|
import os |
|
import json |
|
import datasets |
|
|
|
|
|
|
|
|
|
_DESCRIPTION = """20Q""" |
|
|
|
|
|
_HOMEPAGE = "" |
|
|
|
|
|
_LICENSE = "" |
|
|
|
|
|
|
|
class NewDataset(datasets.GeneratorBasedBuilder): |
|
"""TODO: Short description of my dataset.""" |
|
|
|
VERSION = datasets.Version("1.0.0") |
|
|
|
def _info(self): |
|
features = datasets.Features( |
|
{ |
|
"subject": datasets.Value("string"), |
|
"question": datasets.Value("string"), |
|
"label": datasets.Value("bool"), |
|
"label_fine_grained": datasets.Value("string") |
|
} |
|
) |
|
|
|
return datasets.DatasetInfo( |
|
description=_DESCRIPTION, |
|
features=features, |
|
homepage=_HOMEPAGE, |
|
license=_LICENSE, |
|
citation="_CITATION", |
|
) |
|
|
|
def _split_generators(self, dl_manager): |
|
data_dir = dl_manager.download_and_extract(["data/train.jsonl", "data/test.jsonl"]) |
|
return [ |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TRAIN, |
|
gen_kwargs={ |
|
"filepath": data_dir[0], |
|
"split": "train", |
|
}, |
|
), |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TEST, |
|
gen_kwargs={ |
|
"filepath": data_dir[1], |
|
"split": "test" |
|
}, |
|
), |
|
] |
|
|
|
def _generate_examples(self, filepath, split): |
|
with open(filepath, encoding="utf-8") as f: |
|
for key, row in enumerate(f): |
|
data = json.loads(row) |
|
yield key, { |
|
"subject": data["subject"], |
|
"question": data["question"], |
|
"label": data["label"], |
|
"label_fine_grained": data["label_fine_grained"] |
|
} |
|
|