操作指南：访问 Fabric API

为了便于工作负荷开发人员访问 Fabric 公共 API，我们提供了提取调用的 FabricPlatformAPIClient 客户端。 此客户端主要通过 On-Behalf-Of（OBO）流，利用当前用户的凭据和权限来访问 Fabric API。

主要身份验证方法：代表（OBO）流

扩展性工具包围绕 On-Behalf-Of（OBO）流设计，这是访问 Fabric API 的建议和主要方法。 此流：

• 使用用户凭据：使用当前登录用户的权限进行作
• 维护安全上下文：确保用户只能访问他们有权访问的资源
• 简化开发：无需管理服务主体凭据
• 遵循最佳做法：与 Microsoft Fabric 的安全模型保持一致

创建 FabricPlatformAPIClient

使用代理流（推荐）

import { getWorkloadClient } from "../controller/WorkloadClient";
import { FabricPlatformAPIClient } from "../clients/FabricPlatformAPIClient";

// Create the client using the workload client (uses current user's credentials via OBO flow)
const client = FabricPlatformAPIClient.create(getWorkloadClient());

使用服务主体（仅限高级场景）

import { FabricPlatformAPIClient } from "../clients/FabricPlatformAPIClient";

// Create with service principal authentication (advanced scenarios)
const client = FabricPlatformAPIClient.createWithServicePrincipal(
  "your-client-id",
  "your-client-secret", 
  "your-tenant-id"
);

示例：创建新项

下面是如何使用 OBO 流创建新笔记本项的示例（建议）：

import { FabricPlatformAPIClient } from "../clients/FabricPlatformAPIClient";
import { getWorkloadClient } from "../controller/WorkloadClient";

async function createNewNotebook() {
  try {
    // Create the client using OBO flow with current user's credentials
    const client = FabricPlatformAPIClient.create(getWorkloadClient());
    
    // Get the current workspace ID
    const workspace = await client.workspaces.getCurrentWorkspace();
    const workspaceId = workspace.id;
    
    // Create a new notebook item
    const newItem = await client.items.createItem(workspaceId, {
      displayName: "My New Notebook",
      description: "Created via FabricPlatformAPIClient",
      type: "Notebook", // Item type code
      // Optional: Specify folder ID if you want to create in a specific folder
      // folderId: "your-folder-id",
      
      // Optional: Add item-specific creation properties
      creationPayload: {
        notebookDefaultLanguage: "python"
      }
    });
    
    console.log("Created new item:", newItem);
    return newItem;
  } catch (error) {
    console.error("Error creating item:", error);
    throw error;
  }
}

示例：创建具有定义的项目

对于需要定义（如 Lakehouse）的更复杂的项，可以使用 OBO 流包括定义：

async function createLakehouse() {
  // Create the client using OBO flow with current user's credentials
  const client = FabricPlatformAPIClient.create(getWorkloadClient());
  const workspace = await client.workspaces.getCurrentWorkspace();
  
  const newLakehouse = await client.items.createItem(workspace.id, {
    displayName: "My New Lakehouse",
    description: "Sample Lakehouse created programmatically",
    type: "Lakehouse",
    definition: {
      format: "json",
      parts: [
        {
          path: "definition/settings.json",
          content: JSON.stringify({
            schemaVersion: "1.0",
            // Lakehouse-specific settings
            settings: {
              defaultDatabase: "MyDatabase"
            }
          })
        }
      ]
    }
  });
  
  return newLakehouse;
}
}

与其他 Fabric 资源一起工作

FabricPlatformAPIClient 使用 OBO 流提供对各种 Fabric 资源的访问：

// Create the client using OBO flow with current user's credentials
const client = FabricPlatformAPIClient.create(getWorkloadClient());

// List all workspaces the user has access to
const workspaces = await client.workspaces.listWorkspaces();

// Get workspace capacities
const capacities = await client.capacities.listCapacities();

// Work with OneLake (with user's permissions)
const tables = await client.oneLake.listTables(workspaceId, lakehouseId);

// Work with connections (scoped to user's access)
const connections = await client.connections.listConnections(workspaceId);

将 FabricPlatformAPIClient 与 OBO Flow 配合使用的主要优势

• 安全性：所有作都遵循当前用户的权限和安全上下文
• 简单性：无需管理服务主体凭据或身份验证流
• 一致性：与扩展性工具包的前端专用架构无缝兼容地工作
• 可审核性：所有操作均在实际用户帐户下执行并记录
• 最佳做法：遵循 Microsoft 针对扩展性方案中 Fabric API 访问推荐的模式