google-code-jam.py

from typing import List
import os
import glob

import datasets


_DESCRIPTION = """Given two codes as the input, the task is to do binary classification (0/1), where 1 stands for semantic equivalence and 0 for others."""

_CITATION = """@inproceedings{10.1145/3236024.3236068,
author = {Zhao, Gang and Huang, Jeff},
title = {DeepSim: Deep Learning Code Functional Similarity},
year = {2018},
isbn = {9781450355735},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3236024.3236068},
doi = {10.1145/3236024.3236068},
booktitle = {Proceedings of the 2018 26th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering},
pages = {141–151},
numpages = {11},
keywords = {Classification, Control/Data flow, Code functional similarity, Deep Learning},
location = {Lake Buena Vista, FL, USA},
series = {ESEC/FSE 2018}
}
"""

SPLITS = {
    'test': [5, 6, 7, 8, 12],  # For test in `Language Models are Universal Embedders` https://arxiv.org/pdf/2310.08232.pdf
    'deepsim': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]
}
_URL = "https://huggingface.co/datasets/izhx/google-code-jam/resolve/main/googlejam4.tar.gz"


class GoogleCodeJam(datasets.GeneratorBasedBuilder):
    BUILDER_CONFIGS = [
        datasets.BuilderConfig(name='default', version=datasets.Version("1.0.0"), description=_DESCRIPTION)
    ]
    DEFAULT_CONFIG_NAME = "default"

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "fn1": datasets.Value("string"),
                    "code1": datasets.Value("string"),
                    "fn2": datasets.Value("string"),
                    "code2": datasets.Value("string"),
                    "label": datasets.Value("int32"),
                }
            ),
            homepage="https://github.com/parasol-aser/deepsim",
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
        folder = dl_manager.download_and_extract(_URL)
        folder = os.path.join(folder, 'googlejam4_src')
        return [
            datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"folder": folder, "problems": SPLITS["test"]}),
            datasets.SplitGenerator(name='deepsim', gen_kwargs={"folder": folder, "problems": SPLITS["deepsim"]}),
        ]

    def _generate_examples(self, folder, problems: list):
        raw = dict()
        for i in problems:
            group = list()
            for path in sorted(glob.glob(f'{folder}/{i}/*.java')):
                with open(path) as file:
                    lines = [l for l in file]
                name = os.path.basename(path)
                group.append((name, ''.join(lines[1:])))  # remove name line
            raw[i] = group

        _id = 0
        reverse = False
        for i in range(len(problems)):
            vi = raw[problems[i]]
            for n1, (fn1, code1) in enumerate(vi):
                for j in range(i, len(problems)):
                    vj = raw[problems[j]]
                    match = i == j
                    for n2, (fn2, code2) in enumerate(vj):
                        if match and n2 <= n1:
                            continue
                        ins = {'fn1': fn1, 'code1': code1, 'fn2': fn2, 'code2': code2, 'label': int(match)}
                        if reverse:
                            ins['fn1'], ins['fn2'] = ins['fn2'], ins['fn1']
                            ins['code1'], ins['code2'] = ins['code2'], ins['code1']
                        yield _id, ins
                        _id += 1
                        reverse = not reverse