Finding duplicated test cases in MozTrap https://moztrap.mozilla.org
#Installation
Run install.sh (for Ubuntu)
##Updated:##
Read tests/test_smoke.sh for usage
- Download test cases from Moztrap using
download_sample.sh - Manually mark the duplications as in
input/ground-truth-217.csv - Edit the
config.pyto add your new training data and perdict data files - Run
python finddup.py fitto generate a model (only usetrainLocalJson) - Run
python finddup.py cross-valto do cross validation on feature and model parameters (only usetrainLocalJson) - Run
python finddup.py perdictto create a model usingtrainLocalJson+groundtruth_filenameand use that model to make perdictions onperdictLocalJson. - If you need case ID and links, run
add_case_id_to_output.py, remember to modify the output file path in the script.
If you don't want to call the MozTrap API everytime, use wget/curl to download the this url: https://moztrap.mozilla.org/api/v1/caseversion/?format=json&productversion=217 as json file. And set the localJson variable to its filename. Then change downloadCasevesions() call to loadLocalCaseversion(localJson).
- Define what is a duplication
- Create manully marked groundtruth data
- Tune the features and parameters
- Prettify the network graph
Open a pull request for review. Or submit ideas to slyu@mozilla.org