Plumber

A Python-based CSV manager and manipulator with natural language interface.

Overview

Plumber is a command-line tool that allows you to manipulate CSV files using natural language commands. It uses a local language model to interpret your requests and perform operations like merging, splitting, filtering, and transforming CSV data.

Features

• Natural Language Interface: Describe what you want to do in plain English
• Flexible CSV Operations: Merge, split, filter, count, transform CSV files
• Local LLM Integration: Uses llama-cpp-python with .gguf model files
• Automatic File Discovery: Automatically finds CSV files in the input directory
• Rich Output: Beautiful terminal output with status information

Installation

1. Clone this repository
2. Install dependencies using uv:

   uv sync

Usage

Interactive Mode

Start the interactive CLI:

uv run python -m plumber

Then describe what you want to do:

• "merge all CSV files"
• "split data.csv into 3 parts"
• "count unique values in the name column"
• "filter rows where age > 25"

Other Commands

• Show status: uv run plumber status
• List CSV files: uv run plumber list-files
• Model help: uv run plumber model-help

Setting Up Models

For full natural language processing, download a compatible .gguf model:

1. Install huggingface-hub: pip install huggingface-hub
2. Download a model:

   huggingface-cli download TheBloke/Llama-2-7B-Chat-GGUF llama-2-7b-chat.Q4_K_M.gguf --local-dir models/

3. Update the model name in the configuration if needed

Directory Structure

plumber/
├── input_csvs/     # Place your input CSV files here
├── output_csvs/    # Generated output files appear here
├── models/         # Place .gguf model files here
├── plumber/        # Main application code
└── tests/          # Test files

Development

Running Tests

uv run pytest

Linting and Formatting

./lint.sh