Code-Promptify is a command-line utility for aggregating code repository contents into a single string based on include and exclude patterns. It also provides metadata information, such as token count and billable characters. It is designed to help you prepare prompts for large language models (LLMs). The output is written to a file and also copied to your clipboard automatically.

Note: To count tokens, this tool uses the OpenAI tiktoken tokenizer library and should give you accurate results for most GPT models. For other models (Gemini, Llama, Mistral, etc.) the token count may not be 100% precise but it will be a good approximation. That is because the exact token count depends on the specific tokenizer used by the model, which differs among models.

See example_output.md for what the output looks like for this repo when running with default settings.

Example metadata output:

• Aggregate file contents based on include and exclude patterns (glob strings)
• Automatically ignores virtual environment files
• Option to ignore empty files
• Output results to a file and clipboard

To install promptify, simply run:

pip install code-promptify

To install Promptify, follow these steps:

1. Clone the repository:

   git clone https://github.com/vmehmeri/promptify.git
   cd promptify
   

2. Install the package and its dependencies:

   pip install -e . -r requirements.txt
   

After installation, you can run Promptify using the following command:

promptify [args]

• --include: File patterns to include (default: ["*.py", "*.html", "*.js", "*.css", "*.json", "*.yaml", "*.txt", "*.md"])
• --exclude: File patterns to exclude (default: ["*.pyc", "*egg-info*", "*tmp*"])
• --ignore-empty: Flag to Ignore empty files (default: False)

1. Use default settings:

   promptify
   

2. Include only Python and JavaScript files:

   promptify --include "*.py" "*.js"
   

3. Exclude test files and ignore empty files:

   promptify --exclude "*test*" --ignore-empty
   

See example_output.md

• pyperclip
• tabulate
• token-count

Contributions are welcome! Please feel free to submit a Pull Request.

This project is licensed under the MIT License - see the LICENSE file for details.