music_genre.py

import os
import random
import datasets
import pandas as pd
from datasets.tasks import ImageClassification

# Once upload a new piano brand, please register its name here
_NAMES = [
    "0_none",
    "1_classic",
    "2_non_classic",
    "3_symphony",
    "4_opera",
    "5_solo",
    "6_chamber",
    "7_pop",
    "8_dance_and_house",
    "9_indie",
    "10_soul_or_r_and_b",
    "11_rock"
]

_DBNAME = os.path.basename(__file__).split('.')[0]

_HOMEPAGE = "https://huggingface.co/datasets/ccmusic-database/" + _DBNAME

_CITATION = """\
@dataset{zhaorui_liu_2021_5676893,
  author       = {Zhaorui Liu and Zijin Li},
  title        = {{Music Data Sharing Platform for Computational Musicology Research (CCMUSIC DATASET)}},
  month        = nov,
  year         = 2021,
  publisher    = {Zenodo},
  version      = {1.1},
  doi          = {10.5281/zenodo.5676893},
  url          = {https://doi.org/10.5281/zenodo.5676893}
}
"""

_DESCRIPTION = """\
This database contains about 1700 musical pieces (.mp3 format, downloaded from NetEase Cloud Music) 
with lengths of 270-300s that are divided into 17 genres in total. 
"""

_URL = _HOMEPAGE + "/resolve/main/data/dataset.zip"
_CSV = _HOMEPAGE + "/resolve/main/data/labels.csv"


class music_genre(datasets.GeneratorBasedBuilder):
    def _info(self):
        return datasets.DatasetInfo(
            features=datasets.Features(
                {
                    "image": datasets.Image(),
                    "duration": datasets.Value("string"),
                    "fst_level_label": datasets.features.ClassLabel(names=_NAMES),
                    "sec_level_label": datasets.features.ClassLabel(names=_NAMES),
                    "thr_level_label": datasets.features.ClassLabel(names=_NAMES),
                }
            ),
            supervised_keys=("image", "fst_level_label"),
            homepage=_HOMEPAGE,
            license="mit",
            citation=_CITATION,
            description=_DESCRIPTION,
            task_templates=[
                ImageClassification(
                    task="image-classification",
                    image_column="image",
                    label_column="fst_level_label",
                )
            ]
        )

    # def _set_to_label(self, dataset):
    #     output = []
    #     for path in dataset:
    #         id = int(os.path.basename(path)[:-4])

    def _split_generators(self, dl_manager):
        data_files = dl_manager.download_and_extract(_URL)
        files = dl_manager.iter_files([data_files])
        labels = dl_manager.download(_CSV)

        dataset = []
        for _, path in enumerate(files):
            dataset.append(path)

        random.shuffle(dataset)
        data_count = len(dataset)
        p80 = int(data_count * 0.8)
        p90 = int(data_count * 0.9)

        # tra_set = dataset[:p80]
        # val_set = dataset[p80:p90]
        # tes_set = dataset[p90:]

        # tra_label =

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "files": dataset[:p80],
                    "labels": labels
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={
                    "files": dataset[p80:p90],
                    "labels": labels
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={
                    "files": dataset[p90:],
                    "labels": labels
                },
            ),
        ]

    def _generate_examples(self, files, labels):
        label = pd.read_csv(labels, index_col='id')
        for i, path in enumerate(files):
            file_name = os.path.basename(path)
            if file_name.endswith(".png"):
                yield i, {
                    "image": path,
                    "duration": label.iloc[i]['duration'],
                    "fst_level_label": _NAMES[label.iloc[i]['fst_level_label']],
                    "sec_level_label": _NAMES[label.iloc[i]['sec_level_label']],
                    "thr_level_label": _NAMES[label.iloc[i]['thr_level_label']],
                }