TextAnalyticsClient Class
Definition
Important
Some information relates to prerelease product that may be substantially modified before it’s released. Microsoft makes no warranties, express or implied, with respect to the information provided here.
The client to use for interacting with the Azure Cognitive Service for Language, which includes Text Analytics.
public class TextAnalyticsClient
type TextAnalyticsClient = class
Public Class TextAnalyticsClient
- Inheritance
-
TextAnalyticsClient
Constructors
TextAnalyticsClient() |
Protected constructor to allow mocking. |
TextAnalyticsClient(Uri, AzureKeyCredential) |
Initializes a new instance of the AzureKeyCredential class for the specified service instance. |
TextAnalyticsClient(Uri, AzureKeyCredential, TextAnalyticsClientOptions) |
Initializes a new instance of the AzureKeyCredential class for the specified service instance. |
TextAnalyticsClient(Uri, TokenCredential) |
Initializes a new instance of the TextAnalyticsClient class for the specified service instance. |
TextAnalyticsClient(Uri, TokenCredential, TextAnalyticsClientOptions) |
Initializes a new instance of the TextAnalyticsClient class for the specified service instance. |
Methods
AbstractiveSummarize(WaitUntil, IEnumerable<String>, String, AbstractiveSummarizeOptions, CancellationToken) |
Performs abstractive summarization on a given set of documents, which consists of generating a summary with concise, coherent sentences or words which are not simply extract sentences from the original document. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AbstractiveSummarize(WaitUntil, IEnumerable<TextDocumentInput>, AbstractiveSummarizeOptions, CancellationToken) |
Performs abstractive summarization on a given set of documents, which consists of generating a summary with concise, coherent sentences or words which are not simply extract sentences from the original document. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AbstractiveSummarizeAsync(WaitUntil, IEnumerable<String>, String, AbstractiveSummarizeOptions, CancellationToken) |
Performs abstractive summarization on a given set of documents, which consists of generating a summary with concise, coherent sentences or words which are not simply extract sentences from the original document. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AbstractiveSummarizeAsync(WaitUntil, IEnumerable<TextDocumentInput>, AbstractiveSummarizeOptions, CancellationToken) |
Performs abstractive summarization on a given set of documents, which consists of generating a summary with concise, coherent sentences or words which are not simply extract sentences from the original document. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeActions(WaitUntil, IEnumerable<String>, TextAnalyticsActions, String, AnalyzeActionsOptions, CancellationToken) |
Performs one or more text analysis actions on a set of documents. The list of supported actions includes:
For document length limits, maximum batch size, and supported text encoding, see here. |
AnalyzeActions(WaitUntil, IEnumerable<TextDocumentInput>, TextAnalyticsActions, AnalyzeActionsOptions, CancellationToken) |
Performs one or more text analysis actions on a set of documents. The list of supported actions includes:
For document length limits, maximum batch size, and supported text encoding, see here. |
AnalyzeActionsAsync(WaitUntil, IEnumerable<String>, TextAnalyticsActions, String, AnalyzeActionsOptions, CancellationToken) |
Performs one or more text analysis actions on a set of documents. The list of supported actions includes:
For document length limits, maximum batch size, and supported text encoding, see more information here. |
AnalyzeActionsAsync(WaitUntil, IEnumerable<TextDocumentInput>, TextAnalyticsActions, AnalyzeActionsOptions, CancellationToken) |
Performs one or more text analysis actions on a set of documents. The list of supported actions includes:
For document length limits, maximum batch size, and supported text encoding, see here. |
AnalyzeHealthcareEntities(WaitUntil, IEnumerable<String>, String, AnalyzeHealthcareEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of healthcare entities found in the passed-in document, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeHealthcareEntities(WaitUntil, IEnumerable<TextDocumentInput>, AnalyzeHealthcareEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of healthcare entities found in the passed-in document, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeHealthcareEntitiesAsync(WaitUntil, IEnumerable<String>, String, AnalyzeHealthcareEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of healthcare entities found in the passed-in document, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeHealthcareEntitiesAsync(WaitUntil, IEnumerable<TextDocumentInput>, AnalyzeHealthcareEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of healthcare entities found in the passed-in document, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeSentiment(String, String, AnalyzeSentimentOptions, CancellationToken) |
Runs a predictive model to identify the positive, negative or neutral sentiment contained in the document, as well as a score indicating the model's confidence in the predicted sentiment. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeSentimentAsync(String, String, AnalyzeSentimentOptions, CancellationToken) |
Runs a predictive model to identify the positive, negative, neutral or mixed sentiment contained in the document, as well as a score indicating the model's confidence in the predicted sentiment. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeSentimentBatch(IEnumerable<String>, String, AnalyzeSentimentOptions, CancellationToken) |
Runs a predictive model to identify the positive, negative or neutral sentiment contained in the documents, as well as scores indicating the model's confidence in each of the predicted sentiments. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeSentimentBatch(IEnumerable<TextDocumentInput>, AnalyzeSentimentOptions, CancellationToken) |
Runs a predictive model to identify the positive, negative or neutral sentiment contained in the documents, as well as scores indicating the model's confidence in each of the predicted sentiments. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeSentimentBatchAsync(IEnumerable<String>, String, AnalyzeSentimentOptions, CancellationToken) |
Runs a predictive model to identify the positive, negative or neutral sentiment contained in the documents, as well as scores indicating the model's confidence in each of the predicted sentiments. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
AnalyzeSentimentBatchAsync(IEnumerable<TextDocumentInput>, AnalyzeSentimentOptions, CancellationToken) |
Runs a predictive model to identify the positive, negative or neutral sentiment contained in the documents, as well as scores indicating the model's confidence in each of the predicted sentiments. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
DetectLanguage(String, String, CancellationToken) |
Runs a predictive model to determine the language the passed-in document is written in, and returns the detected language as well as a score indicating the model's confidence that the inferred language is correct. Scores close to 1 indicate high certainty in the result. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
DetectLanguageAsync(String, String, CancellationToken) |
Runs a predictive model to determine the language the passed-in document is written in, and returns the detected language as well as a score indicating the model's confidence that the inferred language is correct. Scores close to 1 indicate high certainty in the result. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
DetectLanguageBatch(IEnumerable<DetectLanguageInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to determine the language the passed-in documents are written in, and returns, for each one, the detected language as well as a score indicating the model's confidence that the inferred language is correct. Scores close to 1 indicate high certainty in the result. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
DetectLanguageBatch(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to determine the language the passed-in documents are written in, and returns, for each one, the detected language as well as a score indicating the model's confidence that the inferred language is correct. Scores close to 1 indicate high certainty in the result. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
DetectLanguageBatchAsync(IEnumerable<DetectLanguageInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to determine the language the passed-in documents are written in, and returns, for each one, the detected language as well as a score indicating the model's confidence that the inferred language is correct. Scores close to 1 indicate high certainty in the result. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
DetectLanguageBatchAsync(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to determine the language the passed-in documents are written in, and returns, for each one, the detected language as well as a score indicating the model's confidence that the inferred language is correct. Scores close to 1 indicate high certainty in the result. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractiveSummarize(WaitUntil, IEnumerable<String>, String, ExtractiveSummarizeOptions, CancellationToken) |
Performs extractive summarization on the given documents, which consists of extracting sentences that collectively represent the most important or relevant information within the original content. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractiveSummarize(WaitUntil, IEnumerable<TextDocumentInput>, ExtractiveSummarizeOptions, CancellationToken) |
Performs extractive summarization on the given documents, which consists of extracting sentences that collectively represent the most important or relevant information within the original content. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractiveSummarizeAsync(WaitUntil, IEnumerable<String>, String, ExtractiveSummarizeOptions, CancellationToken) |
Performs extractive summarization on the given documents, which consists of extracting sentences that collectively represent the most important or relevant information within the original content. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractiveSummarizeAsync(WaitUntil, IEnumerable<TextDocumentInput>, ExtractiveSummarizeOptions, CancellationToken) |
Performs extractive summarization on the given documents, which consists of extracting sentences that collectively represent the most important or relevant information within the original content. For a list of languages supported by this operation, see https://learn.microsoft.com/azure/cognitive-services/language-service/summarization/language-support. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractKeyPhrases(String, String, CancellationToken) |
Runs a model to identify a collection of significant phrases found in the passed-in document. For example, for the document "The food was delicious and there were wonderful staff", the API returns the main talking points: "food" and "wonderful staff". For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractKeyPhrasesAsync(String, String, CancellationToken) |
Runs a model to identify a collection of significant phrases found in the passed-in document. For example, for the document "The food was delicious and there were wonderful staff", the API returns the main talking points: "food" and "wonderful staff". For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractKeyPhrasesBatch(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a model to identify a collection of significant phrases found in the passed-in documents. For example, for the document "The food was delicious and there were wonderful staff", the API returns the main talking points: "food" and "wonderful staff". For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractKeyPhrasesBatch(IEnumerable<TextDocumentInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a model to identify a collection of significant phrases found in the passed-in documents. For example, for the document "The food was delicious and there were wonderful staff", the API returns the main talking points: "food" and "wonderful staff". For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractKeyPhrasesBatchAsync(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a model to identify a collection of significant phrases found in the passed-in documents. For example, for the document "The food was delicious and there were wonderful staff", the API returns the main talking points: "food" and "wonderful staff". For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
ExtractKeyPhrasesBatchAsync(IEnumerable<TextDocumentInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a model to identify a collection of significant phrases found in the passed-in documents. For example, for the document "The food was delicious and there were wonderful staff", the API returns the main talking points: "food" and "wonderful staff". For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
MultiLabelClassify(WaitUntil, IEnumerable<String>, String, String, String, MultiLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with multiple labels in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
MultiLabelClassify(WaitUntil, IEnumerable<TextDocumentInput>, String, String, MultiLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with multiple labels in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
MultiLabelClassifyAsync(WaitUntil, IEnumerable<String>, String, String, String, MultiLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with multiple labels in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
MultiLabelClassifyAsync(WaitUntil, IEnumerable<TextDocumentInput>, String, String, MultiLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with multiple labels in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeCustomEntities(WaitUntil, IEnumerable<String>, String, String, String, RecognizeCustomEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of custom named entities in the passed-in documents, and categorize those entities into custom types such as contracts or financial documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-named-entity-recognition/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeCustomEntities(WaitUntil, IEnumerable<TextDocumentInput>, String, String, RecognizeCustomEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of custom named entities in the passed-in documents, and categorize those entities into custom types such as contracts or financial documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-named-entity-recognition/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeCustomEntitiesAsync(WaitUntil, IEnumerable<String>, String, String, String, RecognizeCustomEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of custom named entities in the passed-in documents, and categorize those entities into custom types such as contracts or financial documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-named-entity-recognition/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeCustomEntitiesAsync(WaitUntil, IEnumerable<TextDocumentInput>, String, String, RecognizeCustomEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of custom named entities in the passed-in documents, and categorize those entities into custom types such as contracts or financial documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-named-entity-recognition/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeEntities(String, String, CancellationToken) |
Runs a predictive model to identify a collection of named entities in the passed-in document, and categorize those entities into types such as person, location, or organization. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/named-entity-recognition/concepts/named-entity-categories. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeEntitiesAsync(String, String, CancellationToken) |
Runs a predictive model to identify a collection of named entities in the passed-in document, and categorize those entities into types such as person, location, or organization. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/named-entity-recognition/concepts/named-entity-categories. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeEntitiesBatch(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of named entities in the passed-in documents, and categorize those entities into types such as person, location, or organization. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/named-entity-recognition/concepts/named-entity-categories. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeEntitiesBatch(IEnumerable<TextDocumentInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of named entities in the passed-in documents, and categorize those entities into types such as person, location, or organization. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/named-entity-recognition/concepts/named-entity-categories. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeEntitiesBatchAsync(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of named entities in the passed-in documents, and categorize those entities into types such as person, location, or organization. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/named-entity-recognition/concepts/named-entity-categories. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeEntitiesBatchAsync(IEnumerable<TextDocumentInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of named entities in the passed-in documents, and categorize those entities into types such as person, location, or organization. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/named-entity-recognition/concepts/named-entity-categories. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeLinkedEntities(String, String, CancellationToken) |
Runs a predictive model to identify a collection of entities found in the passed-in document, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeLinkedEntitiesAsync(String, String, CancellationToken) |
Runs a predictive model to identify a collection of entities found in the passed-in document, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeLinkedEntitiesBatch(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities found in the passed-in documents, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeLinkedEntitiesBatch(IEnumerable<TextDocumentInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities found in the passed-in documents, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeLinkedEntitiesBatchAsync(IEnumerable<String>, String, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities found in the passed-in documents, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizeLinkedEntitiesBatchAsync(IEnumerable<TextDocumentInput>, TextAnalyticsRequestOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities found in the passed-in documents, and include information linking the entities to their corresponding entries in a well-known knowledge base. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizePiiEntities(String, String, RecognizePiiEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities containing Personally Identifiable Information found in the passed-in document, and categorize those entities into types such as US social security number, drivers license number, or credit card number. For more information on available categories, see https://aka.ms/azsdk/language/pii. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizePiiEntitiesAsync(String, String, RecognizePiiEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities containing Personally Identifiable Information found in the passed-in document, and categorize those entities into types such as US social security number, drivers license number, or credit card number. For more information on available categories, see https://aka.ms/azsdk/language/pii. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizePiiEntitiesBatch(IEnumerable<String>, String, RecognizePiiEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities containing Personally Identifiable Information found in the passed-in document, and categorize those entities into types such as US social security number, drivers license number, or credit card number. For more information on available categories, see https://aka.ms/azsdk/language/pii. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizePiiEntitiesBatch(IEnumerable<TextDocumentInput>, RecognizePiiEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities containing Personally Identifiable Information found in the passed-in document, and categorize those entities into types such as US social security number, drivers license number, or credit card number. For more information on available categories, see https://aka.ms/azsdk/language/pii. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizePiiEntitiesBatchAsync(IEnumerable<String>, String, RecognizePiiEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities containing Personally Identifiable Information found in the passed-in document, and categorize those entities into types such as US social security number, drivers license number, or credit card number. For more information on available categories, see https://aka.ms/azsdk/language/pii. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
RecognizePiiEntitiesBatchAsync(IEnumerable<TextDocumentInput>, RecognizePiiEntitiesOptions, CancellationToken) |
Runs a predictive model to identify a collection of entities containing Personally Identifiable Information found in the passed-in document, and categorize those entities into types such as US social security number, drivers license number, or credit card number. For more information on available categories, see https://aka.ms/azsdk/language/pii. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
SingleLabelClassify(WaitUntil, IEnumerable<String>, String, String, String, SingleLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with a single label in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
SingleLabelClassify(WaitUntil, IEnumerable<TextDocumentInput>, String, String, SingleLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with a single label in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
SingleLabelClassifyAsync(WaitUntil, IEnumerable<String>, String, String, String, SingleLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with a single label in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
SingleLabelClassifyAsync(WaitUntil, IEnumerable<TextDocumentInput>, String, String, SingleLabelClassifyOptions, CancellationToken) |
Runs a predictive model to identify a classify each document with a single label in the passed-in documents. For more information on available categories, see https://docs.microsoft.com/azure/cognitive-services/language-service/custom-text-classification/overview. For a list of languages supported by this operation, see https://aka.ms/talangs. For document length limits, maximum batch size, and supported text encoding, see https://aka.ms/azsdk/textanalytics/data-limits. |
Applies to
Azure SDK for .NET