Changelog Generator Tool

This project provides an AI-powered tool that automates the generation of changelogs by collecting information from Git logs and Redmine tickets. It uses a crew of AI agents to improve the quality and efficiency of the final changelog.

Project Overview

The system uses multiple AI agents, each responsible for a specific part of the changelog generation process. This tool is designed to:

• Gather data from Git logs and Redmine tickets.
• Analyze and categorize the collected information.
• Generate a well-formatted changelog or release notes.
• Ensure the output is consistent with project guidelines and requirements.

Key Components

Main Script (main.py)

The main entry point of the tool. It handles command-line arguments, sets up the environment, and initiates the changelog generation process.

Crew Definition (crew.py)

Defines the AI agents and their tasks using the crewai framework. It includes:

• GitlogAnalyst: Analyzes Git commit logs.
• RedmineAnalyst: Retrieves and analyzes Redmine issues.
• WritingAgent: Generates the final changelog or release notes.

Configuration

• config/config.py: Parses configuration settings.
• config/agents.yaml: Defines agent configurations.
• config/tasks.yaml: Defines task configurations.

Tools

• tools/tools.py: Contains custom tools for retrieving Git commits and Redmine issues.

Data

• data/CodeOfConduct.md: Contains the project's code of conduct (used for context).
• data/template2.txt: Changelog template used by the writing agent.

Usage

1. Environment Setup for development:

   • This project uses Devbox for development environment management and Poetry for dependency management.
   • Set up the environment using:

     devbox shell
     poetry install
     

2. Running the Tool:

   • The tool is started using a shell script that manages Docker containers.
   • Basic usage:

     ./create_changelog.sh [options]
     

   • Options:
     • --project: Specify the Redmine project ID
     • --version: Specify the version ID
     • --repo: Path to the Git repository (optional)
     • --type: Type of output (Changelog/ReleaseNotes)
     • --local: Use local LLM instead of OpenAI (optional)

• If an error occurs that the docker volume ollama-local cannot be found, create it with docker volume create ollama-local

3. Example Usage:

./create_changelog.sh -projects 123 -versions 456 -type Changelog -repo /path/to/repo

Dependencies

• Python 3.x
• Poetry for dependency management
• Devbox for development environment
• Docker for running LLM services
• External services:
• Redmine API
• Git repository
• OpenAI API (optional)
• Local LLM (optional, e.g., Ollama)

Configuration

The project uses a TOML configuration file named changelog-config.toml for storing essential settings. This file should contain the following entries:

[DEFAULT]
OPENAI_API_KEY = your_openai_api_key_here
REDMINE_API_KEY = your_redmine_api_key_here
REDMINE_HOST = https://your_redmine_host.com
GIT_REPO_PATH = /path/to/your/git/repo

Configuration File Location

The config.py script searches for the changelog-config.toml file in the following locations, in order:

• Current working directory
• changelog2 subdirectory of the current working directory
• '.config' subdirectory of the current working directory
• User's home directory
• Location specified by the CHANGELOG_CONFIG environment variable

Make sure to create the changelog-config.toml file with your actual API keys and settings before running the tool. Keep this file secure and do not commit it to version control.

However, a copy of changelog-config.toml should be located in the changelog2 subfolder for execution in the docker container.

Development

• To add new dependencies: poetry add <package_name>
• To update dependencies: poetry update

Notes

• The tool can use either OpenAI's models or a local LLM (like Ollama) for text generation.
• Ensure all necessary Docker containers are running before starting the tool.