Datasets:

ArXiv:
License:
holylovenia commited on
Commit
31bb229
1 Parent(s): 50e04d5

Upload xed.py with huggingface_hub

Browse files
Files changed (1) hide show
  1. xed.py +140 -0
xed.py ADDED
@@ -0,0 +1,140 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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 Licenses, Tasks
10
+
11
+ _CITATION = """
12
+ @inproceedings{ohman2020xed,
13
+ title={{XED}: A Multilingual Dataset for Sentiment Analysis and Emotion Detection},
14
+ author={{\"O}hman, Emily and P{`a}mies, Marc and Kajava, Kaisla and Tiedemann, J{\"o}rg},
15
+ booktitle={The 28th International Conference on Computational Linguistics (COLING 2020)},
16
+ year={2020}
17
+ }
18
+ """
19
+ _DATASETNAME = "xed"
20
+
21
+ _DESCRIPTION = """\
22
+ This is the XED dataset. The dataset consists of emotion annotated movie subtitles
23
+ from OPUS. We use Plutchik's 8 core emotions to annotate. The data is multilabel.
24
+ The original annotations have been sourced for mainly English and Finnish, with the
25
+ rest created using annotation projection to aligned subtitles in 41 additional languages,
26
+ with 31 languages included in the final dataset (more than 950 lines of annotated subtitle
27
+ lines). The dataset is an ongoing project with forthcoming additions such as machine translated datasets.
28
+ """
29
+
30
+ _HOMEPAGE = "https://github.com/Helsinki-NLP/XED"
31
+
32
+ _LANGUAGES = ["ind", "vie"]
33
+
34
+ # This License is from the bottom of homepage's README not Unknown (as from Issues)
35
+ _LICENSE = Licenses.CC_BY_4_0.value
36
+
37
+ _LOCAL = False
38
+
39
+ _URLS = {"ind": "https://raw.githubusercontent.com/Helsinki-NLP/XED/master/Projections/id-projections.tsv", "vie": "https://raw.githubusercontent.com/Helsinki-NLP/XED/master/Projections/vi-projections.tsv"}
40
+
41
+ # Because of the multi-label attribute, I choose ASPECT_BASED_SENTIMENT_ANALYSIS than SENTIMENT_ANALYSIS
42
+ _SUPPORTED_TASKS = [Tasks.ASPECT_BASED_SENTIMENT_ANALYSIS]
43
+
44
+ _SOURCE_VERSION = "1.0.0"
45
+
46
+ _SEACROWD_VERSION = "2024.06.20"
47
+
48
+
49
+ class XEDDataset(datasets.GeneratorBasedBuilder):
50
+ """
51
+ This is the XED dataset. The dataset consists of emotion annotated movie subtitles
52
+ from OPUS. We use Plutchik's 8 core emotions to annotate. The data is multilabel.
53
+ The original annotations have been sourced for mainly English and Finnish, with the
54
+ rest created using annotation projection to aligned subtitles in 41 additional languages,
55
+ with 31 languages included in the final dataset (more than 950 lines of annotated subtitle
56
+ lines). The dataset is an ongoing project with forthcoming additions such as machine translated datasets.
57
+ """
58
+
59
+ BUILDER_CONFIGS = [
60
+ SEACrowdConfig(
61
+ name=f"{_DATASETNAME}_{LANG}_source",
62
+ version=datasets.Version(_SOURCE_VERSION),
63
+ description=f"{_DATASETNAME} {LANG} source schema",
64
+ schema="source",
65
+ subset_id=f"{_DATASETNAME}_{LANG}",
66
+ )
67
+ for LANG in _LANGUAGES
68
+ ] + [
69
+ SEACrowdConfig(
70
+ name=f"{_DATASETNAME}_{LANG}_seacrowd_text_multi",
71
+ version=datasets.Version(_SEACROWD_VERSION),
72
+ description=f"{_DATASETNAME} {LANG} SEACrowd schema",
73
+ schema="seacrowd_text_multi",
74
+ subset_id=f"{_DATASETNAME}_{LANG}",
75
+ )
76
+ for LANG in _LANGUAGES
77
+ ]
78
+
79
+ DEFAULT_CONFIG_NAME = f"{_DATASETNAME}_ind_source"
80
+ _LABELS = ["Anger", "Anticipation", "Disgust", "Fear", "Joy", "Sadness", "Surprise", "Trust"]
81
+
82
+ def _info(self) -> datasets.DatasetInfo:
83
+
84
+ if self.config.schema == "source":
85
+ features = datasets.Features({"Sentence": datasets.Value("string"), "Emotions": datasets.Sequence(feature=datasets.ClassLabel(names=self._LABELS))})
86
+
87
+ elif self.config.schema == "seacrowd_text_multi":
88
+ features = schemas.text_multi_features(self._LABELS)
89
+
90
+ return datasets.DatasetInfo(
91
+ description=_DESCRIPTION,
92
+ features=features,
93
+ homepage=_HOMEPAGE,
94
+ license=_LICENSE,
95
+ citation=_CITATION,
96
+ )
97
+
98
+ def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
99
+ """Returns SplitGenerators."""
100
+
101
+ language = self.config.name.split("_")[1]
102
+
103
+ if language in _LANGUAGES:
104
+ data_path = Path(dl_manager.download_and_extract(_URLS[language]))
105
+ else:
106
+ data_path = [Path(dl_manager.download_and_extract(_URLS[language])) for language in _LANGUAGES]
107
+
108
+ return [
109
+ datasets.SplitGenerator(
110
+ name=datasets.Split.TRAIN,
111
+ gen_kwargs={
112
+ "filepath": data_path,
113
+ "split": "train",
114
+ },
115
+ )
116
+ ]
117
+
118
+ def _generate_examples(self, filepath: Path, split: str) -> Tuple[int, Dict]:
119
+ """Yields examples as (key, example) tuples."""
120
+
121
+ emotions_mapping = {1: "Anger", 2: "Anticipation", 3: "Disgust", 4: "Fear", 5: "Joy", 6: "Sadness", 7: "Surprise", 8: "Trust"}
122
+
123
+ df = pd.read_csv(filepath, sep="\t", names=["Sentence", "Emotions"], index_col=None)
124
+ df["Emotions"] = df["Emotions"].apply(lambda x: list(map(int, x.split(", "))))
125
+ df["Emotions"] = df["Emotions"].apply(lambda x: [emotions_mapping[emotion] for emotion in x])
126
+
127
+ for index, row in df.iterrows():
128
+
129
+ if self.config.schema == "source":
130
+ example = row.to_dict()
131
+
132
+ elif self.config.schema == "seacrowd_text_multi":
133
+
134
+ example = {
135
+ "id": str(index),
136
+ "text": str(row["Sentence"]),
137
+ "labels": row["Emotions"],
138
+ }
139
+
140
+ yield index, example