Google Cloud Document AI地域不匹配错误排查求助

问题描述 #

在Mac上使用Python 3.12调用GCP的Document AI PDF分割器时，即便确认处理器部署在eu区域，仍收到错误：Invalid location: 'eu' must match the server deployment 'us'，本地VS Code和Colab运行均失败，尝试获取处理器列表时也触发相同错误。

问题根源 #

Document AI Python客户端默认连接**美国区域（us）**的服务端点，即便你在parent参数中指定了eu区域，客户端仍会向us区域的服务器发送请求，导致区域不匹配验证失败。

修复方案 #

方案1：初始化客户端时指定对应区域的API端点 #

修改客户端初始化代码，通过client_options参数明确指定eu区域的服务端点：

from google.oauth2 import service_account
from google.cloud import documentai_v1 as documentai

def list_document_ai_processors(project_id: str, location: str, service_account_file: str):
    credentials = service_account.Credentials.from_service_account_file(service_account_file)
    
    # 指定对应区域的API端点
    client_options = {"api_endpoint": f"{location}-documentai.googleapis.com"}
    client = documentai.DocumentProcessorServiceClient(
        credentials=credentials, client_options=client_options
    )
    
    parent = f"projects/{project_id}/locations/{location}"
    processors = client.list_processors(parent=parent)
    
    for processor in processors:
        print(f"Processor name: {processor.name}")
        print(f"Processor type: {processor.type_}")
        print(f"Display name: {processor.display_name}")
        print(f"State: {processor.state}\n")

if __name__ == "__main__":
    project_id = "abmtest-429810"
    location = "eu"
    service_account_file = "/Users/XXXXXX/Desktop/DocAI/credentials.json"
    list_document_ai_processors(project_id, location, service_account_file)

方案2：设置环境变量指定默认区域 #

在运行代码前，通过环境变量告诉客户端使用eu区域：

• 终端临时设置：

  export GOOGLE_CLOUD_LOCATION=eu
  

• 或在代码开头添加：

  import os
  os.environ["GOOGLE_CLOUD_LOCATION"] = "eu"
  

额外验证步骤 #

1. 用gcloud命令确认处理器所在区域：

   gcloud documentai processors list --location=eu --project=abmtest-429810
   

2. 检查服务账号是否拥有eu区域的Document AI权限，确保角色包含roles/documentai.admin或roles/documentai.user。

内容的提问来源于stack exchange，提问作者AlexCT