import gradio as gr
import nltk
from nltk.corpus import stopwords
import joblib

nltk.download('punkt')

# Load the trained model and vectorizer outside the function for better performance
loaded_classifier = joblib.load("bible_or_talmud_model.pkl")
vectorizer = joblib.load("bible_or_talmud_vectorizer.pkl")

def parse_text(new_text):
    new_text_tfidf = vectorizer.transform([new_text])
    prediction = loaded_classifier.predict(new_text_tfidf)
    probabilities = loaded_classifier.predict_proba(new_text_tfidf)
    confidence_score = max(probabilities[0])
    labels = {0: 'אחר', 1: 'תנ"ך', 2: 'תלמוד בבלי'}
    predicted_label = labels[prediction[0]]
    return predicted_label, confidence_score

iface = gr.Interface(fn=parse_text, inputs="text", outputs=["text", "number"], title='גילוי פסוקי התנ"ך באמצעות AI', description='הזן טקסט כדי לסווג אם הוא מהתנ"ך או לא.')
iface.launch()