import torch
import gradio as gr
from transformers import Wav2Vec2FeatureExtractor
from datasets import Dataset
import librosa

feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained("superb/hubert-large-superb-er")

def get_emotion(microphone, file_upload, task):
    warn_output = ""
    
    if (microphone is not None) and (file_upload is not None):
        warn_output = (
            "WARNING: You've uploaded an audio file and used the microphone. "
            "The recorded file from the microphone will be used and the uploaded audio will be discarded.\n"
        )
    elif (microphone is None) and (file_upload is None):
        return "ERROR: You have to either use the microphone or upload an audio file"

    file = microphone if microphone is not None else file_upload
    test = feature_extractor(file, sampling_rate=16000, padding=True, return_tensors="pt" ).to(torch.float32)       
    logits = model(**test).logits
    predicted_ids = torch.argmax(logits, dim=-1)
    labels = [model.config.id2label[_id] for _id in predicated_ids.tolist()]
    return labels

demo = gr.Blocks()

mf_transcribe = gr.Interface(
    fn=get_emotion,
    inputs=[
        gr.inputs.Audio(source="microphone", type="filepath", optional=True),
        gr.inputs.Audio(source="upload", type="filepath", optional=True),
    ],
    outputs="text",
    layout="horizontal",
    theme="huggingface",
    title="AER",
    description=(
        "get the emotion"
    ),
    allow_flagging="never",
)

with demo:
    gr.TabbledInterface([mf_transcribe],'Trancribe')
demo.launch(enable_queue=True)