Datasets:

Modalities:
Text
Formats:
parquet
Languages:
English
ArXiv:
Libraries:
Datasets
pandas
License:
azugarini commited on
Commit
a9fa34c
1 Parent(s): ebc4cd2

Upload 2 files

Browse files

update data and script

Files changed (2) hide show
  1. BUSTER.py +114 -0
  2. BUSTER.zip +3 -0
BUSTER.py ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
3
+ #
4
+ # Licensed under the Apache License, Version 2.0 (the "License");
5
+ # you may not use this file except in compliance with the License.
6
+ # You may obtain a copy of the License at
7
+ #
8
+ # http://www.apache.org/licenses/LICENSE-2.0
9
+ #
10
+ # Unless required by applicable law or agreed to in writing, software
11
+ # distributed under the License is distributed on an "AS IS" BASIS,
12
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
+ # See the License for the specific language governing permissions and
14
+ # limitations under the License.
15
+ """BUSTER: a BUSiness Transaction Entity Recognition Dataset"""
16
+
17
+ import os
18
+ import datasets
19
+ from datasets import load_dataset
20
+
21
+ _CITATION = """
22
+ Accepted at EMNLP 2023 - Industry Track.
23
+ TBA
24
+ """
25
+
26
+ _DESCRIPTION = """
27
+ Buster is an Entity Recognition dataset consisting of 3779 manually annotated documents on financial transactions.
28
+ Documents were selected using EDGAR (Electronic Data Gathering, Analysis, and Retrieval system) from the
29
+ U.S. Securities and Exchange Commission (SEC).
30
+ The corpus focuses on the main actors involved in business transactions.
31
+ Overall, there are three families of entities: Parties, Advisors and Generic information, for a total of 6 annotated
32
+ entity types.
33
+ We also released a corpus of 6196 automatically annotated documents.
34
+ """
35
+
36
+ _HOMEPAGE = "https://expert.ai/buster"
37
+ _URL = "buster.zip"
38
+ _VERSION = "1.0.0"
39
+
40
+ logger = datasets.logging.get_logger(__name__)
41
+
42
+
43
+ # --------------------------------------------------------------------------------------------------------
44
+ # Tag set
45
+ _LABELS = [
46
+ "O", # non-entities label
47
+ "B-Parties.BUYING_COMPANY",
48
+ "I-Parties.BUYING_COMPANY",
49
+ "B-Parties.SELLING_COMPANY",
50
+ "I-Parties.SELLING_COMPANY",
51
+ "B-Parties.ACQUIRED_COMPANY",
52
+ "I-Parties.ACQUIRED_COMPANY",
53
+ "B-Advisors.LEGAL_CONSULTING_COMPANY",
54
+ "I-Advisors.LEGAL_CONSULTING_COMPANY",
55
+ "B-Advisors.GENERIC_CONSULTING_COMPANY",
56
+ "I-Advisors.GENERIC_CONSULTING_COMPANY",
57
+ "B-Generic_Info.ANNUAL_REVENUES",
58
+ "I-Generic_Info.ANNUAL_REVENUES"
59
+ ]
60
+
61
+
62
+ class BusterConfig(datasets.BuilderConfig):
63
+ """BuilderConfig for the BUSTER dataset."""
64
+
65
+ def __init__(self, **kwargs):
66
+ """BuilderConfig for the BUSTER dataset.
67
+ Args:
68
+ **kwargs: keyword arguments forwarded to super.
69
+ """
70
+ super(BusterConfig, self).__init__(
71
+ name=f"BUSTER",
72
+ description=_DESCRIPTION,
73
+ version=datasets.Version(_VERSION), # hf dataset script version
74
+ **kwargs,
75
+ )
76
+
77
+
78
+ class Buster(datasets.GeneratorBasedBuilder):
79
+ """The BUSTER dataset."""
80
+
81
+ BUILDER_CONFIGS = [
82
+ BusterConfig()
83
+ ]
84
+
85
+ def _info(self):
86
+ return datasets.DatasetInfo(
87
+ description=_DESCRIPTION,
88
+ features=datasets.Features(
89
+ {
90
+ "document_id": datasets.Value("string"),
91
+ "tokens": datasets.Sequence(datasets.Value("string")),
92
+ "labels": datasets.Sequence(datasets.features.ClassLabel(names=_LABELS)),
93
+ }
94
+ ),
95
+ homepage=_HOMEPAGE,
96
+ citation=_CITATION,
97
+ )
98
+
99
+ def _split_generators(self, dl_manager):
100
+ data_dir = dl_manager.download_and_extract(_URL)
101
+ fold_names = [f"FOLD_{i}" for i in range(5)] + ["SILVER"]
102
+ return [
103
+ datasets.SplitGenerator(
104
+ name=fold_name,
105
+ gen_kwargs={"file_path": os.path.join(data_dir, fold_name)},
106
+ ) for fold_name in fold_names
107
+ ]
108
+
109
+ def _generate_examples(self, file_path):
110
+ dataset = load_dataset("json", data_files=file_path)
111
+ logger.info(f"Generating examples from: {file_path}")
112
+ for idx, example in enumerate(dataset["train"]):
113
+ # example features: document_id, tokens, labels
114
+ yield idx, example
BUSTER.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66082251c87758687c334965fdd3a8f2b3f1e933755c6fdc16e71a522570943f
3
+ size 20824877