About Openflow Connector for Box¶

Workflow¶

1. A Box developer creates a Box Platform app and submits it for authorization.

2. A Box administrator authorizes the app.

3. The Box developer then performs the following tasks:

   1. Shares a Box folder with the app service account.

   2. Shares a Platform app configuration JSON file and a folder ID with a Snowflake account administrator.

4. A Snowflake account administrator performs the following tasks:

   1. Downloads and imports the connector definition file into the Snowflake Openflow canvas.

   2. Configures the connector with Snowflake connection details and the data provided by the Box developer.

   3. Runs the connector flow. The connector does the following:

      1. Creates the required tables, stages, and a Cortex Search service in the specified Snowflake schema.

      2. Fetches Box file content and permissions from the folder specified in the connector configuration.

      3. Runs parsing and chunking on the fetched documents, and saves them in Snowflake tables. The saved chunks are automatically indexed by the Cortex Search service.

5. A Chatbot developer uses the Cortex Search service to build a chatbot application.

Limitations¶

• Cortex Parse Document limitations and requirements

• Cortex Search limitations

• Changes caused by moving folders out of the specified root folder aren’t captured during incremental ingestion.

• The connector ingests only the supported file types and ignores others.

Next steps¶

Set up the Openflow Connector for Box