The artifacts present in our paper, “It’s Not a Feature, It’s a Bug: Fault-Tolerant Model Mining from Noisy Data”, published in the research track of ICSE 2024 are composed of the PMSAT-Inference algorithm, benchmarking sets, use case studies, examples and a diverse set of scripts and wrappers to produce the models and analyze the results.

The repository allows for the usage of our algorithm for model mining from noisy execution traces, in code or directly with our scripts. Additionally, it also contains all information/data required to verify and reproduce the results presented in our paper, including results, models, traces, plots and tables.

We applied for the Artifacts Available and Artifacts Reusable badges and provide detailed step-by-step instructions to reproduce all data, usage, setup and details on where to find everything in our README.

We provide a Docker setup to replicate our exact environment. We tested the Docker setup both on Ubuntu 22.04 and on WSL on Windows 10 and provide instructions for both in our README. Reviewers should have a passing familiarity with command line usage, e.g. bash, as there is no graphical interface for Docker and commands are expected to be run from the shell. Docker was very easy to install and use on Ubuntu and somewhat more tricky on Windows, if both operating systems are available we recommend Ubuntu due to the very easy Docker installation process.