Datasets:
Tasks:
Token Classification
Sub-tasks:
coreference-resolution
Languages:
English
Multilinguality:
monolingual
Annotations Creators:
domain experts
License:
File size: 4,115 Bytes
056956d 18e2798 2b83143 056956d 0b90551 056956d 246b2d2 056956d 2c94766 056956d 37db430 18e2798 3bfa95b 9c1dc3e 151dd98 f08dbf3 151dd98 37db430 ce5868a 056956d 02347ac 4a4522a 3bfa95b 4a4522a 2c94766 |
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 |
"""SciCo"""
import os
from datasets.arrow_dataset import DatasetTransformationNotAllowedError
from datasets.utils import metadata
import jsonlines
import datasets
_CITATION = """\
@inproceedings{
cattan2021scico,
title={SciCo: Hierarchical Cross-Document Coreference for Scientific Concepts},
author={Arie Cattan and Sophie Johnson and Daniel S. Weld and Ido Dagan and Iz Beltagy and Doug Downey and Tom Hope},
booktitle={3rd Conference on Automated Knowledge Base Construction},
year={2021},
url={https://openreview.net/forum?id=OFLbgUP04nC}
}
"""
_DESCRIPTION = """\
SciCo is a dataset for hierarchical cross-document coreference resolution
over scientific papers in the CS domain.
"""
_DATA_URL = "./data.tar"
class Scico(datasets.GeneratorBasedBuilder):
def _info(self):
return datasets.DatasetInfo(
description=_DESCRIPTION,
homepage="https://scico.apps.allenai.org/",
features=datasets.Features(
{
"flatten_tokens": datasets.features.Sequence(datasets.features.Value("string")),
"flatten_mentions": datasets.features.Sequence(datasets.features.Sequence(datasets.features.Value("int32"), length=3)),
"tokens": datasets.features.Sequence(datasets.features.Sequence(datasets.features.Value("string"))),
"doc_ids": datasets.features.Sequence(datasets.features.Value("int32")),
"metadata": datasets.features.Sequence(
{
"title": datasets.features.Value("string"),
"paper_sha": datasets.features.Value("string"),
"fields_of_study": datasets.features.Value("string"),
"Year": datasets.features.Value("string"),
"BookTitle": datasets.features.Value("string"),
"url": datasets.features.Value("string")
}
),
"sentences": datasets.features.Sequence(datasets.features.Sequence(datasets.features.Sequence(datasets.features.Value("int32")))),
"mentions": datasets.features.Sequence(datasets.features.Sequence(datasets.features.Value("int32"), length=4)),
"relations": datasets.features.Sequence(datasets.features.Sequence(datasets.features.Value("int32"), length=2)),
"id": datasets.Value("int32"),
"source": datasets.Value("string"),
"hard_10": datasets.features.Value("bool"),
"hard_20": datasets.features.Value("bool"),
"curated": datasets.features.Value("bool")
}
),
supervised_keys=None,
citation = _CITATION)
def _split_generators(self, dl_manager):
data_dir = dl_manager.download_and_extract(_DATA_URL)
return [
datasets.SplitGenerator(
name=datasets.Split.TEST, gen_kwargs={"filepath": os.path.join(data_dir, "test.jsonl")}
),
datasets.SplitGenerator(
name=datasets.Split.VALIDATION, gen_kwargs={"filepath": os.path.join(data_dir, "dev.jsonl")}
),
datasets.SplitGenerator(
name=datasets.Split.TRAIN, gen_kwargs={"filepath": os.path.join(data_dir, "train.jsonl")}
),
]
def _generate_examples(self, filepath):
"""This function returns the examples in the raw (text) form."""
with jsonlines.open(filepath, 'r') as f:
for i, topic in enumerate(f):
topic['hard_10'] = topic['hard_10'] if 'hard_10' in topic else False
topic['hard_20'] = topic['hard_20'] if 'hard_20' in topic else False
topic["curated"] = topic["curated"] if "curated" in topic else False
yield i, topic
|