你当前正在访问 Microsoft Azure Global Edition 技术文档网站。如果需要访问由世纪互联运营的 Microsoft Azure 中国技术文档网站，请访问 https://docs.azure.cn。

2025-09-08

Azure Conversation Authoring client library for Python - version 1.0.0b1

Conversation Authoring is part of the Conversational Language Understanding (CLU) service. It provides APIs and SDKs to create, manage, train, evaluate, and deploy conversation projects and models. With the ConversationAuthoringClient, you can script everything you’d otherwise do in Language Studio, including:

Project management: Create, update, export, and import projects.
Training jobs: Start, monitor, and cancel model training jobs.
Evaluation: Retrieve model evaluation summaries and per-utterance results.
Deployments: Create, update, swap, and delete deployments.
Snapshots: Load a trained model snapshot back into a project.
Resource assignment: Assign or unassign Azure resources to a deployment.

Getting started

Prerequisites

Python 3.7 or later is required to use this package.
An Azure subscription
A Language service resource

Install the package

Install the Azure Conversation Authoring client library for Python with pip:

pip install azure-ai-language-conversations-authoring

Note: This version of the client library defaults to the 2025-05-15-preview version of the service

Authenticate the client

To interact with the Conversation Authoring service, you'll need to create an instance of the ConversationAuthoringClient. You will need an endpoint and an API key to instantiate a client object. For more information regarding authenticating with Cognitive Services, see Authenticate requests to Azure Cognitive Services.

Get an API key

You can get the endpoint and API key from your Cognitive Services resource in the Azure Portal.

Alternatively, use the Azure CLI command shown below to get the API key from the Cognitive Service resource:

az cognitiveservices account keys list --resource-group <resource-group-name> --name <resource-name>

Create ConversationAuthoringClient

Once you've determined your endpoint and API key, you can instantiate a ConversationAuthoringClient:

from azure.core.credentials import AzureKeyCredential
from azure.ai.language.conversations.authoring import ConversationAuthoringClient

endpoint = "https://<my-custom-subdomain>.cognitiveservices.azure.com/"
credential = AzureKeyCredential("<api-key>")
client = ConversationAuthoringClient(endpoint, credential)

Create a client with an Azure Active Directory Credential

To use an Azure Active Directory (AAD) token credential,
provide an instance of the desired credential type obtained from the azure-identity library.

Note: Regional endpoints do not support AAD authentication.
You must create a custom subdomain for your resource in order to use this type of authentication.

Authentication with AAD requires some initial setup:

Install azure-identity
Register a new AAD application
Grant access to the Language service by assigning the Cognitive Services Language Owner role to your service principal

After setup, you can choose which type of credential to use.
As an example, DefaultAzureCredential can be used to authenticate the client:

Set the values of the client ID, tenant ID, and client secret of the AAD application as environment variables:
AZURE_CLIENT_ID, AZURE_TENANT_ID, AZURE_CLIENT_SECRET

Use the returned token credential to authenticate the client:

from azure.ai.language.conversations.authoring import ConversationAuthoringClient
from azure.identity import DefaultAzureCredential

credential = DefaultAzureCredential()
client = ConversationAuthoringClient(
    endpoint="https://<my-custom-subdomain>.cognitiveservices.azure.com/",
    credential=credential,
)

Key concepts

ConversationAuthoringClient

The ConversationAuthoringClient is the primary entry point for interacting with the Conversation Authoring service.
It provides top-level APIs to manage projects (create, delete, list) and retrieve project-scoped clients.

You can call get_project_client(project_name) to obtain a ConversationAuthoringProjectClient, which exposes operations specific to a given project.

For asynchronous operations, an async version of the client is available in the azure.ai.language.conversations.authoring.aio namespace.

ConversationAuthoringProjectClient

The ConversationAuthoringProjectClient is a project-scoped client returned by ConversationAuthoringClient.get_project_client(project_name).

It organizes project functionality into operation groups:

deployment → [DeploymentOperations]: manage deployments (create, update, delete, swap).
exported_model → [ExportedModelOperations]: handle exported model operations.
project → [ProjectOperations]: manage project-level operations (train, import/export, cancel jobs, assign resources).
trained_model → [TrainedModelOperations]: interact with trained models (evaluation, snapshots, delete models).

This separation ensures you can focus on project-level actions while still using the main ConversationAuthoringClient for higher-level management.

Examples

The azure-ai-language-conversations-authoring client library provides both synchronous and asynchronous APIs.

The following examples show common Conversation Authoring scenarios using the ConversationAuthoringClient or ConversationAuthoringProjectClient (created above).

Create a Conversation Project

# import libraries
import os
from azure.core.credentials import AzureKeyCredential
from azure.ai.language.conversations.authoring import ConversationAuthoringClient
from azure.ai.language.conversations.authoring.models import CreateProjectOptions, ProjectKind

# get secrets
endpoint = os.environ["AZURE_CONVERSATIONS_AUTHORING_ENDPOINT"]
key = os.environ["AZURE_CONVERSATIONS_AUTHORING_KEY"]

project_name = os.environ.get("PROJECT_NAME", "<project-name>")

# create client
client = ConversationAuthoringClient(endpoint, AzureKeyCredential(key))

# create project
body = CreateProjectOptions(
    project_kind=ProjectKind.CONVERSATION,
    project_name=project_name,
    language="<language-tag>",  # e.g. "en-us"
    multilingual=True,
    description="Sample project created via Python SDK",
)

# create project
result = client.create_project(project_name=project_name, body=body)

# print project details (direct attribute access; no getattr)
print("=== Create Project Result ===")
print(f"Project Name: {result.project_name}")
print(f"Language: {result.language}")
print(f"Kind: {result.project_kind}")
print(f"Multilingual: {result.multilingual}")
print(f"Description: {result.description}")

Import a Project

import os
from azure.core.credentials import AzureKeyCredential
from azure.ai.language.conversations.authoring import ConversationAuthoringClient
from azure.ai.language.conversations.authoring.models import (
    ExportedProject, CreateProjectOptions, ProjectKind, ProjectSettings
)

endpoint = os.environ["AZURE_CONVERSATIONS_AUTHORING_ENDPOINT"]
key = os.environ["AZURE_CONVERSATIONS_AUTHORING_KEY"]
project_name = os.environ.get("PROJECT_NAME", "<project-name>")

client = ConversationAuthoringClient(endpoint, AzureKeyCredential(key))
project_client = client.get_project_client(project_name)

# Build assets using objects
intents = [
    ConversationExportedIntent(category="<intent-a>"),
    ConversationExportedIntent(category="<intent-b>"),
]

entities = [
    ConversationExportedEntity(
        category="<entity-a>",
        composition_mode="combineComponents",
    )
]

u1 = ConversationExportedUtterance(
    text="<utterance-1>",
    intent="<intent-b>",
    language="<language-tag>",  # e.g., "en-us"
    dataset="Train",
    entities=[ExportedUtteranceEntityLabel(category="<entity-a>", offset=0, length=5)],
)

u2 = ConversationExportedUtterance(
    text="<utterance-2>",
    intent="<intent-b>",
    language="<language-tag>",
    dataset="Train",
    entities=[ExportedUtteranceEntityLabel(category="<entity-a>", offset=0, length=5)],
)

u3 = ConversationExportedUtterance(
    text="<utterance-3>",
    intent="<intent-b>",
    language="<language-tag>",
    dataset="Train",
    entities=[ExportedUtteranceEntityLabel(category="<entity-a>", offset=0, length=4)],
)

assets = ConversationExportedProjectAsset(
    intents=intents,
    entities=entities,
    utterances=[u1, u2, u3],
)

metadata = CreateProjectOptions(
    project_kind=ProjectKind.CONVERSATION,
    project_name=project_name,  # required
    language="<language-tag>",  # required (e.g., "en-us")
    settings=ProjectSettings(confidence_threshold=0.0),
    multilingual=False,
    description="",
)

exported_project = ExportedProject(
    project_file_version="<project-file-version>",  # e.g., "2025-05-15-preview"
    string_index_type="Utf16CodeUnit",
    metadata=metadata,
    assets=assets,
)

# begin import (long-running operation)
poller = project_client.project.begin_import(
    body=exported_project,
    exported_project_format=ExportedProjectFormat.CONVERSATION,
)

try:
    poller.result()
    print("Import completed.")
    print(f"done: {poller.done()}")
    print(f"status: {poller.status()}")
except HttpResponseError as e:
    msg = getattr(getattr(e, "error", None), "message", str(e))
    print(f"Operation failed: {msg}")

Train a Model

import os
from azure.core.credentials import AzureKeyCredential
from azure.ai.language.conversations.authoring import ConversationAuthoringClient
from azure.ai.language.conversations.authoring.models import (
    TrainingJobDetails, TrainingMode, EvaluationDetails, EvaluationKind
)

endpoint = os.environ["AZURE_CONVERSATIONS_AUTHORING_ENDPOINT"]
key = os.environ["AZURE_CONVERSATIONS_AUTHORING_KEY"]
project_name = os.environ.get("PROJECT_NAME", "<project-name>")

client = ConversationAuthoringClient(endpoint, AzureKeyCredential(key))
project_client = client.get_project_client(project_name)

# build training request
training_job_details = TrainingJobDetails(
    model_label="<model-label>",
    training_mode=TrainingMode.STANDARD,
    training_config_version="<config-version>",
    evaluation_options=EvaluationDetails(
        kind=EvaluationKind.PERCENTAGE,
        testing_split_percentage=20,
        training_split_percentage=80,
    ),
)

# start training job (long-running operation)
poller = project_client.project.begin_train(body=training_job_details)

# wait for job completion and get the result (no explicit type variables)
result = poller.result()

# print result details
print("=== Training Result ===")
print(f"Model Label: {result.model_label}")
print(f"Training Config Version: {result.training_config_version}")
print(f"Training Mode: {result.training_mode}")
print(f"Training Status: {result.training_status}")
print(f"Data Generation Status: {result.data_generation_status}")
print(f"Evaluation Status: {result.evaluation_status}")
print(f"Estimated End: {result.estimated_end_on}")

Optional Configuration

Optional keyword arguments can be passed in at the client and per-operation level. The azure-core reference documentation describes available configurations for retries, logging, transport protocols, and more.

Troubleshooting

General

The Conversation Authoring client will raise exceptions defined in Azure Core.
These exceptions provide consistent error handling across Azure SDK libraries.

Logging

This library uses Python’s built-in logging module for diagnostic logging.

Basic information about HTTP requests (URLs, headers) is logged at INFO level.
Detailed request/response information (including bodies and unredacted headers) is logged at DEBUG level.

You can enable logging when constructing the client by passing logging_enable=True.

import sys
import logging
from azure.core.credentials import AzureKeyCredential
from azure.ai.language.conversations.authoring import ConversationAuthoringClient

# Configure logger
logger = logging.getLogger("azure")
logger.setLevel(logging.DEBUG)
handler = logging.StreamHandler(stream=sys.stdout)
logger.addHandler(handler)

endpoint = "https://<endpoint>.cognitiveservices.azure.com/"
credential = AzureKeyCredential("<api-key>")

# This client will log detailed HTTP information
client = ConversationAuthoringClient(endpoint, credential, logging_enable=True)

Similarly, logging_enable can enable detailed logging for a single operation, even when it isn't enabled for the client:

result = client.create_project(
    project_name="<project-name>",
    body={...},
    logging_enable=True,
)

Next steps

More sample code

See the Sample README for additional examples that demonstrate common Conversation Authoring workflows such as:

Creating and managing projects
Importing and exporting project assets
Training models and retrieving evaluation results
Deploying and swapping models
Assigning or unassigning resources
Loading snapshots and managing trained models

Contributing

See the CONTRIBUTING.md guide for details on building, testing, and contributing to this library.

This 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 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 repositories 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.

反馈

此页面是否有帮助？