Changelog

Changelog - 2.35.0

Released on

Release notes

  • New search_in parameter in analyze/dataset route with the precomputed workflow. Allows to define if the documents extraction has to be done by searching entity keywords in the title and/or in the text. Not available with ner-linking worker.
  • New entity_volume_sentence indicator available in the analyze/{id}/timeseries/{hash}/download route.
  • New extracted parameter for the documents/translate route.
  • A new pending status has been introduced for precomputed instances.
  • Chinese and Japanese languages are now handled by the precomputed workflow.

Comments

  • Use the search_in parameter on the analyze/dataset with {"precompute": true} route to choose if the extraction must be done on title, text or both.

    • Example:
      • {"search_in": ["title"]} The extraction will be done on entities' keywords only on title.
      • {"search_in": ["text"]} The extraction will be done on entities' keywords only on text.
      • {"search_in": ["title", "text"]} The extraction will be done on entities' keywords on title and text.

  • A new indicator entity_volume_sentence is now available when retrieving the timeseries for a given analysis. This indicator represents the volume of sentences where the entity of interest is matched for the aggregation period.

  • A new parameter extracted has been added to the translate route to retrieve a precise version of a document. This date must be the exact date of extraction. If this field is absent, the most recent document will be used.

  • Using the POST analyze/status route and a valid instance id, you can now receive a new status pending. This means your instance is in queue and will be processed.

  • The languages parameter of the dataset/route now supports japanese and chinese languages.