Deprecated cognitive skills in Azure Cognitive Search

This document describes cognitive skills that are considered deprecated. Use the following guide for the contents:

  • Skill Name: The name of the skill that will be deprecated, it maps to the @odata.type attribute.
  • Last available api version: The last version of the Azure Cognitive Search public API through which skillsets containing the corresponding deprecated skill can be created/updated.
  • End of support: The last day after which the corresponding skill is considered unsupported. Previously created skillsets should still continue to function, but users are recommended to migrate away from a deprecated skill.
  • Recommendations: Migration path forward to use a supported skill. Users are advised to follow the recommendations to continue to receive support.

Microsoft.Skills.Text.NamedEntityRecognitionSkill

Last available api version

2017-11-11-Preview

End of support

February 15, 2019

Recommendations

Use Microsoft.Skills.Text.EntityRecognitionSkill instead. It provides most of the functionality of the NamedEntityRecognitionSkill at a higher quality. It also has richer information in its complex output fields.

To migrate to the Entity Recognition Skill, you will have to perform one or more of the following changes to your skill definition. You can update the skill definition using the Update Skillset API.

Note

Currently, confidence score as a concept is not supported. The minimumPrecision parameter exists on the EntityRecognitionSkill for future use and for backwards compatibility.

  1. (Required) Change the @odata.type from "#Microsoft.Skills.Text.NamedEntityRecognitionSkill" to "#Microsoft.Skills.Text.EntityRecognitionSkill".

  2. (Optional) If you are making use of the entities output, use the namedEntities complex collection output from the EntityRecognitionSkill instead. You can use the targetName in the skill definition to map it to an annotation called entities.

  3. (Optional) If you do not explicitly specify the categories, the EntityRecognitionSkill can return different type of categories besides those that were supported by the NamedEntityRecognitionSkill. If this behavior is undesirable, make sure to explicitly set the categories parameter to ["Person", "Location", "Organization"].

    Sample Migration Definitions

    • Simple migration

      (Before) NamedEntityRecognition skill definition

      {
          "@odata.type": "#Microsoft.Skills.Text.NamedEntityRecognitionSkill",
          "categories": [ "Person"],
          "defaultLanguageCode": "en",
          "inputs": [
          {
              "name": "text",
              "source": "/document/content"
          }
          ],
          "outputs": [
          {
              "name": "persons",
              "targetName": "people"
          }
          ]
      }
      

      (After) EntityRecognition skill definition

      {
          "@odata.type": "#Microsoft.Skills.Text.EntityRecognitionSkill",
          "categories": [ "Person"],
          "defaultLanguageCode": "en",
          "inputs": [
          {
              "name": "text",
              "source": "/document/content"
          }
          ],
          "outputs": [
          {
              "name": "persons",
              "targetName": "people"
          }
          ]
      }
      
    • Slightly complicated migration

      (Before) NamedEntityRecognition skill definition

      {
          "@odata.type": "#Microsoft.Skills.Text.NamedEntityRecognitionSkill",
          "defaultLanguageCode": "en",
          "minimumPrecision": 0.1,
          "inputs": [
          {
              "name": "text",
              "source": "/document/content"
          }
          ],
          "outputs": [
          {
              "name": "persons",
              "targetName": "people"
          },
          {
              "name": "entities"
          }
          ]
      }
      

      (After) EntityRecognition skill definition

      {
          "@odata.type": "#Microsoft.Skills.Text.EntityRecognitionSkill",
          "categories": [ "Person", "Location", "Organization" ],
          "defaultLanguageCode": "en",
          "minimumPrecision": 0.1,
          "inputs": [
          {
              "name": "text",
              "source": "/document/content"
          }
          ],
          "outputs": [
          {
              "name": "persons",
              "targetName": "people"
          },
          {
              "name": "namedEntities",
              "targetName": "entities"
          }
          ]
      }
      

See also