Upload netifier.py with huggingface_hub
Browse files- netifier.py +130 -0
netifier.py
ADDED
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from pathlib import Path
|
2 |
+
from typing import Dict, List, Tuple
|
3 |
+
|
4 |
+
import datasets
|
5 |
+
import pandas as pd
|
6 |
+
|
7 |
+
from seacrowd.utils import schemas
|
8 |
+
from seacrowd.utils.configs import SEACrowdConfig
|
9 |
+
from seacrowd.utils.constants import Tasks
|
10 |
+
|
11 |
+
_CITATION = """\
|
12 |
+
"""
|
13 |
+
|
14 |
+
_LANGUAGES = ["ind"] # We follow ISO639-3 language code (https://iso639-3.sil.org/code_tables/639/data)
|
15 |
+
_LOCAL = False
|
16 |
+
|
17 |
+
_DATASETNAME = "netifier"
|
18 |
+
|
19 |
+
_DESCRIPTION = """\
|
20 |
+
Netifier dataset is a collection of scraped posts on famous social media sites in Indonesia,
|
21 |
+
such as Instagram, Twitter, and Kaskus aimed to do multi-label toxicity classification.
|
22 |
+
The dataset consists of 7,773 texts. The author manually labelled ~7k samples into 4 categories:
|
23 |
+
pornography, hate speech, racism, and radicalism.
|
24 |
+
"""
|
25 |
+
|
26 |
+
_HOMEPAGE = "https://github.com/ahmadizzan/netifier"
|
27 |
+
_LICENSE = "Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International"
|
28 |
+
_URLS = {_DATASETNAME: {"train": "https://raw.githubusercontent.com/ahmadizzan/netifier/master/data/processed/train.csv", "test": "https://raw.githubusercontent.com/ahmadizzan/netifier/master/data/processed/test.csv"}}
|
29 |
+
_SUPPORTED_TASKS = [Tasks.ASPECT_BASED_SENTIMENT_ANALYSIS]
|
30 |
+
_SOURCE_VERSION = "1.0.0"
|
31 |
+
_SEACROWD_VERSION = "2024.06.20"
|
32 |
+
|
33 |
+
|
34 |
+
class Netifier(datasets.GeneratorBasedBuilder):
|
35 |
+
"""Netifier dataset is a collection of scraped posts on famous social media sites in Indonesia,
|
36 |
+
such as Instagram, Twitter, and Kaskus aimed to do multi-label toxicity classification."""
|
37 |
+
|
38 |
+
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
39 |
+
SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
|
40 |
+
|
41 |
+
BUILDER_CONFIGS = [
|
42 |
+
SEACrowdConfig(
|
43 |
+
name="netifier_source",
|
44 |
+
version=SOURCE_VERSION,
|
45 |
+
description="Netifier source schema",
|
46 |
+
schema="source",
|
47 |
+
subset_id="netifier",
|
48 |
+
),
|
49 |
+
SEACrowdConfig(
|
50 |
+
name="netifier_seacrowd_text_multi",
|
51 |
+
version=SEACROWD_VERSION,
|
52 |
+
description="Netifier Nusantara schema",
|
53 |
+
schema="seacrowd_text_multi",
|
54 |
+
subset_id="netifier",
|
55 |
+
),
|
56 |
+
]
|
57 |
+
|
58 |
+
DEFAULT_CONFIG_NAME = "netifier_source"
|
59 |
+
|
60 |
+
def _info(self) -> datasets.DatasetInfo:
|
61 |
+
if self.config.schema == "source":
|
62 |
+
features = datasets.Features(
|
63 |
+
{
|
64 |
+
"text": datasets.Value("string"),
|
65 |
+
"pornography": datasets.Value("bool"),
|
66 |
+
"blasphemy_racism_discrimination": datasets.Value("bool"),
|
67 |
+
"radicalism": datasets.Value("bool"),
|
68 |
+
"defamation": datasets.Value("bool"),
|
69 |
+
}
|
70 |
+
)
|
71 |
+
elif self.config.schema == "seacrowd_text_multi":
|
72 |
+
features = schemas.text_multi_features([0, 1])
|
73 |
+
|
74 |
+
return datasets.DatasetInfo(
|
75 |
+
description=_DESCRIPTION,
|
76 |
+
features=features,
|
77 |
+
homepage=_HOMEPAGE,
|
78 |
+
license=_LICENSE,
|
79 |
+
citation=_CITATION,
|
80 |
+
)
|
81 |
+
|
82 |
+
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
83 |
+
"""Returns SplitGenerators."""
|
84 |
+
urls = _URLS[_DATASETNAME]
|
85 |
+
train_data = Path(dl_manager.download(urls["train"]))
|
86 |
+
test_data = Path(dl_manager.download(urls["test"]))
|
87 |
+
|
88 |
+
return [
|
89 |
+
datasets.SplitGenerator(
|
90 |
+
name=datasets.Split.TRAIN,
|
91 |
+
gen_kwargs={
|
92 |
+
"filepath": train_data,
|
93 |
+
"split": "train",
|
94 |
+
},
|
95 |
+
),
|
96 |
+
datasets.SplitGenerator(
|
97 |
+
name=datasets.Split.TEST,
|
98 |
+
gen_kwargs={
|
99 |
+
"filepath": test_data,
|
100 |
+
"split": "test",
|
101 |
+
},
|
102 |
+
),
|
103 |
+
]
|
104 |
+
|
105 |
+
def _generate_examples(self, filepath: Path, split: str) -> Tuple[int, Dict]:
|
106 |
+
"""Yields examples as (key, example) tuples."""
|
107 |
+
# Dataset does not have id, using row index as id
|
108 |
+
label_cols = ["pornography", "blasphemy_racism_discrimination", "radicalism", "defamation"]
|
109 |
+
df = pd.read_csv(filepath, encoding="ISO-8859-1").reset_index()
|
110 |
+
df.columns = ["id", "original_text", "text"] + label_cols
|
111 |
+
|
112 |
+
if self.config.schema == "source":
|
113 |
+
for row in df.itertuples():
|
114 |
+
ex = {
|
115 |
+
"text": row.text,
|
116 |
+
}
|
117 |
+
for label in label_cols:
|
118 |
+
ex[label] = getattr(row, label)
|
119 |
+
yield row.id, ex
|
120 |
+
|
121 |
+
elif self.config.schema == "seacrowd_text_multi":
|
122 |
+
for row in df.itertuples():
|
123 |
+
ex = {
|
124 |
+
"id": str(row.id),
|
125 |
+
"text": row.text,
|
126 |
+
"labels": [label for label in row[4:]],
|
127 |
+
}
|
128 |
+
yield row.id, ex
|
129 |
+
else:
|
130 |
+
raise ValueError(f"Invalid config: {self.config.name}")
|