File size: 4,807 Bytes
ec615ff |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 |
import datasets
import json
import xml.etree.ElementTree as ET
from datasets import BuilderConfig, DatasetInfo, DownloadManager, GeneratorBasedBuilder, Features, SplitGenerator, Value
from typing import Any, Dict, Iterator, List, Tuple, Union
class KakologArchivesConfig(BuilderConfig):
def __init__(self, channel_id: Union[str, None] = None, year: Union[int, None] = None, number_of_files: Union[int, None] = None, **kwargs):
super(KakologArchivesConfig, self).__init__(**kwargs)
self.channel_id = channel_id
self.year = year
self.number_of_files = number_of_files
class KakologArchivesDatasetBuilder(GeneratorBasedBuilder):
VERSION = '1.0.0'
BUILDER_CONFIGS = [
KakologArchivesConfig(
name = 'all',
version = VERSION,
description = '全チャンネル/全期間のすべての過去ログコメントを取得します。180GB 近くあるため注意してください。',
),
KakologArchivesConfig(
name='sample',
version = VERSION,
description = 'サンプルとして、2022年中に投稿された TOKYO MX (実況チャンネル ID: jk9) のすべての過去ログコメントを取得します。1GB ほどあります。',
channel_id = 'jk9',
year = 2022,
),
]
DEFAULT_CONFIG_NAME = "all"
def _info(self) -> DatasetInfo:
return DatasetInfo(
description = 'ニコニコ実況のサービス開始から現在までのすべての過去ログコメントのデータセットです。',
homepage = 'https://jikkyo.tsukumijima.net/',
features = Features({
'thread': Value('string'),
'no': Value('int64'),
'vpos': Value('int64'),
'date': Value('int64'),
'date_usec': Value('int64'),
'user_id': Value('string'),
'mail': Value('string', id=None),
'premium': Value('bool'),
'anonymity': Value('bool'),
'content': Value('string'),
}),
)
def _split_generators(self, dl_manager: DownloadManager) -> List[SplitGenerator]:
def create_relative_paths(json_data: dict[str, Any], current_path: str = "") -> List[str]:
relative_paths = []
for key, value in json_data.items():
new_path = f"{current_path}/{key}" if current_path else key
if value is None:
relative_paths.append(new_path)
elif isinstance(value, dict):
relative_paths.extend(create_relative_paths(value, new_path))
return relative_paths
json_file = dl_manager.download('dataset_structure.json')
with open(json_file, 'r', encoding='utf-8') as f:
relative_paths = create_relative_paths(json.load(f))
if self.config.channel_id is not None:
relative_paths = [path for path in relative_paths if f"{self.config.channel_id}/" in path]
if self.config.year is not None:
relative_paths = [path for path in relative_paths if f"/{self.config.year}/" in path]
if self.config.number_of_files is not None:
relative_paths = relative_paths[:min(self.config.number_of_files, len(relative_paths))]
downloaded_files = dl_manager.download(relative_paths)
return [SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"files": downloaded_files})]
def _generate_examples(self, files: List[str]) -> Iterator[Tuple[int, Dict[str, Any]]]:
for idx_parent, file in enumerate(files):
with open(file, 'r', encoding='utf-8') as f:
contents = '<?xml version="1.0" encoding="UTF-8"?><packet>' + f.read() + '</packet>'
root = ET.fromstring(contents)
for idx, chat in enumerate(root.iter('chat')):
if 'deleted' in chat.attrib:
continue
chat_attrib = {
'thread': chat.attrib.get('thread', 'unknown'),
'no': int(chat.attrib.get('no', 0)),
'vpos': int(chat.attrib.get('vpos', 0)),
'date': int(chat.attrib.get('date', 0)),
'date_usec': int(chat.attrib.get('date_usec', 0)),
'user_id': chat.attrib.get('user_id', 'unknown'),
'mail': chat.attrib.get('mail', None),
'premium': chat.attrib.get('premium', '0') == '1',
'anonymity': chat.attrib.get('anonymity', '0') == '1',
}
yield idx_parent * 1000000 + idx, {**chat_attrib, 'content': chat.text}
|