app.py

import joblib
import gradio as gr
import re



model = joblib.load("./phishing_model.pkl")
vectorizer = joblib.load("./phishing_tfidf_vectorizer.pkl")


def text_clean(text):
    if not isinstance(text, str):
        return ""

    text = text.lower()
    text = re.sub(r"<.*?>", " ", text)
    text = re.sub(r"https?://\S+|www\.\S+", " url ", text)
    text = re.sub(r"\b\d{7,}\b", " ", text)
    text = re.sub(r"[^a-z0-9@.$%\-\s]", " ", text)
    text = re.sub(r"\s+", " ", text).strip()

    return text


def predict(email):
    cleaned = text_clean(email)
    vector = vectorizer.transform([cleaned])

    prediction = model.predict(vector)[0]
    probabilities = model.predict_proba(vector)[0]

    label = "Phishing" if prediction == 1 else "Legitimate"

    return (
        label,
        f"{probabilities[prediction] * 100:.2f}%",
        {
            "Legitimate": float(probabilities[0]),
            "Phishing": float(probabilities[1]),
        },
    )


examples = {
    "Legitimate - Meeting": """Hi Aditya,

Can we meet tomorrow at 3 PM to discuss the internship project?

Thanks,
Rahul""",

    "Legitimate - GitHub": """GitHub

Your pull request has been successfully merged into the main branch.

View changes:
https://github.com/example/repo""",

    "Phishing - Bank": """support@secure-bank-login.xyz
secure-bank-login.xyz

Dear Customer,

Your account has been temporarily suspended.

Click below immediately to verify your identity.

https://secure-bank-login.xyz/login""",

    "Phishing - PayPal": """service@paypal-security.xyz
paypal-security.xyz

We've detected unusual activity on your PayPal account.

Verify your account within 24 hours or it will be permanently limited.

https://paypal-security.xyz"""
}


def load_example(choice):
    return examples[choice]


with gr.Blocks(title="Email Phishing Detector") as demo:

    gr.Markdown(
        """
        # Email Phishing Detector

        Detect whether an email is **Legitimate** or **Phishing** using a Multinomial Naive Bayes model trained on TF-IDF features.
        """
    )

    with gr.Row():

        with gr.Column(scale=3):
            email_input = gr.Textbox(
                label="Email Content",
                lines=18,
                placeholder="Paste the complete email here..."
            )

            predict_btn = gr.Button("Predict", variant="primary")

        with gr.Column(scale=2):
            prediction = gr.Textbox(label="Prediction")
            confidence = gr.Textbox(label="Confidence")
            probabilities = gr.Label(label="Class Probabilities")

    gr.Markdown("### Try an Example")

    example_dropdown = gr.Dropdown(
        choices=list(examples.keys()),
        value=list(examples.keys())[0],
        label="Example Emails"
    )

    example_dropdown.change(
        load_example,
        inputs=example_dropdown,
        outputs=email_input
    )

    predict_btn.click(
        predict,
        inputs=email_input,
        outputs=[
            prediction,
            confidence,
            probabilities
        ]
    )

demo.launch()