Stack-Repo.py

# Copyright 2020 The HuggingFace Datasets Authors and Serivce Now Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""This is a loading script for the Stack-Repo dataset"""


import csv
import json
import os

from pathlib import Path

import datasets


_CITATION = """\
@article{shrivastava2023repofusion,
  title={RepoFusion: Training Code Models to Understand Your Repository},
  author={Shrivastava, Disha and Kocetkov, Denis and de Vries, Harm and Bahdanau, Dzmitry and Scholak, Torsten},
  journal={arXiv preprint arXiv:2306.10998},
  year={2023}
}
"""

_DESCRIPTION = """\
This is the Stack-Repo dataset
"""

_HOMEPAGE = "https://huggingface.co/datasets/RepoFusion/Stack-Repo"

_LICENSE = "other"

_SOURCES = {
    'test': [
        'thekoc', 'JellyBrother', 'shuli495', 'xiao-ren-wu', 'UzcDevelopmentTeam', 'Penguinz22', 'andrewvmail',
        'richard-trotter', 'SlimPay', 'teknux-org', 'FIXTradingCommunity', 'Juliana-Felipe', 'exasol', 'iBiber',
        'sebaslogen', 'billmcchesney1', 'ms-showcase', 'mP1', 'yokotaso', 'ErikHage', 'shadaileng', 'regular-project',
        'wiwilin', '1273091433', 'liangxiong3403', 'wangzlei', 'Vitaliy-Yakovchuk', 'anshul-mohil', 'MarceloCavassani',
        'cwhelan', 'score-server', 'meyer5', 'SeSac-Cloud-Backend-4', 'Hugo-Gao', 'Olivia0202', 'IceIce1ce', '4pxzhou',
        'bboylin', 'acolebourne', '1725762388', 'iViolinSolo', 'diaolingzc', 'trungtuan', 'havenmoney', 'speak2me',
        'VincenzoArceri', 'exadel-inc', 'JohnReeson', 'CyiceK', 'Vooy'
    ], 
    'train': [
        'ValeriyKnyazhev', 'leshiv', 'MfromAzeroth', 'wugjin', 'MarbleDice', 'hivijayreddy', 'sidpatel-github',
        'hellorizon', 'CreativePenguin', 'maulikpandey', 'a630140621', 'luminiusa', 'ot-maksim', 'akash-coded',
        'eddiewgj', 'a616707902', 'oasis9', 'DeyanZhelyazkov', 'yweng47', 'JCThePants', 'Gorlem', 'dongjihui666',
        'mediatrue', 'david2999999', 'comme-il-faut', 'jannal', 'tfjybj', 'DimkaGorhover', 'ZhangTizo', 'CIDARLAB',
        'DwArFeng', 'mbarkley', 'VladRomanchuk', 'vivekmalhotragithub', 'android-little-boy', 'patrik999', 'madanalogy',
        'TheTrueColonel', 'TreeZhiyuan', 'sistcoop', 'java-ea', 'DarwinReforged', 'DougEdey', '402d', 'wangdingfeng',
        'WeiziPlus', 'fabionery', 'famartinrh', 'rghvgrv', 'QuincySx', 'raphaelcohn', 'hYdos', 'tacticalrce',
        'danielkrajnik', 'SeanDecker1', 'bahaahussein', 'Team1772', 'zengfanmao', 'Hopp-Stu', 'My-DIGI-ID', 'slickqa',
        'opentok', 'Manolomon', 'denhatmong', 'baishuo', 'zli78122', 'kenichi-ando', 'pengcash', 'Princelo',
        'RuthDirnfeld', 'HRI-EU', 'jan438', 'sd2438892', 'rsofista', 'mpsitech', 'AvaN0x', 'JDode', 'dreadwail',
        'pcjs156', 'MrWe', 'smallxiongxiong', 'wuxinlingluan', 'FLxmw', 'TransFICC', 'fengpod', 'orient33', 'hamsterxc',
        'diegolemospadilha', 'CheRut', 'ToreAad', 'AndrewHYC', 'Azagwen', 'Ndivhuwo', 'gou02rav32', 'CDZR0',
        'MFunction96', 'google', 'mariodavid', 'jacksontenorio8', 'DieguinhoHR'
    ],
    'val': [
        'malcolmgreaves', 'atlasapi', 'HeXavi8', 'fank243', 'StealWonders', 'asyakasimova', 'truthiswill',
        'swiatecki', 'yanjingfan', 'vamnoize', 'Eugenemdk', 'explodingnuggets', 'MartinItsLinda',
        'Web-Programmer-B-P', 'SergeyZhernovoy', 'sunhaidong', 'marcogrella', 'C0DEC0L0NY', 'JustinZuidgeest',
        'ShareWiFi', 'ConusDaktis', 'lqhshishen', 'vaginessa', 'Juansero29', 'meghbhalerao', 'YunLemon',
        'MeiVinEight', 'Huawei-DTSE-India-Delhi-NCR', 'ym-monkey', 'Selim042', 'ValentinGiboulot', 'jesusbmx',
        'FantasybabyChange', 'zegelin', 'MeasureAuthoringTool', 'death-knight', 'abargnesi', 'cmunilla',
        'sladyn98', 'jhaprabhatt', 'smartsv-dev', 'Lavaong', 'ShahabGT', 'WXB2010', 'digideskio', 'p1r9t3',
        'chathamabate', 'heymysweetheart', 'blockchain-etl', 'Jeniel557'
    ]
}

class StackRepo(datasets.GeneratorBasedBuilder):
    """This is a builder class for the Stack-Repo dataset"""

    VERSION = datasets.Version("1.0.0")

    BUILDER_CONFIGS = [
        datasets.BuilderConfig(
            name="bm25_contexts",
            version=VERSION,
            description="This part of the dataset covers bm25 contexts"
        ),
        datasets.BuilderConfig(
            name="PP_contexts",
            version=VERSION,
            description="This part of the dataset covers PP contexts"
        ),
        datasets.BuilderConfig(
            name="randomNN_contexts",
            version=VERSION,
            description="This part of the dataset covers randomNN contexts"
        ),
        datasets.BuilderConfig(
            name="sources",
            version=VERSION,
            description="This part of the dataset covers java file sources used in building other configurations"
        ),
    ]

    def _info(self):
        if self.config.name == 'sources':
            features = datasets.Features({
                'file': datasets.Value('string'),
                'content': datasets.Value('string')
            })
        else:
            features = datasets.Features({
                'id': datasets.Value('string'),
                'hole_file': datasets.Value('string'),
                'hole_line': datasets.Value('int32'),
                'hole_pos': datasets.Value('int32'),
                'question': datasets.Value('string'),
                'target': datasets.Value('string'),
                'answers': datasets.Sequence(
                    datasets.Value('string')
                ),
                'ctxs': [{
                    'title': datasets.Value('string'),
                    'text': datasets.Value('string'),
                    'score': datasets.Value('float64')
                }]
        })
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    @property
    def manual_download_instructions(self):
        # We are ok if user specifies local path instead of dataset name and data_dir
        # for sources configuration and other configurations can be downloaded from the hub
        # automatically. So, we manipulate return value here to get around manual_dir is None
        # check in _check_manual_download
        # NOTE: this hack can broke if implementaion of the _check_manual_download will change
        # TODO: override _check_manual_download instead
        if self.config.name == 'sources':
            if not self.dl_manager.manual_dir is None:
                base_path = Path(self.dl_manager.manual_dir)
            else:
                base_path = Path(self.base_path)

            if base_path.expanduser().absolute().is_dir():
                return None

            return (
                "first download the dataset to a local folder with "
                "git clone https://huggingface.co/datasets/RepoFusion/Stack-Repo "
                "<path/to/manual/data>"
            )
        else:
            return None
              
    
    def _split_generators(self, dl_manager):
        splits = [
            (datasets.Split.TRAIN, 'train'),
            (datasets.Split.TEST, 'test'),
            (datasets.Split.VALIDATION, 'val')
        ]
        
        if not dl_manager.manual_dir is None:
            base_path = Path(dl_manager.manual_dir)
        else:
            base_path = Path(self.base_path)

        if self.config.name == 'sources':
            src = {
                split[0]: [
                    el.relative_to(base_path)
                    for el in (base_path / f'data/{split[1]}/').rglob('*.java')
                ]
                for split in splits
            }
        else:
            src = {
                split[0]: [
                    f'data/{split[1]}/{el}/hole_and_{self.config.name}.json'
                    for el in _SOURCES[split[1]]
                ]
                for split in splits
            }

        data_dir = dl_manager.download_and_extract(src)
        return [
            datasets.SplitGenerator(
                name=split[0],
                gen_kwargs={
                    "filepaths": data_dir[split[0]],
                    "base_path": base_path
                },
            )
            for split in splits
        ]

    def _generate_examples(self, filepaths, base_path):
        for filepath in filepaths:
            if self.config.name == 'sources':
                filepath = Path(filepath)
                file = str(filepath.relative_to(base_path))
                yield file, {
                    'file': file,
                    'content': filepath.read_text()
                }
            else:
                with open(filepath, encoding="utf-8") as f:
                    for row in f:
                        data = json.loads(row)
                        parts = data['id'].rsplit('_', 2)
                        data['hole_file'] = str(base_path / parts[0])
                        data['hole_line'] = int(parts[1])
                        data['hole_pos'] = int(parts[2])
                        yield data['id'], data