Create hatespeech_filipino.py
Browse files- hatespeech_filipino.py +111 -0
hatespeech_filipino.py
ADDED
@@ -0,0 +1,111 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
+
"""Hate Speech Text Classification Dataset in Filipino."""
|
16 |
+
|
17 |
+
import csv
|
18 |
+
import os
|
19 |
+
|
20 |
+
import datasets
|
21 |
+
from datasets.tasks import TextClassification
|
22 |
+
|
23 |
+
|
24 |
+
_DESCRIPTION = """\
|
25 |
+
Contains 10k tweets (training set) that are labeled as hate speech or non-hate speech. Released with 4,232 validation and 4,232 testing samples. Collected during the 2016 Philippine Presidential Elections.
|
26 |
+
"""
|
27 |
+
|
28 |
+
_CITATION = """\
|
29 |
+
@article{Cabasag-2019-hate-speech,
|
30 |
+
title={Hate speech in Philippine election-related tweets: Automatic detection and classification using natural language processing.},
|
31 |
+
author={Neil Vicente Cabasag, Vicente Raphael Chan, Sean Christian Lim, Mark Edward Gonzales, and Charibeth Cheng},
|
32 |
+
journal={Philippine Computing Journal},
|
33 |
+
volume={XIV},
|
34 |
+
number={1},
|
35 |
+
month={August},
|
36 |
+
year={2019}
|
37 |
+
}
|
38 |
+
"""
|
39 |
+
|
40 |
+
_HOMEPAGE = "https://github.com/jcblaisecruz02/Filipino-Text-Benchmarks"
|
41 |
+
|
42 |
+
# TODO: Add the licence for the dataset here if you can find it
|
43 |
+
_LICENSE = ""
|
44 |
+
|
45 |
+
_URL = "https://huggingface.co/datasets/jcblaise/hatespeech_filipino/resolve/main/hatespeech_raw.zip"
|
46 |
+
|
47 |
+
|
48 |
+
class HateSpeechFilipino(datasets.GeneratorBasedBuilder):
|
49 |
+
"""Hate Speech Text Classification Dataset in Filipino."""
|
50 |
+
|
51 |
+
VERSION = datasets.Version("1.0.0")
|
52 |
+
|
53 |
+
def _info(self):
|
54 |
+
# Labels: 0="Non-hate Speech", 1="Hate Speech"
|
55 |
+
features = datasets.Features(
|
56 |
+
{"text": datasets.Value("string"), "label": datasets.features.ClassLabel(names=["0", "1"])}
|
57 |
+
)
|
58 |
+
return datasets.DatasetInfo(
|
59 |
+
description=_DESCRIPTION,
|
60 |
+
features=features,
|
61 |
+
supervised_keys=None,
|
62 |
+
homepage=_HOMEPAGE,
|
63 |
+
license=_LICENSE,
|
64 |
+
citation=_CITATION,
|
65 |
+
task_templates=[TextClassification(text_column="text", label_column="label")],
|
66 |
+
)
|
67 |
+
|
68 |
+
def _split_generators(self, dl_manager):
|
69 |
+
"""Returns SplitGenerators."""
|
70 |
+
data_dir = dl_manager.download_and_extract(_URL)
|
71 |
+
train_path = os.path.join(data_dir, "hatespeech", "train.csv")
|
72 |
+
test_path = os.path.join(data_dir, "hatespeech", "train.csv")
|
73 |
+
validation_path = os.path.join(data_dir, "hatespeech", "valid.csv")
|
74 |
+
|
75 |
+
return [
|
76 |
+
datasets.SplitGenerator(
|
77 |
+
name=datasets.Split.TRAIN,
|
78 |
+
gen_kwargs={
|
79 |
+
"filepath": train_path,
|
80 |
+
"split": "train",
|
81 |
+
},
|
82 |
+
),
|
83 |
+
datasets.SplitGenerator(
|
84 |
+
name=datasets.Split.TEST,
|
85 |
+
gen_kwargs={
|
86 |
+
"filepath": test_path,
|
87 |
+
"split": "test",
|
88 |
+
},
|
89 |
+
),
|
90 |
+
datasets.SplitGenerator(
|
91 |
+
name=datasets.Split.VALIDATION,
|
92 |
+
gen_kwargs={
|
93 |
+
"filepath": validation_path,
|
94 |
+
"split": "dev",
|
95 |
+
},
|
96 |
+
),
|
97 |
+
]
|
98 |
+
|
99 |
+
def _generate_examples(self, filepath, split):
|
100 |
+
"""Yields examples."""
|
101 |
+
with open(filepath, encoding="utf-8") as csv_file:
|
102 |
+
csv_reader = csv.reader(
|
103 |
+
csv_file, quotechar='"', delimiter=",", quoting=csv.QUOTE_ALL, skipinitialspace=True
|
104 |
+
)
|
105 |
+
next(csv_reader)
|
106 |
+
for id_, row in enumerate(csv_reader):
|
107 |
+
try:
|
108 |
+
text, label = row
|
109 |
+
yield id_, {"text": text, "label": label}
|
110 |
+
except ValueError:
|
111 |
+
pass
|