↑
logo

Strand- dependency/impact analysis on instant-open, single-file graphs

Posted by B3333 |an hour ago |1 comments

B3333 an hour ago

Strand allows you to create a dependency graph within a project and output it as a file. It supports many different languages (Rust, Python, JavaScript, TypeScript, Go and Java), by parsing existing files/lockfiles or source files.

As the graph is contained within a single file, very large projects with tens of millions of connections can be loaded very quickly. No database, backend servers or waiting times required.

Why use this?

Compared to graph databases (Neo4j)

No setup for servers, and no user accounts. No complex querying language needed. Just a file that can be shared, compared or added into a project.

Compared to visualisation tools (Gephi / yEd).

Strand allows for both visualisation, and automated validation. For example, when making changes, if a small change breaks hundreds of other files in your project, you will be alerted before saving.

Compared to language-specific tools (dependency-cruiser).

Most tools for creating dependency graphs only support JavaScript, and perform slowly on large projects. With strand, you can support multi-language projects, and it remains quick even for very large projects.

Compared to spreadsheet/database.

Strand has been structured to allow quick investigation of dependency chains. This means that it can answer questions like "what breaks if I change this?" very quickly.

Additionally, there is a desktop application which allows you to view your project structure, detect cycles in dependencies, and edit the dependency graph directly.

Installable using Cargo (cargo install strand-cli or cargo install strand-gui). Open source under Apache-2.0 license.

I'd love to know what you think! In particular, do you think having a single file containing the dependency graph would be useful in larger teams? Or would people prefer hosted solutions?