import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
import gradio as gr
import random

# Generate synthetic dataset for Bangladesh crops
def generate_synthetic_dataset(num_samples=5000):
    np.random.seed(42)
    
    # Common crops in Bangladesh
    crops = [
        'Rice (Aman)', 'Rice (Boro)', 'Rice (Aus)', 'Jute', 'Wheat', 
        'Maize', 'Potato', 'Sugarcane', 'Pulses (Mungbean)', 'Pulses (Lentil)', 
        'Mustard', 'Sesame', 'Sunflower', 'Tea', 'Mango',
        'Banana', 'Jackfruit', 'Litchi', 'Pineapple', 'Vegetables'
    ]
    
    # Soil types common in Bangladesh
    soil_types = ['Alluvial', 'Loamy', 'Clayey', 'Peaty', 'Sandy']
    
    # Seasons in Bangladesh agriculture
    seasons = ['Kharif-1 (Mar-Jun)', 'Kharif-2 (Jul-Oct)', 'Rabi (Nov-Feb)', 'Whole Year']
    
    # Generate synthetic data
    data = {
        'Temperature (°C)': np.random.uniform(10, 40, num_samples),  # Bangladesh has more moderate temperatures
        'Rainfall (mm)': np.random.uniform(100, 400, num_samples),  # Higher rainfall range
        'Humidity (%)': np.random.uniform(60, 100, num_samples),    # Generally high humidity
        'Soil pH': np.random.uniform(5.0, 8.5, num_samples),        # Slightly acidic to neutral
        'Soil Type': np.random.choice(soil_types, num_samples),
        'Nitrogen (N) Level': np.random.uniform(0, 150, num_samples),
        'Phosphorus (P) Level': np.random.uniform(0, 100, num_samples),
        'Potassium (K) Level': np.random.uniform(0, 200, num_samples),
        'Season': np.random.choice(seasons, num_samples),
        'Crop': np.random.choice(crops, num_samples)
    }
    
    # Add some logical patterns based on real-world knowledge
    df = pd.DataFrame(data)
    
    # Adjust values based on crop preferences
    for idx, row in df.iterrows():
        crop = row['Crop']
        
        # Temperature adjustments
        if 'Rice' in crop:
            df.at[idx, 'Temperature (°C)'] = np.random.uniform(25, 35)
            df.at[idx, 'Humidity (%)'] = np.random.uniform(70, 100)
        elif crop in ['Wheat', 'Mustard', 'Potato']:
            df.at[idx, 'Temperature (°C)'] = np.random.uniform(15, 25)
        elif crop in ['Jute', 'Tea']:
            df.at[idx, 'Temperature (°C)'] = np.random.uniform(20, 30)
            df.at[idx, 'Rainfall (mm)'] = np.random.uniform(200, 400)
        
        # Soil type adjustments
        if crop in ['Jute']:
            df.at[idx, 'Soil Type'] = 'Alluvial'
        elif crop in ['Tea']:
            df.at[idx, 'Soil Type'] = 'Loamy'
        elif crop in ['Rice (Boro)']:
            df.at[idx, 'Soil Type'] = random.choice(['Alluvial', 'Clayey'])
        
        # Season adjustments
        if crop in ['Rice (Aman)', 'Jute']:
            df.at[idx, 'Season'] = 'Kharif-2 (Jul-Oct)'
        elif crop in ['Rice (Boro)', 'Wheat', 'Mustard', 'Potato']:
            df.at[idx, 'Season'] = 'Rabi (Nov-Feb)'
        elif crop in ['Rice (Aus)']:
            df.at[idx, 'Season'] = 'Kharif-1 (Mar-Jun)'
    
    # Add profit estimates (in BDT per acre)
    profit_ranges = {
        'Rice (Aman)': (30000, 60000),
        'Rice (Boro)': (35000, 70000),
        'Rice (Aus)': (25000, 50000),
        'Jute': (40000, 80000),
        'Wheat': (25000, 50000),
        'Maize': (30000, 60000),
        'Potato': (50000, 100000),
        'Sugarcane': (60000, 120000),
        'Pulses (Mungbean)': (20000, 45000),
        'Pulses (Lentil)': (22000, 48000),
        'Mustard': (25000, 55000),
        'Sesame': (18000, 40000),
        'Sunflower': (20000, 45000),
        'Tea': (80000, 150000),
        'Mango': (100000, 250000),
        'Banana': (80000, 180000),
        'Jackfruit': (70000, 150000),
        'Litchi': (90000, 200000),
        'Pineapple': (60000, 120000),
        'Vegetables': (50000, 150000)
    }
    
    df['Profit (BDT/acre)'] = df['Crop'].apply(lambda x: random.randint(*profit_ranges[x]))
    
    return df

# Generate the dataset
df = generate_synthetic_dataset(10000)

# Prepare data for ML model
X = df.drop(['Crop', 'Profit (BDT/acre)'], axis=1)
X = pd.get_dummies(X)  # Convert categorical variables to dummy variables
y = df['Crop']

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

# Train Random Forest Classifier
model = RandomForestClassifier(n_estimators=100, random_state=42)
model.fit(X_train, y_train)

# Crop precautions information
precautions_db = {
    'Rice (Aman)': [
        "Transplant 25-30 day old seedlings",
        "Maintain 2-3 cm standing water during initial stage",
        "Apply 60-80 kg N, 15-20 kg P, and 30-40 kg K per hectare",
        "Control stem borer with proper insecticides"
    ],
    'Rice (Boro)': [
        "Ensure irrigation availability as it's dry season rice",
        "Use cold-tolerant varieties in northern regions",
        "Apply split doses of nitrogen fertilizer",
        "Control rats and birds during ripening stage"
    ],
    'Jute': [
        "Sow in well-prepared land with proper moisture",
        "Retting should be done in clean water for quality fiber",
        "Apply 40-60 kg N, 20-30 kg P, and 20-30 kg K per hectare",
        "Control jute hairy caterpillar with proper measures"
    ],
    'Wheat': [
        "Sow in rows with 20 cm spacing",
        "Apply irrigation at crown root initiation and flowering stages",
        "Use disease-resistant varieties to combat rust",
        "Harvest when moisture content is 20-25%"
    ],
    'Maize': [
        "Sow in rows with 60 cm row to row distance",
        "Apply 150-180 kg N, 35-40 kg P, and 60-70 kg K per hectare",
        "Control fall armyworm with integrated pest management",
        "Harvest when kernels have 20-25% moisture"
    ],
    'Potato': [
        "Use disease-free seed tubers",
        "Apply irrigation at critical growth stages",
        "Control late blight with fungicides",
        "Harvest when vines dry up"
    ],
    'Tea': [
        "Prune bushes regularly for new flush",
        "Apply balanced fertilizer with zinc and magnesium",
        "Control red spider mite with acaricides",
        "Pluck two leaves and a bud for quality"
    ],
    'Mango': [
        "Prune for proper canopy management",
        "Control mango hopper during flowering",
        "Apply irrigation during fruit development",
        "Harvest when shoulders develop"
    ],
    # Default precautions for other crops
    'Default': [
        "Use recommended spacing for the crop",
        "Monitor for pests and diseases regularly",
        "Apply balanced fertilizers as per soil test",
        "Ensure proper irrigation based on weather conditions"
    ]
}

# Function to get top precautions based on input features
def get_precautions(crop, temperature, rainfall, humidity, soil_type):
    precautions = precautions_db.get(crop, precautions_db['Default'])
    
    # Add weather-specific precautions
    if temperature > 35:
        precautions.append("Provide mulch to reduce soil temperature")
        precautions.append("Increase irrigation frequency during hot days")
    if rainfall > 300:
        precautions.append("Ensure proper drainage to prevent waterlogging")
    if humidity > 85:
        precautions.append("Watch for fungal diseases and apply preventive sprays")
    
    # Add soil-specific precautions
    if soil_type == 'Alluvial':
        precautions.append("Apply organic matter to maintain soil fertility")
    elif soil_type == 'Peaty':
        precautions.append("Apply lime to reduce acidity if needed")
    
    return precautions[:5]  # Return top 5 precautions

# Function to predict crop and details
def predict_crop(temperature, rainfall, humidity, soil_ph, soil_type, nitrogen, phosphorus, potassium, season):
    # Create input dataframe
    input_data = {
        'Temperature (°C)': [temperature],
        'Rainfall (mm)': [rainfall],
        'Humidity (%)': [humidity],
        'Soil pH': [soil_ph],
        'Nitrogen (N) Level': [nitrogen],
        'Phosphorus (P) Level': [phosphorus],
        'Potassium (K) Level': [potassium],
        'Season': [season]
    }
    
    # Add soil type columns (one-hot encoding)
    for st in ['Alluvial', 'Loamy', 'Clayey', 'Peaty', 'Sandy']:
        input_data[f'Soil Type_{st}'] = [1 if soil_type == st else 0]
    
    # Add season columns (one-hot encoding)
    for s in ['Kharif-1 (Mar-Jun)', 'Kharif-2 (Jul-Oct)', 'Rabi (Nov-Feb)', 'Whole Year']:
        input_data[f'Season_{s}'] = [1 if season == s else 0]
    
    input_df = pd.DataFrame(input_data)
    
    # Ensure columns are in same order as training data
    input_df = input_df.reindex(columns=X.columns, fill_value=0)
    
    # Predict crop
    crop = model.predict(input_df)[0]
    
    # Get profit range
    profit = df[df['Crop'] == crop]['Profit (BDT/acre)'].mean()
    
    # Get precautions
    precautions = get_precautions(crop, temperature, rainfall, humidity, soil_type)
    
    # Get similar crops (top 3 alternatives)
    probas = model.predict_proba(input_df)[0]
    top3_idx = np.argsort(probas)[-3:][::-1]
    similar_crops = [model.classes_[i] for i in top3_idx if model.classes_[i] != crop][:2]
    
    # Prepare output
    output = {
        "Recommended Crop": crop,
        "Expected Profit (BDT per acre)": f"৳{int(profit):,}",
        "Top Precautions": precautions,
        "Alternative Crops": similar_crops,
        "Best Season": season
    }
    
    return output

# Custom CSS for farmer-friendly interface
custom_css = """
/* Main container styling */
.agrismart-container {
    background: linear-gradient(135deg, #f5f7fa 0%, #e4efe9 100%);
    border-radius: 15px;
    padding: 20px;
    box-shadow: 0 10px 20px rgba(0,0,0,0.1);
    font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
}

/* Header styling */
.agrismart-header {
    background: linear-gradient(to right, #4CAF50, #2E8B57);
    color: white;
    padding: 15px 20px;
    border-radius: 10px;
    text-align: center;
    margin-bottom: 20px;
    box-shadow: 0 4px 8px rgba(0,0,0,0.1);
}

/* Input section styling */
.agrismart-input {
    background-color: rgba(255, 255, 255, 0.9);
    padding: 20px;
    border-radius: 10px;
    margin-bottom: 20px;
    box-shadow: 0 2px 5px rgba(0,0,0,0.05);
}

/* Output section styling */
.agrismart-output {
    background-color: #ffffff;
    padding: 20px;
    border-radius: 10px;
    box-shadow: 0 2px 10px rgba(0,0,0,0.1);
    border-left: 5px solid #4CAF50;
}

/* Button styling */
.agrismart-button {
    background: linear-gradient(to right, #4CAF50, #2E8B57) !important;
    color: white !important;
    border: none !important;
    padding: 12px 25px !important;
    border-radius: 8px !important;
    font-size: 16px !important;
    cursor: pointer !important;
    transition: all 0.3s !important;
    box-shadow: 0 4px 6px rgba(0,0,0,0.1) !important;
}

.agrismart-button:hover {
    transform: translateY(-2px) !important;
    box-shadow: 0 6px 8px rgba(0,0,0,0.15) !important;
}

/* Slider styling */
.agrismart-slider .gr-slider {
    background: #e0e0e0 !important;
    height: 10px !important;
    border-radius: 5px !important;
}

.agrismart-slider .gr-slider .gr-slider-selection {
    background: linear-gradient(to right, #4CAF50, #2E8B57) !important;
}

/* Label styling */
.agrismart-label {
    font-weight: bold !important;
    color: #2E8B57 !important;
    margin-bottom: 5px !important;
    font-size: 16px !important;
}

/* Dropdown styling */
.agrismart-dropdown {
    border: 1px solid #ddd !important;
    border-radius: 8px !important;
    padding: 8px 12px !important;
    box-shadow: inset 0 1px 3px rgba(0,0,0,0.1) !important;
}

/* Result card styling */
.agrismart-result-card {
    background: #f9f9f9;
    border-radius: 10px;
    padding: 15px;
    margin: 10px 0;
    border-left: 4px solid #4CAF50;
}

.agrismart-result-title {
    color: #2E8B57;
    font-weight: bold;
    margin-bottom: 10px;
}

.agrismart-result-value {
    font-size: 18px;
    color: #333;
}

/* Precautions list styling */
.agrismart-precautions {
    list-style-type: none;
    padding-left: 0;
}

.agrismart-precautions li {
    background: url('data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" viewBox="0 0 24 24" fill="%234CAF50"><path d="M9 16.17L4.83 12l-1.42 1.41L9 19 21 7l-1.41-1.41z"/></svg>') no-repeat left center;
    padding-left: 25px;
    margin-bottom: 8px;
    line-height: 1.5;
}

/* Responsive design */
@media (max-width: 768px) {
    .agrismart-container {
        padding: 10px;
    }
}
"""

# Function to format outputs
def format_outputs(output):
    crop_md = f"**Recommended Crop:** {output['Recommended Crop']}"
    profit_md = f"**Expected Profit (BDT per acre):** {output['Expected Profit (BDT per acre)']}"
    season_md = f"**Best Season:** {output['Best Season']}"
    alt_md = f"**Alternative Crops:** {', '.join(output['Alternative Crops'])}"
    
    prec_html = """
    <ul class="agrismart-precautions">
    """ + "\n".join([f"<li>{p}</li>" for p in output['Top Precautions']]) + """
    </ul>
    """
    
    return crop_md, profit_md, prec_html, alt_md, season_md

# Create Gradio interface
with gr.Blocks(css=custom_css) as demo:
    with gr.Column(elem_classes="agrismart-container"):
        with gr.Row(elem_classes="agrismart-header"):
            gr.Markdown("""
            # 🌱 বাংলাদেশের জন্য ফসল সুপারিশকারী
            ### আপনার জমির অবস্থা অনুযায়ী উপযুক্ত ফসলের পরামর্শ পান
            """)
        
        with gr.Row():
            with gr.Column(elem_classes="agrismart-input"):
                gr.Markdown("### 🌦️ আপনার জমির তথ্য দিন", elem_classes="agrismart-label")
                
                with gr.Row():
                    temperature = gr.Slider(10, 40, label="1. তাপমাত্রা (°C)", 
                                          info="ছায়াযুক্ত স্থানে বায়ুর তাপমাত্রা মাপুন", 
                                          elem_classes="agrismart-slider")
                    rainfall = gr.Slider(100, 400, label="2. বৃষ্টিপাত (mm)", 
                                       info="আপনার এলাকার বার্ষিক বৃষ্টিপাতের পরিমাণ", 
                                       elem_classes="agrismart-slider")
                
                with gr.Row():
                    humidity = gr.Slider(60, 100, label="3. আর্দ্রতা (%)", 
                                       info="বাতাসে আর্দ্রতার পরিমাণ", 
                                       elem_classes="agrismart-slider")
                    soil_ph = gr.Slider(5, 8.5, label="4. মাটির pH মান", 
                                      info="৭ হলো নিরপেক্ষ, ৭ এর নিচে অম্লীয়, ৭ এর উপরে ক্ষারীয়", 
                                      elem_classes="agrismart-slider")
                
                with gr.Row():
                    soil_type = gr.Dropdown(
                        ["Alluvial", "Loamy", "Clayey", "Peaty", "Sandy"], 
                        label="5. মাটির ধরন",
                        info="বাংলাদেশের সাধারণ মাটির ধরন",
                        elem_classes="agrismart-dropdown"
                    )
                    season = gr.Dropdown(
                        ["Kharif-1 (Mar-Jun)", "Kharif-2 (Jul-Oct)", "Rabi (Nov-Feb)", "Whole Year"], 
                        label="6. মৌসুম",
                        elem_classes="agrismart-dropdown"
                    )
                
                with gr.Row():
                    nitrogen = gr.Slider(0, 150, label="7. মাটিতে নাইট্রোজেনের পরিমাণ (N)", 
                                        info="গাছের পাতার বৃদ্ধির জন্য প্রয়োজনীয় (kg/ha)", 
                                        elem_classes="agrismart-slider")
                    phosphorus = gr.Slider(0, 100, label="8. মাটিতে ফসফরাসের পরিমাণ (P)", 
                                         info="শিকড়ের উন্নতির জন্য গুরুত্বপূর্ণ (kg/ha)", 
                                         elem_classes="agrismart-slider")
                    potassium = gr.Slider(0, 200, label="9. মাটিতে পটাশিয়ামের পরিমাণ (K)", 
                                        info="ফলের গুণগত মানের জন্য সহায়ক (kg/ha)", 
                                        elem_classes="agrismart-slider")
                
                submit_btn = gr.Button("ফসলের সুপারিশ পান", elem_classes="agrismart-button")
            
            with gr.Column(elem_classes="agrismart-output"):
                gr.Markdown("### 📊 সুপারিশকৃত ফসলের বিবরণ", elem_classes="agrismart-label")
                
                with gr.Column(elem_classes="agrismart-result-card"):
                    crop = gr.Markdown("**সুপারিশকৃত ফসল:** ", elem_classes="agrismart-result-value")
                    profit = gr.Markdown("**আনুমানিক লাভ (প্রতি একরে):** ", elem_classes="agrismart-result-value")
                    season_out = gr.Markdown("**উপযুক্ত মৌসুম:** ", elem_classes="agrismart-result-value")
                    alternatives = gr.Markdown("**বিকল্প ফসল:** ", elem_classes="agrismart-result-value")
                
                gr.Markdown("### 🛡️ প্রয়োজনীয় সতর্কতা", elem_classes="agrismart-result-title")
                precautions = gr.HTML("""
                <ul class="agrismart-precautions">
                    <li>আপনার জমির তথ্য প্রদান করে বাটনে ক্লিক করুন</li>
                </ul>
                """)
                
                # Example images of common Bangladeshi crops
                gr.Markdown("### 🌾 বাংলাদেশের প্রধান ফসল")
                gr.HTML("""
                <div style="display: flex; flex-wrap: wrap; gap: 10px; justify-content: center;">
                    <div style="text-align: center;">
                        <div style="background: #e3f2fd; padding: 10px; border-radius: 10px; width: 100px;">
                            <div style="font-size: 40px;">🌾</div>
                            <div>ধান</div>
                        </div>
                    </div>
                    <div style="text-align: center;">
                        <div style="background: #e8f5e9; padding: 10px; border-radius: 10px; width: 100px;">
                            <div style="font-size: 40px;">🧶</div>
                            <div>পাট</div>
                        </div>
                    </div>
                    <div style="text-align: center;">
                        <div style="background: #fff3e0; padding: 10px; border-radius: 10px; width: 100px;">
                            <div style="font-size: 40px;">🥔</div>
                            <div>আলু</div>
                        </div>
                    </div>
                    <div style="text-align: center;">
                        <div style="background: #f3e5f5; padding: 10px; border-radius: 10px; width: 100px;">
                            <div style="font-size: 40px;">🍌</div>
                            <div>কলা</div>
                        </div>
                    </div>
                </div>
                """)
    
    # Define button click action
    submit_btn.click(
        fn=lambda temp, rain, hum, ph, soil, n, p, k, seas: format_outputs(
            predict_crop(temp, rain, hum, ph, soil, n, p, k, seas)
        ),
        inputs=[temperature, rainfall, humidity, soil_ph, soil_type, nitrogen, phosphorus, potassium, season],
        outputs=[crop, profit, precautions, alternatives, season_out]
    )

# Launch the application
if __name__ == "__main__":
    demo.launch()