Closed SophieHYe closed 6 years ago
Hi @SophieHYe Great, but I have a question: How do you classify the patches from the folder "patches" (i.e., all patches reported by the publications) between "correct" and "incorrect (plausible)". Do you use a script?
Hi @martinezmatias,
I did not use a script. I collected the " correct" patches from the corresponding paper which claims their tool correctly generate patches after manual analysis the patch correctness. For the rest patches are not claimed as "correct", I will put them in the "incorrect" patches. WDYT?
Hi @SophieHYe
So the separation of patches into two folders are based on the classifications done by the papers's authors. However, It did not get the finality of splitting the patches in two folders. I am probably missing something. Is there any experiment you need to carry out that needs the two folders?
I mean, I afraid that a misclassification of a patch correctness (from part of the original author's papers) creates a confusion in the organization of our data. I would thing that we could have all patches in one folder (our analysis dataset) and one table where rows are patches and columns are correctness assessment methods (the first columns could be "Manual assessment done by the paper' authors "). (i.e., a way of splitting row data -patches- from meta-data -patch classification-) WDYT?
decision:
Hi @martinezmatias I have updated the patches folder name from patches to -> correct_patches as I would also like to collect the plausible patches to perform the experiment. I have combined these two folders, could you pls check if they are correct, then we could remove patches folder:) 👍
Thanks a lot! I have script to run the ASE15 tests, will update the main table soon :)