import os
import pandas as pd
from datasets import DatasetDict, Dataset, Audio
from sklearn.model_selection import train_test_split

def load_data(data_dir, test_size=0.2, seed=42):

    metadata = pd.read_csv(os.path.join(data_dir, "metadata.csv"))

    # Split the data into train and test subsets
    train_data, test_data = train_test_split(metadata, test_size=0.2, random_state=seed)

    # Convert to Hugging Face datasets
    train_dataset = Dataset.from_pandas(train_data)
    test_dataset = Dataset.from_pandas(test_data)

    # Cast the 'path' column to Audio type
    train_dataset = train_dataset.cast_column("path", Audio())
    test_dataset = test_dataset.cast_column("path", Audio())

    # Combine into a DatasetDict
    data = DatasetDict({
        "train": train_dataset,
        "test": test_dataset
    })

    return data

# Load and split the dataset
data = load_data("my_dataset")
print(data)