About the Snowflake Connector for SharePoint¶

This topic describes the basic concepts of Snowflake Connector for SharePoint, its use cases and benefits, key features, how it works, and limitations.

The Snowflake Connector for SharePoint connector connects a Microsoft 365 SharePoint site and Snowflake to ingest files and user permissions and keeps them up to date. Snowflake Connector for SharePoint also supports the Cortex Search service and can make ingested files ready for conversational analysis for use in AI Assistants using SQL, Python or REST APIs.

优势

• Frictionless ingestion: The connector is easy to set up and configure. You can use files from SharePoint with Cortex Search in your chat interface of choice.
• 默认提供安全保护：连接器通过 Cortex Search 筛选器遵守 SharePoint 中的最终用户访问控制。
• Scalable by design: Built on the Snowflake Native App framework, the connector leverages Snowflake’s built-in security, scalability and reliability capabilities.
• Saves costs: The connector saves you cost by eliminating the need to manually transfer files or integrate against API endpoints or manage third-party solutions.

用例

如果您希望实现以下功能，可以使用此连接器：

• 为组织 SharePoint 站点内的公共文档创建 AI 助手
• 使您的 AI 助手遵守组织 SharePoint 站点中指定的访问控制

工作原理

本节介绍该连接器在前面提到的两种使用情况下的工作原理。

为组织 SharePoint 站点内的公共文档创建 AI 助手¶

Working with Snowflake Connector for SharePoint for this use case can be broadly divided into four phases, each associated with a specific user persona. The following workflow describes these phases, the associated user journey, and how this connector works:

1. An Azure or Office 365 account administrator (https://learn.microsoft.com/en-us/microsoft-365/admin/add-users/about-admin-roles) in your organization configures Microsoft Graph (https://learn.microsoft.com/en-us/graph/overview) to enable OAuth authentication as described in Get access without a user (https://learn.microsoft.com/en-us/graph/auth-v2-service?tabs=http). They then share the required credentials with the organization’s data engineer.

2. A data engineer or data scientist in your organization installs the SharePoint Connector for Snowflake from the Snowflake marketplace into their Snowflake account. They then configure the connector with the following information:

   • 指定在步骤 1 中获得的 SharePoint OAuth 凭据（ClientID、客户端密钥和 TenantID）。
   • 指定其 SharePoint 站点的 URL。一般来说，这是 SharePoint 站点中的一个特定子站点。
   • Choosing whether to ingest files from all folders or a specific folder in the SharePoint URL. Note that the files from subfolders are always included.

连接器验证配置后，会执行以下操作：

1. 从指定源引入支持的文件和用户权限。

2. Uses the PARSE_DOCUMENT function of Cortex AI to parse and chunk the ingested files.

3. Creates a Cortex Search service to serve as a RAG engine for your own AI assistants with your parsed and chunked data.

4. An IT developer in the organization creates a chatbot in their chat interface of choice, such as bot extensions in Slack, Teams or a web page, and hosts it as appropriate within their environment. The IT Developer configures roles, permissions and authentication in Snowflake to use the Cortex Search REST API endpoint available in the suite of Snowflake REST APIs.

5. After your AI assistant is up and running, business users in your organization can interact with it to ask questions and see responses based on the files ingested from your SharePoint site into your Snowflake account. All responses have citations that are links to the source documents from your SharePoint site.

使您的 AI 助手遵守组织 SharePoint 站点中指定的访问控制¶

Working with Snowflake Connector for SharePoint for this use case can be broadly divided into four phases, each associated with a specific user persona. The following workflow describes these phases, the associated user journey, and how this connector works:

1. An Azure or Office 365 account administrator (https://learn.microsoft.com/en-us/microsoft-365/admin/add-users/about-admin-roles) in your organization configures Microsoft Graph (https://learn.microsoft.com/en-us/graph/overview) to enable OAuth authentication as described in Get access without a user (https://learn.microsoft.com/en-us/graph/auth-v2-service?tabs=http). They then share the required credentials with the organization’s data engineer or data scientist.

2. A data engineer or data scientist in your organization installs the SharePoint Connector for Snowflake from the Snowflake marketplace into their Snowflake account. They then configure the connector by:

   • 指定在步骤 1 中获得的 SharePoint OAuth 凭据（客户端 ID、客户端密钥和租户 ID）。
   • 指定其 SharePoint 站点的 URL。一般来说，这是 SharePoint 站点中的一个特定子站点。
   • Choosing whether to ingest files from all folders or a specific folder in the SharePoint URL. Note that the files from subfolders are always included.

连接器验证配置后，会执行以下操作：

1. 从指定源引入支持的文件和用户权限。

2. 使用 Cortex AI 的 PARSE_DOCUMENT 函数对引入的文件进行解析和分块。

3. 创建一个 Cortex Search 服务，将解析和分块后的数据作为 RAG 引擎，供自己的 AI 助手使用。

4. An IT developer in the organization creates a chatbot in their chat interface of choice, such as bot extensions in Slack, Teams or a web page, and hosts it as appropriate within their environment.

   1. 他们在 Snowflake 中配置角色、权限和身份验证，以使用 Snowflake REST APIs 套件中的 Cortex Search REST API 端点。
   2. They specify a filter containing the email ID of the SharePoint user when the AI assistant queries the Cortex Search REST API, for example filter.@contains.user_ids or filter.@contains.user_emails. This restricts responses from Cortex Search to documents that the specified business user has access to in your organization’s SharePoint.

5. After your AI assistant is up and running, when business users in your organization interact with it to ask questions, they will only see information from files in your SharePoint thay have access to because of the filter specified in Step 3(b). All responses have citations that are links to the source documents from your SharePoint site.

限制

• Cortex Parse Document limitations and requirements
• Cortex Search limitations
• 增量引入过程中，不会捕获因移动或重命名文件夹而造成的更改。
• 该连接器仅支持 Microsoft 365 组。
• 该连接器仅引入支持的文件类型，会忽略其他文件类型。

区域可用性

The Snowflake Connector for SharePoint depends on Cortex Parse document and Cortex Search.

The Snowflake Connector for SharePoint is currently available in the regions listed in Cortex Parse Document regional availability.

后续步骤

Set up the Snowflake Connector for SharePoint.