使用AtlasClient在Purview指定集合创建实体遇授权问题

运行代码时遇到AtlasException：

{"requestId":"36fb4fc3-bc6b-4583-b6a1-faac6a5723b4","errorCode":"ATLAS-403-00-001","errorMessage":"5133fd99-b486-4723-ac91-c8acc375a4ed is not authorized to perform create entity: type=test file"}

注意：仅拥有Purview中特定集合的访问权限，无法在根集合中创建实体。

运行的代码 #

from azure.identity import ClientSecretCredential
from pyapacheatlas.auth import ServicePrincipalAuthentication
from pyapacheatlas.core import AtlasClient, collections
from pyapacheatlas.core import AtlasEntity, RelationshipTypeDef
import requests
import os
import json
import pandas as pd
import pydicom

# Azure AD credentials
tenant_id = "xxxxxxxxxxxxxxxxxxxx"
client_id = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
client_secret = "xxxxxxxxxxxxxxxxxxxxxxxxxxx"

# Purview account details
purview_account_name = "account-name"
collection_name = "abcde"

# Authenticate using AAD
credential = ClientSecretCredential(tenant_id, client_id, client_secret)
auth = ServicePrincipalAuthentication(
       tenant_id=tenant_id,
       client_id=client_id,
       client_secret=client_secret
   )

# Create an Atlas client
client = AtlasClient(
       endpoint_url=f"https://{purview_account_name}.purview.azure.com/catalog/api/atlas/v2",
       authentication=auth
   )

entity_up = AtlasEntity(
       name="API_TEST_NEW",
       typeName='test file',
       qualified_name="TESTING_API_NEW",
       attributes={
        "description": "A sample entity with custom attributes",
        "businessAttributes": {
            "businessAttribute1": "Business Value 1",
            "businessAttribute2": "Business Value 2"
        },
        "customAttributes": {
            "customKey1": "Custom Value 1",
            "customKey2": "Custom Value 2"
        }
    },
    collectionId = 'abcde', 
    domainId = 'account-name',
       contacts = {
      "Expert": [
        {
          "id": "xxxxxxxxxxxxxxxxx",
          "info": "Example Expert Info"
        }
      ],
      "Owner": [
        {
          "id": "xxxxxxxxxxxxxxxxxxxxxxxxxx",
          "info": "Example Owner Info"
        }
      ]
    },
   )
client.upload_entities(entity_up)

提问 #

如何使用AtlasClient在特定集合（而非根集合）中创建实体？已能通过PurviewClient实现，但使用AtlasClient时遇到问题。

问题核心是未正确指定集合上下文，且AtlasEntity中误用了无效参数。按以下步骤修正：

1. 获取目标集合的唯一ID
   不能直接使用集合名称abcde，需先通过AtlasClient查询集合的实际ID：

# 获取指定集合的详细信息，提取唯一ID
collection_details = client.get_collection(collection_name)
target_collection_id = collection_details["collectionId"]

2. 在上传时指定集合
   AtlasEntity本身不支持直接设置collectionId属性，需通过upload_entities接口的collection参数显式指定，或在初始化AtlasClient时设置默认集合：

• 方式一：上传时指定集合

client.upload_entities(entity_up, collection=target_collection_id)

• 方式二：初始化客户端时设置默认集合

client = AtlasClient(
    endpoint_url=f"https://{purview_account_name}.purview.azure.com/catalog/api/atlas/v2",
    authentication=auth,
    default_collection=target_collection_id
)

3. 清理无效参数
   删除AtlasEntity中的collectionId和domainId参数，这两个不属于AtlasEntity的合法属性，会被接口忽略。

修改后的完整代码示例 #

from azure.identity import ClientSecretCredential
from pyapacheatlas.auth import ServicePrincipalAuthentication
from pyapacheatlas.core import AtlasClient
from pyapacheatlas.core import AtlasEntity
import requests
import os
import json
import pandas as pd
import pydicom

# Azure AD credentials
tenant_id = "xxxxxxxxxxxxxxxxxxxx"
client_id = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
client_secret = "xxxxxxxxxxxxxxxxxxxxxxxxxxx"

# Purview account details
purview_account_name = "account-name"
collection_name = "abcde"

# Authenticate using AAD
auth = ServicePrincipalAuthentication(
       tenant_id=tenant_id,
       client_id=client_id,
       client_secret=client_secret
   )

# Create an Atlas client
client = AtlasClient(
       endpoint_url=f"https://{purview_account_name}.purview.azure.com/catalog/api/atlas/v2",
       authentication=auth
   )

# 获取目标集合的唯一ID
collection_details = client.get_collection(collection_name)
target_collection_id = collection_details["collectionId"]

entity_up = AtlasEntity(
       name="API_TEST_NEW",
       typeName='test file',
       qualified_name="TESTING_API_NEW",
       attributes={
        "description": "A sample entity with custom attributes",
        "businessAttributes": {
            "businessAttribute1": "Business Value 1",
            "businessAttribute2": "Business Value 2"
        },
        "customAttributes": {
            "customKey1": "Custom Value 1",
            "customKey2": "Custom Value 2"
        }
    },
       contacts = {
      "Expert": [
        {
          "id": "xxxxxxxxxxxxxxxxx",
          "info": "Example Expert Info"
        }
      ],
      "Owner": [
        {
          "id": "xxxxxxxxxxxxxxxxxxxxxxxxxx",
          "info": "Example Owner Info"
        }
      ]
    },
   )

# 上传实体时指定目标集合
client.upload_entities(entity_up, collection=target_collection_id)

关键说明 #

• 权限错误本质是上传时默认使用了根集合，而当前账号无此权限，指定正确的目标集合即可解决。
• 必须使用集合的唯一ID而非名称，避免因名称重复导致的错误。
• AtlasEntity的属性需符合PyApacheAtlas的定义，自定义集合参数应通过客户端或上传接口传递。

内容的提问来源于stack exchange，提问作者Naveen Chandraiah