public async Task RecognizePiiEntitiesBatchWithStatisticsTest()
{
TextAnalyticsClient client = GetClient();
var inputs = new List <TextDocumentInput>
{
new TextDocumentInput("1", "A developer with SSN 555-55-5555 whose phone number is 555-555-5555 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",
}
};
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(inputs, new TextAnalyticsRequestOptions { IncludeStatistics = true });
foreach (RecognizePiiEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 2);
}
Assert.IsNotNull(results.Statistics.DocumentCount);
Assert.IsNotNull(results.Statistics.InvalidDocumentCount);
Assert.IsNotNull(results.Statistics.TransactionCount);
Assert.IsNotNull(results.Statistics.ValidDocumentCount);
}
public async Task RecognizePiiEntitiesBatchTest()
{
TextAnalyticsClient client = GetClient();
List <TextDocumentInput> documents = batchDocuments;
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(documents);
foreach (RecognizePiiEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 2);
}
}
public async Task RecognizePiiEntitiesBatchWithStatisticsTest()
{
TextAnalyticsClient client = GetClient();
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(s_batchDocuments, new RecognizePiiEntitiesOptions { IncludeStatistics = true });
var expectedOutput = new Dictionary <string, List <string> >()
{
{ "1", s_document1ExpectedOutput },
{ "2", s_document2ExpectedOutput },
};
ValidateBatchDocumentsResult(results, expectedOutput, includeStatistics: true);
}
public async Task RecognizePiiEntitiesBatchTest()
{
TextAnalyticsClient client = GetClient();
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(s_batchDocuments);
var expectedOutput = new Dictionary <string, List <string> >()
{
{ "1", s_document1ExpectedOutput },
{ "2", s_document2ExpectedOutput },
};
ValidateBatchDocumentsResult(results, expectedOutput);
}
public async Task RecognizePiiEntitiesBatchConvenienceTest()
{
TextAnalyticsClient client = GetClient();
var inputs = new List <string>
{
"A developer with SSN 555-55-5555 whose phone number is 555-555-5555 is building tools with our APIs.",
"Your ABA number - 111000025 - is the first 9 digits in the lower left hand corner of your personal check."
};
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(inputs);
foreach (RecognizePiiEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 2);
}
}
public async Task RecognizePiiEntitiesBatchWithStatisticsTest()
{
TextAnalyticsClient client = GetClient();
List <TextDocumentInput> documents = batchDocuments;
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(documents, new RecognizePiiEntitiesOptions { IncludeStatistics = true });
foreach (RecognizePiiEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 2);
}
Assert.Greater(results.Statistics.DocumentCount, 0);
Assert.AreEqual(0, results.Statistics.InvalidDocumentCount);
Assert.Greater(results.Statistics.TransactionCount, 0);
Assert.Greater(results.Statistics.ValidDocumentCount, 0);
}
public async Task RecognizePiiEntitiesBatchWithDomainTest()
{
TextAnalyticsClient client = GetClient();
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(s_batchDocuments, new RecognizePiiEntitiesOptions()
{
DomainFilter = PiiEntityDomainType.ProtectedHealthInformation
});
var expectedOutput = new Dictionary <string, List <string> >()
{
{ "1", s_document1ExpectedOutput },
{ "2", s_document2ExpectedOutput },
};
ValidateBatchDocumentsResult(results, expectedOutput);
}
public async Task RecognizePiiEntitiesBatchWithErrorTest()
{
TextAnalyticsClient client = GetClient();
var inputs = new List <string>
{
"A developer with SSN 555-55-5555 whose phone number is 555-555-5555 is building tools with our APIs.",
"",
"Your ABA number - 111000025 - is the first 9 digits in the lower left hand corner of your personal check.",
};
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(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 RecognizePiiEntitiesBatchWithErrorTest()
{
TextAnalyticsClient client = GetClient();
var documents = new List <string>
{
"A developer with SSN 859-98-0987 whose phone number is 800-102-1100 is building tools with our APIs.",
"",
"Your ABA number - 111000025 - is the first 9 digits in the lower left hand corner of your personal check.",
};
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(documents);
Assert.IsFalse(results[0].HasError);
Assert.IsFalse(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 RecognizePiiEntitiesBatchTest()
{
TextAnalyticsClient client = GetClient();
var documents = new List <TextDocumentInput>
{
new TextDocumentInput("1", "A developer with SSN 555-55-5555 whose phone number is 555-555-5555 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",
}
};
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(documents);
foreach (RecognizePiiEntitiesResult result in results)
{
Assert.GreaterOrEqual(result.Entities.Count(), 2);
}
}
public async Task RecognizePiiEntitiesBatchWitCategoryTest()
{
TextAnalyticsClient client = GetClient();
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(s_batchDocuments, new RecognizePiiEntitiesOptions()
{
CategoriesFilter = { PiiEntityCategory.PhoneNumber }
});
var expectedOutput = new Dictionary <string, List <string> >()
{
{ "1", new List <string>()
{
"800-102-1100"
} },
{ "2", new List <string>()
{
"111000025"
} },
};
ValidateBatchDocumentsResult(results, expectedOutput);
}
public async Task RecognizePiiEntitiesBatchConvenienceAsync()
{
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));
var documents = new List <string>
{
"A developer with SSN 859-98-0987 whose phone number is 800-102-1100 is building tools with our APIs.",
"Your ABA number - 111000025 - is the first 9 digits in the lower left hand corner of your personal check."
};
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(documents);
int i = 0;
foreach (RecognizePiiEntitiesResult result in results)
{
Console.WriteLine($"For document: {documents[i++]}");
if (result.Entities.Count > 0)
{
Console.WriteLine($"Redacted Text: {result.Entities.RedactedText}");
Console.WriteLine($"The following {result.Entities.Count} PII entit{(result.Entities.Count > 1 ? "ies were" : "y was")} found:");
foreach (PiiEntity entity in result.Entities)
{
Console.WriteLine($" Text: {entity.Text}, Category: {entity.Category}, SubCategory: {entity.SubCategory}, Confidence score: {entity.ConfidenceScore}");
}
}
else
{
Console.WriteLine("No entities were found.");
}
}
}
public async Task RecognizePiiEntitiesBatchConvenienceAsync()
{
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 = @"Parker Doe has repaid all of their loans as of 2020-04-25.
Their SSN is 859-98-0987. To contact them, use their phone number 800-102-1100.
They are originally from Brazil and have document ID number 998.214.865-68";
string documentB = @"Yesterday, Dan Doe was asking where they could find the ABA number. I explained
that it is the first 9 digits in the lower left hand corner of their personal check.
After looking at their account they confirmed the number was 111000025";
string documentC = string.Empty;
var documents = new List <string>
{
documentA,
documentB,
documentC
};
Response <RecognizePiiEntitiesResultCollection> response = await client.RecognizePiiEntitiesBatchAsync(documents);
RecognizePiiEntitiesResultCollection entititesPerDocuments = response.Value;
int i = 0;
Console.WriteLine($"Results of \"PII Entity Recognition\" Model, version: \"{entititesPerDocuments.ModelVersion}\"");
Console.WriteLine("");
foreach (RecognizePiiEntitiesResult piiEntititesInDocument in entititesPerDocuments)
{
Console.WriteLine($"On document with Text: \"{documents[i++]}\"");
Console.WriteLine("");
if (piiEntititesInDocument.HasError)
{
Console.WriteLine(" Error!");
Console.WriteLine($" Document error code: {piiEntititesInDocument.Error.ErrorCode}.");
Console.WriteLine($" Message: {piiEntititesInDocument.Error.Message}");
}
else
{
Console.WriteLine($" Redacted Text: {piiEntititesInDocument.Entities.RedactedText}");
Console.WriteLine("");
Console.WriteLine($" Recognized {piiEntititesInDocument.Entities.Count} PII entities:");
foreach (PiiEntity piiEntity in piiEntititesInDocument.Entities)
{
Console.WriteLine($" Text: {piiEntity.Text}");
Console.WriteLine($" Category: {piiEntity.Category}");
if (!string.IsNullOrEmpty(piiEntity.SubCategory))
{
Console.WriteLine($" SubCategory: {piiEntity.SubCategory}");
}
Console.WriteLine($" Confidence score: {piiEntity.ConfidenceScore}");
Console.WriteLine("");
}
}
Console.WriteLine("");
}
}
public async Task RecognizePiiEntitiesBatchAsync()
{
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 = @"Parker Doe has repaid all of their loans as of 2020-04-25.
Their SSN is 859-98-0987. To contact them, use their phone number 800-102-1100.
They are originally from Brazil and have document ID number 998.214.865-68";
string documentB = @"Hoy recibí una llamada al medio día del usuario Juanito Perez, quien preguntaba
cómo acceder a su nuevo correo electrónico. Este trabaja en Microsoft y su correo es
[email protected]. El usuario accedió a compartir su número para futuras comunicaciones.
El número es 800-102-1101";
string documentC = @"Yesterday, Dan Doe was asking where they could find the ABA number. I explained
that it is the first 9 digits in the lower left hand corner of their personal check.
After looking at their account they confirmed the number was 111000025";
var documents = new List <TextDocumentInput>
{
new TextDocumentInput("1", documentA)
{
Language = "en",
},
new TextDocumentInput("2", documentB)
{
Language = "es",
},
new TextDocumentInput("3", documentC)
{
Language = "en",
}
};
var options = new RecognizePiiEntitiesOptions {
IncludeStatistics = true
};
Response <RecognizePiiEntitiesResultCollection> response = await client.RecognizePiiEntitiesBatchAsync(documents, options);
RecognizePiiEntitiesResultCollection entititesPerDocuments = response.Value;
int i = 0;
Console.WriteLine($"Results of \"PII Entity Recognition\" Model, version: \"{entititesPerDocuments.ModelVersion}\"");
Console.WriteLine("");
foreach (RecognizePiiEntitiesResult piiEntititesInDocument in entititesPerDocuments)
{
TextDocumentInput document = documents[i++];
Console.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\"):");
if (piiEntititesInDocument.HasError)
{
Console.WriteLine(" Error!");
Console.WriteLine($" Document error code: {piiEntititesInDocument.Error.ErrorCode}.");
Console.WriteLine($" Message: {piiEntititesInDocument.Error.Message}");
}
else
{
Console.WriteLine($" Redacted Text: {piiEntititesInDocument.Entities.RedactedText}");
Console.WriteLine("");
Console.WriteLine($" Recognized {piiEntititesInDocument.Entities.Count} PII entities:");
foreach (PiiEntity piiEntity in piiEntititesInDocument.Entities)
{
Console.WriteLine($" Text: {piiEntity.Text}");
Console.WriteLine($" Category: {piiEntity.Category}");
if (!string.IsNullOrEmpty(piiEntity.SubCategory))
{
Console.WriteLine($" SubCategory: {piiEntity.SubCategory}");
}
Console.WriteLine($" Confidence score: {piiEntity.ConfidenceScore}");
Console.WriteLine("");
}
Console.WriteLine($" Document statistics:");
Console.WriteLine($" Character count: {piiEntititesInDocument.Statistics.CharacterCount}");
Console.WriteLine($" Transaction count: {piiEntititesInDocument.Statistics.TransactionCount}");
}
Console.WriteLine("");
}
Console.WriteLine($"Batch operation statistics:");
Console.WriteLine($" Document count: {entititesPerDocuments.Statistics.DocumentCount}");
Console.WriteLine($" Valid document count: {entititesPerDocuments.Statistics.ValidDocumentCount}");
Console.WriteLine($" Invalid document count: {entititesPerDocuments.Statistics.InvalidDocumentCount}");
Console.WriteLine($" Transaction count: {entititesPerDocuments.Statistics.TransactionCount}");
Console.WriteLine("");
}
public async Task RecognizePiiEntitiesBatchAsync()
{
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));
var documents = new List <TextDocumentInput>
{
new TextDocumentInput("1", "A developer with SSN 859-98-0987 whose phone number is 800-102-1100 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",
}
};
RecognizePiiEntitiesResultCollection results = await client.RecognizePiiEntitiesBatchAsync(documents, new RecognizePiiEntitiesOptions { IncludeStatistics = true });
int i = 0;
Console.WriteLine($"Results of Azure Text Analytics \"Pii Entity Recognition\" Model, version: \"{results.ModelVersion}\"");
Console.WriteLine("");
foreach (RecognizePiiEntitiesResult result in results)
{
TextDocumentInput document = documents[i++];
Console.WriteLine($"On document (Id={document.Id}, Language=\"{document.Language}\", Text=\"{document.Text}\"):");
if (result.HasError)
{
Console.WriteLine($" Document error code: {result.Error.ErrorCode}.");
Console.WriteLine($" Message: {result.Error.Message}.");
}
else
{
if (result.Entities.Count > 0)
{
Console.WriteLine($" Redacted Text: {result.Entities.RedactedText}");
Console.WriteLine($" Recognized the following {result.Entities.Count} PII entit{(result.Entities.Count > 1 ? "ies" : "y ")}:");
foreach (PiiEntity entity in result.Entities)
{
Console.WriteLine($" Text: {entity.Text}, Category: {entity.Category}, SubCategory: {entity.SubCategory}, Confidence score: {entity.ConfidenceScore}");
}
}
else
{
Console.WriteLine("No entities were found.");
}
Console.WriteLine($" Document statistics:");
Console.WriteLine($" Character count (in Unicode graphemes): {result.Statistics.CharacterCount}");
Console.WriteLine($" 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("");
}
