emotion_id_opinion.py

from pathlib import Path
from typing import List

import datasets
import pandas as pd

from seacrowd.utils import schemas
from seacrowd.utils.configs import SEACrowdConfig
from seacrowd.utils.constants import DEFAULT_SEACROWD_VIEW_NAME, DEFAULT_SOURCE_VIEW_NAME, Tasks

_DATASETNAME = "emotion_id_opinion"
_SOURCE_VIEW_NAME = DEFAULT_SOURCE_VIEW_NAME
_UNIFIED_VIEW_NAME = DEFAULT_SEACROWD_VIEW_NAME

_LANGUAGES = ["ind"]  # We follow ISO639-3 langauge code (https://iso639-3.sil.org/code_tables/639/data)
_LOCAL = False
_CITATION = """\
@article{RICCOSAN2022108465,
title = {Emotion dataset from Indonesian public opinion},
journal = {Data in Brief},
volume = {43},
pages = {108465},
year = {2022},
issn = {2352-3409},
doi = {https://doi.org/10.1016/j.dib.2022.108465},
url = {https://www.sciencedirect.com/science/article/pii/S2352340922006588},
author = { Riccosan and Karen Etania Saputra and Galih Dea Pratama and Andry Chowanda},
keywords = {Emotion classification, Dataset, Tweet, Indonesia},
abstract = {An opinion is a type of judgment or a person's point of view about something. Twitter is a popular social media platform that includes a lot of public opinions and would be a suitable location to mine data in text form. With its vast population and active Twitter user base, Indonesia has the potential to be a source of opinion data mining. An opinion may be processed and result in the form of a person's emotional response towards something, such as whether they like, hate, love, or are happy about it. Upon that basis, a dataset of Indonesian-language tweets conveying public opinion on various topics was formed. The fact that there are only limited publicly available emotions text datasets in the Indonesian language supports our basis in this research to form our emotion dataset. The gathered data was cleaned and normalized in the pre-processing stage to the necessary form for study on the task of classifying emotions in Indonesian. The data collected is annotated with six emotional labels: anger, fear, joy, love, sad, and neutral.}
}
"""

_DESCRIPTION = """\
Emotion ID Opinion is a dataset of Indonesian-language tweets conveying public opinion on a variety of topics.
It comtains 7080 indunesian tweets and a person's emotion response towards each tweet.
The data is annotated with six emotional labels, namely anger, fear, joy, love, sad, and neutral.
"""

_HOMEPAGE = "https://github.com/Ricco48/Emotion-Dataset-from-Indonesian-Public-Opinion"

_LICENSE = "Creative Commons Attribution Share-Alike 4.0 International"

_URLs = {
    "anger": "https://raw.githubusercontent.com/Ricco48/Emotion-Dataset-from-Indonesian-Public-Opinion/main/Emotion%20Dataset%20from%20Indonesian%20Public%20Opinion/AngerData.csv",
    "fear": "https://raw.githubusercontent.com/Ricco48/Emotion-Dataset-from-Indonesian-Public-Opinion/main/Emotion%20Dataset%20from%20Indonesian%20Public%20Opinion/FearData.csv",
    "joy": "https://raw.githubusercontent.com/Ricco48/Emotion-Dataset-from-Indonesian-Public-Opinion/main/Emotion%20Dataset%20from%20Indonesian%20Public%20Opinion/JoyData.csv",
    "love": "https://raw.githubusercontent.com/Ricco48/Emotion-Dataset-from-Indonesian-Public-Opinion/main/Emotion%20Dataset%20from%20Indonesian%20Public%20Opinion/LoveData.csv",
    "sad": "https://raw.githubusercontent.com/Ricco48/Emotion-Dataset-from-Indonesian-Public-Opinion/main/Emotion%20Dataset%20from%20Indonesian%20Public%20Opinion/SadData.csv",
    "neutral": "https://raw.githubusercontent.com/Ricco48/Emotion-Dataset-from-Indonesian-Public-Opinion/main/Emotion%20Dataset%20from%20Indonesian%20Public%20Opinion/NeutralData.csv"
}

_SUPPORTED_TASKS = [Tasks.EMOTION_CLASSIFICATION]

_SOURCE_VERSION = "1.0.0"
_SEACROWD_VERSION = "2024.06.20"


class EmoIdOpinion(datasets.GeneratorBasedBuilder):

    BUILDER_CONFIGS = [
        SEACrowdConfig(
            name="emotion_id_opinion_source",
            version=datasets.Version(_SOURCE_VERSION),
            description="EmoIdOpinion source schema",
            schema="source",
            subset_id="emotion_id_opinion",
        ),
        SEACrowdConfig(
            name="emotion_id_opinion_seacrowd_text",
            version=datasets.Version(_SEACROWD_VERSION),
            description="EmoIdOpinion Nusantara schema",
            schema="seacrowd_text",
            subset_id="emotion_id_opinion",
        ),
    ]

    DEFAULT_CONFIG_NAME = "emotion_id_opinion_source"

    def _info(self):
        if self.config.schema == "source":
            features = datasets.Features(
                {
                    "index": datasets.Value("string"),
                    "tweet": datasets.Value("string"),
                    "label": datasets.Value("string"),
                }
            )
        elif self.config.schema == "seacrowd_text":
            features = schemas.text_features(["Joy", "Love", "Fear", "Anger", "Sad", "Neutral"])

        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
        anger_tsv_path = Path(dl_manager.download_and_extract(_URLs["anger"]))
        fear_tsv_path = Path(dl_manager.download_and_extract(_URLs["fear"]))
        joy_tsv_path = Path(dl_manager.download_and_extract(_URLs["joy"]))
        love_tsv_path = Path(dl_manager.download_and_extract(_URLs["love"]))
        neutral_tsv_path = Path(dl_manager.download_and_extract(_URLs["neutral"]))
        sad_tsv_path = Path(dl_manager.download_and_extract(_URLs["sad"]))

        data_files = {
            "anger": anger_tsv_path,
            "fear": fear_tsv_path,
            "joy": joy_tsv_path,
            "love": love_tsv_path,
            "neutral": neutral_tsv_path,
            "sad": sad_tsv_path
        }

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={"filepath": [
                    data_files["anger"],
                    data_files["fear"],
                    data_files["joy"],
                    data_files["love"],
                    data_files["neutral"],
                    data_files["sad"]
                ]},
            ),
        ]

    def _generate_examples(self, filepath: List[Path]):
        increment = 0
        for i, fp in enumerate(filepath):
        # df = pd.concat([pd.read_csv(fp, sep="\t", header="infer").reset_index() for fp in filepath])
            df = pd.read_csv(fp, sep="\t", header="infer").reset_index()
            df.columns = ["id", "Tweet", "Label"]

            if self.config.schema == "source":
                for row in df.itertuples():
                    ex = {"index": str(increment + row.id), "tweet": row.Tweet, "label": row.Label}
                    yield increment + row.id, ex
            elif self.config.schema == "seacrowd_text":
                for row in df.itertuples():
                    ex = {"id": str(increment + row.id), "text": row.Tweet, "label": row.Label}
                    yield increment + row.id, ex
            else:
                raise ValueError(f"Invalid config: {self.config.name}")

            increment += row.id + 1