Skip to main content

Sharepoint

Use the SharePoint connector to bring files and pages from SharePoint into the Knowledge Repository. This connector supports SharePoint site libraries, pages, and attachments where allowed by SharePoint permissions.

When to use​

  • Organizations that keep documentation, policies, or knowledge in SharePoint site libraries.
  • Consolidating content from SharePoint into a searchable Knowledge Repository for QA, search, or knowledge ops.

Notes​

  • Connector permissions must be configured to allow reading the target files and pages.
  • SVAHNAR does not store your SharePoint files permanently; it reads and ingests data during import according to repository policies.

Usage​

  1. Register an Azure AD application (service principal) with the necessary API permissions (see "How to find these" below).
  2. Build the SharePointData configuration (below) with your tenant_id, client_id, client_secret, and site_url.
  3. Call the import endpoint of the Knowledge Repository connector (or run the connector tool) providing the SharePointData payload.
  4. Monitor logs for items that could not be fetched due to permissions, throttling, or unsupported formats.

Typical flow​

  • The connector authenticates using OAuth2 client credentials (tenant_id, client_id, client_secret) to receive a token for Microsoft Graph or SharePoint API.
  • It enumerates lists and libraries under the provided site_url (pages, document libraries) and downloads files/pages according to the configured flags.
  • The connector normalizes content (optionally preserving structure where possible) and sends extracted text and metadata into the Knowledge Repository ingestion pipeline.

Parameter reference​

  • tenant_id (string, required): Your Azure Active Directory tenant identifier (Directory ID). This identifies the Azure AD tenant that owns the SharePoint tenant.

  • client_id (string, required): The Application (client) ID of the Azure AD app registration used by the connector.

  • client_secret (string, required): A client secret (value) generated under the Azure AD app registration (Certificates & secrets). Treat this like a password — store it securely.

  • site_url (string, required): The full URL of the SharePoint site you want to import, for example: https://contoso.sharepoint.com/sites/Engineering or https://contoso.sharepoint.com/teams/HR.

How to find tenant_id, client_id, client_secret, and site_url​

1) Tenant ID (Directory ID)

  • Azure Portal: Sign in to the Azure Portal (portal.azure.com) → Azure Active Directory → Overview → copy the Tenant ID (also called Directory ID).
  • Azure CLI: az account show --query tenantId -o tsv (if you use the CLI).

2) Register an app to get client_id and create client_secret

  • Azure Portal: Sign in → Azure Active Directory → App registrations → New registration.

    • Give the app a name (e.g., svc-knowledge-importer).
    • For a backend/service connector, choose Accounts in this organizational directory only (single tenant) or other option matching your org.
    • Redirect URI is not required for client credentials flow.

  • After registering:

    • Application (client) ID is shown on the app's Overview — this is your client_id.
    • Go to Certificates & secrets → New client secret → add a description and expiry → Add. Copy the Value immediately — this is the client_secret (you cannot view it again).

3) API permissions & admin consent

  • Under the registered app → API permissions → Add a permission → Microsoft Graph (or SharePoint) → choose Application permissions (for app-only access) or Delegated permissions (if the connector will act on behalf of a user).

  • Typical application permissions for read-only imports:

    • Sites.Read.All (Microsoft Graph) — read items and lists across sites.
    • Sites.ReadWrite.All only if you need write; prefer least privilege.

  • After adding application permissions, click Grant admin consent (requires an admin).

4) Site URL (site_url)

  • Navigate to the SharePoint site in your browser and copy the top-level URL. Example site URLs:

    • https://yourtenant.sharepoint.com/sites/Engineering
    • https://yourtenant.sharepoint.com/teams/HR

  • If you need to import a sub-site or a specific site collection, use that exact URL.

Notes on app-only vs delegated

  • App-only (client credentials): The connector authenticates using tenant_id + client_id + client_secret and behaves as the application identity. Use this for server-to-server imports and grant Application permissions such as Sites.Read.All and give admin consent.
  • Delegated: If you prefer the connector to act as a user, use delegated permissions and an OAuth flow where a user signs in; this is less common for automated imports.

Example payload​

{
  "tenant_id": "<TENANT_ID>",
  "client_id": "<CLIENT_ID>",
  "client_secret": "<CLIENT_SECRET>",
  "site_url": "https://yourtenant.sharepoint.com/sites/Engineering",
}

Authentication & permissions​

  • Use an Azure AD app registration with the minimum required permissions (prefer Sites.Read.All for read-only import).
  • Grant Admin consent for application permissions so app-only (client credentials) tokens can access site content.
  • Make sure the SharePoint site is within the same tenant and the site permissions don't block the app's access.

File & page handling​

  • The connector will attempt to fetch site pages (modern site pages) and files from libraries. Files will be downloaded subject to size limits and ingestion policies.
  • Binary attachments (large videos, very large binaries) may be skipped or stored separately depending on repository ingestion limits.

Limitations & caveats​

  • Throttling: Microsoft Graph and SharePoint APIs enforce throttling. Large imports should respect retry/backoff semantics.
  • Complex web parts or custom scripts on pages may not translate perfectly to plain text. The connector will extract the textual content and common web-part outputs but may leave placeholders for custom or unsupported web parts.
  • Permissions: App-only permissions require admin consent; delegated flows require a user with the right permissions.

Troubleshooting​

  • 401/403 errors: Verify tenant_id, client_id, and client_secret, and confirm admin consent for application permissions.
  • Missing files/pages: Confirm that the app permission scope includes the target site and that the site URL is correct.
  • Throttling: Implement exponential backoff and consider importing large sites in batches.