sesge.py

import csv
import os
import json

import datasets
from datasets.utils.py_utils import size_str
from tqdm import tqdm

from scipy.io.wavfile import read, write
import io

#from .release_stats import STATS

_CITATION = """\
@inproceedings{demint2024,
  author = {Pérez-Ortiz, Juan Antonio and 
    Esplà-Gomis, Miquel and 
    Sánchez-Cartagena, Víctor M. and 
    Sánchez-Martínez, Felipe and 
    Chernysh, Roman and 
    Mora-Rodríguez, Gabriel and 
    Berezhnoy, Lev},
  title = {{DeMINT}: Automated Language Debriefing for English Learners via {AI} 
           Chatbot Analysis of Meeting Transcripts},
  booktitle = {Proceedings of the 13th Workshop on NLP for Computer Assisted Language Learning},
  month = october,
  year = {2024},
  url = {https://aclanthology.org/volumes/2024.nlp4call-1/},
}
"""

class SesgeConfig(datasets.BuilderConfig):
    def __init__(self, name, version, **kwargs):
        self.language = kwargs.pop("language", None)
        self.release_date = kwargs.pop("release_date", None)
        description = (
            "A dataset containing English speech with grammatical errors, along with the corresponding transcriptions."
            "Utterances are synthesized using a text-to-speech model, whereas the grammatically incorrect texts come from the C4_200M synthetic dataset."
        )

        super(SesgeConfig, self).__init__(
            name=name,
            **kwargs,
        )

class Sesge():

    BUILDER_CONFIGS = [
        SesgeConfig(
            name="sesge",
            version=1.0,
            language='eng',
            release_date="2024-10-8",
        )
    ]

    def _info(self):
        total_languages = 1
        total_valid_hours = 1
        description = (
            "A dataset containing English speech with grammatical errors, along with the corresponding transcriptions."
            "Utterances are synthesized using a text-to-speech model, whereas the grammatically incorrect texts come from the C4_200M synthetic dataset."
        )
        features = datasets.Features(
            {
                "audio": datasets.features.Audio(sampling_rate=48_000),
                "sentence": datasets.Value("string"),
            }
        )

    def _generate_examples(self, local_extracted_archive_paths, archives, meta_path, split):
        archives = os.listdir(archives)
        metadata = {}
        with open(meta_path, encoding="utf-8") as f:
            reader = csv.DictReader(f, delimiter=";", quoting=csv.QUOTE_NONE)
            for row in tqdm(reader):
                metadata[row["file_name"]] = row

        for i, path in enumerate(archives):
            #for path, file in audio_archive:
            _, filename = os.path.split(path)
            file = os.path.join("data", split, filename)
            if file in metadata:
                result = dict(metadata[file])
                print("Result: ", result)
                with open(os.path.join(local_extracted_archive_paths, filename), 'rb') as wavfile:
                    input_wav = wavfile.read()

                rate, data = read(io.BytesIO(input_wav))
                
                path = os.path.join(local_extracted_archive_paths[i], path)
                result["audio"] = {"path": path, "bytes": data}
                result["path"] = path
                yield path, result
            else:
                print("No file found")
                yield None, None