Graphviz (Python)

The graphviz package gives Python programs a thin, ergonomic wrapper around the Graphviz graph-drawing software’s DOT language. Callers build a graph object, add nodes and edges through a small API, and either retrieve the generated DOT source as a string or render it to PDF, PNG, SVG, and other formats using a locally installed Graphviz toolchain. It also integrates with Jupyter notebooks and the Jupyter QtConsole, displaying rendered graphs inline during interactive exploration.

What You Get

• Graph and Digraph classes for building undirected and directed graphs via add-node/add-edge calls
• Automatic DOT source generation, retrievable as a string or written straight to a file
• One-call rendering to PDF, PNG, SVG, and other formats using the system’s installed Graphviz binaries
• Built-in Jupyter notebook and QtConsole integration for inline graph display during interactive work
• A view option/method that opens the rendered file in the OS’s default viewer application

Common Use Cases

• Visualizing data structures, dependency graphs, or state machines directly from Python code
• Generating architecture or pipeline diagrams programmatically as part of documentation tooling
• Rendering graphs inline inside Jupyter notebooks for exploratory data analysis or teaching
• Producing DOT source files for downstream processing by other Graphviz-based tools

Under The Hood

Architecture The package is organized around Graph and Digraph classes in the graphviz module that accumulate node/edge statements and serialize them to DOT syntax, then shell out to the system’s installed dot executable to perform actual rendering. Tech Stack It’s pure Python (3.10+) with no runtime dependencies beyond the standard library, built with setuptools, tested with pytest via run-tests.py, and requires a separate system-level Graphviz installation for rendering to work. Code Quality The repo maintains a detailed CHANGES.rst, uses Codecov for coverage tracking, runs a lint-code.py and typecheck-code.py pass in CI, and includes doctest-driven examples embedded directly in the README. API Design The Graph/Digraph builder pattern closely mirrors DOT syntax itself, so anyone with basic Graphviz familiarity can pick it up quickly, and the view convenience method removes the need to manually open rendered output files.