Taskbase Feedback API Changelog

• Additions to feedback-model:
  • The result field includes now allAspects which are all detected aspects for the provided answer.

• New endpoint for computing relevant aspects given a task.

• Additions to task-model:
  • New task types MULTIPLE_CHOICE and MULTIPLE_RESPONSE.

• Additions to feedback-request and feedback:
  • New resource object that holds information of audio file for FREEEFORM_TEXT

• Remove tokenizationStrategy from FREEFORM_TEXT task-model

• Additions to task-model:
  • New task type IN_TEXT_CHOICES. Multiple choice embedded in text.

• Additions to task-model:
  • title can be used to create tasks with a custom title, if not given it will fallback to taskId.
  • language can be used to specify the task language, if not given will default to en.

• Added a tenantId to the feedback-request. This field allows to specify a different tenant than the domain's default. As a result, one can create tasks outside of the default tenant.

• overallCorrectness was removed.
• Introduced a metaData field in the feedback-request. This object contains a timeOnTask field which replaces the duration field that was on the root level before. The new field is also optional and not required as the duration was.
• Renaming:
  • input is now answer
  • blankText is not textWithBlanks
  • blankElements are now blanks
  • All task specific prefixes are dropped, e.g. freeformTextTaskDefinition is now simply taskDefinition.
  • part within a feedback in the feedback-response is now context
  • solution in the feedback-response is now sampleSolution
• blanks are now an array of objects instead of an array of string
• context within feedback in the feedback-response is now an array of objects instead of an object

• Fixed the task definition examples for the feedback request that did not show the correct type field depending on the type of request.

🎉 Initial release 🎉

• Supported task types:

  • FREEFORM_TEXT
  • IN_TEXT_FILL_IN_BLANKS

• Supported operations:

  • Compute Feedback