Add Jaccard as a fourth SBFL ranking metric

[Sonzakir]

Implements the standard Jaccard similarity coefficient, ef / (ef + nf + ep),
alongside Tarantula, Ochiai, and DStar.

• New MetricJaccard class following the existing metric interface, with the
  shared epsilon guard against division by zero
• Registered in RankingMetricManager; added Jaccard column + index to the
  Score table and included it in both ranked-entity queries

No downstream changes needed, reporting iterates the metrics generically, so
Jaccard gets its own score table and report file automatically.

[mohrez86]

Hi @Sonzakir ,

Thank you for your contribution. Adding Jaccard to SBFL is actually a nice idea, and I really appreciate how you cleanly adapted the codebase's style. I am a bit tied up at the moment, but I will run some end-to-end tests asap and merge the PR once everything looks good.

[Sonzakir]

Hi @mohrez86 ,

Thank you so much for the quick response and for this excellent tool. FauxPy has been incredibly useful for the Repair Framework I’m currently working on.

In addition, I ran into a few bugs when trying to run FauxPy in bug-isolated containers. I seem to have resolved them for now, but I haven’t tested the framework thoroughly enough yet. Once I’ve finished reviewing everything on my end and the tests pass, I’ll open a pull request for that as well.

[mohrez86]

@Sonzakir ,

Thank you for the feature! Merged.
I am glad to hear that FauxPy has been useful for the repair framework you are working on.
Please do open that next PR once you have finished testing.