Python 3.9.0
For parse.py, create a credentials.json file with the following:
{
"gh_user": "Github Username",
"gh_token": "API_token"
}
search.py
search.py [csvFile] [resultsFolder] > saveFile.txt
- csvFile: Downloaded from https://seart-ghs.si.usi.ch/
- resultsFolder: Path to folder where everything will be saved
- saveFile: Pipe output to file
This script first take the dump csvFile of GitHubSearch containing most popular python projets.
Then it sorts them by interesting features and creates a top list of project to consider.
Then it calls SourceGraph searching for "flaky" in those projects' commit messages.
Then, it saves queries needed to get results in "saveFile.txt"
sh saveFile.txt
Execute all queries 1 by 1.
Save response as sourcegraphsearch.json in the corresponding folder
parse.py
parse.py [resultsFolder]
Script that goes through all folders in resultsFolder
A folder is a project containing a sourcegraphsearch.json
In this file, there is a list of "Results", match for "flaky" in any commit messages
This script will query the GitHub API for all commits containing "flaky" and save JSON response in the project folder
commitAnalysis.py
commitAnalysis.py [resultsFolder]
This script takes the path of the resultsFolder
It analyses all commits containing "flaky" keyword
Playing aroung fix kerwords, attempts to find fixes in test and/or code