|
import datasets |
|
import json |
|
|
|
_CITATION = """\ |
|
@inproceedings{TBD |
|
} |
|
""" |
|
|
|
_DESCRIPTION = """ |
|
NorSumm is a Norwegian manually created summarization dataset for benchmarking models. |
|
""" |
|
|
|
_LICENSE = "CC0 1.0 Universal" |
|
|
|
_HOMEPAGE= "https://github.com/SamiaTouileb/NorSumm/" |
|
_URL = "https://huggingface.co/datasets/SamiaT/NorSumm/raw/main/" |
|
_URLS = { |
|
"dev": _URL + "NorSumm_dev.json", |
|
"test": _URL + "NorSumm_test.json", |
|
} |
|
|
|
|
|
|
|
class NorSumm(datasets.GeneratorBasedBuilder): |
|
"""NorSumm 1.0 dataset""" |
|
|
|
VERSION = datasets.Version("1.0.0") |
|
BUILDER_CONFIGS = [ |
|
datasets.BuilderConfig( |
|
name="norsumm_v1", |
|
version=VERSION, |
|
description=_DESCRIPTION, |
|
), |
|
] |
|
|
|
def _info(self): |
|
return datasets.DatasetInfo( |
|
description=_DESCRIPTION, |
|
features=datasets.Features( |
|
{ |
|
"id": datasets.Value("string"), |
|
"article": datasets.Value("string"), |
|
"summaries_nb": datasets.Sequence( |
|
{ |
|
"summary1": datasets.Value("string"), |
|
"summary2": datasets.Value("string"), |
|
"summary3": datasets.Value("string"), |
|
} |
|
), |
|
"summaries_nn": datasets.Sequence( |
|
{ |
|
"summary1": datasets.Value("string"), |
|
"summary2": datasets.Value("string"), |
|
"summary3": datasets.Value("string"), |
|
} |
|
), |
|
} |
|
), |
|
supervised_keys=None, |
|
homepage=_HOMEPAGE, |
|
license=_LICENSE, |
|
citation=_CITATION, |
|
) |
|
|
|
def _split_generators(self, dl_manager): |
|
"""Returns SplitGenerators.""" |
|
|
|
urls_to_download = _URLS |
|
downloaded_files = dl_manager.download_and_extract(urls_to_download) |
|
|
|
return [ |
|
datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["dev"]}), |
|
datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}), |
|
] |
|
|
|
def _generate_examples(self, filepath): |
|
|
|
with open(filepath, 'r', encoding="utf-8") as f: |
|
norsumm = json.load(f) |
|
for example in norsumm: |
|
id_ = example["id"] |
|
article = example["article"] |
|
summary1_nb = example["summaries_nb"][0] |
|
summary2_nb = example["summaries_nb"][1] |
|
summary3_nb = example["summaries_nb"][2] |
|
|
|
summary1_nn = example["summaries_nn"][0] |
|
summary2_nn = example["summaries_nn"][1] |
|
summary3_nn = example["summaries_nn"][2] |
|
|
|
yield id_, { |
|
"article": article, |
|
"id": id_, |
|
"summaries_nb": { |
|
"summary1": summary1_nb, |
|
"summary2": summary2_nb, |
|
"summary3": summary3_nb, |
|
}, |
|
"summaries_nn": { |
|
"summary1": summary1_nn, |
|
"summary2": summary1_nn, |
|
"summary3": summary1_nn, |
|
}, |
|
} |
