holylovenia commited on
Commit
fdfbdae
1 Parent(s): 1408622

Upload nusax_senti.py with huggingface_hub

Browse files
Files changed (1) hide show
  1. nusax_senti.py +172 -0
nusax_senti.py ADDED
@@ -0,0 +1,172 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from pathlib import Path
2
+ from typing import Dict, List, Tuple
3
+
4
+ import datasets
5
+ import pandas as pd
6
+
7
+ from nusacrowd.utils import schemas
8
+ from nusacrowd.utils.configs import NusantaraConfig
9
+ from nusacrowd.utils.constants import (DEFAULT_NUSANTARA_VIEW_NAME,
10
+ DEFAULT_SOURCE_VIEW_NAME, Tasks)
11
+
12
+ _LOCAL = False
13
+
14
+ _DATASETNAME = "nusax_senti"
15
+ _SOURCE_VIEW_NAME = DEFAULT_SOURCE_VIEW_NAME
16
+ _UNIFIED_VIEW_NAME = DEFAULT_NUSANTARA_VIEW_NAME
17
+
18
+ _LANGUAGES = ["ind", "ace", "ban", "bjn", "bbc", "bug", "jav", "mad", "min", "nij", "sun", "eng"] # We follow ISO639-3 language code (https://iso639-3.sil.org/code_tables/639/data)
19
+
20
+ _CITATION = """\
21
+ @misc{winata2022nusax,
22
+ title={NusaX: Multilingual Parallel Sentiment Dataset for 10 Indonesian Local Languages},
23
+ author={Winata, Genta Indra and Aji, Alham Fikri and Cahyawijaya,
24
+ Samuel and Mahendra, Rahmad and Koto, Fajri and Romadhony,
25
+ Ade and Kurniawan, Kemal and Moeljadi, David and Prasojo,
26
+ Radityo Eko and Fung, Pascale and Baldwin, Timothy and Lau,
27
+ Jey Han and Sennrich, Rico and Ruder, Sebastian},
28
+ year={2022},
29
+ eprint={2205.15960},
30
+ archivePrefix={arXiv},
31
+ primaryClass={cs.CL}
32
+ }
33
+ """
34
+
35
+ _DESCRIPTION = """\
36
+ NusaX is a high-quality multilingual parallel corpus that covers 12 languages, Indonesian, English, and 10 Indonesian local languages, namely Acehnese, Balinese, Banjarese, Buginese, Madurese, Minangkabau, Javanese, Ngaju, Sundanese, and Toba Batak.
37
+
38
+ NusaX-Senti is a 3-labels (positive, neutral, negative) sentiment analysis dataset for 10 Indonesian local languages + Indonesian and English.
39
+ """
40
+
41
+ _HOMEPAGE = "https://github.com/IndoNLP/nusax/tree/main/datasets/sentiment"
42
+
43
+ _LICENSE = "Creative Commons Attribution Share-Alike 4.0 International"
44
+
45
+ _SUPPORTED_TASKS = [Tasks.SENTIMENT_ANALYSIS]
46
+
47
+ _SOURCE_VERSION = "1.0.0"
48
+
49
+ _NUSANTARA_VERSION = "1.0.0"
50
+
51
+ _URLS = {
52
+ "train": "https://raw.githubusercontent.com/IndoNLP/nusax/main/datasets/sentiment/{lang}/train.csv",
53
+ "validation": "https://raw.githubusercontent.com/IndoNLP/nusax/main/datasets/sentiment/{lang}/valid.csv",
54
+ "test": "https://raw.githubusercontent.com/IndoNLP/nusax/main/datasets/sentiment/{lang}/test.csv",
55
+ }
56
+
57
+
58
+ def nusantara_config_constructor(lang, schema, version):
59
+ """Construct NusantaraConfig with nusax_senti_{lang}_{schema} as the name format"""
60
+ if schema != "source" and schema != "nusantara_text":
61
+ raise ValueError(f"Invalid schema: {schema}")
62
+
63
+ if lang == "":
64
+ return NusantaraConfig(
65
+ name="nusax_senti_{schema}".format(schema=schema),
66
+ version=datasets.Version(version),
67
+ description="nusax_senti with {schema} schema for all 12 languages".format(schema=schema),
68
+ schema=schema,
69
+ subset_id="nusax_senti",
70
+ )
71
+ else:
72
+ return NusantaraConfig(
73
+ name="nusax_senti_{lang}_{schema}".format(lang=lang, schema=schema),
74
+ version=datasets.Version(version),
75
+ description="nusax_senti with {schema} schema for {lang} language".format(lang=lang, schema=schema),
76
+ schema=schema,
77
+ subset_id="nusax_senti",
78
+ )
79
+
80
+
81
+ LANGUAGES_MAP = {
82
+ "ace": "acehnese",
83
+ "ban": "balinese",
84
+ "bjn": "banjarese",
85
+ "bug": "buginese",
86
+ "eng": "english",
87
+ "ind": "indonesian",
88
+ "jav": "javanese",
89
+ "mad": "madurese",
90
+ "min": "minangkabau",
91
+ "nij": "ngaju",
92
+ "sun": "sundanese",
93
+ "bbc": "toba_batak",
94
+ }
95
+
96
+
97
+ class NusaXSenti(datasets.GeneratorBasedBuilder):
98
+ """NusaX-Senti is a 3-labels (positive, neutral, negative) sentiment analysis dataset for 10 Indonesian local languages + Indonesian and English."""
99
+
100
+ BUILDER_CONFIGS = (
101
+ [nusantara_config_constructor(lang, "source", _SOURCE_VERSION) for lang in LANGUAGES_MAP]
102
+ + [nusantara_config_constructor(lang, "nusantara_text", _NUSANTARA_VERSION) for lang in LANGUAGES_MAP]
103
+ + [nusantara_config_constructor("", "source", _SOURCE_VERSION), nusantara_config_constructor("", "nusantara_text", _NUSANTARA_VERSION)]
104
+ )
105
+
106
+ DEFAULT_CONFIG_NAME = "nusax_senti_ind_source"
107
+
108
+ def _info(self) -> datasets.DatasetInfo:
109
+ if self.config.schema == "source":
110
+ features = datasets.Features(
111
+ {
112
+ "id": datasets.Value("string"),
113
+ "text": datasets.Value("string"),
114
+ "label": datasets.Value("string"),
115
+ }
116
+ )
117
+ elif self.config.schema == "nusantara_text":
118
+ features = schemas.text_features(["negative", "neutral", "positive"])
119
+
120
+ return datasets.DatasetInfo(
121
+ description=_DESCRIPTION,
122
+ features=features,
123
+ homepage=_HOMEPAGE,
124
+ license=_LICENSE,
125
+ citation=_CITATION,
126
+ )
127
+
128
+ def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
129
+ """Returns SplitGenerators."""
130
+ if self.config.name == "nusax_senti_source" or self.config.name == "nusax_senti_nusantara_text":
131
+ # Load all 12 languages
132
+ train_csv_path = dl_manager.download_and_extract([_URLS["train"].format(lang=LANGUAGES_MAP[lang]) for lang in LANGUAGES_MAP])
133
+ validation_csv_path = dl_manager.download_and_extract([_URLS["validation"].format(lang=LANGUAGES_MAP[lang]) for lang in LANGUAGES_MAP])
134
+ test_csv_path = dl_manager.download_and_extract([_URLS["test"].format(lang=LANGUAGES_MAP[lang]) for lang in LANGUAGES_MAP])
135
+ else:
136
+ lang = self.config.name[12:15]
137
+ train_csv_path = Path(dl_manager.download_and_extract(_URLS["train"].format(lang=LANGUAGES_MAP[lang])))
138
+ validation_csv_path = Path(dl_manager.download_and_extract(_URLS["validation"].format(lang=LANGUAGES_MAP[lang])))
139
+ test_csv_path = Path(dl_manager.download_and_extract(_URLS["test"].format(lang=LANGUAGES_MAP[lang])))
140
+
141
+ return [
142
+ datasets.SplitGenerator(
143
+ name=datasets.Split.TRAIN,
144
+ gen_kwargs={"filepath": train_csv_path},
145
+ ),
146
+ datasets.SplitGenerator(
147
+ name=datasets.Split.VALIDATION,
148
+ gen_kwargs={"filepath": validation_csv_path},
149
+ ),
150
+ datasets.SplitGenerator(
151
+ name=datasets.Split.TEST,
152
+ gen_kwargs={"filepath": test_csv_path},
153
+ ),
154
+ ]
155
+
156
+ def _generate_examples(self, filepath: Path) -> Tuple[int, Dict]:
157
+ if self.config.schema != "source" and self.config.schema != "nusantara_text":
158
+ raise ValueError(f"Invalid config: {self.config.name}")
159
+
160
+ if self.config.name == "nusax_senti_source" or self.config.name == "nusax_senti_nusantara_text":
161
+ ldf = []
162
+ for fp in filepath:
163
+ ldf.append(pd.read_csv(fp))
164
+ df = pd.concat(ldf, axis=0, ignore_index=True).reset_index()
165
+ # Have to use index instead of id to avoid duplicated key
166
+ df = df.drop(columns=["id"]).rename(columns={"index": "id"})
167
+ else:
168
+ df = pd.read_csv(filepath).reset_index()
169
+
170
+ for row in df.itertuples():
171
+ ex = {"id": str(row.id), "text": row.text, "label": row.label}
172
+ yield row.id, ex