Delete loading script
Browse files- qangaroo.py +0 -126
qangaroo.py
DELETED
@@ -1,126 +0,0 @@
|
|
1 |
-
"""TODO(qangaroo): Add a description here."""
|
2 |
-
|
3 |
-
|
4 |
-
import json
|
5 |
-
import os
|
6 |
-
|
7 |
-
import datasets
|
8 |
-
|
9 |
-
|
10 |
-
# TODO(qangaroo): BibTeX citation
|
11 |
-
|
12 |
-
_CITATION = """
|
13 |
-
"""
|
14 |
-
|
15 |
-
# TODO(quangaroo):
|
16 |
-
_DESCRIPTION = """\
|
17 |
-
We have created two new Reading Comprehension datasets focussing on multi-hop (alias multi-step) inference.
|
18 |
-
|
19 |
-
Several pieces of information often jointly imply another fact. In multi-hop inference, a new fact is derived by combining facts via a chain of multiple steps.
|
20 |
-
|
21 |
-
Our aim is to build Reading Comprehension methods that perform multi-hop inference on text, where individual facts are spread out across different documents.
|
22 |
-
|
23 |
-
The two QAngaroo datasets provide a training and evaluation resource for such methods.
|
24 |
-
"""
|
25 |
-
|
26 |
-
_MEDHOP_DESCRIPTION = """\
|
27 |
-
With the same format as WikiHop, this dataset is based on research paper abstracts from PubMed, and the queries are about interactions between pairs of drugs.
|
28 |
-
The correct answer has to be inferred by combining information from a chain of reactions of drugs and proteins.
|
29 |
-
"""
|
30 |
-
_WIKIHOP_DESCRIPTION = """\
|
31 |
-
With the same format as WikiHop, this dataset is based on research paper abstracts from PubMed, and the queries are about interactions between pairs of drugs.
|
32 |
-
The correct answer has to be inferred by combining information from a chain of reactions of drugs and proteins.
|
33 |
-
"""
|
34 |
-
|
35 |
-
_URL = "qangaroo_v1.1.zip"
|
36 |
-
|
37 |
-
|
38 |
-
class QangarooConfig(datasets.BuilderConfig):
|
39 |
-
def __init__(self, data_dir, **kwargs):
|
40 |
-
"""BuilderConfig for qangaroo dataset
|
41 |
-
|
42 |
-
Args:
|
43 |
-
data_dir: directory for the given dataset name
|
44 |
-
**kwargs: keyword arguments forwarded to super.
|
45 |
-
|
46 |
-
"""
|
47 |
-
|
48 |
-
super(QangarooConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
|
49 |
-
|
50 |
-
self.data_dir = data_dir
|
51 |
-
|
52 |
-
|
53 |
-
class Qangaroo(datasets.GeneratorBasedBuilder):
|
54 |
-
"""TODO(qangaroo): Short description of my dataset."""
|
55 |
-
|
56 |
-
# TODO(qangaroo): Set up version.
|
57 |
-
VERSION = datasets.Version("0.1.0")
|
58 |
-
BUILDER_CONFIGS = [
|
59 |
-
QangarooConfig(name="medhop", description=_MEDHOP_DESCRIPTION, data_dir="medhop"),
|
60 |
-
QangarooConfig(name="masked_medhop", description=_MEDHOP_DESCRIPTION, data_dir="medhop"),
|
61 |
-
QangarooConfig(name="wikihop", description=_WIKIHOP_DESCRIPTION, data_dir="wikihop"),
|
62 |
-
QangarooConfig(name="masked_wikihop", description=_WIKIHOP_DESCRIPTION, data_dir="wikihop"),
|
63 |
-
]
|
64 |
-
|
65 |
-
def _info(self):
|
66 |
-
# TODO(qangaroo): Specifies the datasets.DatasetInfo object
|
67 |
-
return datasets.DatasetInfo(
|
68 |
-
# This is the description that will appear on the datasets page.
|
69 |
-
description=_DESCRIPTION,
|
70 |
-
# datasets.features.FeatureConnectors
|
71 |
-
features=datasets.Features(
|
72 |
-
{
|
73 |
-
# These are the features of your dataset like images, labels ...
|
74 |
-
"query": datasets.Value("string"),
|
75 |
-
"supports": datasets.features.Sequence(datasets.Value("string")),
|
76 |
-
"candidates": datasets.features.Sequence(datasets.Value("string")),
|
77 |
-
"answer": datasets.Value("string"),
|
78 |
-
"id": datasets.Value("string")
|
79 |
-
# These are the features of your dataset like images, labels ...
|
80 |
-
}
|
81 |
-
),
|
82 |
-
# If there's a common (input, target) tuple from the features,
|
83 |
-
# specify them here. They'll be used if as_supervised=True in
|
84 |
-
# builder.as_dataset.
|
85 |
-
supervised_keys=None,
|
86 |
-
# Homepage of the dataset for documentation
|
87 |
-
homepage="http://qangaroo.cs.ucl.ac.uk/index.html",
|
88 |
-
citation=_CITATION,
|
89 |
-
)
|
90 |
-
|
91 |
-
def _split_generators(self, dl_manager):
|
92 |
-
"""Returns SplitGenerators."""
|
93 |
-
# TODO(qangaroo): Downloads the data and defines the splits
|
94 |
-
# dl_manager is a datasets.download.DownloadManager that can be used to
|
95 |
-
# download and extract URLs
|
96 |
-
dl_dir = dl_manager.download_and_extract(_URL)
|
97 |
-
data_dir = os.path.join(dl_dir, "qangaroo_v1.1")
|
98 |
-
train_file = "train.masked.json" if "masked" in self.config.name else "train.json"
|
99 |
-
dev_file = "dev.masked.json" if "masked" in self.config.name else "dev.json"
|
100 |
-
return [
|
101 |
-
datasets.SplitGenerator(
|
102 |
-
name=datasets.Split.TRAIN,
|
103 |
-
# These kwargs will be passed to _generate_examples
|
104 |
-
gen_kwargs={"filepath": os.path.join(data_dir, self.config.data_dir, train_file)},
|
105 |
-
),
|
106 |
-
datasets.SplitGenerator(
|
107 |
-
name=datasets.Split.VALIDATION,
|
108 |
-
# These kwargs will be passed to _generate_examples
|
109 |
-
gen_kwargs={"filepath": os.path.join(data_dir, self.config.data_dir, dev_file)},
|
110 |
-
),
|
111 |
-
]
|
112 |
-
|
113 |
-
def _generate_examples(self, filepath):
|
114 |
-
"""Yields examples."""
|
115 |
-
# TODO(quangaroo): Yields (key, example) tuples from the dataset
|
116 |
-
with open(filepath, encoding="utf-8") as f:
|
117 |
-
data = json.load(f)
|
118 |
-
for example in data:
|
119 |
-
id_ = example["id"]
|
120 |
-
yield id_, {
|
121 |
-
"id": example["id"],
|
122 |
-
"query": example["query"],
|
123 |
-
"supports": example["supports"],
|
124 |
-
"candidates": example["candidates"],
|
125 |
-
"answer": example["answer"],
|
126 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|