I’m happy to share DependencyAtlas.jl, a software package for turning Julia codebases into explorable dependency graphs.
DependencyAtlas focuses on static and compiler-time source analysis. It builds graph views from three engines: DepAtlasSource (source structure), StaticLint (semantic tooling), and JET (compiler-time evidence), so you can inspect a project as connected methods, files, and modules instead of reading it only as a directory tree.
Existing Julia packages already cover important parts of this space. StaticLint.jl, CSTParser.jl, and LanguageServer.jl support static and binding-aware analysis; JET.jl, SnoopCompile.jl, and Cthulhu.jl provide compiler-time and inference-oriented inspection; and ProfileView.jl, PProf.jl, and TraceFuns.jl target runtime behavior. The gap is that these tools usually expose only one layer of evidence at a time through editor, REPL, or profiler-specific workflows, rather than through a unified project-level dependency model. DependencyAtlas aims to fill that gap by combining source structure, semantic evidence, and compiler-time call evidence into a unified graph interface with method, file, and module views, bounded reachability, overload-site inspection, and source previews.
What it does
DependencyAtlas can help you:
- explore method, file, and module dependency views
- inspect callers, callees, and bounded reachability
- preview source snippets and overload sites
- surface Base/stdlib and external package calls
Why you may want it
When trying to understand a Julia codebase, you often want answers like:
- which functions or files call a particular method?
- what does the full call chain of a particular function look like?
- which files define the structure of this subsystem?
- what does the file dependency chain of the project look like?
Julia already has many excellent tooling components, but this software package brings these relationships together in one place through an interactive graph UI, making them easier to inspect as connected structure.
Tech Stack
DependencyAtlas is a mixed-language project.
Backend: Julia
Frontend: TypeScript, React, Vite, React Flow, Tailwind CSS
Docs: Docusaurus, React, TypeScript
Documentation
Website:
Repository:
Screenshots
File dependency graph
Node path highlighting
Reachability mode (dark theme)
