import gradio as gr
import pandas as pd
import re
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.preprocessing import LabelBinarizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score, f1_score

# Define your functions and logic here
def load_and_prepare_data():
    try:
        file_path = 'WELFake_Dataset.csv'  # Ensure this is the correct path
        dataset = pd.read_csv(file_path)
        print(f"Dataset loaded with {dataset.shape[0]} records")
        dataset = dataset.drop(columns=['Unnamed: 0'])
        dataset = dataset.dropna(subset=['title', 'text'])
        dataset['clean_text'] = dataset['text'].apply(clean_text)
        print(f"Dataset cleaned. Records after cleaning: {dataset.shape[0]}")
        return dataset
    except Exception as e:
        return f"Error loading and preparing data: {e}"

def clean_text(text):
    try:
        text = re.sub(r'\W', ' ', text)
        text = re.sub(r'\s+', ' ', text)
        text = re.sub(r'\d', '', text)
        text = text.lower().strip()
        return text
    except Exception as e:
        return f"Error cleaning text: {e}"

def train_model(dataset):
    try:
        X_train, X_test, y_train, y_test = train_test_split(dataset['clean_text'], dataset['label'], test_size=0.2, random_state=42)
        print(f"Training data size: {X_train.shape[0]}, Test data size: {X_test.shape[0]}")
        
        vectorizer = TfidfVectorizer(max_features=10000)
        X_train_tfidf = vectorizer.fit_transform(X_train)
        X_test_tfidf = vectorizer.transform(X_test)
        
        lb = LabelBinarizer()
        y_train_binary = lb.fit_transform(y_train)
        y_test_binary = lb.transform(y_test)
        
        log_reg_model = LogisticRegression(max_iter=1000)
        log_reg_model.fit(X_train_tfidf, y_train)
        
        y_pred_log_reg_train = log_reg_model.predict(X_train_tfidf)
        train_accuracy_log_reg = accuracy_score(y_train, y_pred_log_reg_train)
        train_f1_log_reg = f1_score(y_train, y_pred_log_reg_train)
        
        y_pred_log_reg = log_reg_model.predict(X_test_tfidf)
        accuracy_log_reg = accuracy_score(y_test, y_pred_log_reg)
        f1_log_reg = f1_score(y_test, y_pred_log_reg)
        
        print(f"Train Accuracy: {train_accuracy_log_reg}, Train F1 Score: {train_f1_log_reg}")
        print(f"Test Accuracy: {accuracy_log_reg}, Test F1 Score: {f1_log_reg}")
        
        return vectorizer, lb, log_reg_model, accuracy_log_reg, f1_log_reg
    except Exception as e:
        return f"Error training model: {e}"

def fake_news_detection(text):
    try:
        dataset = load_and_prepare_data()
        if isinstance(dataset, str):  # Check if there was an error in loading data
            return dataset
        vectorizer, lb, log_reg_model, accuracy_log_reg, f1_log_reg = train_model(dataset)
        if isinstance(vectorizer, str):  # Check if there was an error in training models
            return vectorizer
        
        clean_text_input = clean_text(text)
        text_tfidf = vectorizer.transform([clean_text_input])
        prediction = log_reg_model.predict_proba(text_tfidf)
        result = "Real" if prediction[0][1] >= 0.5 else "Fake"
        
        return f"Prediction: {result}"
    except Exception as e:
        return f"Error in fake news detection: {e}"

iface = gr.Interface(
    fn=fake_news_detection, 
    inputs=gr.Textbox(lines=2, placeholder="Enter news text here..."), 
    outputs="text",
    title="Fake News Detector",
    description="Enter a news headline or article text to check if it is fake or real."
)

if __name__ == "__main__":
    iface.launch()