azure-ai-translation-document-py
用 Azure Document Translation SDK 批次翻譯文件並保留格式
中文翻譯
---
name: azure-ai-translation-document-py
description: |
Azure AI Document Translation SDK，用來批次翻譯文件並保留格式。適合大量翻譯 Word、PDF、Excel、PowerPoint 等文件格式。
Triggers: "document translation", "batch translation", "translate documents", "DocumentTranslationClient".
license: MIT
metadata:
author: Microsoft
version: "1.0.0"
package: azure-ai-translation-document
---
# Azure AI Document Translation SDK for Python
Azure AI Translator 文件翻譯服務的 client library，用來批次翻譯文件並保留格式。
## Installation
```bash
pip install azure-ai-translation-document
```
## Environment Variables
```bash
AZURE_DOCUMENT_TRANSLATION_ENDPOINT=https://<resource>.cognitiveservices.azure.com # 所有驗證方式都需要
# 來源與目標文件的儲存空間
AZURE_SOURCE_CONTAINER_URL=https://<storage>.blob.core.windows.net/<container>?<sas> # 所有驗證方式都需要
AZURE_TARGET_CONTAINER_URL=https://<storage>.blob.core.windows.net/<container>?<sas> # 所有驗證方式都需要
AZURE_TOKEN_CREDENTIALS=prod # 只有在 production 使用 DefaultAzureCredential 時才需要
AZURE_DOCUMENT_TRANSLATION_KEY=<your-api-key> # 只有下面舊版 API key 驗證路徑才需要
```
## Authentication & Lifecycle
> **🔑 下面每個程式範例都適用兩條規則：**
>
> 1. **優先用 `DefaultAzureCredential`。** 它在本機（Azure CLI / VS Code / Developer CLI）和 Azure 上（managed identity、workload identity）都能跑，程式碼不用改。避免用 connection string、account/API key，因為它們會繞過 Entra 的稽核和金鑰輪替。
> - 本機開發：`DefaultAzureCredential` 直接可用。
> - Production：設定 `AZURE_TOKEN_CREDENTIALS=prod`（或 `AZURE_TOKEN_CREDENTIALS=<specific_credential>`），把 credential chain 限縮在 production 安全的 credential。
> 2. **每個 client 都要包在 context manager 裡**，這樣 HTTP transport、socket 和 token cache 會確實被釋放：
> - 同步：`with <Client>(...) as client:`
> - 非同步：`async with <Client>(...) as client:` **而且** `async with DefaultAzureCredential() as credential:`（來自 `azure.identity.aio`）
>
> 範例程式碼可能會省略這部分設定，但 production 程式碼一定要兩條規則都遵守。
```python
import os
from azure.identity import DefaultAzureCredential, ManagedIdentityCredential
from azure.ai.translation.document import DocumentTranslationClient
# 本機開發：DefaultAzureCredential。Production：設定 AZURE_TOKEN_CREDENTIALS=prod 或 AZURE_TOKEN_CREDENTIALS=<specific_credential>
credential = DefaultAzureCredential()
# 或在 production 直接用指定的 credential：
# 參考 https://learn.microsoft.com/python/api/overview/azure/identity-readme?view=azure-python#credential-classes
# credential = ManagedIdentityCredential()
with DocumentTranslationClient(
endpoint=os.environ["AZURE_DOCUMENT_TRANSLATION_ENDPOINT"],
credential=credential,
) as client:
statuses = list(client.list_translation_statuses())
```
### Legacy: API Key (existing keyed deployments)
新程式碼請用上面的 `DefaultAzureCredential`。只有在既有的 key 驗證部署還沒遷移到 Entra ID 時才用 `AzureKeyCredential`，例如還在進行 Entra 上線的受管制環境。
```python
import os
from azure.core.credentials import AzureKeyCredential
from azure.ai.translation.document import DocumentTranslationClient, SingleDocumentTranslationClient
with DocumentTranslationClient(
endpoint=os.environ["AZURE_DOCUMENT_TRANSLATION_ENDPOINT"],
credential=AzureKeyCredential(os.environ["AZURE_DOCUMENT_TRANSLATION_KEY"]),
) as client:
statuses = list(client.list_translation_statuses())
# SingleDocumentTranslationClient 也接受同樣的 key 型 credential。
```
## Basic Document Translation
```python
import os
from azure.ai.translation.document import DocumentTranslationClient, DocumentTranslationInput, TranslationTarget
from azure.core.exceptions import HttpResponseError
from azure.identity import DefaultAzureCredential
credential = DefaultAzureCredential()
with DocumentTranslationClient(
endpoint=os.environ["AZURE_DOCUMENT_TRANSLATION_ENDPOINT"],
credential=credential,
) as client:
source_url = os.environ["AZURE_SOURCE_CONTAINER_URL"]
target_url = os.environ["AZURE_TARGET_CONTAINER_URL"]
try:
# 啟動翻譯工作
poller = client.begin_translation(
inputs=[
DocumentTranslationInput(
source_url=source_url,
targets=[
TranslationTarget(
target_url=target_url,
language="es" # 翻成西班牙文
)
]
)
]
)
# 等待完成
result = poller.result()
print(f"Status: {poller.status()}")
print(f"Documents translated: {poller.details.documents_succeeded_count}")
print(f"Documents failed: {poller.details.documents_failed_count}")
except HttpResponseError as e:
print(f"Translation failed: {e.message}")
raise
```
## Multiple Target Languages
```python
poller = client.begin_translation(
inputs=[
DocumentTranslationInput(
source_url=source_url,
targets=[
TranslationTarget(target_url=target_url_es, language="es"),
TranslationTarget(target_url=target_url_fr, language="fr"),
TranslationTarget(target_url=target_url_de, language="de")
]
)
]
)
```
## Translate Single Document
```python
from azure.ai.translation.document import SingleDocumentTranslationClient
from azure.identity import DefaultAzureCredential
with open("document.docx", "rb") as f:
document_content = f.read()
with SingleDocumentTranslationClient(endpoint, DefaultAzureCredential()) as single_client:
result = single_client.translate(
body=document_content,
target_language="es",
content_type="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
)
# 儲存翻譯後的文件
with open("document_es.docx", "wb") as f:
f.write(result)
```
## Check Translation Status
```python
# 取得所有翻譯作業
operations = client.list_translation_statuses()
for op in operations:
print(f"Operation ID: {op.id}")
print(f"Status: {op.status}")
print(f"Created: {op.created_on}")
print(f"Total documents: {op.documents_total_count}")
print(f"Succeeded: {op.documents_succeeded_count}")
print(f"Failed: {op.documents_failed_count}")
```
## List Document Statuses
```python
# 取得工作中每份文件的狀態
operation_id = poller.id
document_statuses = client.list_document_statuses(operation_id)
for doc in document_statuses:
print(f"Document: {doc.source_document_url}")
print(f" Status: {doc.status}")
print(f" Translated to: {doc.translated_to}")
if doc.error:
print(f" Error: {doc.error.message}")
```
## Cancel Translation
```python
# 取消執行中的翻譯
client.cancel_translation(operation_id)
```
## Using Glossary
```python
from azure.ai.translation.document import TranslationGlossary
poller = client.begin_translation(
inputs=[
DocumentTranslationInput(
source_url=source_url,
targets=[
TranslationTarget(
target_url=target_url,
language="es",
glossaries=[
TranslationGlossary(
glossary_url="https://<storage>.blob.core.windows.net/glossary/terms.csv?<sas>",
file_format="csv"
)
]
)
]
)
]
)
```
## Supported Document Formats
```python
# 取得支援的格式
formats = client.get_supported_document_formats()
for fmt in formats:
print(f"Format: {fmt.format}")
print(f" Extensions: {fmt.file_extensions}")
print(f" Content types: {fmt.content_types}")
```
## Supported Languages
```python
# 取得支援的語言
languages = client.get_supported_languages()
for lang in languages:
print(f"Language: {lang.name} ({lang.code})")
```
## Async Client
```python
from azure.ai.translation.document.aio import DocumentTranslationClient
from azure.identity.aio import DefaultAzureCredential
async def translate_documents():
async with DefaultAzureCredential() as credential:
async with DocumentTranslationClient(
endpoint=endpoint,
credential=credential,
) as client:
poller = await client.begin_translation(inputs=[...])
result = await poller.result()
```
## Supported Formats
| Category | Formats |
|----------|---------|
| Documents | DOCX, PDF, PPTX, XLSX, HTML, TXT, RTF |
| Structured | CSV, TSV, JSON, XML |
| Localization | XLIFF, XLF, MHTML |
## Storage Requirements
- 來源和目標 container 必須是 Azure Blob Storage
- 使用權限剛好夠用的 SAS token：
- 來源：Read, List
- 目標：Write, List
## Best Practices
1. **同步或非同步擇一，並保持一致。** 同一條呼叫路徑裡不要混用 `azure.xxx` 同步 client 和 `azure.xxx.aio` 非同步 client。每個 module 只選一種模式。
2. **client 和非同步 credential 一律用 context manager。** 每個 client 都包在 `with Client(...) as client:`（同步）或 `async with Client(...) as client:`（非同步）裡。非同步的 `DefaultAzureCredential`（來自 `azure.identity.aio`）也要用 `async with credential:`，這樣 token 和 transport 才會被清掉。
3. **使用 SAS token**，權限給最小必要範圍
4. **用 `poller.status()` 監控**長時間執行的作業
5. **逐一檢查文件狀態**，處理文件層級的錯誤
6. **使用 glossary** 處理特定領域的術語
7. **每種語言用獨立的目標 container**
8. **多個並行工作時用非同步 client**
9. **送出文件前先確認支援的格式**
## Reference Files
| File | Contents |
|------|----------|
| [references/capabilities.md](references/capabilities.md) | 其他非主打功能、operation group 涵蓋範圍，以及 production 檢查清單。 |
| [references/non-hero-scenarios.md](references/non-hero-scenarios.md) | 次要與進階情境的專屬非主打範例。 |英文原文
---
name: azure-ai-translation-document-py
description: |
Azure AI Document Translation SDK for batch translation of documents with format preservation. Use for translating Word, PDF, Excel, PowerPoint, and other document formats at scale.
Triggers: "document translation", "batch translation", "translate documents", "DocumentTranslationClient".
license: MIT
metadata:
author: Microsoft
version: "1.0.0"
package: azure-ai-translation-document
---
# Azure AI Document Translation SDK for Python
Client library for Azure AI Translator document translation service for batch document translation with format preservation.
## Installation
```bash
pip install azure-ai-translation-document
```
## Environment Variables
```bash
AZURE_DOCUMENT_TRANSLATION_ENDPOINT=https://<resource>.cognitiveservices.azure.com # Required for all auth methods
# Storage for source and target documents
AZURE_SOURCE_CONTAINER_URL=https://<storage>.blob.core.windows.net/<container>?<sas> # Required for all auth methods
AZURE_TARGET_CONTAINER_URL=https://<storage>.blob.core.windows.net/<container>?<sas> # Required for all auth methods
AZURE_TOKEN_CREDENTIALS=prod # Required only if DefaultAzureCredential is used in production
AZURE_DOCUMENT_TRANSLATION_KEY=<your-api-key> # Only required for the legacy API-key auth path below
```
## Authentication & Lifecycle
> **🔑 Two rules apply to every code sample below:**
>
> 1. **Prefer `DefaultAzureCredential`.** It works locally (Azure CLI / VS Code / Developer CLI) and in Azure (managed identity, workload identity) with no code change. Avoid connection strings, account/API keys — they bypass Entra audit and rotation.
> - Local dev: `DefaultAzureCredential` works as-is.
> - Production: set `AZURE_TOKEN_CREDENTIALS=prod` (or `AZURE_TOKEN_CREDENTIALS=<specific_credential>`) to constrain the credential chain to production-safe credentials.
> 2. **Wrap every client in a context manager** so HTTP transports, sockets, and token caches are released deterministically:
> - Sync: `with <Client>(...) as client:`
> - Async: `async with <Client>(...) as client:` **and** `async with DefaultAzureCredential() as credential:` (from `azure.identity.aio`)
>
> Snippets may abbreviate this setup, but production code should always follow both rules.
```python
import os
from azure.identity import DefaultAzureCredential, ManagedIdentityCredential
from azure.ai.translation.document import DocumentTranslationClient
# Local dev: DefaultAzureCredential. Production: set AZURE_TOKEN_CREDENTIALS=prod or AZURE_TOKEN_CREDENTIALS=<specific_credential>
credential = DefaultAzureCredential()
# Or use a specific credential directly in production:
# See https://learn.microsoft.com/python/api/overview/azure/identity-readme?view=azure-python#credential-classes
# credential = ManagedIdentityCredential()
with DocumentTranslationClient(
endpoint=os.environ["AZURE_DOCUMENT_TRANSLATION_ENDPOINT"],
credential=credential,
) as client:
statuses = list(client.list_translation_statuses())
```
### Legacy: API Key (existing keyed deployments)
New code should use `DefaultAzureCredential` above. Use `AzureKeyCredential` only if you have an existing keyed deployment that hasn't been migrated to Entra ID yet — for example, regulated environments still completing their Entra rollout.
```python
import os
from azure.core.credentials import AzureKeyCredential
from azure.ai.translation.document import DocumentTranslationClient, SingleDocumentTranslationClient
with DocumentTranslationClient(
endpoint=os.environ["AZURE_DOCUMENT_TRANSLATION_ENDPOINT"],
credential=AzureKeyCredential(os.environ["AZURE_DOCUMENT_TRANSLATION_KEY"]),
) as client:
statuses = list(client.list_translation_statuses())
# SingleDocumentTranslationClient accepts the same key-based credential.
```
## Basic Document Translation
```python
import os
from azure.ai.translation.document import DocumentTranslationClient, DocumentTranslationInput, TranslationTarget
from azure.core.exceptions import HttpResponseError
from azure.identity import DefaultAzureCredential
credential = DefaultAzureCredential()
with DocumentTranslationClient(
endpoint=os.environ["AZURE_DOCUMENT_TRANSLATION_ENDPOINT"],
credential=credential,
) as client:
source_url = os.environ["AZURE_SOURCE_CONTAINER_URL"]
target_url = os.environ["AZURE_TARGET_CONTAINER_URL"]
try:
# Start translation job
poller = client.begin_translation(
inputs=[
DocumentTranslationInput(
source_url=source_url,
targets=[
TranslationTarget(
target_url=target_url,
language="es" # Translate to Spanish
)
]
)
]
)
# Wait for completion
result = poller.result()
print(f"Status: {poller.status()}")
print(f"Documents translated: {poller.details.documents_succeeded_count}")
print(f"Documents failed: {poller.details.documents_failed_count}")
except HttpResponseError as e:
print(f"Translation failed: {e.message}")
raise
```
## Multiple Target Languages
```python
poller = client.begin_translation(
inputs=[
DocumentTranslationInput(
source_url=source_url,
targets=[
TranslationTarget(target_url=target_url_es, language="es"),
TranslationTarget(target_url=target_url_fr, language="fr"),
TranslationTarget(target_url=target_url_de, language="de")
]
)
]
)
```
## Translate Single Document
```python
from azure.ai.translation.document import SingleDocumentTranslationClient
from azure.identity import DefaultAzureCredential
with open("document.docx", "rb") as f:
document_content = f.read()
with SingleDocumentTranslationClient(endpoint, DefaultAzureCredential()) as single_client:
result = single_client.translate(
body=document_content,
target_language="es",
content_type="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
)
# Save translated document
with open("document_es.docx", "wb") as f:
f.write(result)
```
## Check Translation Status
```python
# Get all translation operations
operations = client.list_translation_statuses()
for op in operations:
print(f"Operation ID: {op.id}")
print(f"Status: {op.status}")
print(f"Created: {op.created_on}")
print(f"Total documents: {op.documents_total_count}")
print(f"Succeeded: {op.documents_succeeded_count}")
print(f"Failed: {op.documents_failed_count}")
```
## List Document Statuses
```python
# Get status of individual documents in a job
operation_id = poller.id
document_statuses = client.list_document_statuses(operation_id)
for doc in document_statuses:
print(f"Document: {doc.source_document_url}")
print(f" Status: {doc.status}")
print(f" Translated to: {doc.translated_to}")
if doc.error:
print(f" Error: {doc.error.message}")
```
## Cancel Translation
```python
# Cancel a running translation
client.cancel_translation(operation_id)
```
## Using Glossary
```python
from azure.ai.translation.document import TranslationGlossary
poller = client.begin_translation(
inputs=[
DocumentTranslationInput(
source_url=source_url,
targets=[
TranslationTarget(
target_url=target_url,
language="es",
glossaries=[
TranslationGlossary(
glossary_url="https://<storage>.blob.core.windows.net/glossary/terms.csv?<sas>",
file_format="csv"
)
]
)
]
)
]
)
```
## Supported Document Formats
```python
# Get supported formats
formats = client.get_supported_document_formats()
for fmt in formats:
print(f"Format: {fmt.format}")
print(f" Extensions: {fmt.file_extensions}")
print(f" Content types: {fmt.content_types}")
```
## Supported Languages
```python
# Get supported languages
languages = client.get_supported_languages()
for lang in languages:
print(f"Language: {lang.name} ({lang.code})")
```
## Async Client
```python
from azure.ai.translation.document.aio import DocumentTranslationClient
from azure.identity.aio import DefaultAzureCredential
async def translate_documents():
async with DefaultAzureCredential() as credential:
async with DocumentTranslationClient(
endpoint=endpoint,
credential=credential,
) as client:
poller = await client.begin_translation(inputs=[...])
result = await poller.result()
```
## Supported Formats
| Category | Formats |
|----------|---------|
| Documents | DOCX, PDF, PPTX, XLSX, HTML, TXT, RTF |
| Structured | CSV, TSV, JSON, XML |
| Localization | XLIFF, XLF, MHTML |
## Storage Requirements
- Source and target containers must be Azure Blob Storage
- Use SAS tokens with appropriate permissions:
- Source: Read, List
- Target: Write, List
## Best Practices
1. **Pick sync OR async and stay consistent.** Do not mix `azure.xxx` sync clients with `azure.xxx.aio` async clients in the same call path. Choose one mode per module.
2. **Always use context managers for clients and async credentials.** Wrap every client in `with Client(...) as client:` (sync) or `async with Client(...) as client:` (async). For async `DefaultAzureCredential` from `azure.identity.aio`, also use `async with credential:` so tokens and transports are cleaned up.
3. **Use SAS tokens** with minimal required permissions
4. **Monitor long-running operations** with `poller.status()`
5. **Handle document-level errors** by iterating document statuses
6. **Use glossaries** for domain-specific terminology
7. **Separate target containers** for each language
8. **Use async client** for multiple concurrent jobs
9. **Check supported formats** before submitting documents
## Reference Files
| File | Contents |
|------|----------|
| [references/capabilities.md](references/capabilities.md) | Additional non-hero capabilities, operation-group coverage, and production checklists. |
| [references/non-hero-scenarios.md](references/non-hero-scenarios.md) | Dedicated non-hero examples for secondary/advanced scenarios. |