src/streamlit_app.py

import streamlit as st
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
from datasets import load_dataset

# -------------------------- Title --------------------------
st.title("🍷 Wine Quality Prediction")
st.write("Using Random Forest on the famous Wine Quality dataset")

# -------------------------- Load Data --------------------------
@st.cache_data
def get_data():
    ds = load_dataset("codesignal/wine-quality")
    
    df = ds[list(ds.keys())[0]].to_pandas()  
   
    return df
    ds = load_dataset("codesignal/wine-quality")
    df = ds['test'].to_pandas()
    return df

df = get_data()
st.write("Dataset loaded! Here's a preview:")
st.dataframe(df.head())
# -------------------------- Preprocessing --------------------------
X = df.drop("quality", axis=1)       # ← fixed: no "Id" column exists
y = df["quality"]

# Make it binary classification: good (≥6) vs bad (<6)
y = (y >= 6).astype(int)


# Train-test split
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42, stratify=y
)

# Scale features
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)

# -------------------------- Train Model --------------------------
@st.cache_resource
def train_model():
    model = RandomForestClassifier(
        n_estimators=200,
        random_state=42,
        n_jobs=-1
    )
    model.fit(X_train_scaled, y_train)
    return model

model = train_model()

# Predictions & accuracy
y_pred = model.predict(X_test_scaled)
accuracy = accuracy_score(y_test, y_pred)

st.success(f"Model Accuracy: *{accuracy:.4f}* ({accuracy*100:.2f}%)")

# -------------------------- Interactive Prediction --------------------------
st.header("Predict quality of a new wine")
cols = st.columns(3)

input_data = {}
features = X.columns.tolist()

for i, feature in enumerate(features):
    with cols[i % 3]:
        val = st.slider(
            feature,
            float(X[feature].min()),
            float(X[feature].max()),
            float(X[feature].mean())
        )
        input_data[feature] = val

if st.button("Predict Quality"):
    input_df = pd.DataFrame([input_data])
    input_scaled = scaler.transform(input_df)
    pred = model.predict(input_scaled)[0]
    prob = model.predict_proba(input_scaled)[0]
    
    if pred == 1:
        st.balloons()
        st.success(f"*Good wine!* 🍾 (confidence: {prob[1]:.2%})")
    else:
        st.error(f"*Not great wine* 😢 (confidence: {prob[0]:.2%})")