Repo-to-Text-AI

Turn an entire Git repository into a single, organized text file designed for providing context to Large Language Models (LLMs).

repo-to-text-ai is a versatile command-line tool that intelligently scans a local repository and consolidates all relevant files into one output file. It's designed to be the perfect first step for pasting a complete project context into an AI prompt for code analysis, documentation generation, refactoring, or feature development.

The tool is smart, respecting .gitignore rules, and highly customizable, allowing you to fine-tune the included context to be as precise as you need.

Key Features

• Single File Output: Combines all relevant source files into one .txt file, ready for copy-pasting.
• Project Tree Summary: Automatically generates a directory tree at the beginning of the output for a high-level overview of the repository structure.
• Intelligent File Filtering:
  • Automatically respects rules in all .gitignore files, including those in subdirectories.
  • Skips binary files, the .git directory, and other meta-files.
• Highly Customizable Context:
  • Blacklist (.context_ignore): Use a .context_ignore file to exclude additional files or directories from the output without modifying your main .gitignore.
  • Whitelist (.context): For maximum precision, use a .context file to specify exactly which files and directories should be included.
• User-Friendly CLI: A simple and intuitive command-line interface with a progress bar for large repositories.

Installation

You can install repo-to-text-ai directly from the Python Package Index (PyPI):

pip install repo-to-text-ai

Usage

Basic Usage

Navigate to the root directory of the Git repository you want to process and run the command:

repo-to-text-ai .

This will create a context_output.txt file in the current directory containing the consolidated context.

Command-Line Options

Example with options:

# Run on a different project and save the output to a specific file
repo-to-text-ai /path/to/your/project -o my_project.txt

Customizing the Context

You have two powerful methods to control exactly what content goes into the output file.

1. The Blacklist Method: .context_ignore

This is the best method for excluding a few extra files or directories that are not in your main .gitignore.

Create a file named .context_ignore in the root of your repository. It uses the exact same syntax as a .gitignore file.

Use Case: You want to exclude test files or examples from the AI's context without adding them to your project's primary .gitignore.

Example .context_ignore:

# Exclude all test files from the AI context
tests/
examples/

# Exclude specific large data files
data/large_dataset.csv

2. The Whitelist Method: .context

This is the most precise method. If a .context file is found in the root of your repository, only the files and directories listed in it will be included. All other files will be ignored.

Create a file named .context in the root of your repository. List the paths you want to include, one per line.

Use Case: You are working on a specific feature and only want to provide the AI with context from two directories and one configuration file.

Example .context:

# This is a comment, it will be ignored

# Include the main application file and the entire 'utils' directory
src/app.py
src/utils/

# Also include the main configuration file
config/settings.yml

Note: Even when using a .context whitelist, the exclusion rules from your .gitignore and .context_ignore files are still respected. For example, if src/utils/ contains a temp.log file and your .gitignore excludes *.log, that file will not be included in the output.

Development

Interested in contributing? We use pytest for testing and ruff/black for formatting and linting.

Prerequisites

• Python 3.8+
• Docker (Recommended for a consistent, isolated environment)

Local Setup

1. Clone the repository:

   git clone https://github.com/your-username/repo-to-text-ai.git
   cd repo-to-text-ai
   

2. Create and activate a virtual environment:

   </code></pre>
   </li>
   </ol>
   <p>sh
   python3 -m venv .venv
   source .venv/bin/activate
   ```</p>
   <ol start="3">
   <li>
   <p><strong>Install in editable mode with test dependencies:</strong>
   This command links the package to your source files and installs development tools like <code>pytest</code>.</p>
   <pre><code class="language-bash">pip install -e ".[test]"
   

3. Run the tests:

   pytest