Update app.py

app.py

@@ -1,86 +1,67 @@
-import streamlit as st
-import prediction_pipeline as pp
-import pandas as pd
-import os
 
-# -----------------------------
-# PAGE CONFIG
-# Must be first Streamlit command
-st.set_page_config(
-    page_title="Tourism Package Prediction",
-    page_icon="✈️"
-)
+from fastapi import FastAPI
+from pydantic import BaseModel
+import pandas as pd
+import numpy as np
 
-st.title("✈️ Tourism Package Prediction")
-st.write("Predict whether a customer will purchase a tourism package.")
+from prediction_pipeline import load_artifacts, preprocess_input, MODEL, X_TRAIN_ENCODED_COLUMNS
 
-# -----------------------------
-# LOAD MODEL
-@st.cache_resource
-def load_model():
-    pp.load_artifacts()
-    return pp.MODEL, pp.X_TRAIN_ENCODED_COLUMNS
+# Define the input data model using Pydantic BaseModel
+class PredictionRequest(BaseModel):
+    Age: float
+    TypeofContact: str
+    CityTier: int
+    DurationOfPitch: float
+    Occupation: str
+    Gender: str
+    NumberOfPersonVisiting: int
+    NumberOfFollowups: float
+    ProductPitched: str
+    PreferredPropertyStar: float
+    MaritalStatus: str
+    NumberOfTrips: float
+    Passport: int
+    PitchSatisfactionScore: int
+    OwnCar: int
+    NumberOfChildrenVisiting: float
+    Designation: str
+    MonthlyIncome: float
 
-model, columns = load_model()
+# Initialize FastAPI app
+app = FastAPI()
 
-if model is None or columns is None:
-    st.error("❌ Model failed to load. Check logs for details.")
-    st.stop()
+# Load model and column names when the application starts
+# This ensures artifacts are loaded only once
+@app.on_event("startup")
+async def startup_event():
+    print("Loading model and column names on startup...")
+    load_artifacts()
+    if MODEL is None or X_TRAIN_ENCODED_COLUMNS is None:
+        raise RuntimeError("Failed to load model or column names during startup.")
+    print("Model and column names loaded successfully.")
 
-# -----------------------------
-# UI INPUTS
-Age = st.number_input("Age", min_value=0.0, value=35.0)
-TypeofContact = st.selectbox("Type of Contact", ["Self Enquiry", "Company Invited"])
-CityTier = st.selectbox("City Tier", [1, 2, 3])
-DurationOfPitch = st.number_input("Duration Of Pitch", min_value=0.0, value=10.0)
-Occupation = st.selectbox("Occupation", ["Salaried", "Free Lancer", "Small Business", "Large Business"])
-Gender = st.selectbox("Gender", ["Male", "Female"])
-NumberOfPersonVisiting = st.number_input("Number Of Person Visiting", min_value=1, value=2)
-NumberOfFollowups = st.number_input("Number Of Followups", min_value=0.0, value=3.0)
-ProductPitched = st.selectbox("Product Pitched", ["Basic", "Standard", "Deluxe", "Super Deluxe"])
-PreferredPropertyStar = st.number_input("Preferred Property Star", min_value=1.0, max_value=5.0, value=3.0)
-MaritalStatus = st.selectbox("Marital Status", ["Married", "Single", "Divorced"])
-NumberOfTrips = st.number_input("Number Of Trips", min_value=0.0, value=2.0)
-Passport = st.selectbox("Passport", [0, 1])
-PitchSatisfactionScore = st.slider("Pitch Satisfaction Score", 1, 5, 4)
-OwnCar = st.selectbox("Own Car", [0, 1])
-NumberOfChildrenVisiting = st.number_input("Number Of Children Visiting", min_value=0.0, value=1.0)
-Designation = st.selectbox("Designation", ["Executive", "Manager", "Senior Manager", "AVP", "VP"])
-MonthlyIncome = st.number_input("Monthly Income", min_value=0.0, value=25000.0)
+@app.get("/")
+async def root():
+    return {"message": "Welcome to the Tourism Package Prediction API!"}
 
-# -----------------------------
-# PREDICT BUTTON
-if st.button("Predict"):
-    input_data = {
-        "Age": Age,
-        "TypeofContact": TypeofContact,
-        "CityTier": CityTier,
-        "DurationOfPitch": DurationOfPitch,
-        "Occupation": Occupation,
-        "Gender": Gender,
-        "NumberOfPersonVisiting": NumberOfPersonVisiting,
-        "NumberOfFollowups": NumberOfFollowups,
-        "ProductPitched": ProductPitched,
-        "PreferredPropertyStar": PreferredPropertyStar,
-        "MaritalStatus": MaritalStatus,
-        "NumberOfTrips": NumberOfTrips,
-        "Passport": Passport,
-        "PitchSatisfactionScore": PitchSatisfactionScore,
-        "OwnCar": OwnCar,
-        "NumberOfChildrenVisiting": NumberOfChildrenVisiting,
-        "Designation": Designation,
-        "MonthlyIncome": MonthlyIncome
-    }
+@app.post("/predict")
+async def predict(request: PredictionRequest):
+    # Convert incoming request data to a dictionary
+    input_data_dict = request.model_dump()
 
-    try:
-        df = pp.preprocess_input(input_data)
-        pred = model.predict(df)[0]
-        proba = model.predict_proba(df)[0]
+    # Preprocess the input data
+    preprocessed_df = preprocess_input(input_data_dict)
 
-        st.success(f"Prediction: {'Purchased' if pred == 1 else 'Not Purchased'}")
-        st.write(f"Probability Purchased: **{proba[1]:.2%}**")
-        st.write(f"Probability Not Purchased: **{proba[0]:.2%}**")
-    except Exception as e:
-        st.error(f"Prediction failed: {e}")
+    # Make prediction
+    prediction = MODEL.predict(preprocessed_df)[0]
+    prediction_proba = MODEL.predict_proba(preprocessed_df)[0].tolist()
 
+    # Map prediction to a human-readable label if desired
+    prediction_label = "Purchased" if prediction == 1 else "Not Purchased"
 
+    return {
+        "prediction": int(prediction),
+        "prediction_label": prediction_label,
+        "probability_not_purchased": prediction_proba[0],
+        "probability_purchased": prediction_proba[1]
+    }