SYNC

Implementation for SYNC: Structurally guided Hard Negative Curricula for Generalizable Neural Code Search

Environment

To recreate our computational environment we provide a YML file export of our conda environment (py3.7.yml)

conda env create -f py3.7.yml

Data:

You can download all the files from here: https://drive.google.com/drive/folders/1LFm1lrjKxYHAm9Dq7fKx8m0ZNtRv88wI?usp=sharing (put in data folder).

and

triple files from here: https://drive.google.com/drive/folders/18796d0_F9hpv6sV7-2RBQA4DMMVxLlVL?usp=sharing (put in triples folder).

Download the pre-generated negative samples for SYNC () and DISCO ().

Running Instructions:

Running instructions for training models and baselines.
(NOTE: for the <model_name>_ood.sh scripts just adjust the experiment folder path or -en argument)

Baselines:

Simple baselines:

nBOW:

bash scripts/train_nbow.sh
bash scripts/nbow_ood.sh

RNN:

bash scripts/train_rnn.sh
bash scripts/rnn_ood.sh

CNN:

bash scripts/train_cnn.sh
bash scripts/cnn_ood.sh

Regular triplet training:

CodeBERT:

bash scripts/codebert.sh
bash scripts/codebert_ood.sh

GraphCodeBERT:

bash scripts/graphcodebert.sh
bash scripts/graphcodebert_ood.sh

UniXcoder:

bash scripts/unixcoder.sh
bash scripts/unixcoder_ood.sh

DNS:

CodeBERT:

bash scripts/codebert_dns.sh
bash scripts/codebert_ood.sh

GraphCodeBERT:

bash scripts/graphcodebert_dns.sh
bash scripts/graphcodebert_ood.sh

UniXcoder:

bash scripts/unixcoder_dns.sh
bash scripts/unixcoder_ood.sh

DISCO:

CodeBERT:

bash scripts/codebert_disco.sh
bash scripts/codebert_ood.sh

GraphCodeBERT:

bash scripts/graphcodebert_disco.sh
bash scripts/graphcodebert_ood.sh

UniXcoder:

bash scripts/unixcoder_disco.sh
bash scripts/unixcoder_ood.sh

CodeRetriever:

CodeBERT:

bash scripts/codebert_code_retriever.sh
bash scripts/codebert_ood.sh

GraphCodeBERT:

bash scripts/graphcodebert_code_retriever.sh
bash scripts/graphcodebert_ood.sh

UniXcoder:

bash scripts/unixcoder_code_retriever.sh
bash scripts/unixcoder_ood.sh

SYNC experiments:

CodeBERT:

bash scripts/codebert_sync.sh
bash scripts/codebert_ood.sh

GraphCodeBERT:

bash scripts/graphcodebert_sync.sh
bash scripts/graphcodebert_ood.sh

UniXcoder:

bash scripts/unixcoder_sync.sh
bash scripts/unixcoder_ood.sh

Citation:

@inproceedings{naik23sync,
    title = {SYNC: A Structurally guided Hard Negative Curricula for
Generalizable Neural Code Search},
    author = {Atharva Naik and Soumtira Das and Jyothi Vedurada and Somak Aditya},
    booktitle = {The Asia-Pacific Chapter of the Association for Computational Linguistics},
    month = {November},
    year = {2023}
}