger API for programmatic labeling, connecting to data pipelines and more
• Slack integration. Assigning a Text or Tabular project to a Slack channel and all users in the channel can start the annotation task in Slack.