Delete _scigen.py
Browse files- _scigen.py +0 -67
_scigen.py
DELETED
@@ -1,67 +0,0 @@
|
|
1 |
-
"""Scigen: dataset for reasoning-aware data-to-text generation from scientific tables"""
|
2 |
-
|
3 |
-
import json
|
4 |
-
import datasets
|
5 |
-
import glob
|
6 |
-
import os
|
7 |
-
|
8 |
-
_CITATION = """\
|
9 |
-
@article{moosavi:2021:SciGen,
|
10 |
-
author = {Nafise Sadat Moosavi, Andreas R{\"u}ckl{\'e}, Dan Roth, Iryna Gurevych},
|
11 |
-
title = {Learning to Reason for Text Generation from Scientific Tables},
|
12 |
-
journal = {arXiv preprint arXiv:2104.08296},
|
13 |
-
year = {2021},
|
14 |
-
url = {https://arxiv.org/abs/2104.08296}
|
15 |
-
}
|
16 |
-
"""
|
17 |
-
_DESCRIPTION = """\
|
18 |
-
SciGen is dataset for the task of reasoning-aware data-to-text generation consisting of tables from scientific articles and their corresponding descriptions.
|
19 |
-
"""
|
20 |
-
|
21 |
-
_URL = "https://github.com/UKPLab/SciGen"
|
22 |
-
_LICENSE = "CC BY-NC-SA 4.0"
|
23 |
-
|
24 |
-
class SciGen(datasets.GeneratorBasedBuilder):
|
25 |
-
VERSION = "1.0.0"
|
26 |
-
def _info(self):
|
27 |
-
return datasets.DatasetInfo(
|
28 |
-
description=_DESCRIPTION,
|
29 |
-
features=datasets.Features({
|
30 |
-
'paper': datasets.Value(dtype='string'),
|
31 |
-
'paper_id': datasets.Value(dtype='string'),
|
32 |
-
'table_caption': datasets.Value(dtype='string'),
|
33 |
-
'table_column_names': datasets.Value(dtype='large_string'),
|
34 |
-
'table_content_values': datasets.Value(dtype='large_string'),
|
35 |
-
'text' : datasets.Value(dtype='large_string'),
|
36 |
-
}),
|
37 |
-
supervised_keys=None,
|
38 |
-
homepage=_URL,
|
39 |
-
citation=_CITATION,
|
40 |
-
license=_LICENSE
|
41 |
-
)
|
42 |
-
|
43 |
-
def _split_generators(self, dl_manager):
|
44 |
-
"""Returns SplitGenerators."""
|
45 |
-
return [
|
46 |
-
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": "dataset", "split" : "train"}),
|
47 |
-
datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": "dataset", "split" : "dev"}),
|
48 |
-
datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": "dataset", "split" : "test"}),
|
49 |
-
]
|
50 |
-
|
51 |
-
def _generate_examples(self, filepath, split):
|
52 |
-
data_dir = "development" if split == "dev" else split
|
53 |
-
|
54 |
-
if split in ["train", "dev"]:
|
55 |
-
file_path = os.path.join(filepath, data_dir, "medium", f"{split}.json")
|
56 |
-
else:
|
57 |
-
# there is also "test-Other.json", should be looked into
|
58 |
-
file_path = os.path.join(filepath, data_dir, f"test-CL.json")
|
59 |
-
|
60 |
-
with open(file_path) as f:
|
61 |
-
j = json.load(f)
|
62 |
-
for example_idx, entry in enumerate(list(j.values())):
|
63 |
-
yield example_idx, {key: str(value) for key, value in entry.items()}
|
64 |
-
|
65 |
-
if __name__ == '__main__':
|
66 |
-
dataset = datasets.load_dataset(__file__)
|
67 |
-
dataset.push_to_hub("kasnerz/scigen")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|