from datasets import load_dataset
from IPython.display import Audio

# 1. Load the dataset in 'streaming' mode
# Configuration 'fil_ph' is for Filipino (Tagalog)
dataset = load_dataset("google/fleurs", "fil_ph", split="train", streaming=True, trust_remote_code=True)

# 2. Grab just the first sample
sample = next(iter(dataset))

# 3. View the metadata
print(f"Transcription: {sample['transcription']}")
print(f"Sampling Rate: {sample['audio']['sampling_rate']}")

# 4. Listen to the audio (if in a Jupyter/Colab notebook)
Audio(sample["audio"]["array"], rate=sample["audio"]["sampling_rate"])