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 RecognizeEntitiesBatchConvenienceTest()
{
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);
foreach (RecognizeEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 1);
}
}
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 <List <RecognizeEntitiesResult> > PredictNerBatchAsync(List <string> queries)
{
// verify text analytics char limit
VerifyCharLimit(queries);
// paginate throw chunks to abide by "DocumentCountLimit" of the api call
var listPaginator = new Paginator <string>(queries, Constants.TextAnaylticsApiCallDocumentLimit);
var result = new List <RecognizeEntitiesResult>();
while (listPaginator.HasNext())
{
var subList = (listPaginator.GetNextPage()).ToList();
var response = await _textAnalyticsClient.RecognizeEntitiesBatchAsync(subList, language : _predictionLanguage);
HandleError(response.Value);
result.AddRange(response.Value);
}
return(result);
}
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 void RecognizeEntitiesBatchWithInvalidDocumentBatch()
{
TextAnalyticsClient client = GetClient();
var documents = new List <string>
{
"document 1",
"document 2",
"document 3",
"document 4",
"document 5",
"document 6"
};
RequestFailedException ex = Assert.ThrowsAsync <RequestFailedException>(
async() => await client.RecognizeEntitiesBatchAsync(documents));
Assert.AreEqual(400, ex.Status);
Assert.AreEqual(TextAnalyticsErrorCode.InvalidDocumentBatch, ex.ErrorCode);
}
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 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 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);
}
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 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 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 RecognizeEntitiesBatchAsync()
{
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), CreateSampleOptions());
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 = await client.RecognizeEntitiesBatchAsync(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($" Offset: {entity.Offset}");
Console.WriteLine($" Length: {entity.Length}");
Console.WriteLine($" Category: {entity.Category}");
if (!string.IsNullOrEmpty(entity.SubCategory))
{
Console.WriteLine($" SubCategory: {entity.SubCategory}");
}
Console.WriteLine($" Confidence score: {entity.ConfidenceScore}");
Console.WriteLine("");
}
Console.WriteLine($" 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($" Transaction count: {entitiesInDocuments.Statistics.TransactionCount}");
Console.WriteLine("");
}
public async Task RecognizeEntitiesBatchConvenienceAsync()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
// Instantiate a client that will be used to call the service.
var client = new TextAnalyticsClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
string documentA = @"We love this trail and make the trip every year. The views are breathtaking and well
worth the hike! Yesterday was foggy though, so we missed the spectacular views.
We tried again today and it was amazing. Everyone in my family liked the trail although
it was too challenging for the less athletic among us.
Not necessarily recommended for small children.
A hotel close to the trail offers services for childcare in case you want that.";
string documentB = @"Last week we stayed at Hotel Foo to celebrate our anniversary. The staff knew about
our anniversary so they helped me organize a little surprise for my partner.
The room was clean and with the decoration I requested. It was perfect!";
string documentC = @"That was the best day of my life! We went on a 4 day trip where we stayed at Hotel Foo.
They had great amenities that included an indoor pool, a spa, and a bar.
The spa offered couples massages which were really good.
The spa was clean and felt very peaceful. Overall the whole experience was great.
We will definitely come back.";
string documentD = string.Empty;
var documents = new List <string>
{
documentA,
documentB,
documentC,
documentD
};
Response <RecognizeEntitiesResultCollection> response = await client.RecognizeEntitiesBatchAsync(documents);
RecognizeEntitiesResultCollection entititesPerDocuments = response.Value;
int i = 0;
Console.WriteLine($"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 void RecognizeEntitiesBatchWithNullIdTest()
{
TextAnalyticsClient client = GetClient();
var documents = new List <TextDocumentInput> {
new TextDocumentInput(null, "Hello world")
};
RequestFailedException ex = Assert.ThrowsAsync <RequestFailedException>(async() => await client.RecognizeEntitiesBatchAsync(documents));
Assert.AreEqual(TextAnalyticsErrorCode.InvalidDocument, ex.ErrorCode);
}
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("");
}
