from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech, SpeechT5HifiGan
from datasets import load_dataset
import torch
import soundfile as sf
from datasets import load_dataset
from transformers import pipeline
import gradio as gr
import tempfile


summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
speech = pipeline("text-to-speech", model="microsoft/speecht5_tts")     

# code from the Model card
processor = SpeechT5Processor.from_pretrained("microsoft/speecht5_tts")
model = SpeechT5ForTextToSpeech.from_pretrained("microsoft/speecht5_tts")
vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
speaker_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
speaker_embeddings = torch.tensor(speaker_dataset[0]["xvector"]).unsqueeze(0)

def summarize_text_and_speak(prompt):
    summary = summarizer(prompt, max_length=150, min_length=30, do_sample=False)
    summary_text = summary[0]['summary_text']
    
    #inputs = processor(text="Hello, my dog is cute.", return_tensors="pt")
    inputs = processor(text=summary_text, return_tensors="pt")
    #speech = model.generate_speech(inputs["input_ids"], speaker_embeddings, vocoder=vocoder)
    speech_audio  = model.generate_speech(inputs["input_ids"], speaker_embeddings, vocoder=vocoder)
    
    #sf.write("speech.wav", speech.numpy(), samplerate=16000)
    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
        sf.write(tmp_file.name, speech_audio.numpy(), samplerate=16000)
        audio_path = tmp_file.name
        
    return summary_text, audio_path
    
interface = gr.Interface(
    fn=summarize_text_and_speak,
    inputs=gr.Textbox(lines=10, label="Input text"),
    outputs=[gr.Textbox(label="Summary"), gr.Audio(label="Audio")]
)

interface.launch(share=True)