Code2Prompt - Convert Codebase into LLM Prompt - Install Locally

---

---

AI Nuggets

Installation and Usage of code to prompt CLI Tool

Installation of Rust and code to prompt

1. Install Rust:

   curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh  

   • Follow the prompts for the default installation.
   • After installation, either restart your shell or source the environment file to update your path.

2. Check Rust version:

   rustc --version  

3. Install code to prompt using Cargo:

   cargo install code_to_prompt  

Using code to prompt

1. Create a simple git repository for demonstration.

2. Generate a prompt from your code base:

   code_to_prompt <project_path>  

   • The prompt is automatically copied to the clipboard and can be pasted into a text file or any desired location.

3. To use custom templates from the code to prompt GitHub repository:

   • Clone the repository (URL provided in the video description).
   • Use the -t flag to specify a template:

     code_to_prompt -t <template_path> <project_path>  

Features and Tips

• Converts code base into a single LLM prompt.
• Respects .gitignore files.
• Can filter and exclude files using glob patterns.
• Displays the token count of generated prompts.
• Optionally includes git diff output in the generated prompt.
• Saves the generated prompt to an output file.
• Can exclude files and folders by name or path.
• Adds line numbers to source code blocks.
• Customizable prompt template.
• Templates for documentation, finding security vulnerabilities, and bugs are available in the GitHub repository.

Additional Information

• The video is sponsored by MK compute, which offers GPU rentals at affordable prices.
• A coupon code for a 50% discount on already discounted offerings is mentioned (the code itself is not provided in the transcript).
• The presenter is using an NVIDIA RTX A6000 with 48 GB of VRAM for the demonstration.

GitHub Repository and GPU Rental

• GitHub Repository: The link is provided in the video’s description.
• MK compute Website: The link is provided in the video’s description for GPU rental services.

Recommendations

• Instead of using Cargo to install the latest version, consider cloning the GitHub repo and building from there to get the latest, less buggy release:

  git clone <repo_url>  
  cd <cloned_repo_directory>  
  cargo build --release  

Conclusion

• The presenter finds code to prompt to be a useful tool for documentation and code analysis.
• Encourages viewers to subscribe to the channel and share the content.

(Note: The exact URLs for the GitHub repository and MK compute website are not provided in the transcript and should be obtained from the video’s description as mentioned.)