Detecting Code Clones with Graph Neural Network and Flow-Augmented Abstract Syntax Tree

Code for paper "Detecting Code Clones with Graph Neural Network and Flow-Augmented Abstract Syntax Tree", SANER 2020
Requires:
pytorch
javalang
pytorch-geometric

Data

Google Code Jam snippets in googlejam4_src.zip
Google Code Jam clone pairs in javadata.zip
BigCloneBench snippets and clone pairs in BCB.zip

Running

Run experiments on Google Code Jam:
python run_java.py
For BigCloneBench:
python run_bcb.py

This operation include training, validation, testing and writing test results to files.

Arguments:
nextsib, ifedge, whileedge, foredge, blockedge, nexttoken, nextuse: whether to include these edge types in FA-AST
data_setting: whether to perform data balance on training set
'0': no data balance
'11': pos:neg = 1:1
'13': pos:neg = 1:3
'0'/'11'/'13'/+'small': use a smaller version of the training set