Specification mining is fundamental to verifying area-specific code invariants with runtime verification. However, the computational overhead increases monstrously for non-trivially sized repositories (> 100 MB). Similar problems are found in runtime monitoring, for which evolutionary-aware tools have been developed in an attempt to ameliorate complexity. However, in the context of specification mining, current evolution-aware analysis just compounds complex- ity further, due to the need for re-mining specifications at each evolution in a repository’s lifespan. We propose the first model for an incremental specification mining pipeline (ISM). By making use of ISM, specification miners can reap the benefits from an evolu- tionary aware system while getting rid of unnecessary overhead. We also present alternative ideas for integrating ISM into existing specification mining framework
The Dockerfile can used to build locally. This docker image takes a rather long time to build. A pre-built image is available here.
You can run
docker pull cyankaet/specminer
docker run -it -v shared:/usr/src/app/run-spec-miners/ws cyankaet/specminerto create an interactive session. This will plop you into the main folder to run
javert.sh, bddminer.sh, and dsm-manual.sh at will.