Skip to main content
Version: Current

Cognitive Services Models

Overview

This page describes object types returned by Azure AI Foundry workflow activities.

Use these models when binding activity outputs in the Process Data drawer or when mapping results in downstream activities.


AnalyzeSentimentResult

Returned by Analyze Sentiment as Analyze Sentiment Result.

Properties

PropertyTypeDescription
SentimentStringOverall sentiment: Positive, Negative, Neutral, or Mixed.
PositiveScoreDoubleConfidence that the sentiment is positive (0 to 1).
NegativeScoreDoubleConfidence that the sentiment is negative (0 to 1).
NeutralScoreDoubleConfidence that the sentiment is neutral (0 to 1).
MixedScoreDoubleConfidence that the sentiment is mixed (0 to 1).
SentencesList<SentimentSentence>Sentence-level sentiment details.

SentimentSentence

Represents sentiment information for a single sentence within AnalyzeSentimentResult.Sentences.

Properties

PropertyTypeDescription
SentenceStringText of the sentence.
SentimentStringSentiment detected for the sentence.
PositiveScoreDoubleConfidence that the sentiment is positive (0 to 1).
NegativeScoreDoubleConfidence that the sentiment is negative (0 to 1).
NeutralScoreDoubleConfidence that the sentiment is neutral (0 to 1).
MixedScoreDoubleConfidence that the sentiment is mixed (0 to 1).
LengthInt32Length of the sentence text.
OffsetInt32Starting position of the sentence within the input text.

DetectLanguageResult

Returned by Detect Language as Detect Language Result.

Properties

PropertyTypeDescription
LanguageStringDetected language name (for example, English).
LanguageCodeStringISO 639-1 language code (for example, en).
ConfidenceScoreDoubleConfidence of the language detection (0 to 1).

Entity

Base type for extracted entities.

Used by:

  • Recognize Entities
  • Recognize PII Entities

Properties

PropertyTypeDescription
TextStringDetected entity text.
ConfidenceScoreDoubleConfidence of the detection (0 to 1).
LengthInt32Length of the entity text.
OffsetInt32Starting position of the entity within the input text.

CategorizedEntity

Returned in Categorized Entities and PII Entities collections.

Extends Entity with category information.

Properties

PropertyTypeDescription
TextStringDetected entity text.
ConfidenceScoreDoubleConfidence of the detection (0 to 1).
LengthInt32Length of the entity text.
OffsetInt32Starting position of the entity within the input text.
CategoryStringEntity category.
SubCategoryStringEntity subcategory, when available.

AnalyzeResult

Returned by Analyze Document as Analyze Result.

Represents the top-level output of document analysis.

Properties

PropertyTypeDescription
ApiVersionStringAPI version used for the analysis.
ModelIdStringModel identifier used for the analysis.
ContentStringExtracted text content from the document.
ContentFormatStringFormat of the extracted content.
PagesListDocument pages with layout and text details.
ParagraphsListParagraphs detected in the document.
TablesListTables detected in the document.
FiguresListFigures detected in the document, when available.
SectionsListDocument sections, when available.
KeyValuePairsListKey-value pairs detected when the feature is enabled.
StylesListStyle information detected when the feature is enabled.
LanguagesListLanguages detected when the feature is enabled.
DocumentsListStructured fields extracted for the selected document type.

Usage Notes

The structure of nested page, table, paragraph, figure, and document field objects is defined by Azure AI Foundry Document Intelligence.

Refer to the Azure AI Foundry documentation for detailed schema information.