This package contains a client library for the de-identification service in Azure Health Data Services which enables users to tag, redact, or surrogate health data containing Protected Health Information (PHI). For more on service functionality and important usage considerations, see the de-identification service overview.
This library support API versions 2024-11-15 and earlier.
Use the client library for the de-identification service to:
Source code | Package (PyPI) | API reference documentation | Product documentation | Samples
Getting started Prequisitespython -m pip install azure-health-deidentification
To authenticate with the de-identification service, install azure-identity:
python -m pip install azure.identity
You can use DefaultAzureCredential to automatically find the best credential to use at runtime.
You will need a service URL to instantiate a client object. You can find the service URL for a particular resource in the Azure portal, or using the Azure CLI:
# Get the service URL for the resource
az deidservice show --name "<resource-name>" --resource-group "<resource-group-name>" --query "properties.serviceUrl"
Optionally, save the service URL as an environment variable named AZURE_HEALTH_DEIDENTIFICATION_ENDPOINT for the sample client initialization code.
Create a client with the endpoint and credential:
endpoint = os.environ["AZURE_HEALTH_DEIDENTIFICATION_ENDPOINT"]
credential = DefaultAzureCredential()
client = DeidentificationClient(endpoint, credential)
Given an input text, the de-identification service can perform three main operations:
Tag returns the category and location within the text of detected PHI entities.Redact returns output text where detected PHI entities are replaced with placeholder text. For example John replaced with [name].Surrogate returns output text where detected PHI entities are replaced with realistic replacement values. For example, My name is John Smith could become My name is Tom Jones.There are two ways to interact with the de-identification service. You can send text directly, or you can create jobs to de-identify documents in Azure Storage.
You can de-identify text directly using the DeidentificationClient:
body = DeidentificationContent(input_text="Hello, my name is John Smith.")
result: DeidentificationResult = client.deidentify_text(body)
print(f'\nOriginal Text: "{body.input_text}"')
print(f'Surrogated Text: "{result.output_text}"') # Surrogated output: Hello, my name is <synthetic name>.
To de-identify documents in Azure Storage, see Tutorial: Configure Azure Storage to de-identify documents for prerequisites and configuration options.
To run the sample code below, populate the following environment variables:
AZURE_STORAGE_ACCOUNT_LOCATION: an Azure Storage container endpoint, like https://<storageaccount>.blob.core.windows.net/<container>.INPUT_PREFIX: the prefix of the input document name(s) in the container. For example, providing folder1 would create a job that would process documents like https://<storageaccount>.blob.core.windows.net/<container>/folder1/document1.txtThe client exposes a begin_deidentify_documents method that returns a LROPoller instance. You can get the result of the operation by calling result(), optionally passing in a timeout value in seconds:
endpoint = os.environ["AZURE_HEALTH_DEIDENTIFICATION_ENDPOINT"]
storage_location = os.environ["AZURE_STORAGE_ACCOUNT_LOCATION"]
inputPrefix = os.environ["INPUT_PREFIX"]
outputPrefix = "_output"
credential = DefaultAzureCredential()
client = DeidentificationClient(endpoint, credential)
jobname = f"sample-job-{uuid.uuid4().hex[:8]}"
job = DeidentificationJob(
source_location=SourceStorageLocation(
location=storage_location,
prefix=inputPrefix,
),
target_location=TargetStorageLocation(location=storage_location, prefix=outputPrefix, overwrite=True),
)
finished_job: DeidentificationJob = client.begin_deidentify_documents(jobname, job).result(timeout=60)
print(f"Job Name: {finished_job.job_name}")
print(f"Job Status: {finished_job.status}")
print(f"File Count: {finished_job.summary.total_count if finished_job.summary is not None else 0}")
The following sections provide code samples covering some of the most common client use cases, including:
See the samples for code files illustrating common patterns, including creating and managing jobs to de-identify documents in Azure Storage.
Discover PHI in unstructured textWhen you specify the TAG operation, the service will return information about the PHI entities it detects. You can use this information to customize your de-identification workflow:
body = DeidentificationContent(
input_text="Hello, I'm Dr. John Smith.", operation_type=DeidentificationOperationType.TAG
)
result: DeidentificationResult = client.deidentify_text(body)
print(f'\nOriginal Text: "{body.input_text}"')
if result.tagger_result and result.tagger_result.entities:
print(f"Tagged Entities:")
for entity in result.tagger_result.entities:
print(
f'\tEntity Text: "{entity.text}", Entity Category: "{entity.category}", Offset: "{entity.offset.code_point}", Length: "{entity.length.code_point}"'
)
else:
print("\tNo tagged entities found.")
When you specify the REDACT operation, the service will replace the PHI entities it detects with placeholder values. You can learn more about redaction customization.
body = DeidentificationContent(
input_text="It's great to work at Contoso.", operation_type=DeidentificationOperationType.REDACT
)
result: DeidentificationResult = client.deidentify_text(body)
print(f'\nOriginal Text: "{body.input_text}"')
print(f'Redacted Text: "{result.output_text}"') # Redacted output: "It's great to work at [organization]."
The default operation is the SURROGATE operation. Using this operation, the service will replace the PHI entities it detects with realistic surrogate values:
body = DeidentificationContent(input_text="Hello, my name is John Smith.")
result: DeidentificationResult = client.deidentify_text(body)
print(f'\nOriginal Text: "{body.input_text}"')
print(f'Surrogated Text: "{result.output_text}"') # Surrogated output: Hello, my name is <synthetic name>.
The DeidentificationClient raises various AzureError exceptions. For example, if you provide an invalid service URL, an ServiceRequestError would be raised with a message indicating the failure cause. In the following code snippet, the error is handled and displayed:
error_client = DeidentificationClient("https://contoso.deid.azure.com", credential)
body = DeidentificationContent(input_text="Hello, I'm Dr. John Smith.")
try:
error_client.deidentify_text(body)
except AzureError as e:
print("\nError: " + e.message)
If you encounter an error indicating that the service is unable to access source or target storage in a de-identification job:
Find a bug, or have feedback? Raise an issue with the Health Deidentification label.
TroubleshootingThis project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.microsoft.com.
When you submit a pull request, a CLA-bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., label, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.
This project has adopted the Microsoft Open Source Code of Conduct. For more information, see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4