import streamlit as st
import numpy as np
import pickle
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.compose import ColumnTransformer
import os

# Set page config first
st.set_page_config(
    page_title="Crop Prediction App",
    page_icon="🌾",
    layout="centered",
    initial_sidebar_state="expanded"
)

# Custom CSS
st.markdown("""
    <style>
        .title {
            color: #2c3e50;
            text-align: center;
            margin-bottom: 30px;
        }
        .stButton>button {
            background-color: #27ae60;
            color: white;
            border-radius: 8px;
            padding: 10px 20px;
            width: 100%;
            transition: all 0.3s;
        }
        .stButton>button:hover {
            background-color: #2ecc71;
            transform: scale(1.02);
        }
        .input-section {
            background-color: #f8f9fa;
            padding: 20px;
            border-radius: 10px;
            margin-bottom: 20px;
        }
        .prediction-section {
            background-color: #e8f5e9;
            padding: 20px;
            border-radius: 10px;
            margin-top: 20px;
        }
        .step-card {
            background-color: #ffffff;
            border-radius: 10px;
            padding: 15px;
            margin-bottom: 10px;
            box-shadow: 0 2px 4px rgba(0,0,0,0.1);
        }
    </style>
""", unsafe_allow_html=True)

# Load model (with error handling)
@st.cache_resource
def load_model():
    try:
        with open("lor_f.pkl", "rb") as f:
            model = pickle.load(f)
        return model
    except FileNotFoundError:
        st.error("Model file not found. Please ensure 'lor_f.pkl' exists.")
        return None
    except Exception as e:
        st.error(f"Error loading model: {str(e)}")
        return None

model = load_model()

# App title
st.markdown("<h1 class='title'>🌾 Smart Crop Prediction</h1>", unsafe_allow_html=True)

# Main app sections
tab1, tab2 = st.tabs(["Prediction", "Project Overview"])

with tab1:
    st.subheader("Enter Soil and Weather Conditions")
    
    with st.container():
        st.markdown("<div class='input-section'>", unsafe_allow_html=True)
        
        col1, col2 = st.columns(2)
        with col1:
            nitrogen = st.slider("Nitrogen (N) level", 1, 140, 50, 
                               help="Nitrogen content in soil (1-140 kg/ha)")
            phosphorus = st.slider("Phosphorus (P) level", 5, 145, 50,
                                  help="Phosphorus content in soil (5-145 kg/ha)")
            potassium = st.slider("Potassium (K) level", 5, 205, 50,
                                 help="Potassium content in soil (5-205 kg/ha)")
            ph_value = st.slider("Soil pH Value", 3.0, 9.9, 6.5, 0.1,
                               help="Soil acidity/alkalinity (3.0-9.9 pH)")
        
        with col2:
            temperature = st.slider("Temperature (°C)", 8.0, 43.0, 25.0, 0.1,
                                  help="Average temperature (8-43°C)")
            humidity = st.slider("Humidity (%)", 14, 99, 60,
                               help="Relative humidity (14-99%)")
            rainfall = st.slider("Rainfall (mm)", 20, 298, 100,
                               help="Annual rainfall (20-298 mm)")
        
        st.markdown("</div>", unsafe_allow_html=True)
    
    if st.button("Predict Optimal Crop", key="predict_btn"):
        if model is None:
            st.error("Model not available. Please check the model file.")
        else:
            try:
                user_data = np.array([[nitrogen, phosphorus, potassium, temperature, 
                                     humidity, ph_value, rainfall]])
                prediction = model.predict(user_data)
                
                with st.container():
                    st.markdown("<div class='prediction-section'>", unsafe_allow_html=True)
                    st.success(f"### Recommended Crop: **{prediction[0]}**")
                    
                    # Add some visual feedback
                    st.write("Based on your inputs, the optimal crop for these conditions is:")
                    st.markdown(f"<h3 style='text-align: center; color: #27ae60;'>{prediction[0]}</h3>", 
                               unsafe_allow_html=True)
                    
                    # Add some additional information
                    st.markdown("""
                    **Tips for better yield:**
                    - Maintain proper irrigation
                    - Monitor soil nutrients regularly
                    - Follow recommended crop rotation practices
                    """)
                    
                    st.markdown("</div>", unsafe_allow_html=True)
            except Exception as e:
                st.error(f"Prediction error: {str(e)}")

with tab2:
    st.header("Machine Learning Project Steps")
    st.write("""
    This crop prediction system was developed following these key machine learning steps:
    """)
    
    steps = {
        "1. Problem Definition 🧠": {
            "description": "Identify the agricultural challenge and define objectives for crop prediction.",
            "actions": [
                "Determine key factors affecting crop growth",
                "Define success metrics for the model"
            ]
        },
        "2. Data Collection 📊": {
            "description": "Gather comprehensive agricultural datasets.",
            "actions": [
                "Collect soil nutrient data (N, P, K, pH)",
                "Gather weather and climate data",
                "Obtain historical crop yield information"
            ]
        },
        "3. Data Cleaning 🧹": {
            "description": "Prepare raw data for analysis by addressing quality issues.",
            "actions": [
                "Handle missing values and outliers",
                "Correct measurement inconsistencies",
                "Remove duplicate entries"
            ]
        },
        "4. Exploratory Analysis 🔍": {
            "description": "Understand data patterns and relationships.",
            "actions": [
                "Analyze feature distributions",
                "Identify correlations between variables",
                "Visualize data patterns"
            ]
        },
        "5. Feature Engineering ⚙️": {
            "description": "Select and transform relevant features.",
            "actions": [
                "Normalize numerical features",
                "Create derived features if needed",
                "Select most predictive features"
            ]
        },
        "6. Model Selection 🤖": {
            "description": "Choose appropriate machine learning algorithms.",
            "actions": [
                "Compare classification algorithms",
                "Evaluate based on accuracy and performance",
                "Select final model (Logistic Regression)"
            ]
        },
        "7. Model Training 🏋️‍♂️": {
            "description": "Train the selected model with prepared data.",
            "actions": [
                "Split data into training and validation sets",
                "Train model with optimal parameters",
                "Validate model performance"
            ]
        },
        "8. Model Evaluation 📈": {
            "description": "Assess model performance rigorously.",
            "actions": [
                "Calculate precision, recall, and F1-score",
                "Analyze confusion matrix",
                "Test on unseen data"
            ]
        },
        "9. Deployment 🚀": {
            "description": "Implement the model in a production environment.",
            "actions": [
                "Develop user-friendly interface",
                "Create API endpoints if needed",
                "Ensure scalability"
            ]
        },
        "10. Monitoring 🔄": {
            "description": "Continuously track and improve the system.",
            "actions": [
                "Monitor prediction accuracy",
                "Update model with new data",
                "Address concept drift"
            ]
        }
    }
    
    for step, content in steps.items():
        with st.expander(step):
            st.write(content["description"])
            st.markdown("**Key Actions:**")
            for action in content["actions"]:
                st.markdown(f"- {action}")
    
    st.markdown("---")
    st.write("This application helps farmers make data-driven decisions for optimal crop selection.")

# Add footer
st.markdown("---")
st.markdown(
    """
    <div style="text-align: center; color: #777; font-size: 0.9em;">
        Agricultural Decision Support System • Powered by Machine Learning
    </div>
    """,
    unsafe_allow_html=True
)