public async Task RecognizeEntitiesBatchConvenienceTest()
{
TextAnalyticsClient client = GetClient();
var documents = new List <string>
{
"Microsoft was founded by Bill Gates and Paul Allen.",
"My cat and my dog might need to see a veterinarian."
};
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(documents);
foreach (RecognizeEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 1);
}
}
public async Task RecognizeEntitiesBatchWithNullTextTest()
{
TextAnalyticsClient client = GetClient();
var documents = new List <TextDocumentInput> {
new TextDocumentInput("1", null)
};
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(documents);
var exceptionMessage = "Cannot access result for document 1, due to error InvalidDocument: Document text is empty.";
Assert.IsTrue(results[0].HasError);
InvalidOperationException ex = Assert.Throws <InvalidOperationException>(() => results[0].Entities.Count());
Assert.AreEqual(exceptionMessage, ex.Message);
}
public async Task RecognizeEntitiesBatchWithStatisticsTest()
{
TextAnalyticsClient client = GetClient();
List <TextDocumentInput> documents = batchDocuments;
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(documents, new TextAnalyticsRequestOptions { IncludeStatistics = true });
foreach (RecognizeEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 1);
}
Assert.IsNotNull(results.Statistics.DocumentCount);
Assert.IsNotNull(results.Statistics.InvalidDocumentCount);
Assert.IsNotNull(results.Statistics.TransactionCount);
Assert.IsNotNull(results.Statistics.ValidDocumentCount);
}
public async Task RecognizeEntitiesResultsSorted_WithErrors()
{
var mockResults = new List <RecognizeEntitiesResult>()
{
new RecognizeEntitiesResult("2", new TextDocumentStatistics(), new List <CategorizedEntity>()
{
new CategorizedEntity("EntityText0", "EntityCategory0", "EntitySubCategory0", 0, 1, 0.5),
new CategorizedEntity("EntityText1", "EntityCategory1", "EntitySubCategory1", 0, 1, 0.5),
}),
new RecognizeEntitiesResult("3", new TextDocumentStatistics(), new List <CategorizedEntity>()
{
new CategorizedEntity("EntityText0", "EntityCategory0", "EntitySubCategory0", 0, 1, 0.5),
new CategorizedEntity("EntityText1", "EntityCategory1", "EntitySubCategory1", 0, 1, 0.5),
}),
new RecognizeEntitiesResult("4", new TextAnalyticsError("InvalidDocument", "Document is invalid.")),
new RecognizeEntitiesResult("5", new TextAnalyticsError("InvalidDocument", "Document is invalid.")),
};
var mockResultCollection = new RecognizeEntitiesResultCollection(mockResults,
new TextDocumentBatchStatistics(2, 2, 2, 2),
"modelVersion");
var mockResponse = new MockResponse(200);
mockResponse.SetContent(SerializationHelpers.Serialize(mockResultCollection, SerializeRecognizeEntitiesResultCollection));
var mockTransport = new MockTransport(mockResponse);
TextAnalyticsClient client = CreateTestClient(mockTransport);
var documents = new List <TextDocumentInput>()
{
new TextDocumentInput("4", "TextDocument1"),
new TextDocumentInput("5", "TextDocument2"),
new TextDocumentInput("2", "TextDocument3"),
new TextDocumentInput("3", "TextDocument4"),
};
var response = await client.RecognizeEntitiesBatchAsync(documents, new TextAnalyticsRequestOptions());
var resultCollection = response.Value;
Assert.AreEqual("4", resultCollection[0].Id);
Assert.AreEqual("5", resultCollection[1].Id);
Assert.AreEqual("2", resultCollection[2].Id);
Assert.AreEqual("3", resultCollection[3].Id);
}
public async Task RecognizeEntitiesBatchWithErrorTest()
{
TextAnalyticsClient client = GetClient();
var inputs = new List <string>
{
"Microsoft was founded by Bill Gates and Paul Allen.",
"",
"My cat might need to see a veterinarian."
};
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(inputs);
Assert.IsTrue(!results[0].HasError);
Assert.IsTrue(!results[2].HasError);
Assert.IsTrue(results[1].HasError);
Assert.Throws <InvalidOperationException>(() => results[1].Entities.GetType());
}
public async Task RecognizeEntitiesBatchWithStatisticsTest()
{
TextAnalyticsRequestOptions options = new TextAnalyticsRequestOptions {
IncludeStatistics = true
};
TextAnalyticsClient client = GetClient();
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(s_batchDocuments, options);
var expectedOutput = new Dictionary <string, List <string> >()
{
{ "1", s_document1ExpectedOutput },
{ "2", s_document1ExpectedOutput },
};
ValidateBatchDocumentsResult(results, expectedOutput, includeStatistics: true);
// Assert the options classes since overloads were added and the original now instantiates a RecognizeEntitiesOptions.
Assert.IsTrue(options.IncludeStatistics);
Assert.IsNull(options.ModelVersion);
}
public async Task EntitiesCategories()
{
TextAnalyticsClient client = GetClient();
const string document = "Bill Gates | Microsoft | New Mexico";
RecognizeEntitiesResultCollection response = await client.RecognizeEntitiesBatchAsync(new List <string>() { document }, "en", new TextAnalyticsRequestOptions()
{
ModelVersion = "2020-02-01"
});
var entities = response.FirstOrDefault().Entities.ToList();
Assert.AreEqual(3, entities.Count);
Assert.AreEqual(EntityCategory.Person, entities[0].Category);
Assert.AreEqual(EntityCategory.Organization, entities[1].Category);
Assert.AreEqual(EntityCategory.Location, entities[2].Category);
}
public async Task RecognizeEntitiesBatchConvenienceWithRecognizeEntitiesOptionsStatisticsTest()
{
RecognizeEntitiesOptions options = new RecognizeEntitiesOptions {
IncludeStatistics = true
};
TextAnalyticsClient client = GetClient();
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(s_batchConvenienceDocuments, "en", options);
var expectedOutput = new Dictionary <string, List <string> >()
{
{ "0", s_document1ExpectedOutput },
{ "1", s_document2ExpectedOutput },
};
ValidateBatchDocumentsResult(results, expectedOutput, includeStatistics: true);
// Assert the options classes since overloads were added and the original now instantiates a RecognizeEntitiesOptions.
Assert.IsTrue(options.IncludeStatistics);
Assert.IsNull(options.ModelVersion);
Assert.AreEqual(StringIndexType.Utf16CodeUnit, options.StringIndexType);
}
public async Task RecognizeEntitiesBatchConvenienceWithStatisticsTest()
{
TextAnalyticsClient client = GetClient();
var inputs = new List <string>
{
"Microsoft was founded by Bill Gates and Paul Allen.",
"My cat and my dog might need to see a veterinarian."
};
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(inputs, "en", new TextAnalyticsRequestOptions { IncludeStatistics = true });
foreach (RecognizeEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 1);
}
Assert.IsNotNull(results.Statistics.DocumentCount);
Assert.IsNotNull(results.Statistics.InvalidDocumentCount);
Assert.IsNotNull(results.Statistics.TransactionCount);
Assert.IsNotNull(results.Statistics.ValidDocumentCount);
}
private void ValidateBatchDocumentsResult(RecognizeEntitiesResultCollection results, Dictionary <string, List <string> > minimumExpectedOutput, bool includeStatistics = default)
{
Assert.That(results.ModelVersion, Is.Not.Null.And.Not.Empty);
if (includeStatistics)
{
Assert.IsNotNull(results.Statistics);
Assert.Greater(results.Statistics.DocumentCount, 0);
Assert.Greater(results.Statistics.TransactionCount, 0);
Assert.GreaterOrEqual(results.Statistics.InvalidDocumentCount, 0);
Assert.GreaterOrEqual(results.Statistics.ValidDocumentCount, 0);
}
else
{
Assert.IsNull(results.Statistics);
}
foreach (RecognizeEntitiesResult entitiesInDocument in results)
{
Assert.That(entitiesInDocument.Id, Is.Not.Null.And.Not.Empty);
Assert.False(entitiesInDocument.HasError);
//Even though statistics are not asked for, TA 5.0.0 shipped with Statistics default always present.
Assert.IsNotNull(entitiesInDocument.Statistics);
if (includeStatistics)
{
Assert.GreaterOrEqual(entitiesInDocument.Statistics.CharacterCount, 0);
Assert.Greater(entitiesInDocument.Statistics.TransactionCount, 0);
}
else
{
Assert.AreEqual(0, entitiesInDocument.Statistics.CharacterCount);
Assert.AreEqual(0, entitiesInDocument.Statistics.TransactionCount);
}
ValidateInDocumenResult(entitiesInDocument.Entities, minimumExpectedOutput[entitiesInDocument.Id]);
}
}
public async Task RecognizeEntitiesBatchTest()
{
TextAnalyticsClient client = GetClient();
var inputs = new List <TextDocumentInput>
{
new TextDocumentInput("1", "Microsoft was founded by Bill Gates and Paul Allen.")
{
Language = "en",
},
new TextDocumentInput("2", "Mi perro y mi gato tienen que ir al veterinario.")
{
Language = "es",
}
};
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(inputs);
foreach (RecognizeEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 1);
}
}
public async Task RecognizeEntitiesBatchWithErrorTest()
{
TextAnalyticsClient client = GetClient();
var inputs = new List <string>
{
"Microsoft was founded by Bill Gates and Paul Allen.",
"",
"My cat might need to see a veterinarian."
};
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(inputs);
Assert.IsTrue(!results[0].HasError);
Assert.IsTrue(!results[2].HasError);
var exceptionMessage = "Cannot access result for document 1, due to error InvalidDocument: Document text is empty.";
Assert.IsTrue(results[1].HasError);
InvalidOperationException ex = Assert.Throws <InvalidOperationException>(() => results[1].Entities.GetType());
Assert.AreEqual(exceptionMessage, ex.Message);
}
public async Task RecognizeEntitiesResultsSorted_NoErrors()
{
var mockResults = new List <RecognizeEntitiesResult>()
{
new RecognizeEntitiesResult("1", new TextDocumentStatistics(), new List <NamedEntity>()
{
new NamedEntity("EntityText0", "EntityType0", "EntitySubType0", 0, 1, 0.5),
new NamedEntity("EntityText1", "EntityType1", "EntitySubType1", 0, 1, 0.5),
}),
new RecognizeEntitiesResult("2", new TextDocumentStatistics(), new List <NamedEntity>()
{
new NamedEntity("EntityText0", "EntityType0", "EntitySubType0", 0, 1, 0.5),
new NamedEntity("EntityText1", "EntityType1", "EntitySubType1", 0, 1, 0.5),
}),
};
var mockResultCollection = new RecognizeEntitiesResultCollection(mockResults,
new TextDocumentBatchStatistics(2, 2, 0, 2),
"modelVersion");
var mockResponse = new MockResponse(200);
mockResponse.SetContent(SerializationHelpers.Serialize(mockResultCollection, SerializeRecognizeEntitiesResultCollection));
var mockTransport = new MockTransport(mockResponse);
TextAnalyticsClient client = CreateTestClient(mockTransport);
var inputs = new List <TextDocumentInput>()
{
new TextDocumentInput("1", "TextDocument1"),
new TextDocumentInput("2", "TextDocument2"),
};
var response = await client.RecognizeEntitiesAsync(inputs, new TextAnalyticsRequestOptions());
var resultCollection = response.Value;
Assert.AreEqual("1", resultCollection[0].Id);
Assert.AreEqual("2", resultCollection[1].Id);
}
public async Task TextWithStringIndexType()
{
TextAnalyticsClient client = GetClient();
string document = "👨 Microsoft the company.";
RecognizeEntitiesResultCollection responseWithUnicodeCodePoint = await client.RecognizeEntitiesBatchAsync(new List <string>() { document }, "en", new TextAnalyticsRequestOptions()
{
StringIndexType = StringIndexType.UnicodeCodePoint
});
RecognizeEntitiesResultCollection responseWithUtf16CodeUnit = await client.RecognizeEntitiesBatchAsync(new List <string>() { document }, "en");
var entitiesWithUnicodeCodePoint = responseWithUnicodeCodePoint.FirstOrDefault().Entities;
var entitiesWithUtf16CodeUnit = responseWithUtf16CodeUnit.FirstOrDefault().Entities;
Assert.AreEqual(1, entitiesWithUnicodeCodePoint.Count);
Assert.AreEqual("Microsoft", entitiesWithUnicodeCodePoint.FirstOrDefault().Text);
Assert.AreEqual(2, entitiesWithUnicodeCodePoint.FirstOrDefault().Offset);
Assert.AreEqual(1, entitiesWithUtf16CodeUnit.Count);
Assert.AreEqual("Microsoft", entitiesWithUtf16CodeUnit.FirstOrDefault().Text);
Assert.AreEqual(3, entitiesWithUtf16CodeUnit.FirstOrDefault().Offset);
}
public async Task RecognizeEntitiesWithSubCategoryTest()
{
TextAnalyticsClient client = GetClient();
string document = "I had a wonderful trip to Seattle last week.";
RecognizeEntitiesResultCollection result = await client.RecognizeEntitiesBatchAsync(new List <string>() { document }, options : new TextAnalyticsRequestOptions()
{
ModelVersion = "2020-04-01"
});
var documentResult = result.FirstOrDefault();
Assert.IsFalse(documentResult.HasError);
Assert.GreaterOrEqual(documentResult.Entities.Count, 3);
foreach (CategorizedEntity entity in documentResult.Entities)
{
if (entity.Text == "last week")
{
Assert.AreEqual("DateRange", entity.SubCategory);
}
}
}
public async Task AnalyzeOperationWithMultipleActions()
{
TextAnalyticsClient client = GetClient();
var batchDocuments = new List <TextDocumentInput>
{
new TextDocumentInput("1", "Microsoft was founded by Bill Gates and Paul Allen.")
{
Language = "en",
},
new TextDocumentInput("2", "Mi perro y mi gato tienen que ir al veterinario.")
{
Language = "es",
}
};
TextAnalyticsActions batchActions = new TextAnalyticsActions()
{
ExtractKeyPhrasesOptions = new List <ExtractKeyPhrasesOptions>()
{
new ExtractKeyPhrasesOptions()
},
RecognizeEntitiesOptions = new List <RecognizeEntitiesOptions>()
{
new RecognizeEntitiesOptions()
},
RecognizePiiEntitiesOptions = new List <RecognizePiiEntitiesOptions>()
{
new RecognizePiiEntitiesOptions()
},
RecognizeLinkedEntitiesOptions = new List <RecognizeLinkedEntitiesOptions>()
{
new RecognizeLinkedEntitiesOptions()
},
DisplayName = "AnalyzeOperationWithMultipleTasks"
};
AnalyzeBatchActionsOperation operation = await client.StartAnalyzeBatchActionsAsync(batchDocuments, batchActions);
Assert.AreEqual(0, operation.ActionsFailed);
Assert.AreEqual(0, operation.ActionsSucceeded);
Assert.AreEqual(0, operation.ActionsInProgress);
Assert.AreEqual(0, operation.TotalActions);
await operation.WaitForCompletionAsync(PollingInterval);
Assert.AreEqual(0, operation.ActionsFailed);
Assert.AreEqual(4, operation.ActionsSucceeded);
Assert.AreEqual(0, operation.ActionsInProgress);
Assert.AreEqual(4, operation.TotalActions);
Assert.AreNotEqual(new DateTimeOffset(), operation.CreatedOn);
Assert.AreNotEqual(new DateTimeOffset(), operation.LastModified);
Assert.AreNotEqual(new DateTimeOffset(), operation.ExpiresOn);
//Take the first page
AnalyzeBatchActionsResult resultCollection = operation.Value.ToEnumerableAsync().Result.FirstOrDefault();
RecognizeEntitiesResultCollection entitiesResult = resultCollection.RecognizeEntitiesActionsResults.ElementAt(0).Result;
ExtractKeyPhrasesResultCollection keyPhrasesResult = resultCollection.ExtractKeyPhrasesActionsResults.ElementAt(0).Result;
RecognizePiiEntitiesResultCollection piiResult = resultCollection.RecognizePiiEntitiesActionsResults.ElementAt(0).Result;
RecognizeLinkedEntitiesResultCollection entityLinkingResult = resultCollection.RecognizeLinkedEntitiesActionsResults.ElementAt(0).Result;
Assert.IsNotNull(keyPhrasesResult);
Assert.IsNotNull(entitiesResult);
Assert.IsNotNull(piiResult);
Assert.IsNotNull(entityLinkingResult);
Assert.AreEqual("AnalyzeOperationWithMultipleTasks", operation.DisplayName);
// Keyphrases
Assert.AreEqual(2, keyPhrasesResult.Count);
var keyPhrasesListId1 = new List <string> {
"Bill Gates", "Paul Allen", "Microsoft"
};
var keyPhrasesListId2 = new List <string> {
"gato", "perro", "veterinario"
};
foreach (string keyphrase in keyPhrasesResult[0].KeyPhrases)
{
Assert.IsTrue(keyPhrasesListId1.Contains(keyphrase));
}
foreach (string keyphrase in keyPhrasesResult[1].KeyPhrases)
{
Assert.IsTrue(keyPhrasesListId2.Contains(keyphrase));
}
// Entities
Assert.AreEqual(2, entitiesResult.Count);
Assert.AreEqual(3, entitiesResult[0].Entities.Count);
var entitiesList = new List <string> {
"Bill Gates", "Microsoft", "Paul Allen"
};
foreach (CategorizedEntity entity in entitiesResult[0].Entities)
{
Assert.IsTrue(entitiesList.Contains(entity.Text));
Assert.IsNotNull(entity.Category);
Assert.IsNotNull(entity.Offset);
Assert.IsNotNull(entity.ConfidenceScore);
}
// PII
Assert.AreEqual(2, entitiesResult.Count);
Assert.AreEqual(3, entitiesResult[0].Entities.Count);
Assert.IsNotNull(entitiesResult[0].Id);
Assert.IsNotNull(entitiesResult[0].Entities);
Assert.IsNotNull(entitiesResult[0].Error);
// Entity Linking
Assert.AreEqual(2, entityLinkingResult.Count);
Assert.AreEqual(3, entityLinkingResult[0].Entities.Count);
Assert.IsNotNull(entityLinkingResult[0].Id);
Assert.IsNotNull(entityLinkingResult[0].Entities);
Assert.IsNotNull(entityLinkingResult[0].Error);
foreach (LinkedEntity entity in entityLinkingResult[0].Entities)
{
if (entity.Name == "Bill Gates")
{
Assert.AreEqual("Bill Gates", entity.DataSourceEntityId);
Assert.AreEqual("Wikipedia", entity.DataSource);
}
if (entity.Name == "Microsoft")
{
Assert.AreEqual("Microsoft", entity.DataSourceEntityId);
Assert.AreEqual("Wikipedia", entity.DataSource);
}
}
}
public void RecognizePiiEntitiesBatchAdvanced()
{
string endpoint = Environment.GetEnvironmentVariable("TEXT_ANALYTICS_ENDPOINT");
string subscriptionKey = Environment.GetEnvironmentVariable("TEXT_ANALYTICS_SUBSCRIPTION_KEY");
// Instantiate a client that will be used to call the service.
var client = new TextAnalyticsClient(new Uri(endpoint), subscriptionKey);
var inputs = new List <TextDocumentInput>
{
new TextDocumentInput("1", "A developer with SSN 859-98-0987 whose phone number is 206-867-5309 is building tools with our APIs.")
{
Language = "en",
},
new TextDocumentInput("2", "Your ABA number - 111000025 - is the first 9 digits in the lower left hand corner of your personal check.")
{
Language = "en",
},
new TextDocumentInput("3", "Is 998.214.865-68 your Brazilian CPF number?")
{
Language = "en",
}
};
RecognizeEntitiesResultCollection results = client.RecognizePiiEntities(inputs, new TextAnalysisOptions {
IncludeStatistics = true
});
int i = 0;
Debug.WriteLine($"Results of Azure Text Analytics \"Pii Entity Recognition\" Model, version: \"{results.ModelVersion}\"");
Debug.WriteLine("");
foreach (var result in results)
{
var document = inputs[i++];
Debug.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\", Text=\"{document.Text}\"):");
if (result.ErrorMessage != default)
{
Debug.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\", Text=\"{document.Text}\"):");
}
else
{
Debug.WriteLine($" Recognized the following {result.NamedEntities.Count()} PII entit{(result.NamedEntities.Count() > 1 ? "ies" : "y ")}:");
foreach (var entity in result.NamedEntities)
{
Debug.WriteLine($" Text: {entity.Text}, Type: {entity.Type}, SubType: {entity.SubType ?? "N/A"}, Score: {entity.Score:0.00}, Offset: {entity.Offset}, Length: {entity.Length}");
}
Debug.WriteLine($" Document statistics:");
Debug.WriteLine($" Character count: {result.Statistics.CharacterCount}");
Debug.WriteLine($" Transaction count: {result.Statistics.TransactionCount}");
Debug.WriteLine("");
}
}
Debug.WriteLine($"Batch operation statistics:");
Debug.WriteLine($" Document count: {results.Statistics.DocumentCount}");
Debug.WriteLine($" Valid document count: {results.Statistics.ValidDocumentCount}");
Debug.WriteLine($" Invalid document count: {results.Statistics.InvalidDocumentCount}");
Debug.WriteLine($" Transaction count: {results.Statistics.TransactionCount}");
Debug.WriteLine("");
}
public async Task AnalyzeOperationWithMultipleActions()
{
TextAnalyticsClient client = GetClient();
var batchDocuments = new List <TextDocumentInput>
{
new TextDocumentInput("1", "Microsoft was founded by Bill Gates and Paul Allen.")
{
Language = "en",
},
new TextDocumentInput("2", "Mi perro y mi gato tienen que ir al veterinario.")
{
Language = "es",
}
};
TextAnalyticsActions batchActions = new TextAnalyticsActions()
{
ExtractKeyPhrasesActions = new List <ExtractKeyPhrasesAction>()
{
new ExtractKeyPhrasesAction()
},
RecognizeEntitiesActions = new List <RecognizeEntitiesAction>()
{
new RecognizeEntitiesAction()
},
RecognizePiiEntitiesActions = new List <RecognizePiiEntitiesAction>()
{
new RecognizePiiEntitiesAction()
},
RecognizeLinkedEntitiesActions = new List <RecognizeLinkedEntitiesAction>()
{
new RecognizeLinkedEntitiesAction()
},
AnalyzeSentimentActions = new List <AnalyzeSentimentAction>()
{
new AnalyzeSentimentAction()
},
DisplayName = "AnalyzeOperationWithMultipleTasks"
};
AnalyzeActionsOperation operation = await client.StartAnalyzeActionsAsync(batchDocuments, batchActions);
Assert.AreEqual(0, operation.ActionsFailed);
Assert.AreEqual(0, operation.ActionsSucceeded);
Assert.AreEqual(0, operation.ActionsInProgress);
Assert.AreEqual(0, operation.ActionsTotal);
await operation.WaitForCompletionAsync();
Assert.AreEqual(0, operation.ActionsFailed);
Assert.AreEqual(5, operation.ActionsSucceeded);
Assert.AreEqual(0, operation.ActionsInProgress);
Assert.AreEqual(5, operation.ActionsTotal);
Assert.AreNotEqual(new DateTimeOffset(), operation.CreatedOn);
Assert.AreNotEqual(new DateTimeOffset(), operation.LastModified);
Assert.AreNotEqual(new DateTimeOffset(), operation.ExpiresOn);
//Take the first page
AnalyzeActionsResult resultCollection = operation.Value.ToEnumerableAsync().Result.FirstOrDefault();
IReadOnlyCollection <RecognizeEntitiesActionResult> entitiesActionsResults = resultCollection.RecognizeEntitiesResults;
IReadOnlyCollection <ExtractKeyPhrasesActionResult> keyPhrasesActionsResults = resultCollection.ExtractKeyPhrasesResults;
IReadOnlyCollection <RecognizePiiEntitiesActionResult> piiActionsResults = resultCollection.RecognizePiiEntitiesResults;
IReadOnlyCollection <RecognizeLinkedEntitiesActionResult> entityLinkingActionsResults = resultCollection.RecognizeLinkedEntitiesResults;
IReadOnlyCollection <AnalyzeSentimentActionResult> analyzeSentimentActionsResults = resultCollection.AnalyzeSentimentResults;
Assert.IsNotNull(keyPhrasesActionsResults);
Assert.IsNotNull(entitiesActionsResults);
Assert.IsNotNull(piiActionsResults);
Assert.IsNotNull(entityLinkingActionsResults);
Assert.IsNotNull(analyzeSentimentActionsResults);
Assert.AreEqual("AnalyzeOperationWithMultipleTasks", operation.DisplayName);
// Keyphrases
ExtractKeyPhrasesResultCollection keyPhrasesDocumentsResults = keyPhrasesActionsResults.FirstOrDefault().DocumentsResults;
Assert.AreEqual(2, keyPhrasesDocumentsResults.Count);
var keyPhrasesListId1 = new List <string> {
"Bill Gates", "Paul Allen", "Microsoft"
};
var keyPhrasesListId2 = new List <string> {
"Mi", "gato", "perro", "veterinario"
};
foreach (string keyphrase in keyPhrasesDocumentsResults[0].KeyPhrases)
{
Assert.IsTrue(keyPhrasesListId1.Contains(keyphrase));
}
foreach (string keyphrase in keyPhrasesDocumentsResults[1].KeyPhrases)
{
Assert.IsTrue(keyPhrasesListId2.Contains(keyphrase));
}
// Entities
RecognizeEntitiesResultCollection entitiesDocumentsResults = entitiesActionsResults.FirstOrDefault().DocumentsResults;
Assert.AreEqual(2, entitiesDocumentsResults.Count);
Assert.AreEqual(3, entitiesDocumentsResults[0].Entities.Count);
var entitiesList = new List <string> {
"Bill Gates", "Microsoft", "Paul Allen"
};
foreach (CategorizedEntity entity in entitiesDocumentsResults[0].Entities)
{
Assert.IsTrue(entitiesList.Contains(entity.Text));
Assert.IsNotNull(entity.Category);
Assert.IsNotNull(entity.Offset);
Assert.IsNotNull(entity.ConfidenceScore);
}
// PII
RecognizePiiEntitiesResultCollection piiDocumentsResults = piiActionsResults.FirstOrDefault().DocumentsResults;
Assert.AreEqual(2, piiDocumentsResults.Count);
Assert.AreEqual(3, piiDocumentsResults[0].Entities.Count);
Assert.IsNotNull(piiDocumentsResults[0].Id);
Assert.IsNotNull(piiDocumentsResults[0].Entities);
Assert.IsNotNull(piiDocumentsResults[0].Error);
// Entity Linking
RecognizeLinkedEntitiesResultCollection entityLinkingDocumentsResults = entityLinkingActionsResults.FirstOrDefault().DocumentsResults;
// Disable because of bug https://github.com/Azure/azure-sdk-for-net/issues/22648
//Assert.AreEqual(2, entityLinkingDocumentsResults.Count);
Assert.AreEqual(3, entityLinkingDocumentsResults[0].Entities.Count);
Assert.IsNotNull(entityLinkingDocumentsResults[0].Id);
Assert.IsNotNull(entityLinkingDocumentsResults[0].Entities);
Assert.IsNotNull(entityLinkingDocumentsResults[0].Error);
foreach (LinkedEntity entity in entityLinkingDocumentsResults[0].Entities)
{
if (entity.Name == "Bill Gates")
{
Assert.AreEqual("Bill Gates", entity.DataSourceEntityId);
Assert.AreEqual("Wikipedia", entity.DataSource);
}
if (entity.Name == "Microsoft")
{
Assert.AreEqual("Microsoft", entity.DataSourceEntityId);
Assert.AreEqual("Wikipedia", entity.DataSource);
}
}
// Analyze sentiment
AnalyzeSentimentResultCollection analyzeSentimentDocumentsResults = analyzeSentimentActionsResults.FirstOrDefault().DocumentsResults;
Assert.AreEqual(2, analyzeSentimentDocumentsResults.Count);
Assert.AreEqual(TextSentiment.Neutral, analyzeSentimentDocumentsResults[0].DocumentSentiment.Sentiment);
Assert.AreEqual(TextSentiment.Neutral, analyzeSentimentDocumentsResults[1].DocumentSentiment.Sentiment);
}
public async Task AnalyzeOperationAsync_AutomaticPollingWithSetInterval()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
#region Snippet:AnalyzeOperationBatchAsyncAutomaticPolling
string document = @"We went to Contoso Steakhouse located at midtown NYC last week for a dinner party,
and we adore the spot! They provide marvelous food and they have a great menu. The
chief cook happens to be the owner (I think his name is John Doe) and he is super
nice, coming out of the kitchen and greeted us all. We enjoyed very much dining in
the place! The Sirloin steak I ordered was tender and juicy, and the place was impeccably
clean. You can even pre-order from their online menu at www.contososteakhouse.com,
call 312-555-0176 or send email to [email protected]! The only complaint
I have is the food didn't come fast enough. Overall I highly recommend it!";
var batchDocuments = new List <TextDocumentInput>
{
new TextDocumentInput("1", document)
{
Language = "en",
}
};
AnalyzeOperationOptions operationOptions = new AnalyzeOperationOptions()
{
KeyPhrasesTaskParameters = new KeyPhrasesTaskParameters(),
EntitiesTaskParameters = new EntitiesTaskParameters(),
PiiTaskParameters = new PiiTaskParameters(),
DisplayName = "AnalyzeOperationSample"
};
AnalyzeOperation operation = await client.StartAnalyzeOperationBatchAsync(batchDocuments, operationOptions);
TimeSpan pollingInterval = new TimeSpan(1000);
await operation.WaitForCompletionAsync(pollingInterval);
await foreach (AnalyzeOperationResult documentsInPage in operation.Value)
{
RecognizeEntitiesResultCollection entitiesResult = documentsInPage.Tasks.EntityRecognitionTasks[0].Results;
ExtractKeyPhrasesResultCollection keyPhrasesResult = documentsInPage.Tasks.KeyPhraseExtractionTasks[0].Results;
RecognizePiiEntitiesResultCollection piiResult = documentsInPage.Tasks.EntityRecognitionPiiTasks[0].Results;
Console.WriteLine("Recognized Entities");
foreach (RecognizeEntitiesResult result in entitiesResult)
{
Console.WriteLine($" Recognized the following {result.Entities.Count} entities:");
foreach (CategorizedEntity entity in result.Entities)
{
Console.WriteLine($" Entity: {entity.Text}");
Console.WriteLine($" Category: {entity.Category}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" ConfidenceScore: {entity.ConfidenceScore}");
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine("");
}
Console.WriteLine("Recognized PII Entities");
foreach (RecognizePiiEntitiesResult result in piiResult)
{
Console.WriteLine($" Recognized the following {result.Entities.Count} PII entities:");
foreach (PiiEntity entity in result.Entities)
{
Console.WriteLine($" Entity: {entity.Text}");
Console.WriteLine($" Category: {entity.Category}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" ConfidenceScore: {entity.ConfidenceScore}");
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine("");
}
Console.WriteLine("Key Phrases");
foreach (ExtractKeyPhrasesResult result in keyPhrasesResult)
{
Console.WriteLine($" Recognized the following {result.KeyPhrases.Count} Keyphrases:");
foreach (string keyphrase in result.KeyPhrases)
{
Console.WriteLine($" {keyphrase}");
}
Console.WriteLine("");
}
}
}
private void SerializeRecognizeEntitiesResultCollection(ref Utf8JsonWriter json, RecognizeEntitiesResultCollection resultCollection)
{
json.WriteStartObject();
json.WriteStartArray("documents");
if (resultCollection.FirstOrDefault(r => r.Entities.Count > 0) != default)
{
foreach (var result in resultCollection)
{
if (!result.HasError)
{
json.WriteStartObject();
json.WriteString("id", result.Id);
json.WriteStartArray("entities");
foreach (var entity in result.Entities)
{
json.WriteStartObject();
json.WriteString("text", entity.Text);
json.WriteString("category", JsonSerializer.Serialize(entity.Category));
json.WriteString("subcategory", JsonSerializer.Serialize(entity.SubCategory));
json.WriteNumber("confidenceScore", entity.ConfidenceScore);
json.WriteEndObject();
}
json.WriteEndArray();
json.WriteStartArray("warnings");
foreach (var warning in result.Entities.Warnings)
{
json.WriteStartObject();
json.WriteString("code", warning.WarningCode.ToString());
json.WriteString("message", warning.Message);
json.WriteEndObject();
}
json.WriteEndArray();
json.WriteEndObject();
}
}
}
json.WriteEndArray();
json.WriteStartArray("errors");
if (resultCollection.FirstOrDefault(r => r.HasError) != default)
{
foreach (var result in resultCollection)
{
if (result.HasError)
{
json.WriteStartObject();
json.WriteString("id", result.Id);
json.WriteStartObject("error");
json.WriteStartObject("innererror");
json.WriteString("code", result.Error.ErrorCode.ToString());
json.WriteString("message", result.Error.Message);
json.WriteEndObject();
json.WriteEndObject();
json.WriteEndObject();
}
}
}
json.WriteEndArray();
json.WriteString("modelVersion", resultCollection.ModelVersion);
json.WriteEndObject();
// TODO: add statistics if needed
}
public void RecognizeEntitiesBatch()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
// Instantiate a client that will be used to call the service.
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
#region Snippet:TextAnalyticsSample4RecognizeEntitiesBatch
string documentA = @"We love this trail and make the trip every year. The views are breathtaking and well
worth the hike! Yesterday was foggy though, so we missed the spectacular views.
We tried again today and it was amazing. Everyone in my family liked the trail although
it was too challenging for the less athletic among us.
Not necessarily recommended for small children.
A hotel close to the trail offers services for childcare in case you want that.";
string documentB = @"Nos hospedamos en el Hotel Foo la semana pasada por nuestro aniversario. La gerencia
sabía de nuestra celebración y me ayudaron a tenerle una sorpresa a mi pareja.
La habitación estaba limpia y decorada como yo había pedido. Una gran experiencia.
El próximo año volveremos.";
string documentC = @"That was the best day of my life! We went on a 4 day trip where we stayed at Hotel Foo.
They had great amenities that included an indoor pool, a spa, and a bar.
The spa offered couples massages which were really good.
The spa was clean and felt very peaceful. Overall the whole experience was great.
We will definitely come back.";
var documents = new List <TextDocumentInput>
{
new TextDocumentInput("1", documentA)
{
Language = "en",
},
new TextDocumentInput("2", documentB)
{
Language = "es",
},
new TextDocumentInput("3", documentC)
{
Language = "en",
},
new TextDocumentInput("4", string.Empty)
};
var options = new TextAnalyticsRequestOptions {
IncludeStatistics = true
};
Response <RecognizeEntitiesResultCollection> response = client.RecognizeEntitiesBatch(documents, options);
RecognizeEntitiesResultCollection entitiesInDocuments = response.Value;
int i = 0;
Console.WriteLine($"Results of Azure Text Analytics \"Named Entity Recognition\" Model, version: \"{entitiesInDocuments.ModelVersion}\"");
Console.WriteLine("");
foreach (RecognizeEntitiesResult entitiesInDocument in entitiesInDocuments)
{
TextDocumentInput document = documents[i++];
Console.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\"):");
if (entitiesInDocument.HasError)
{
Console.WriteLine(" Error!");
Console.WriteLine($" Document error code: {entitiesInDocument.Error.ErrorCode}.");
Console.WriteLine($" Message: {entitiesInDocument.Error.Message}");
}
else
{
Console.WriteLine($" Recognized the following {entitiesInDocument.Entities.Count()} entities:");
foreach (CategorizedEntity entity in entitiesInDocument.Entities)
{
Console.WriteLine($" Text: {entity.Text}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" Category: {entity.Category}");
if (!string.IsNullOrEmpty(entity.SubCategory))
{
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine($" Confidence score: {entity.ConfidenceScore}");
Console.WriteLine("");
}
Console.WriteLine($" Document statistics:");
Console.WriteLine($" Character count: {entitiesInDocument.Statistics.CharacterCount}");
Console.WriteLine($" Transaction count: {entitiesInDocument.Statistics.TransactionCount}");
}
Console.WriteLine("");
}
Console.WriteLine($"Batch operation statistics:");
Console.WriteLine($" Document count: {entitiesInDocuments.Statistics.DocumentCount}");
Console.WriteLine($" Valid document count: {entitiesInDocuments.Statistics.ValidDocumentCount}");
Console.WriteLine($" Invalid document count: {entitiesInDocuments.Statistics.InvalidDocumentCount}");
Console.WriteLine($" Transaction count: {entitiesInDocuments.Statistics.TransactionCount}");
Console.WriteLine("");
#endregion
}
public void RecognizeEntitiesBatch()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
// Instantiate a client that will be used to call the service.
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
#region Snippet:TextAnalyticsSample4RecognizeEntitiesBatch
var documents = new List <TextDocumentInput>
{
new TextDocumentInput("1", "Microsoft was founded by Bill Gates and Paul Allen.")
{
Language = "en",
},
new TextDocumentInput("2", "Text Analytics is one of the Azure Cognitive Services.")
{
Language = "en",
},
new TextDocumentInput("3", "A key technology in Text Analytics is Named Entity Recognition (NER).")
{
Language = "en",
}
};
RecognizeEntitiesResultCollection results = client.RecognizeEntitiesBatch(documents, new TextAnalyticsRequestOptions {
IncludeStatistics = true
});
#endregion
int i = 0;
Debug.WriteLine($"Results of Azure Text Analytics \"Named Entity Recognition\" Model, version: \"{results.ModelVersion}\"");
Debug.WriteLine("");
foreach (RecognizeEntitiesResult result in results)
{
TextDocumentInput document = documents[i++];
Debug.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\", Text=\"{document.Text}\"):");
if (result.HasError)
{
Debug.WriteLine($" Document error code: {result.Error.ErrorCode}.");
Debug.WriteLine($" Message: {result.Error.Message}.");
}
else
{
Debug.WriteLine($" Recognized the following {result.Entities.Count()} entities:");
foreach (CategorizedEntity entity in result.Entities)
{
Debug.WriteLine($" Text: {entity.Text}, Offset (in UTF-16 code units): {entity.Offset}, Length (in UTF-16 code units): {entity.Length}");
Debug.WriteLine($" Text: {entity.Text}, Category: {entity.Category}, SubCategory: {entity.SubCategory}, Confidence score: {entity.ConfidenceScore}");
}
Debug.WriteLine($" Document statistics:");
Debug.WriteLine($" Character count (in Unicode graphemes): {result.Statistics.CharacterCount}");
Debug.WriteLine($" Transaction count: {result.Statistics.TransactionCount}");
Debug.WriteLine("");
}
}
Debug.WriteLine($"Batch operation statistics:");
Debug.WriteLine($" Document count: {results.Statistics.DocumentCount}");
Debug.WriteLine($" Valid document count: {results.Statistics.ValidDocumentCount}");
Debug.WriteLine($" Invalid document count: {results.Statistics.InvalidDocumentCount}");
Debug.WriteLine($" Transaction count: {results.Statistics.TransactionCount}");
Debug.WriteLine("");
}
public async Task RecognizeEntitiesBatchConvenienceAsync()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
// Instantiate a client that will be used to call the service.
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
string documentA = @"We love this trail and make the trip every year. The views are breathtaking and well
worth the hike! Yesterday was foggy though, so we missed the spectacular views.
We tried again today and it was amazing. Everyone in my family liked the trail although
it was too challenging for the less athletic among us.
Not necessarily recommended for small children.
A hotel close to the trail offers services for childcare in case you want that.";
string documentB = @"Last week we stayed at Hotel Foo to celebrate our anniversary. The staff knew about
our anniversary so they helped me organize a little surprise for my partner.
The room was clean and with the decoration I requested. It was perfect!";
string documentC = @"That was the best day of my life! We went on a 4 day trip where we stayed at Hotel Foo.
They had great amenities that included an indoor pool, a spa, and a bar.
The spa offered couples massages which were really good.
The spa was clean and felt very peaceful. Overall the whole experience was great.
We will definitely come back.";
string documentD = string.Empty;
var documents = new List <string>
{
documentA,
documentB,
documentC,
documentD
};
Response <RecognizeEntitiesResultCollection> response = await client.RecognizeEntitiesBatchAsync(documents);
RecognizeEntitiesResultCollection entititesPerDocuments = response.Value;
int i = 0;
Console.WriteLine($"Results of Azure Text Analytics \"Named Entity Recognition\" Model, version: \"{entititesPerDocuments.ModelVersion}\"");
Console.WriteLine("");
foreach (RecognizeEntitiesResult entitiesInDocument in entititesPerDocuments)
{
Console.WriteLine($"On document with Text: \"{documents[i++]}\"");
Console.WriteLine("");
if (entitiesInDocument.HasError)
{
Console.WriteLine(" Error!");
Console.WriteLine($" Document error code: {entitiesInDocument.Error.ErrorCode}.");
Console.WriteLine($" Message: {entitiesInDocument.Error.Message}");
}
else
{
Console.WriteLine($" Recognized the following {entitiesInDocument.Entities.Count()} entities:");
foreach (CategorizedEntity entity in entitiesInDocument.Entities)
{
Console.WriteLine($" Text: {entity.Text}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" Length: {entity.Length}");
Console.WriteLine($" Category: {entity.Category}");
if (!string.IsNullOrEmpty(entity.SubCategory))
{
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine($" Confidence score: {entity.ConfidenceScore}");
Console.WriteLine("");
}
}
Console.WriteLine("");
}
}
public async Task RecognizeEntitiesBatchAsync()
{
string endpoint = Environment.GetEnvironmentVariable("TEXT_ANALYTICS_ENDPOINT");
string apiKey = Environment.GetEnvironmentVariable("TEXT_ANALYTICS_API_KEY");
// Instantiate a client that will be used to call the service.
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
var documents = new List <TextDocumentInput>
{
new TextDocumentInput("1", "Microsoft was founded by Bill Gates and Paul Allen.")
{
Language = "en",
},
new TextDocumentInput("2", "Text Analytics is one of the Azure Cognitive Services.")
{
Language = "en",
},
new TextDocumentInput("3", "A key technology in Text Analytics is Named Entity Recognition (NER).")
{
Language = "en",
}
};
RecognizeEntitiesResultCollection results = await client.RecognizeEntitiesBatchAsync(documents, new TextAnalyticsRequestOptions { IncludeStatistics = true });
int i = 0;
Console.WriteLine($"Results of Azure Text Analytics \"Named Entity Recognition\" Model, version: \"{results.ModelVersion}\"");
Console.WriteLine("");
foreach (RecognizeEntitiesResult result in results)
{
TextDocumentInput document = documents[i++];
Console.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\", Text=\"{document.Text}\"):");
if (result.HasError)
{
Console.WriteLine($" Document error code: {result.Error.Code}.");
Console.WriteLine($" Message: {result.Error.Message}.");
}
else
{
Console.WriteLine($" Recognized the following {result.Entities.Count()} entities:");
foreach (CategorizedEntity entity in result.Entities)
{
Console.WriteLine($" Text: {entity.Text}, Category: {entity.Category}, SubCategory: {entity.SubCategory}, Confidence score: {entity.ConfidenceScore}");
}
Console.WriteLine($" Document statistics:");
Console.WriteLine($" Character count (in Unicode graphemes): {result.Statistics.GraphemeCount}");
Console.WriteLine($" Transaction count: {result.Statistics.TransactionCount}");
Console.WriteLine("");
}
}
Console.WriteLine($"Batch operation statistics:");
Console.WriteLine($" Document count: {results.Statistics.DocumentCount}");
Console.WriteLine($" Valid document count: {results.Statistics.ValidDocumentCount}");
Console.WriteLine($" Invalid document count: {results.Statistics.InvalidDocumentCount}");
Console.WriteLine($" Transaction count: {results.Statistics.TransactionCount}");
Console.WriteLine("");
}
public async Task AnalyzeOperationAsync()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
#region Snippet:AnalyzeOperationBatchAsync
string documentA = @"We love this trail and make the trip every year. The views are breathtaking and well
worth the hike! Yesterday was foggy though, so we missed the spectacular views.
We tried again today and it was amazing. Everyone in my family liked the trail although
it was too challenging for the less athletic among us.
Not necessarily recommended for small children.
A hotel close to the trail offers services for childcare in case you want that.";
string documentB = @"Last week we stayed at Hotel Foo to celebrate our anniversary. The staff knew about
our anniversary so they helped me organize a little surprise for my partner.
The room was clean and with the decoration I requested. It was perfect!";
string documentC = @"That was the best day of my life! We went on a 4 day trip where we stayed at Hotel Foo.
They had great amenities that included an indoor pool, a spa, and a bar.
The spa offered couples massages which were really good.
The spa was clean and felt very peaceful. Overall the whole experience was great.
We will definitely come back.";
var batchDocuments = new List <TextDocumentInput>
{
new TextDocumentInput("1", documentA)
{
Language = "en",
},
new TextDocumentInput("2", documentB)
{
Language = "en",
},
new TextDocumentInput("3", documentC)
{
Language = "en",
}
};
TextAnalyticsActions batchActions = new TextAnalyticsActions()
{
ExtractKeyPhrasesOptions = new List <ExtractKeyPhrasesOptions>()
{
new ExtractKeyPhrasesOptions()
},
RecognizeEntitiesOptions = new List <RecognizeEntitiesOptions>()
{
new RecognizeEntitiesOptions()
},
RecognizePiiEntitiesOptions = new List <RecognizePiiEntitiesOptions>()
{
new RecognizePiiEntitiesOptions()
},
DisplayName = "AnalyzeOperationSample"
};
AnalyzeBatchActionsOperation operation = await client.StartAnalyzeBatchActionsAsync(batchDocuments, batchActions);
await operation.WaitForCompletionAsync();
await foreach (AnalyzeBatchActionsResult documentsInPage in operation.Value)
{
RecognizeEntitiesResultCollection entitiesResult = documentsInPage.RecognizeEntitiesActionsResults.FirstOrDefault().Result;
ExtractKeyPhrasesResultCollection keyPhrasesResult = documentsInPage.ExtractKeyPhrasesActionsResults.FirstOrDefault().Result;
RecognizePiiEntitiesResultCollection piiResult = documentsInPage.RecognizePiiEntitiesActionsResults.FirstOrDefault().Result;
Console.WriteLine("Recognized Entities");
foreach (RecognizeEntitiesResult result in entitiesResult)
{
Console.WriteLine($" Recognized the following {result.Entities.Count} entities:");
foreach (CategorizedEntity entity in result.Entities)
{
Console.WriteLine($" Entity: {entity.Text}");
Console.WriteLine($" Category: {entity.Category}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" ConfidenceScore: {entity.ConfidenceScore}");
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine("");
}
Console.WriteLine("Recognized PII Entities");
foreach (RecognizePiiEntitiesResult result in piiResult)
{
Console.WriteLine($" Recognized the following {result.Entities.Count} PII entities:");
foreach (PiiEntity entity in result.Entities)
{
Console.WriteLine($" Entity: {entity.Text}");
Console.WriteLine($" Category: {entity.Category}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" ConfidenceScore: {entity.ConfidenceScore}");
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine("");
}
Console.WriteLine("Key Phrases");
foreach (ExtractKeyPhrasesResult result in keyPhrasesResult)
{
Console.WriteLine($" Recognized the following {result.KeyPhrases.Count} Keyphrases:");
foreach (string keyphrase in result.KeyPhrases)
{
Console.WriteLine($" {keyphrase}");
}
Console.WriteLine("");
}
}
}
public void RecognizeEntitiesBatchAdvanced()
{
string endpoint = Environment.GetEnvironmentVariable("TEXT_ANALYTICS_ENDPOINT");
string subscriptionKey = Environment.GetEnvironmentVariable("TEXT_ANALYTICS_SUBSCRIPTION_KEY");
// Instantiate a client that will be used to call the service.
var client = new TextAnalyticsClient(new Uri(endpoint), subscriptionKey);
var inputs = new List <TextDocumentInput>
{
new TextDocumentInput("1", "Microsoft was founded by Bill Gates and Paul Allen.")
{
Language = "en",
},
new TextDocumentInput("2", "Text Analytics is one of the Azure Cognitive Services.")
{
Language = "en",
},
new TextDocumentInput("3", "A key technology in Text Analytics is Named Entity Recognition (NER).")
{
Language = "en",
}
};
RecognizeEntitiesResultCollection results = client.RecognizeEntities(inputs, new TextAnalyticsRequestOptions {
IncludeStatistics = true
});
int i = 0;
Debug.WriteLine($"Results of Azure Text Analytics \"Named Entity Recognition\" Model, version: \"{results.ModelVersion}\"");
Debug.WriteLine("");
foreach (var result in results)
{
var document = inputs[i++];
Debug.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\", Text=\"{document.Text}\"):");
if (result.ErrorMessage != default)
{
Debug.WriteLine($" Document error: {result.ErrorMessage}.");
}
else
{
Debug.WriteLine($" Recognized the following {result.NamedEntities.Count()} entities:");
foreach (var entity in result.NamedEntities)
{
Debug.WriteLine($" Text: {entity.Text}, Type: {entity.Type}, SubType: {entity.SubType ?? "N/A"}, Score: {entity.Score:0.00}, Offset: {entity.Offset}, Length: {entity.Length}");
}
Debug.WriteLine($" Document statistics:");
Debug.WriteLine($" Character count: {result.Statistics.CharacterCount}");
Debug.WriteLine($" Transaction count: {result.Statistics.TransactionCount}");
Debug.WriteLine("");
}
}
Debug.WriteLine($"Batch operation statistics:");
Debug.WriteLine($" Document count: {results.Statistics.DocumentCount}");
Debug.WriteLine($" Valid document count: {results.Statistics.ValidDocumentCount}");
Debug.WriteLine($" Invalid document count: {results.Statistics.InvalidDocumentCount}");
Debug.WriteLine($" Transaction count: {results.Statistics.TransactionCount}");
Debug.WriteLine("");
}
private void SerializeRecognizeEntitiesResultCollection(ref Utf8JsonWriter json, RecognizeEntitiesResultCollection resultCollection)
{
json.WriteStartObject();
json.WriteStartArray("documents");
if (resultCollection.FirstOrDefault(r => r.Entities.Count > 0) != default)
{
foreach (var result in resultCollection)
{
if (!result.HasError)
{
json.WriteStartObject();
json.WriteString("id", result.Id);
json.WriteStartArray("entities");
foreach (var entity in result.Entities)
{
json.WriteStartObject();
json.WriteString("text", entity.Text);
json.WriteString("type", JsonSerializer.Serialize(entity.Category));
json.WriteString("subtype", JsonSerializer.Serialize(entity.SubCategory));
json.WriteNumber("offset", entity.Offset);
json.WriteNumber("length", entity.Length);
json.WriteNumber("score", entity.Score);
json.WriteEndObject();
}
json.WriteEndArray();
json.WriteEndObject();
}
}
}
json.WriteEndArray();
json.WriteStartArray("errors");
if (resultCollection.FirstOrDefault(r => r.HasError) != default)
{
foreach (var result in resultCollection)
{
if (result.HasError)
{
json.WriteStartObject();
json.WriteString("id", result.Id);
json.WriteStartObject("error");
json.WriteStartObject("innerError");
json.WriteString("message", result.Error.Message);
json.WriteEndObject();
json.WriteEndObject();
json.WriteEndObject();
}
}
}
json.WriteEndArray();
json.WriteString("modelVersion", resultCollection.ModelVersion);
json.WriteEndObject();
// TODO: add statistics if needed
}
public void AnalyzeOperationBatchConvenience()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
#region Snippet:AnalyzeOperationBatchConvenience
string documentA = @"We love this trail and make the trip every year. The views are breathtaking and well
worth the hike! Yesterday was foggy though, so we missed the spectacular views.
We tried again today and it was amazing. Everyone in my family liked the trail although
it was too challenging for the less athletic among us.
Not necessarily recommended for small children.
A hotel close to the trail offers services for childcare in case you want that.";
string documentB = @"Last week we stayed at Hotel Foo to celebrate our anniversary. The staff knew about
our anniversary so they helped me organize a little surprise for my partner.
The room was clean and with the decoration I requested. It was perfect!";
string documentC = @"That was the best day of my life! We went on a 4 day trip where we stayed at Hotel Foo.
They had great amenities that included an indoor pool, a spa, and a bar.
The spa offered couples massages which were really good.
The spa was clean and felt very peaceful. Overall the whole experience was great.
We will definitely come back.";
var batchDocuments = new List <string>
{
documentA,
documentB,
documentC
};
TextAnalyticsActions actions = new TextAnalyticsActions()
{
ExtractKeyPhrasesOptions = new List <ExtractKeyPhrasesOptions>()
{
new ExtractKeyPhrasesOptions()
},
RecognizeEntitiesOptions = new List <RecognizeEntitiesOptions>()
{
new RecognizeEntitiesOptions()
},
RecognizePiiEntitiesOptions = new List <RecognizePiiEntitiesOptions>()
{
new RecognizePiiEntitiesOptions()
},
RecognizeLinkedEntitiesOptions = new List <RecognizeLinkedEntitiesOptions>()
{
new RecognizeLinkedEntitiesOptions()
},
DisplayName = "AnalyzeOperationSample"
};
AnalyzeBatchActionsOperation operation = client.StartAnalyzeBatchActions(batchDocuments, actions);
TimeSpan pollingInterval = new TimeSpan(1000);
while (!operation.HasCompleted)
{
Thread.Sleep(pollingInterval);
operation.UpdateStatus();
Console.WriteLine($"Status: {operation.Status}");
//If operation has not started, all other fields are null
if (operation.Status != TextAnalyticsOperationStatus.NotStarted)
{
Console.WriteLine($"Expires On: {operation.ExpiresOn}");
Console.WriteLine($"Last modified: {operation.LastModified}");
if (!string.IsNullOrEmpty(operation.DisplayName))
{
Console.WriteLine($"Display name: {operation.DisplayName}");
}
Console.WriteLine($"Total actions: {operation.TotalActions}");
Console.WriteLine($" Succeeded actions: {operation.ActionsSucceeded}");
Console.WriteLine($" Failed actions: {operation.ActionsFailed}");
Console.WriteLine($" In progress actions: {operation.ActionsInProgress}");
}
}
foreach (AnalyzeBatchActionsResult documentsInPage in operation.GetValues())
{
RecognizeEntitiesResultCollection entitiesResult = documentsInPage.RecognizeEntitiesActionsResults.FirstOrDefault().Result;
ExtractKeyPhrasesResultCollection keyPhrasesResult = documentsInPage.ExtractKeyPhrasesActionsResults.FirstOrDefault().Result;
RecognizePiiEntitiesResultCollection piiResult = documentsInPage.RecognizePiiEntitiesActionsResults.FirstOrDefault().Result;
RecognizeLinkedEntitiesResultCollection linkedEntitiesResult = documentsInPage.RecognizeLinkedEntitiesActionsResults.FirstOrDefault().Result;
Console.WriteLine("Recognized Entities");
foreach (RecognizeEntitiesResult result in entitiesResult)
{
Console.WriteLine($" Recognized the following {result.Entities.Count} entities:");
foreach (CategorizedEntity entity in result.Entities)
{
Console.WriteLine($" Entity: {entity.Text}");
Console.WriteLine($" Category: {entity.Category}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" Length: {entity.Length}");
Console.WriteLine($" ConfidenceScore: {entity.ConfidenceScore}");
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine("");
}
Console.WriteLine("Recognized PII Entities");
foreach (RecognizePiiEntitiesResult result in piiResult)
{
Console.WriteLine($" Recognized the following {result.Entities.Count} PII entities:");
foreach (PiiEntity entity in result.Entities)
{
Console.WriteLine($" Entity: {entity.Text}");
Console.WriteLine($" Category: {entity.Category}");
Console.WriteLine($" Offset: {entity.Offset}");
Console.WriteLine($" Length: {entity.Length}");
Console.WriteLine($" ConfidenceScore: {entity.ConfidenceScore}");
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine("");
}
Console.WriteLine("Key Phrases");
foreach (ExtractKeyPhrasesResult result in keyPhrasesResult)
{
Console.WriteLine($" Recognized the following {result.KeyPhrases.Count} Keyphrases:");
foreach (string keyphrase in result.KeyPhrases)
{
Console.WriteLine($" {keyphrase}");
}
Console.WriteLine("");
}
Console.WriteLine("Recognized Linked Entities");
foreach (RecognizeLinkedEntitiesResult result in linkedEntitiesResult)
{
Console.WriteLine($" Recognized the following {result.Entities.Count} linked entities:");
foreach (LinkedEntity entity in result.Entities)
{
Console.WriteLine($" Entity: {entity.Name}");
Console.WriteLine($" DataSource: {entity.DataSource}");
Console.WriteLine($" DataSource EntityId: {entity.DataSourceEntityId}");
Console.WriteLine($" Language: {entity.Language}");
Console.WriteLine($" DataSource Url: {entity.Url}");
Console.WriteLine($" Total Matches: {entity.Matches.Count()}");
foreach (LinkedEntityMatch match in entity.Matches)
{
Console.WriteLine($" Match Text: {match.Text}");
Console.WriteLine($" ConfidenceScore: {match.ConfidenceScore}");
Console.WriteLine($" Offset: {match.Offset}");
Console.WriteLine($" Length: {match.Length}");
}
Console.WriteLine("");
}
Console.WriteLine("");
}
}
}
public async Task AnalyzeOperationWithMultipleTasks()
{
TextAnalyticsClient client = GetClient();
var batchDocuments = new List <TextDocumentInput>
{
new TextDocumentInput("1", "Microsoft was founded by Bill Gates and Paul Allen.")
{
Language = "en",
},
new TextDocumentInput("2", "Mi perro y mi gato tienen que ir al veterinario.")
{
Language = "es",
}
};
AnalyzeOperationOptions operationOptions = new AnalyzeOperationOptions()
{
KeyPhrasesTaskParameters = new KeyPhrasesTaskParameters(),
EntitiesTaskParameters = new EntitiesTaskParameters(),
PiiTaskParameters = new PiiTaskParameters(),
DisplayName = "AnalyzeOperationWithMultipleTasks"
};
AnalyzeOperation operation = await client.StartAnalyzeOperationBatchAsync(batchDocuments, operationOptions);
await operation.WaitForCompletionAsync();
AnalyzeOperationResult resultCollection = operation.Value;
RecognizeEntitiesResultCollection entitiesResult = resultCollection.Tasks.EntityRecognitionTasks[0].Results;
ExtractKeyPhrasesResultCollection keyPhrasesResult = resultCollection.Tasks.KeyPhraseExtractionTasks[0].Results;
RecognizePiiEntitiesResultCollection piiResult = resultCollection.Tasks.EntityRecognitionPiiTasks[0].Results;
Assert.IsNotNull(keyPhrasesResult);
Assert.IsNotNull(entitiesResult);
Assert.IsNotNull(piiResult);
Assert.AreEqual("AnalyzeOperationWithMultipleTasks", resultCollection.DisplayName);
// Keyphrases
Assert.AreEqual(2, keyPhrasesResult.Count);
var keyPhrasesListId1 = new List <string> {
"Bill Gates", "Paul Allen", "Microsoft"
};
var keyPhrasesListId2 = new List <string> {
"gato", "perro", "veterinario"
};
foreach (string keyphrase in keyPhrasesResult[0].KeyPhrases)
{
Assert.IsTrue(keyPhrasesListId1.Contains(keyphrase));
}
foreach (string keyphrase in keyPhrasesResult[1].KeyPhrases)
{
Assert.IsTrue(keyPhrasesListId2.Contains(keyphrase));
}
// Entities
Assert.AreEqual(2, entitiesResult.Count);
Assert.AreEqual(3, entitiesResult[0].Entities.Count);
var entitiesList = new List <string> {
"Bill Gates", "Microsoft", "Paul Allen"
};
foreach (CategorizedEntity entity in entitiesResult[0].Entities)
{
Assert.IsTrue(entitiesList.Contains(entity.Text));
Assert.IsNotNull(entity.Category);
Assert.IsNotNull(entity.Offset);
Assert.IsNotNull(entity.ConfidenceScore);
}
// PII
Assert.AreEqual(2, entitiesResult.Count);
Assert.AreEqual(3, entitiesResult[0].Entities.Count);
Assert.IsNotNull(entitiesResult[0].Id);
Assert.IsNotNull(entitiesResult[0].Entities);
Assert.IsNotNull(entitiesResult[0].Error);
}
