app.py

import gradio as gr
import pickle
from sklearn.feature_extraction.text import TfidfVectorizer
import pandas as pd

# def predict_text(text):
#     with open('model.pkl', 'rb') as file:
#         loaded_model, vectorizer = pickle.load(file)
#     vect_input=vectorizer.transform([text])
#     # with open('model.pkl', 'rb') as file:
#     #     loaded_model = pickle.load(file)
#     print("=== Model Loading ===", loaded_model)
#     text_label = loaded_model.predict(vect_input)
#     print(loaded_model.predict(vect_input))
#     return text , "is" , text_label[0], "generated"

# demo = gr.Interface(fn=predict_text, inputs="text", outputs="text")
# demo.launch(share=True)


import pickle
import gradio as gr

# Define the prediction function
def predict_text(text):
    # Load the model and vectorizer
    with open('model.pkl', 'rb') as file:
        loaded_model, vectorizer = pickle.load(file)
    
    # Transform the input text using the loaded vectorizer
    vect_input = vectorizer.transform([text])
    
    # Make predictions using the loaded model
    text_label = loaded_model.predict(vect_input)
    
    # Return the prediction result
    if text_label[0] == "chatgpt":
        result = "AI-generated text"
    else:
        result = "Human-written text"
    
    return f"The text you entered is: {result}"

# Build a simple Gradio interface
demo = gr.Interface(
    fn=predict_text,          # Function to call for predictions
    inputs=gr.Textbox(lines=5, placeholder="Enter your text here..."),  # Input textbox for the text
    outputs="text",           # Output as a text label
    title="AI vs Human Text Classifier",  # Title of the interface
    description="Enter a piece of text to find out if it was written by AI or a human.",  # Description
    theme="compact",          # Gradio theme for simplicity
    #live=True                 # Make it live without refreshing
)

# Launch the interface
demo.launch(share=True)