import gradio as gr
import librosa
from transformers import AutoFeatureExtractor, pipeline


def load_and_fix_data(input_file, model_sampling_rate):
    speech, sample_rate = librosa.load(input_file)
    if len(speech.shape) > 1:
        speech = speech[:, 0] + speech[:, 1]
    if sample_rate != model_sampling_rate:
        speech = librosa.resample(speech, sample_rate, model_sampling_rate)
    return speech


feature_extractor = AutoFeatureExtractor.from_pretrained("jonatasgrosman/wav2vec2-large-xlsr-53-spanish")
sampling_rate = feature_extractor.sampling_rate

asr = pipeline("automatic-speech-recognition", model="jonatasgrosman/wav2vec2-large-xlsr-53-spanish")


def predict_and_ctc_lm_decode(input_file):
    speech = load_and_fix_data(input_file, sampling_rate)
    transcribed_text = asr(speech, chunk_length_s=5, stride_length_s=1)["text"]
    pipe2 = pipeline("text-classification", model = "hackathon-pln-es/twitter_sexismo-finetuned-robertuito-exist2021")
    sexism_detection = pipe2(transcribed_text)[0]['label']

    return sexism_detection


gr.Interface(
    predict_and_ctc_lm_decode,
    inputs=[gr.inputs.Audio(source="microphone", type="filepath", label="Record your audio")],
    #outputs=[gr.outputs.Label(num_top_classes=2),gr.outputs.Label(num_top_classes=2), gr.outputs.Label(num_top_classes=2)],
    outputs=[gr.outputs.Textbox(label="Predicción")],
    examples=[["audio1.wav"], ["audio2.wav"], ["audio3.wav"], ["audio4.wav"], ["sample_audio.wav"]],
    title="Spanish-Audio-Transcription-based-Sexism-Detection",
    description="This is a Gradio demo for Sentiment Analysis of Transcribed Spanish Audio",
    layout="horizontal",
    theme="huggingface",
).launch(enable_queue=True, cache_examples=True)