AI 原生开放教材

外观

Microsoft Agent Skills技能与配置库·英文原文 ↗仓库 ↗

Azure AI Content Safety SDK Acceptance Criteria ​

SDK: azure-ai-contentsafetyRepository: https://github.com/Azure/azure-sdk-for-pythonCommit: mainPurpose: Skill testing acceptance criteria for validating generated code correctness

1. Correct Import Patterns ​

1.1 Client Imports ​

✅ CORRECT: ContentSafetyClient and BlocklistClient ​

python
from azure.ai.contentsafety import ContentSafetyClient, BlocklistClient

✅ CORRECT: Credentials ​

python
from azure.core.credentials import AzureKeyCredential
from azure.identity import DefaultAzureCredential

1.2 Model Imports ​

✅ CORRECT: Text analysis models ​

python
from azure.ai.contentsafety.models import AnalyzeTextOptions, TextCategory, AnalyzeTextOutputType

✅ CORRECT: Image analysis models ​

python
from azure.ai.contentsafety.models import AnalyzeImageOptions, ImageData, ImageCategory

✅ CORRECT: Blocklist models ​

python
from azure.ai.contentsafety.models import (
    TextBlocklist,
    TextBlocklistItem,
    AddOrUpdateTextBlocklistItemsOptions,
    RemoveTextBlocklistItemsOptions,
)

1.3 Anti-Patterns (ERRORS) ​

❌ INCORRECT: Importing clients from models ​

python
# WRONG - clients are in azure.ai.contentsafety, not models
from azure.ai.contentsafety.models import ContentSafetyClient
from azure.ai.contentsafety.models import BlocklistClient

❌ INCORRECT: Wrong credential import path ​

python
# WRONG - AzureKeyCredential is in azure.core.credentials
from azure.ai.contentsafety import AzureKeyCredential

2. Authentication Patterns ​

2.1 ✅ CORRECT: API Key Authentication ​

python
import os
from azure.ai.contentsafety import ContentSafetyClient
from azure.core.credentials import AzureKeyCredential

client = ContentSafetyClient(
    endpoint=os.environ["CONTENT_SAFETY_ENDPOINT"],
    credential=AzureKeyCredential(os.environ["CONTENT_SAFETY_KEY"]),
)

2.2 ✅ CORRECT: Microsoft Entra ID Authentication ​

python
import os
from azure.ai.contentsafety import ContentSafetyClient
from azure.identity import DefaultAzureCredential

client = ContentSafetyClient(
    endpoint=os.environ["CONTENT_SAFETY_ENDPOINT"],
    credential=DefaultAzureCredential(),
)

2.3 Anti-Patterns (ERRORS) ​

❌ INCORRECT: Hardcoding secrets or endpoint ​

python
# WRONG - hardcoded secrets and endpoint
client = ContentSafetyClient(
    endpoint="https://example.cognitiveservices.azure.com",
    credential=AzureKeyCredential("my-secret-key"),
)

❌ INCORRECT: Passing key directly as credential ​

python
# WRONG - credential must be AzureKeyCredential
client = ContentSafetyClient(
    endpoint=os.environ["CONTENT_SAFETY_ENDPOINT"],
    credential=os.environ["CONTENT_SAFETY_KEY"],
)

3. Text Analysis Patterns ​

3.1 ✅ CORRECT: Analyze Text with Categories ​

python
from azure.ai.contentsafety.models import AnalyzeTextOptions, TextCategory

request = AnalyzeTextOptions(text="Your text content to analyze")
response = client.analyze_text(request)

for category in [
    TextCategory.HATE,
    TextCategory.SELF_HARM,
    TextCategory.SEXUAL,
    TextCategory.VIOLENCE,
]:
    result = next(
        (item for item in response.categories_analysis if item.category == category),
        None,
    )
    if result:
        print(f"{category}: severity {result.severity}")

3.2 ❌ INCORRECT: Using image models for text ​

python
# WRONG - ImageCategory is for image analysis
from azure.ai.contentsafety.models import AnalyzeTextOptions, ImageCategory

request = AnalyzeTextOptions(text="text")
response = client.analyze_text(request)
if any(item.category == ImageCategory.HATE for item in response.categories_analysis):
    print("Wrong category enum")

4. Image Analysis Patterns ​

4.1 ✅ CORRECT: Analyze Image from File Bytes ​

python
from azure.ai.contentsafety.models import AnalyzeImageOptions, ImageData

with open("image.jpg", "rb") as file:
    request = AnalyzeImageOptions(image=ImageData(content=file.read()))

response = client.analyze_image(request)

4.2 ✅ CORRECT: Analyze Image from Blob URL ​

python
from azure.ai.contentsafety.models import AnalyzeImageOptions, ImageData

request = AnalyzeImageOptions(
    image=ImageData(blob_url="https://example.com/image.jpg")
)
response = client.analyze_image(request)

4.3 ❌ INCORRECT: Using text models for image ​

python
# WRONG - AnalyzeTextOptions is for text analysis only
from azure.ai.contentsafety.models import AnalyzeTextOptions

request = AnalyzeTextOptions(text="image.jpg")
response = client.analyze_image(request)

4.4 ❌ INCORRECT: Wrong ImageData field ​

python
# WRONG - ImageData uses blob_url, not url
request = AnalyzeImageOptions(
    image=ImageData(url="https://example.com/image.jpg")
)

5. Blocklist Management Patterns ​

5.1 ✅ CORRECT: Create or Update Blocklist ​

python
from azure.ai.contentsafety import BlocklistClient
from azure.ai.contentsafety.models import TextBlocklist

blocklist_client = BlocklistClient(endpoint, credential)

blocklist = TextBlocklist(
    blocklist_name="my-blocklist",
    description="Custom terms to block",
)

blocklist_client.create_or_update_text_blocklist(
    blocklist_name="my-blocklist",
    options=blocklist,
)

5.2 ✅ CORRECT: Add Items to Blocklist ​

python
from azure.ai.contentsafety.models import AddOrUpdateTextBlocklistItemsOptions, TextBlocklistItem

items = AddOrUpdateTextBlocklistItemsOptions(
    blocklist_items=[
        TextBlocklistItem(text="blocked-term-1"),
        TextBlocklistItem(text="blocked-term-2"),
    ]
)

blocklist_client.add_or_update_blocklist_items(
    blocklist_name="my-blocklist",
    options=items,
)

5.3 ✅ CORRECT: Analyze Text with Blocklist ​

python
from azure.ai.contentsafety.models import AnalyzeTextOptions

request = AnalyzeTextOptions(
    text="Text containing blocked-term-1",
    blocklist_names=["my-blocklist"],
    halt_on_blocklist_hit=True,
)

response = client.analyze_text(request)
if response.blocklists_match:
    for match in response.blocklists_match:
        print(f"Blocked: {match.blocklist_item_text}")

5.4 ✅ CORRECT: Remove Blocklist Items ​

python
from azure.ai.contentsafety.models import RemoveTextBlocklistItemsOptions

blocklist_client.remove_blocklist_items(
    blocklist_name="my-blocklist",
    options=RemoveTextBlocklistItemsOptions(blocklist_item_ids=["item-id"]),
)

5.5 ❌ INCORRECT: Using ContentSafetyClient for blocklist operations ​

python
# WRONG - blocklist operations are on BlocklistClient
client = ContentSafetyClient(endpoint, credential)
client.create_or_update_text_blocklist(...)

6. Multi-Severity Classification Patterns ​

6.1 ✅ CORRECT: Eight severity levels for text ​

python
from azure.ai.contentsafety.models import AnalyzeTextOptions, AnalyzeTextOutputType

request = AnalyzeTextOptions(
    text="Your text",
    output_type=AnalyzeTextOutputType.EIGHT_SEVERITY_LEVELS,
)
response = client.analyze_text(request)

6.2 ❌ INCORRECT: Using string output type ​

python
# WRONG - output_type must be AnalyzeTextOutputType enum
request = AnalyzeTextOptions(
    text="Your text",
    output_type="EightSeverityLevels",
)