Language detection cognitive skill

Статия
08/28/2024

The Language Detection skill detects the language of input text and reports a single language code for every document submitted on the request. The language code is paired with a score indicating the strength of the analysis. This skill uses the machine learning models provided in Azure AI Language.

This capability is especially useful when you need to provide the language of the text as input to other skills (for example, the Sentiment Analysis skill or Text Split skill).

See supported languages for Language Detection. If you have content expressed in an unsupported language, the response is (Unknown).

Note

This skill is bound to Azure AI services and requires a billable resource for transactions that exceed 20 documents per indexer per day. Execution of built-in skills is charged at the existing Azure AI services pay-as-you go price.

@odata.type

Microsoft.Skills.Text.LanguageDetectionSkill

Data limits

The maximum size of a record should be 50,000 characters as measured by String.Length. If you need to break up your data before sending it to the language detection skill, you can use the Text Split skill.

Skill parameters

Parameters are case-sensitive.

Inputs	Description
`defaultCountryHint`	(Optional) An ISO 3166-1 alpha-2 two letter country code can be provided to use as a hint to the language detection model if it can't disambiguate the language. Specifically, the `defaultCountryHint` parameter is used with documents that don't specify the `countryHint` input explicitly.
`modelVersion`	(Optional) Specifies the version of the model to use when calling language detection. It defaults to the latest available when not specified. We recommend you don't specify this value unless it's necessary.

Skill inputs

Parameters are case-sensitive.

Inputs	Description
`text`	The text to be analyzed.
`countryHint`	An ISO 3166-1 alpha-2 two letter country code to use as a hint to the language detection model if it can't disambiguate the language.

Skill outputs

Output Name	Description
`languageCode`	The ISO 6391 language code for the language identified. For example, "en".
`languageName`	The name of language. For example, "English".
`score`	A value between 0 and 1. The likelihood that language is correctly identified. The score can be lower than 1 if the sentence has mixed languages.

Sample definition

 {
    "@odata.type": "#Microsoft.Skills.Text.LanguageDetectionSkill",
    "inputs": [
      {
        "name": "text",
        "source": "/document/text"
      },
      {
        "name": "countryHint",
        "source": "/document/countryHint"
      }
    ],
    "outputs": [
      {
        "name": "languageCode",
        "targetName": "myLanguageCode"
      },
      {
        "name": "languageName",
        "targetName": "myLanguageName"
      },
      {
        "name": "score",
        "targetName": "myLanguageScore"
      }

    ]
  }

Sample input

{
    "values": [
      {
        "recordId": "1",
        "data":
           {
             "text": "Glaciers are huge rivers of ice that ooze their way over land, powered by gravity and their own sheer weight. "
           }
      },
      {
        "recordId": "2",
        "data":
           {
             "text": "Estamos muy felices de estar con ustedes."
           }
      },
      {
        "recordId": "3",
        "data":
           {
             "text": "impossible",
             "countryHint": "fr"
           }
      }
    ]

Sample output

{
    "values": [
      {
        "recordId": "1",
        "data":
            {
              "languageCode": "en",
              "languageName": "English",
              "score": 1,
            }
      },
      {
        "recordId": "2",
        "data":
            {
              "languageCode": "es",
              "languageName": "Spanish",
              "score": 1,
            }
      },
      {
        "recordId": "3",
        "data":
            {
              "languageCode": "fr",
              "languageName": "French",
              "score": 1,
            }
      }
    ]
}

Споделяне чрез