CP-Detector: Using Configuration-related Performance Properties to Expose Performance Bugs
To run the property suggester,
Install requirements:
pip install -U prefixspan t
pip install spacy
pip install https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-2.2.0/en_core_web_sm-2.2.0.tar.gz
and check all directory name in the three pyhon script fit the envirnment
Generate universal unique id for each normalized word.
e.g., (NN, Resource) --> 58 (NN, Volume) --> 89 etc.
python MapWordAbstraction2UniqueNodeID.py
- Normalizing Words
- Mining Association Rules (CARs)
- Calculate the quality for each (CARs)
python Preprocessing & Mining.py
This will generate 10^6 - 10^7 CARs, we save it to disk (FreqSeq
), and the quality of each CAR (S_C_L
, stands for support, confidence, lift).
- Selecting Optimal rules.
- build a voting classifier
python OptimalSelection & Classification.py
you can choose to use the Random-N
or Top-N
by changing RandomNOrTopN
to 1
or 0
.
The result will be shown in python console, or we provide simple visualization code to help you.