This artefact contains all evaluation resources related to the paper: Tileria, Marcos, Jorge Blasco, and Santanu Kumar Dash. “DocFlow: Extracting Taint Specifications from Software Documentation.” in 2024 IEEE/ACM 46th International Conference on Software Engineering ICSE, Lisbon, Portugal, 2024 pp. 718-729. The following resources are included: - Docflow source code, our tool to generate taint specifications. - notebook to reproduce the paper’s result. - Script to run standalone operations such as taint specification or method classification. - Raw and processed Android AOSP and Google Play Services documentation.

The artefact is based on a docker file and should run on any OS that supports this technology. People evaluating this artefact should be able to use docker, python, and preferably some experience using machine learning packages for python. More instructions can be found in the readme file.

We submit this artefact to claim the available and reusable badge. Link for permanent repository: https://archive.softwareheritage.org/browse/origin/directory/?origin_url=https://gitlab.com/s3lab-rhul/android/docflow.git