import gradio as gr
import tensorflow as tf
import numpy as np
from PIL import Image
import json
import requests
import io
import base64
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"  # Disable GPU

# ===== CONFIGURACIÓN =====
PLANTNET_CONFIG = {
    "base_url": "https://my-api.plantnet.org/v2/identify/all",
    "api_key": "2b10GtUDt6p1whX94wlEiR3CG",
    "timeout": 10,
    "lang": "es"  # Español
}

# ===== VARIABLES GLOBALES =====
MODEL_LOADED = False
model = None
labels = []

# ===== FUNCIONES DE UTILIDAD =====
def load_model():
    global model, labels, MODEL_LOADED
    try:
        model = tf.keras.models.load_model("flores_modelo (2).h5")
        with open("clases_orden_oxford.json", "r") as f:
            class_indices = json.load(f)
        labels = [None] * len(class_indices)
        for class_name, idx in class_indices.items():
            labels[idx] = class_name.replace("_", " ").title()
        MODEL_LOADED = True
    except Exception as e:
        print(f"Error al cargar el modelo: {e}")
        MODEL_LOADED = False

def image_to_base64(image_path):
    try:
        with open(image_path, "rb") as img_file:
            return base64.b64encode(img_file.read()).decode('utf-8')
    except:
        return None

def preprocess_image(image):
    image = image.resize((224, 224))
    img_array = np.array(image) / 255.0
    if img_array.shape[-1] == 4:
        img_array = img_array[..., :3]
    return np.expand_dims(img_array, axis=0)

# ===== INTEGRACIÓN CON PLANTNET =====
def get_flower_info_from_plantnet(flower_name, image_array):
    try:
        image_pil = Image.fromarray((image_array[0] * 255).astype(np.uint8))
        img_byte_arr = io.BytesIO()
        image_pil.save(img_byte_arr, format='JPEG')
        img_byte_arr.seek(0)

        url = f"{PLANTNET_CONFIG['base_url']}?api-key={PLANTNET_CONFIG['api_key']}&lang={PLANTNET_CONFIG['lang']}"
        files = {'images': ('image.jpg', img_byte_arr, 'image/jpeg')}
        data = {'organs': 'flower'}

        response = requests.post(url, files=files, data=data, timeout=PLANTNET_CONFIG['timeout'])
        if response.status_code == 200:
            return parse_plantnet_response(response.json(), flower_name)
        else:
            print(f"Error en PlantNet: {response.status_code}")
            return get_fallback_info(flower_name)
    except requests.exceptions.Timeout:
        print("Tiempo de espera agotado en PlantNet")
        return get_fallback_info(flower_name)
    except Exception as e:
        print(f"PlantNet error: {e}")
        return get_fallback_info(flower_name)

def parse_plantnet_response(data, flower_name):
    if 'results' in data and len(data['results']) > 0:
        results = data['results'][:3]
        info = f"""
<div class="bg-[rgba(255,255,255,0.1)] border-2 border-[rgba(255,165,0,0.3)] rounded-2xl p-5 m-[15px_0] shadow-[0_8px_32px_rgba(0,0,0,0.2)]">
    <div class="flex items-center justify-center">
        <h2 class="text-yellow-400 font-bold text-2xl">🌸 {flower_name}</h2>
    </div>
    <div>
        <h3 class="text-yellow-400 font-bold text-xl mt-4">📊 Identificación Científica</h3>
"""
        for i, result in enumerate(results, 1):
            species = result.get('species', {})
            scientific_name = species.get('scientificNameWithoutAuthor', 'N/A')
            authorship = species.get('scientificNameAuthorship', '')
            common_names = species.get('commonNames', [])
            family = species.get('family', {}).get('scientificNameWithoutAuthor', 'N/A')
            genus = species.get('genus', {}).get('scientificNameWithoutAuthor', 'N/A')
            score = result.get('score', 0)
            common_names_str = ', '.join(common_names[:3]) if common_names else 'No disponible'

            info += f"""
        <div class="flex items-start mt-4">
            <div class="text-black mr-4">#{i}</div>
            <div>
                <h4 class="text-black font-semibold">{scientific_name} {authorship}</h4>
                <div class="relative h-2 bg-gray-200 rounded mt-2">
                    <div class="absolute h-full bg-yellow-400 rounded" style="width: {score*100}%"></div>
                    <span class="text-black text-sm mt-2 block">{score:.1%} confianza</span>
                </div>
                <div class="grid grid-cols-1 gap-2 mt-2">
                    <div class="text-black"><strong>Nombres comunes:</strong> {common_names_str}</div>
                    <div class="text-black"><strong>Familia:</strong> {family}</div>
                    <div class="text-black"><strong>Género:</strong> {genus}</div>
                </div>
            </div>
        </div>
"""
        info += """
    </div>
    <div class="mt-6">
        <h3 class="text-yellow-400 font-bold text-xl">🌿 Cuidados Generales</h3>
        <div class="grid grid-cols-2 gap-4 mt-2">
            <div class="flex items-start"><div class="text-2xl mr-2">☀️</div><div class="text-black"><strong>Luz:</strong> Luz solar directa o indirecta</div></div>
            <div class="flex items-start"><div class="text-2xl mr-2">💧</div><div class="text-black"><strong>Riego:</strong> Mantener húmedo, evitar exceso</div></div>
            <div class="flex items-start"><div class="text-2xl mr-2">🌡️</div><div class="text-black"><strong>Temperatura:</strong> Evitar cambios bruscos</div></div>
            <div class="flex items-start"><div class="text-2xl mr-2">🌱</div><div class="text-black"><strong>Suelo:</strong> Bien drenado y rico en nutrientes</div></div>
        </div>
    </div>
</div>
"""
        return info
    return get_fallback_info(flower_name)

def get_fallback_info(flower_name):
    return f"""
<div class="bg-[rgba(255,255,255,0.1)] border-2 border-[rgba(255,165,0,0.3)] rounded-2xl p-5 m-[15px_0] shadow-[0_8px_32px_rgba(0,0,0,0.2)]">
    <div class="flex items-center justify-between">
        <h2 class="text-yellow-400 font-bold text-2xl">🌸 {flower_name}</h2>
        <div class="bg-red-600 text-black px-2 py-1 rounded">PlantNet no disponible</div>
    </div>
    <div class="mt-4">
        <h3 class="text-yellow-400 font-bold text-xl">📖 Información General</h3>
        <p class="text-black">Identificado por nuestro modelo de IA entrenado en el conjunto de datos Oxford 102 Flowers.</p>
        <div class="grid grid-cols-2 gap-4 mt-4">
            <div class="text-black">
                <h4 class="font-semibold">🌺 Características</h4>
                <ul class="list-disc ml-5">
                    <li>Estructuras reproductivas de la planta</li>
                    <li>Varios colores y formas</li>
                    <li>Evolucionadas para atraer polinizadores</li>
                </ul>
            </div>
            <div class="text-black">
                <h4 class="font-semibold">🎯 Cuidados Básicos</h4>
                <ul class="list-disc ml-5">
                    <li>Buena iluminación según la especie</li>
                    <li>Riego regular sin exceso</li>
                    <li>Temperatura estable</li>
                    <li>Fertilización adecuada</li>
                </ul>
            </div>
        </div>
    </div>
</div>
"""

# ===== LÓGICA DE PREDICCIÓN =====
def predict(image):
    if not image:
        return "No se cargó ninguna imagen", "0%", """
<div class="bg-red-600 text-black rounded-2xl p-5 m-[15px_0]">
    <h3 class="text-xl font-bold">⚠️ Imagen requerida</h3>
    <p>Por favor, carga una imagen de una flor para iniciar la identificación.</p>
</div>
"""
    if not MODEL_LOADED:
        return "Error en el modelo", "0%", """
<div class="bg-red-600 text-black rounded-2xl p-5 m-[15px_0]">
    <h3 class="text-xl font-bold">🚫 Modelo no disponible</h3>
    <p>No se pudo cargar el modelo de clasificación. Verifica los archivos del modelo.</p>
</div>
"""
    try:
        img_array = preprocess_image(image)
        preds = model.predict(img_array)
        class_idx = np.argmax(preds[0])
        confidence = preds[0][class_idx]
        label_name = labels[class_idx]
        flower_details = get_flower_info_from_plantnet(label_name, img_array)
        return label_name, f"{confidence:.2%}", flower_details
    except Exception as e:
        return "Error", "0%", f"""
<div class="bg-red-600 text-black rounded-2xl p-5 m-[15px_0]">
    <h3 class="text-xl font-bold">❌ Error de predicción</h3>
    <p>Error durante el procesamiento: {str(e)}</p>
</div>
"""

# ===== CSS PARA EL FONDO DE LA APLICACIÓN =====
custom_css = """
.gradio-container {
    background: #1E2A44; 
    min-height: 100vh;
}
"""

# ===== JAVASCRIPT PARA OPTIMIZACIÓN DE CÁMARA =====
camera_js_improved = """
<script>
function initCameraOptimization() {
    function ensureButtonsVisible() {
        document.querySelectorAll('[data-testid="image"]').forEach(container => {
            container.querySelectorAll('button').forEach(button => {
                button.style.display = 'flex';
                button.style.visibility = 'visible';
                button.style.opacity = '1';
                button.style.zIndex = '999';
            });
        });
    }
    ensureButtonsVisible();
    setInterval(ensureButtonsVisible, 5000);
}
if (document.readyState === 'loading') {
    document.addEventListener('DOMContentLoaded', initCameraOptimization);
} else {
    initCameraOptimization();
}
</script>
"""

# ===== APLICACIÓN PRINCIPAL =====
# Cargar el modelo al inicio
load_model()
img3_b64 = image_to_base64("img3.png")
img2_b64 = image_to_base64("img2.png")

with gr.Blocks(theme=gr.themes.Soft(), css=custom_css, title="🌸 Flower ") as demo:
    header_html = f'''
    <div style="background: rgba(37, 58, 105); border-radius: 16px; padding: 16px; margin-bottom: 24px; border: 1px solid #e5e7eb; box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1); max-width: 1024px; margin-left: auto; margin-right: auto;">
        <div style="display: flex; align-items: center; justify-content: space-between; gap: 16px;">
            {'<div style="flex-shrink: 0; width: 150px; height: 150px;"><img src="data:image/png;base64,' + img3_b64 + '" alt="Logo" style="width: 100%; height: 100%; object-fit: contain; border-radius: 8px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"></div>' if img3_b64 else '<div style="width: 80px; height: 80px;"></div>'}
            <div style="text-align: center;">
                <h1 style="color: #FFCC00; font-size: 24px; font-weight: bold; margin-bottom: 4px;">IDENTIFICADOR DE FLORES</h1>
                <p style="color: #FFFFFF; font-weight: 500; font-size: 14px; margin-bottom: 4px;">Identifica cualquier flor en cuestión de segundos</p>
                <p style="color: #22c55e; font-size: 12px;">Oxford 102 Flowers + PlantNet</p>
            </div>
            {'<div style="flex-shrink: 0; width: 150px; height: 150px;"><img src="data:image/png;base64,' + img2_b64 + '" alt="Logo" style="width: 100%; height: 100%; object-fit: contain; border-radius: 8px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"></div>' if img2_b64 else '<div style="width: 48px; height: 48px;"></div>'}
        </div>
    </div>
    '''
    gr.HTML(header_html)

    with gr.Row():
        with gr.Column(scale=1):
            image_input = gr.Image(type="pil", label="📷 Cargar Imagen de Flor", height=350, sources=["upload", "webcam", "clipboard"])
            predict_btn = gr.Button("🔍 Identificar Flor", variant="primary")
            gr.HTML(f'<div class="bg-[rgba(255,255,255,0.1)] border-2 border-[rgba(255,165,0,0.3)] rounded-2xl p-5 m-[15px_0] shadow-[0_8px_32px_rgba(0,0,0,0.2)]"><h3 class="text-yellow-400 font-bold text-xl">📊 Estado del Sistema</h3><p class="text-black">Modelo: {"✅ Activo" if MODEL_LOADED else "❌ Error"}</p><p class="text-black">Clases: {len(labels)}</p></div>')

        with gr.Column(scale=1):
            result_label = gr.Textbox(label="🌼 Flor Identificada", interactive=False, placeholder="El nombre de la flor aparecerá aquí...")
            result_conf = gr.Textbox(label="📊 Confianza", interactive=False, placeholder="El nivel de confianza aparecerá aquí...")
            flower_info_output = gr.HTML(value='<div class="bg-[rgba(255,255,255,0.1)] border-2 border-[rgba(255,165,0,0.3)] rounded-2xl p-5 m-[15px_0] shadow-[0_8px_32px_rgba(0,0,0,0.2)]"><h3 class="text-yellow-400 font-bold text-xl">🌸 ¡Bienvenido!</h3><p class="text-black">Carga una imagen de una flor para iniciar la identificación.</p></div>')

    predict_btn.click(fn=predict, inputs=image_input, outputs=[result_label, result_conf, flower_info_output])
    gr.HTML(camera_js_improved)

if __name__ == "__main__":
    demo.launch(share=False)