app.py

import streamlit as st
import pandas as pd
import numpy as np
import sklearn
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import MultinomialNB
from sklearn.tree import DecisionTreeClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.metrics import accuracy_score
from fastapi import FastAPI
from fastapi.responses import JSONResponse
import threading

# Read dataset
df = pd.read_csv(r"spam.csv")

# Initialize Streamlit app
st.title("Identifying Spam and Ham Emails")

# Define feature and target variables
x = df["Message"]
y = df["Category"]

# Create a Bag of Words (BoW) model
bow = CountVectorizer(stop_words="english")
final_data = pd.DataFrame(bow.fit_transform(x).toarray(), columns=bow.get_feature_names_out())

# Train-test split
x_train, x_test, y_train, y_test = train_test_split(final_data, y, test_size=0.2, random_state=20)

# Initialize models
models = {
    "Naive Bayes": MultinomialNB(),
    "KNN": KNeighborsClassifier(),
    "Logistic Regression": LogisticRegression(),
    "Decision Tree": DecisionTreeClassifier(),
    "SVM": SVC()
}

# Model selection
model_choice = st.selectbox("Choose a Classification Algorithm", list(models.keys()))

# Train the selected model
obj = models[model_choice]
obj.fit(x_train, y_train)
y_pred = obj.predict(x_test)
accuracy = accuracy_score(y_test, y_pred)

# Display accuracy
if st.button("Show Accuracy"):
    st.write(f"Accuracy of {model_choice}: {accuracy:.4f}")

# Email input and prediction function
email_input = st.text_input("Enter an Email for Prediction")

def predict_email(email):
    data = bow.transform([email]).toarray() 
    prediction = obj.predict(data)[0]
    st.write(f"Prediction: {prediction}")

if st.button("Predict Email"):
    if email_input:
        predict_email(email_input)
    else:
        st.write(":red[Please enter an email to classify]")

# FastAPI app to handle GET requests
app = FastAPI()

@app.get("/predict/")
def predict_spam(email: str):
    """
    This endpoint predicts whether the email is Spam or Ham.
    Query parameter: email (str) - The email text to be classified.
    """
    data = bow.transform([email]).toarray()
    prediction = obj.predict(data)[0]
    return JSONResponse(content={"prediction": prediction})

# Running FastAPI in a separate thread to work alongside Streamlit
def run_api():
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)

# Start FastAPI in a separate thread
api_thread = threading.Thread(target=run_api, daemon=True)
api_thread.start()

# You can also check API response using the link below:
# http://localhost:8000/predict/?email=Your_email_text_here