Skip to main content

CLI tool to summarize code in a repo.

Project description

👻 repoGhost

repoGhost is a command-line tool to scan a local code repository, split files into chunks, and summarize each chunk using an LLM (e.g., GPT-4o). Summaries are stored in a dedicated summary directory, and repeated runs skip unchanged files to save cost.

Features

  • Hash-based caching: Skips unchanged files (no repeated LLM calls).
  • Auto .gitignore: Automatically adds the summary directory to .gitignore if found.
  • Dedicated Summary Directory: Creates a summary folder for all outputs.
  • Clipboard: Copies the last summary to your clipboard for easy reference.
  • Configurable chunk size: Choose how many lines per chunk.
  • Repository Map: Generates a hierarchical view of your repository structure at the top of the summary.
  • CWD Defaults: Defaults to analyzing the current working directory if no path is specified.

Installation

pip install repoGhost

Usage

Simply run in your project directory:

repoGhost

Or specify a different repository path:

repoGhost /path/to/your/repo

Parameters

  • --repo_path: Optional path to the local repo (defaults to current directory).
  • --lines_per_chunk: Lines per chunk for summarizing (default 30).
  • --context-size, -c: Controls the snippet size extracted from files. The default value is 8192 characters, but users can set it higher up to a maximum of 32768. If an unsupported value (e.g., negative or too large) is supplied, the application exits with an error message.
  • --api-key: Provide the OpenAI API key. If not provided, the tool checks the OPENAI_API_KEY environment variable or the config file at ~/.repoghostconfig.json.

Example

# Analyze current directory
repoGhost

# Analyze specific directory with custom chunk size
repoGhost /path/to/project --lines_per_chunk 50

# Analyze specific directory with custom context size
repoGhost /path/to/project -c 16384

This generates a summary directory containing:

  • hash_cache.json: Contains file hashes and chunk summaries (used to skip unchanged files).
  • summaries.json: Contains all chunk summaries (the final output).

The last chunk’s summary is copied to your clipboard automatically.

API Key Configuration

The tool will look for your OpenAI API key in the following order:

  1. Via the command-line argument --api-key.
  2. Via the environment variable OPENAI_API_KEY.
  3. Via a configuration file located at ~/.repoghostconfig.json.

If no API key is found, the tool will exit with an error message. This allows you to avoid passing the API key every time you run the command.

Constants

In the code, the following constants can be modified to suit your needs:

EXCLUDED_DIRS = {
    "migrations",
    "static",
    "media",
    "__pycache__",
    ".git",
    "venv",
    "node_modules",
    "summary",  # Excludes the summary directory itself
}
EXCLUDED_EXTENSIONS = {
    ".pyc", ".css", ".scss", ".png", ".jpg", ".jpeg", ".svg", ".sqlite3"
}
EXCLUDED_FILES = {"manage.py", "wsgi.py", "asgi.py", "package-lock.json"}
VALID_EXTENSIONS = {".py", ".js", ".html", ".json"}

Feel free to add or remove items based on the files you want to skip or process.

Customizing the Prompt & Model

Inside the script, the summarize_chunk function calls an OpenAI model:

openai.ChatCompletion.create(
    model="gpt-4o",  # or your custom model name
    messages=[
        {"role": "user", "content": f"Please summarize this code chunk concisely:\n\n{chunk}"}
    ],
    temperature=0.1,
    max_tokens=1000
)

You can modify:

  • The model parameter (e.g., gpt-3.5-turbo, gpt-4, gpt-4o-mini, etc.).
  • The prompt text (if you want a different style of summary).
  • The temperature or max_tokens values.

OpenAI API Key

You need an OPENAI_API_KEY set in your environment variables for the script to call OpenAI’s API. For instance:

export OPENAI_API_KEY="sk-1234..."

Then run:

repoGhost --repo_path /path/to/repo

Requirements

See requirements.txt. Python 3.7+ recommended.

  • openai
  • pyperclip
  • rich

Development / Local Install

  1. Clone this repo:
    git clone https://github.com/georgestander/repoGhost.git
    cd repoGhost
    
  2. Install in editable mode:
    pip install -e .
    
  3. Verify the CLI is installed:
    repoGhost --help
    

License

MIT License.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

repoghost-0.2.4.tar.gz (11.5 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

repoGhost-0.2.4-py3-none-any.whl (10.3 kB view details)

Uploaded Python 3

File details

Details for the file repoghost-0.2.4.tar.gz.

File metadata

  • Download URL: repoghost-0.2.4.tar.gz
  • Upload date:
  • Size: 11.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.8.18

File hashes

Hashes for repoghost-0.2.4.tar.gz
Algorithm Hash digest
SHA256 28025a3c8d2706c057ec45647b55873e6903c742d70dd75aa0c2f315157bf530
MD5 dd5bc0309474c662f13e88edd6989afa
BLAKE2b-256 b3a455523075286e9aedecaa130fbfa2276b62b3448d27e67a7de760653c8881

See more details on using hashes here.

File details

Details for the file repoGhost-0.2.4-py3-none-any.whl.

File metadata

  • Download URL: repoGhost-0.2.4-py3-none-any.whl
  • Upload date:
  • Size: 10.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.8.18

File hashes

Hashes for repoGhost-0.2.4-py3-none-any.whl
Algorithm Hash digest
SHA256 9475adc8896d2811f3832dbd7105e59bc565d774bfa33053ea27d10ea377e594
MD5 e2db6a5994209fdcb69a98efe6398db0
BLAKE2b-256 1b33e01e68cd53183fabea9dd3fea3820c66f891823772fc126365df831793b5

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page