app.py

from transformers import pipeline
import gradio as gr

trans = pipeline("automatic-speech-recognition", model ="facebook/wav2vec2-large-xlsr-53-spanish")
traductor = pipeline("translation", model = "Helsinki-NLP/opus-mt-es-en")
ner = pipeline("ner", model = "d4data/biomedical-ner-all")

def audio2text(audio):
    text = trans(audio)["text"]
    return text

def text2eng(text):
    return traductor(text)[0]["translation_text"]

def eng2ner(text):
    output = ner(text)
    return {"text": text, "entities": output}

demo = gr.Blocks()

with demo:
    gr.Markdown("Demo sobre historia clínica en español a entidades en ingles")
    audio = gr.Audio(sources="microphone", type="filepath")

    b_text = gr.Button("Transcribir")   
    texto = gr.Textbox() 
    b_text.click(audio2text, inputs=audio, outputs=texto)

    b_trans = gr.Button("Traducir historia")
    transcripcion = gr.Textbox()
    b_trans.click(text2eng, inputs=texto, outputs=transcripcion)

    b_ner =gr.Button("Search entities")
    entidades = gr.HighlightedText()
    b_ner.click(eng2ner, inputs=transcripcion, outputs=entidades)   

demo.launch()