ggrab 0.1.0a6

A CLI tool for extracting and sharing code files and functions. A manual context builder for pasting into llms.

ggrab - Grug Grab Tool

A CLI tool for manually building context to paste into llm chats. ggrab gathers functions, classes, or entire files wraps them with xml and copies it to the clipboard.

Key Features

• Smart Extraction: Grab entire files or specific functions/classes
• Intelligent Autocompletion: File paths and function/class names
• Token Counting: Built-in token counting for AI context limits
• Language Support:
  • Python (AST-based extraction)
  • JavaScript (Esprima-based extraction)
  • Other files (full file extraction)

Quick Start

1. Install:

pip install ggrab    # Basic install
# or
pipx install ggrab   # Recommended

Why pipx? The pipx install is recommended because it:

• Won't interfere with other Python projects (creates its own venv)
• Makes ggrab available globally in your terminal

2. Setup:

Add the following to your shell config file (~/.bashrc, ~/.zshrc, or equivalent):

# Enable ggrab autocompletion
eval "$(register-python-argcomplete ggrab)"

Then reload your shell configuration:

# For bash
source ~/.bashrc

# For zsh
source ~/.zshrc

Usage

# Grab entire files
ggrab file.py

# Extract specific functions
ggrab file.py function1 function2

# Mix multiple files and functions
ggrab file1.py func1 func2 file2.js func3

Press TAB at any time for smart autocompletion:

• After ggrab: Shows file suggestions
• After selecting a file: Shows available functions/classes

Examples

# Extract multiple functions from a Python file
$ ggrab utils.py parse_data validate_input
Gathered code from:
  - utils.py, functions: parse_data, validate_input (203 tokens)

===== Combined Snippets (copied to clipboard) =====

# Extract from multiple files
$ ggrab api.py handle_request db.js queryDatabase
Gathered code from:
  - api.py, functions: handle_request (156 tokens)
  - db.js, functions: queryDatabase (178 tokens)

Total tokens across all files: 334

Performance Features

• Lazy Loading: Heavy dependencies load only when needed
• Smart Caching: Function/class listings cached for quick access
• Hybrid Parsing: Fast regex for completion, accurate AST/Esprima for extraction

Requirements

• Python 3.7+
• Required: argcomplete
• Optional but recommended:
  • pyperclip (clipboard support)
  • tiktoken (token counting)
  • esprima (JavaScript parsing)

Advanced Setup

Disable completion sorting (optional):

complete -o nosort -F _python_argcomplete ggrab

For zsh users:

autoload -U compinit && compinit
autoload -U bashcompinit && bashcompinit
eval "$(register-python-argcomplete ggrab)"

Acknowledgments

• Developed -- ok, prompted into existance by Claude 3.5, o1 pro. Fixed up to working order by Keizo Gates making GrugNotes :)

License

MIT License

---

For bug reports and feature requests, please open an issue on GitHub.