Skip to main content

LASSO

A Large-Scale Software Observatorium for the Dynamic Selection, Analysis and Comparison of Software

Explore Pipelines 🧪5 Minute Tutorial 📖What is LASSO ❔
Explore LSL Pipelines
See LSL Study Pipelines in Action
Go to TDSEHub, or Search/Generate Code
TDSEHubSearchGenerate

Recent Blog Posts

4/14/2025

Presentation Slides - ICST 2025

3/31/2025

Preview of LSLFlow

2/17/2025

Preview of Arena Test Driver for Python


Try LASSO in 5 Minutes

To get started, run the following two commands in a local directory on your machine (requires docker compose):

curl https://raw.githubusercontent.com/SoftwareObservatorium/lasso/refs/heads/develop/docker/compose/docker-compose-embedded.yml -o docker-compose.yml
docker compose up

Wait until all services started (LASSO platform, Code Search Index and Artifact Repository) and then open LASSO's dashboard at http://localhost:10222/webui/ (login: admin / admin123). See 5 Minute Tutorial for details, and Labs (Playground) for more options.

Key Features of LASSO

Test-Driven Software Experimentation with LSL

Utilize LSL (LASSO Scripting Language) for customizable code analysis services, including Code Search and Generation (via LLMs), Test Generation, and Study Designs, enabling reproducible research in software engineering.

Data Structures - Storing Observations for Big Code

LASSO's core building blocks include: Use Sequence Sheets (SSNs) as test specifications, Stimulus Response Matrices (SRMs) for configurations of tests and implementations, and Stimulus Response Hypercubes (SRHs) for software analytics of behavior.

Behavior-aware Software Analytics

Conduct extensive, behavior-aware software analytics based on observations stored in SRHs in your favorite data analytics tools. This includes establishing functional correctness across code candidates as well as behavioral clustering based on observed behavior (i.e., outputs).

Open, Extensible Platform that Fosters Sharability

The platform is extensible. Extend it with your own tools and techniques via well-defined APIs and state-of-the-art tools (i.e., containerization). The platform and its data complies with open research principles and fosters sharability of pipelines and results.