Skip to content

Latest commit

 

History

History
59 lines (49 loc) · 2.06 KB

README.md

File metadata and controls

59 lines (49 loc) · 2.06 KB

Analytics Assistant CLI

The Data Analytics Assistant is a command-line interface (CLI) tool designed to assist users in obtaining insights from their MySQL databases using natural language questions. By leveraging the power of OpenAI's API, users can easily ask questions about their data, and the assistant will provide answers based on the information stored in the database.

DEMO_1

Requirements

  • Go 1.18 or later
  • A MySQL database
  • An OpenAI API key

Installation

curl -s https://raw.githubusercontent.com/FonsecaGoncalo/analytics-assistant/main/installer.sh | bash

Install From Sources

  • Clone the repository:
git clone https://github.com/FonsecaGoncalo/analytics-assistant.git

Navigate to the project directory:

cd analytics-assistant

Build the CLI tool:

go build -o analytics-assistant

This will create an executable named analytics-assistant in the current directory.

Usage

Usage:
analytics-assistant session [flags]

Flags:
-c, --context-file string   Path to a file containing business context
-p, --db-password string    MySQL database password
-u, --db-url string         MySQL database URL
-n, --db-username string    MySQL database username
-h, --help                  help for session
-s, --log-sql               Log SQL

To use the CLI tool, you will first need to set and environment variable with the openai API key:

export OPENAI_API_KEY="{you_openai_api_key}"

And then run:

analytics-assistant session --db-url "tcp(localhost:3306)/your_database_name" --db-username "your_username" --db-password "your_password"

Replace your_database_name, your_username, and your_password with the appropriate values for your MySQL database. This will initiate an interactive session with the data analyst assistant.

To provide the assistant with additional context when responding to questions, supply a file containing business rules or database field descriptions using the --context-file flag.

License

This project is licensed under the MIT License.