🎄 Leveraging Graphviz for Programmatic Visualization in Software Engineering 🌟

Dec 4, 2023

Back to 2023 Advent Calendar

Graphviz, a robust visualization tool, is not only useful for static diagram creation but also excels in dynamic, programmatically generated visualizations.

On day 2 of Advent, we already talked about MermaidJS, a tool for generating diagrams from text. However, we recommend to use Graphviz inside scripts and command line, often when you just want to understand a data structure or a process.

And MermaidJS is better at web-based applications, where you already know what you want to visualize.

Let's go over couple examples where Graphviz would be a great fit.

1. Dynamic Dependency Analysis

Consider a Python script that parses the source code to extract dependencies between classes or modules.

As developers commit new code, this script runs, updating the Graphviz-generated dependency graph. This real-time visualization helps in identifying complex dependencies and architectural drifts.

digraph G {
  "ClassA" -> "ClassB";
  "ClassB" -> "ClassC";
  "ClassC" -> "ClassA";

This markup describes a simple circular dependency between three classes, showcasing the relationship visually.

2. Tracking Service Call Dependencies in Microservices

A tool can be developed to parse the access logs of various microservices. This tool identifies which services are calling others and how frequently. Graphviz is then used to create a diagram representing these interactions, providing a clear view of the service mesh and its dependencies.

digraph G {
  "ServiceA" -> "ServiceB" [label="calls: 10/day"];
  "ServiceB" -> "ServiceC" [label="calls: 5/day"];
  "ServiceC" -> "ServiceA" [label="calls: 2/day"];

In this markup, directed edges between services are annotated with call frequencies, offering insights into the communication patterns and potential bottlenecks.