Graphical Code Tracer

Graphical Code Tracer (gct) is the world's first visual static code analyzer.

Within seconds it can tell you how your functions and classes are connected to each other!

GCT web app • Getting started • Installation • Configuration

Getting started

Usecases:

1. Onboard to new codebases faster.
2. Debug code faster Twitter thread
3. Create share-able versions of your code. Using gctpy.com, you can instantly share your UML diagrams across teams.

Some examples you can run GCT on:

Simple python file • karpathy/MinGPT • scikit-learn/cluster • geohot/tinygrad • PyTorch/autograd • Flask Web App

Generate graph for any python file

python -m gct -i path/to/file.py # run gct on a local file
python -m gct -i https://github.com/user_name/path/to/file.py # run gct on a file hosted on a web server

Installation

Step 1: Install GCT Python package

pip install gct-py

Step 2: Install Graphviz executable

GCT generates graphs using graphviz. To get accurate graphs, we highly recommend using latest graphviz version available for your OS.

Skip this step if you've already installed graphviz executable version. Check the dot version by running: dot -V.

Windows

Install graphviz by downloading executable (version >=6.0.1) from graphviz website.

MacOS

(Optional) more details: graphviz.

  brew install graphviz

Linux

(optional) more details: graphviz.

  sudo apt install graphviz

Configuration

Customize the experience by aliasing gct:

alias gct='python -m gct -i'
gct path/to/file.py