app.py

import gradio as gr
from transformers import pipeline, T5Tokenizer, T5ForConditionalGeneration, AutoTokenizer, AutoModelForSeq2SeqLM
import os
import requests

# Load environment variable for Hugging Face API token
token = os.getenv("HF_TOKEN")
headers = {"Authorization": f"Bearer {token}"}

# Load summarization model and tokenizer
tokenizer = T5Tokenizer.from_pretrained("sumedh/t5-base-amazonreviews", clean_up_tokenization_spaces=True)
model = T5ForConditionalGeneration.from_pretrained("sumedh/t5-base-amazonreviews")
summarizer = pipeline("summarization", model=model, tokenizer=tokenizer)

# Translation API details
API_URL = "https://api-inference.huggingface.co/models/Helsinki-NLP/opus-mt-en-es"

# Summarization and Translation Function
def texto_sum(text):
    # Summarize the input text
    summary = summarizer(text, do_sample=False)[0]['summary_text']
    
    # Translate summary using the Hugging Face API
    response = requests.post(API_URL, headers=headers, json={"inputs": summary})
    translation = response.json()
    
    # Check if translation is successful
    if 'error' in translation:
        return f"Error in translation: {translation['error']}"
    
    return translation[0]['translation_text']

# Gradio interface
demo = gr.Interface(
    fn=texto_sum,
    inputs=gr.Textbox(label="Texto a introducir:", placeholder="Introduce el texto a resumir aquí..."),
    outputs="text"
)

# Launch the interface
demo.launch()