Función
Solución
docs
Blog
Compañía
GPTBots Documentación > TUTORIAL > Database > Overview
GPTBots Documentación > TUTORIAL > Database > Overview
Overview
Última actualización：2024-09-20

Overview

The GPTBots database is designed for business scenarios that require precise data queries and analysis. Users can perform natural language queries and analysis of reliable data, helping business personnel quickly mine data and gain insights into data value.
The database supports extracting structured entity data from unstructured data (such as PDF, Word, Excel, TXT documents) and also allows adding entity data through template import and API, facilitating easy data storage. ![Database]alt text

Data Reliability

When the Bot calls the database capabilities, it transparently displays SQL query statements and data table sources, making each query verifiable and reviewable, ensuring the reliability and accuracy of the result data.

Natural Language Query

Users do not need to master complex SQL languages. Through LLM, natural language user queries are accurately converted into SQL statements, enabling precise and efficient multi-table linkage queries.

Data Visualization

Based on the queried data, the Bot supports the automatic generation of various types of dynamic interactive charts (line charts, bar charts, scatter plots, etc.), helping users intuitively understand and analyze data.

Data Table Types

The GPTBots platform supports local data tables, remote data tables, and three types, allowing the selection of the appropriate type based on business needs.

  • Local Data Tables: Data is stored in the data tables of the GPTBots platform, supporting data addition in both "structured" and "unstructured" ways.
  • Remote Data Tables (coming soon): Only metadata of the data table is obtained after authentication, with data still stored in the remote database.

Data Addition Methods

  • Unstructured Data Entry Supports entity extraction from unstructured data (such as PDF, Word, Excel, TXT documents), facilitating the recognition, cleaning, extraction, and entry of unstructured data.

  • Structured Data Entry Supports importing existing structured data into the database through CSV template import and API, achieving rapid data entry and management. It will soon support interfacing and accessing external databases through standardized interfaces (JDBC, RESTful, etc.), enabling data import and remote connection.