Data App Usage

Starting with version 0.5.0, the DB-GPT project has natively integrated the management and distribution of data-centric applications. The application management repository is also part of the open-source dbgpts project. The dbgpts project manages and shares resources categorized into the following:

• apps: These are native intelligent data applications developed using the DB-GPT framework.
• workflow: Workflows constructed using the AWEL (Agentic Workflow Expression Language).
• agents: Intelligent agents that can perform various tasks.
• operators: Basic operational operators (or symbols) that can be used within workflows.

NOTE

Please note that this tutorial primarily focuses on the installation and use of intelligent agent workflows. For the development of applications, you should refer to the Development Guide.

Support for these capabilities is provided from version V0.5.0 onwards. For developers and teams looking to build and distribute their applications through DB-GPT, this structured approach provides both a framework and ecosystem for creating, sharing, and managing data applications effectively.

Here we introduce the creation of a data intelligence analysis assistant application. This tutorial utilizes the auto-planning capability of Multi-Agents.

The effect is as follows:

In the application panel, click on Create Application and fill in the parameters as shown in the image. It is important to note that the work mode selected here is auto_plan. This involves the collaboration of two dependent Agents: 1. DataScientist and 2. Reporter. Both of these agents depend on the resource database, and for testing, you can use the default database and data provided in the official tutorial.

Special Note: Currently, in auto-plan mode, the building of applications is conducted through multiple Agents. This project has a number of built-in Agents, which currently include:

• CodeEngineer
• Reporter
• DataScientist
• ToolExpert
• RetrieveSummarizer
• Summarizer

If you wish to expand and implement customized Agents, you can refer to the Agents Development Guide.

After adding the necessary information, choose to submit to complete the creation of the application. In the application panel, click the dialogue button to enter the dialogue interface.

Summary

This tutorial is just a simple introduction to application construction. If you are interested in more complex applications, you can achieve more intricate scenarios by orchestrating AWEL workflows and customizing the expansion of Agents.