hoyo_pianos.py

import os
import shutil
import random
import hashlib
import datasets
from midi2abc import midi2abc


_HOMEPAGE = f"https://huggingface.co/datasets/MuGeminorum/{os.path.basename(__file__).split('.')[0]}"

_CITATION = """\
@dataset{mihoyo_pianos,
  author       = {MuGeminorum Studio},
  title        = {mihoyo game piano songs},
  month        = {nov},
  year         = {2023},
  publisher    = {HF},
  version      = {1.1},
  url          = {https://huggingface.co/datasets/MuGeminorum/hoyo_pianos}
}
"""

_DESCRIPTION = """\
This database contains mihoyo game piano songs downloaded from musescore
"""

_URLS = {
    "genshin": f"{_HOMEPAGE}/resolve/main/data/genshin.zip",
    "starail": f"{_HOMEPAGE}/resolve/main/data/starail.zip"
}


class hoyo_pianos(datasets.GeneratorBasedBuilder):
    def _info(self):
        return datasets.DatasetInfo(
            features=datasets.Features(
                {
                    "midi": datasets.Value("string"),
                    "abc": datasets.Value("string"),
                    "tag": datasets.Value("string")
                }
            ),
            supervised_keys=("abc", "tags"),
            homepage=_HOMEPAGE,
            license="mit",
            citation=_CITATION,
            description=_DESCRIPTION
        )

    def _calculate_hash(self, file_path):
        # 计算文件的哈希值
        with open(file_path, 'rb') as midi_file:
            content = midi_file.read()
            return hashlib.md5(content).hexdigest()

    def _rm_duplicates_in_folder(self, input_folder):
        # 用于存储文件哈希值的字典
        hash_dict = {}
        duplist = []
        # 遍历输入文件夹
        for root, _, files in os.walk(input_folder):
            for file in files:
                file_path = os.path.join(root, file)
                file_hash = self._calculate_hash(file_path)

                # 检查文件哈希值是否已存在
                if file_hash in hash_dict:
                    print(f"Duplicates found: {file}")
                    # 将重复文件直接删除
                    duplist.append(file_path)
                    shutil.rmtree(file_path)
                else:
                    # 存储文件哈希值
                    hash_dict[file_hash] = file_path

        return duplist

    def _split_generators(self, dl_manager):
        dataset = []
        for key in _URLS.keys():
            data_files = dl_manager.download_and_extract(_URLS[key])
            files = dl_manager.iter_files([data_files])
            subset = []

            extract_dir = f'{data_files}\\{key}'
            duplist = self._rm_duplicates_in_folder(extract_dir)

            for path in files:
                if (not path in duplist) and (os.path.basename(path).endswith(".mid")):
                    subset.append(path)

            random.shuffle(subset)
            dataset.append(
                datasets.SplitGenerator(
                    name=key,
                    gen_kwargs={
                        "files": subset
                    }
                )
            )

        return dataset

    def _generate_examples(self, files):
        for i, path in enumerate(files):
            yield i, {
                "midi": path,
                "abc": midi2abc(path),
                "tag": os.path.basename(path)[:-4].encode('cp437').decode('gbk')
            }