import streamlit as st
from transformers import T5Tokenizer, T5ForConditionalGeneration
from pdfminer.high_level import extract_text

def main():
    st.title("PDF Translation")
    
    # Upload the pdf
    uploaded_file = st.file_uploader("Upload a PDF file and we will translate the text inside to German and French.", type=["pdf"])

    if uploaded_file is not None:
        # Extract text from pdf
        text = extract_text(uploaded_file)
        tokenizer = T5Tokenizer.from_pretrained("t5-small")
        model = T5ForConditionalGeneration.from_pretrained("t5-small")

        # Define translation prefixes for each language
        translation_prefixes = {
            "german": "translate English to German: ",
            "french": "translate English to French: "
        }

        # Generate translations for each language
        translations = {}

        # Buttons to trigger translation
        translate_german = st.button("Translate to German")
        translate_french = st.button("Translate to French")

        for language, prefix in translation_prefixes.items():
            # Translate the entire text, not page by page
            text_to_translate = prefix + text
            input_ids = tokenizer(text_to_translate, return_tensors="pt").input_ids
            outputs = model.generate(input_ids=input_ids, max_length=150, num_beams=4, no_repeat_ngram_size=2)
            translated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
            translations[language] = translated_text

        # Display the translations based on the button clicked
        if translate_german:
            display_translation(translations["german"], "German")

        if translate_french:
            display_translation(translations["french"], "French")


def display_translation(translation, language):
    st.write(f"\nLanguage: {language}")
    st.write(f"Translation: {translation}")


if __name__ == "__main__":
    main()