search

mfejzer / tracking_buggy_files

Tracking Buggy Files: New Efficient Adaptive Bug Localization Algorithm
MIT License
6 stars 6 forks source link

"views" and "bug_fixing" missing field in aspectj.json #8

Open hangdj opened 3 years ago

hangdj commented 3 years ago

On line 115 of retrieve_features_5_6.py, it needs to find the "views" and "bug_fixing" fields in bug_report, but they are not in the aspect.json provided by the author( i get it from https://mat.umk.pl/mfejzer/aspectj_data.tar.bz2). It is not clear whether the aspectj.json provided by the author is the final processed version, so how can I solve this problem?

Do i need to execute these commands？But i didn't find code which geneate such fields and Where is the Aspectj.xml used by the author downloaded from? ./process_bug_reports.py AspectJ.xml ../tracking_buggy_files_aspectj_dataset/ aspectj_base.json ./fix_and_augment.py aspectj_base.json ../tracking_buggy_files_aspectj_dataset/ > aspectj_aug.json ./pick_bug_freq.py aspectj_aug.json ../tracking_buggy_files_aspectj_dataset/ > aspectj.json

Hope to get help from the author, thank you.

mfejzer commented 3 years ago

Hello,

I downloaded the original AspectJ.xml files and other xml files from website which contained the data from Xin Ye et al publications. Probably it was from http://figshare.com/articles/The_dataset_of_six_open_source_Java_projects/951967 or one of its mirrors. I've attached backup copy of AspectJ.xml and other xml files which I found on my disk xmls.zip

If aspectj.json does not contain "views" key with the data, it needs to be then processed but the scripts adding it like "pick_bug_freq". The data in https://mat.umk.pl/mfejzer/aspectj_data.tar.bz2 and related tar.bz2 archives contains partial results to prepare final data - so should work with following scripts even with partial (lacking some filelds) aspectj.json: ./load_data_to_joblib_memmap.py aspectj ./train_adaptive.py aspectj

Sorry for any inconvenience.

hangdj commented 3 years ago

Hello,

I downloaded the original AspectJ.xml files and other xml files from website which contained the data from Xin Ye et al publications. Probably it was from http://figshare.com/articles/The_dataset_of_six_open_source_Java_projects/951967 or one of its mirrors. I've attached backup copy of AspectJ.xml and other xml files which I found on my disk xmls.zip

If aspectj.json does not contain "views" key with the data, it needs to be then processed but the scripts adding it like "pick_bug_freq". The data in https://mat.umk.pl/mfejzer/aspectj_data.tar.bz2 and related tar.bz2 archives contains partial results to prepare final data - so should work with following scripts even with partial (lacking some filelds) aspectj.json: ./load_data_to_joblib_memmap.py aspectj ./train_adaptive.py aspectj

Sorry for any inconvenience.

Yes,Thank you for your help ,I have obtained the final result through these scripts(./load_data_to_joblib_memmap.py aspectj > ./train_adaptive.py aspectj) and partial aspectj.json.

When I run this script (pick_bug_freq) to generate new aspcetj.json with these missing fields,I have encountered so many problems(mainly the environment problems) that I cannot carry out the following experiment.I don’t know if you still have files with these fields，If convenient, can you send me a copy to make it easier for me to carry out the following experiment.If not,I will paste the error for further help.

I truly appreciate your timely help.

mfejzer commented 3 years ago

I cannot find the backup copy of this file, please provide the errors and we will try to fix the code.

ph0810 commented 2 years ago

can you public code train_adptive using decision tree. Thank you!!

mfejzer commented 2 years ago

Hello, we are not adding new methods to the existing algorithm, as this repository is for replication of paper results.