Published May 5, 2023 | Version v1

Dataset and Source Code for ASE23 #296

Authors/Creators

Description

The dataset including the train, valid, and test data is in Dataset.zip. The crawled script is shown in atcoder_crawl_server.py.

The graph construction, optimization and linear transformation code are in graph_construct_transform.py.

The training process of the pre-trained model with contrastive learning are shown in model.py and train.py.

The commands for Joern to generate PDG:

This command generates a cpg.bin file in the current directory for the code in the test_code folder:
joern-parse /home/test_code/
This command processes the generated cpg.bin file and outputs the corresponding DOT files in the outdir_java folder:
joern-export cpg.bin --repr pdg --out outdir_java
This command converts the corresponding DOT file to a PNG file, which represents the PDG of the source code:
dot -Tpng -o 0.png 0-pdg.dot

Files

Dataset.zip

Files (146.5 MB)

Name Size
md5:b5bb408f01fbc223ad5338affd6f2fd6
3.7 kB Download
md5:d7da347e6191a84aa17f10afbceb2674
146.5 MB Preview Download
md5:a43fdbaa1d57918c8aae360bf15e7182
3.9 kB Download
md5:d93d7844d1e296aefa326356862f648f
2.8 kB Download
md5:44cf5228abced1eee984603f1396011a
14.2 kB Download