src/streamlit_app.py

import streamlit as st
import json
import requests
from PIL import Image
import io
import base64
from datetime import datetime
import time

# Page configuration
st.set_page_config(
    page_title="AdForge AI - AI Advertisement Generator",
    page_icon="🎨",
    layout="wide",
    initial_sidebar_state="expanded"
)

# Custom CSS for professional styling
st.markdown("""
<style>
    .main-header {
        font-size: 3rem;
        font-weight: bold;
        background: linear-gradient(120deg, #667eea 0%, #764ba2 100%);
        -webkit-background-clip: text;
        -webkit-text-fill-color: transparent;
        text-align: center;
        padding: 1rem 0;
    }
    .sub-header {
        text-align: center;
        color: #666;
        font-size: 1.2rem;
        margin-bottom: 2rem;
    }
    .stButton>button {
        width: 100%;
        background: linear-gradient(120deg, #667eea 0%, #764ba2 100%);
        color: white;
        border: none;
        padding: 0.75rem 1.5rem;
        font-size: 1.1rem;
        font-weight: bold;
        border-radius: 10px;
        transition: all 0.3s;
    }
    .stButton>button:hover {
        transform: translateY(-2px);
        box-shadow: 0 10px 20px rgba(0,0,0,0.2);
    }
    .success-box {
        padding: 1rem;
        border-radius: 10px;
        background: #d4edda;
        border: 1px solid #c3e6cb;
        color: #155724;
        margin: 1rem 0;
    }
    .info-box {
        padding: 1rem;
        border-radius: 10px;
        background: #d1ecf1;
        border: 1px solid #bee5eb;
        color: #0c5460;
        margin: 1rem 0;
    }
    .feature-card {
        padding: 1.5rem;
        border-radius: 10px;
        background: white;
        box-shadow: 0 4px 6px rgba(0,0,0,0.1);
        margin: 1rem 0;
    }
</style>
""", unsafe_allow_html=True)

# Initialize session state
if 'generated_images' not in st.session_state:
    st.session_state.generated_images = []
if 'generation_history' not in st.session_state:
    st.session_state.generation_history = []
if 'api_token' not in st.session_state:
    st.session_state.api_token = ""

# Helper Functions
def validate_json(json_string):
    """Validate and parse JSON input"""
    try:
        data = json.loads(json_string)
        return data, None
    except json.JSONDecodeError as e:
        return None, f"Invalid JSON format: {str(e)}"

def generate_advanced_prompt(ad_data):
    """Generate sophisticated advertising prompt from JSON data"""
    product = ad_data.get('product_name', 'product')
    description = ad_data.get('description', '')
    style = ad_data.get('style', 'professional')
    mood = ad_data.get('mood', 'energetic')
    target_audience = ad_data.get('target_audience', 'general audience')
    colors = ad_data.get('colors', [])
    setting = ad_data.get('setting', 'studio')
    composition = ad_data.get('composition', 'centered')
    lighting = ad_data.get('lighting', 'soft studio lighting')
    
    # Build sophisticated prompt
    prompt_parts = []
    
    # Main subject
    prompt_parts.append(f"Professional advertising photography of {product}")
    
    if description:
        prompt_parts.append(f"{description}")
    
    # Style and mood
    prompt_parts.append(f"{style} style, {mood} mood")
    
    # Target audience context
    if target_audience and target_audience != 'general audience':
        prompt_parts.append(f"appealing to {target_audience}")
    
    # Visual elements
    if colors:
        color_palette = ", ".join(colors)
        prompt_parts.append(f"color palette: {color_palette}")
    
    prompt_parts.append(f"{setting} setting")
    prompt_parts.append(f"{composition} composition")
    prompt_parts.append(f"{lighting}")
    
    # Quality tags
    quality_tags = [
        "high-end commercial photography",
        "8k resolution",
        "sharp focus",
        "professional color grading",
        "dramatic lighting",
        "photorealistic",
        "award-winning advertisement"
    ]
    
    prompt = ", ".join(prompt_parts) + ", " + ", ".join(quality_tags)
    
    return prompt

def generate_negative_prompt(ad_data):
    """Generate negative prompt to avoid unwanted elements"""
    base_negative = [
        "low quality", "blurry", "distorted", "ugly", "amateur",
        "poor composition", "bad lighting", "oversaturated",
        "watermark", "text", "signature", "logo overlay",
        "cropped", "out of frame", "deformed", "disfigured",
        "extra limbs", "poorly rendered", "cartoon", "anime"
    ]
    
    # Add custom negative prompts if specified
    if 'avoid' in ad_data and ad_data['avoid']:
        base_negative.extend(ad_data['avoid'])
    
    return ", ".join(base_negative)

def call_huggingface_api(prompt, negative_prompt, api_token):
    """Call Hugging Face Inference API for FLUX.1-schnell"""
    API_URL = "https://api-inference.huggingface.co/models/black-forest-labs/FLUX.1-schnell"
    
    headers = {
        "Authorization": f"Bearer {api_token}",
        "Content-Type": "application/json"
    }
    
    payload = {
        "inputs": prompt,
        "parameters": {
            "negative_prompt": negative_prompt,
            "num_inference_steps": 4,  # Schnell optimized for 1-4 steps
            "guidance_scale": 0.0,  # Schnell doesn't use guidance
            "width": 1024,
            "height": 1024
        }
    }
    
    try:
        response = requests.post(API_URL, headers=headers, json=payload, timeout=60)
        
        if response.status_code == 200:
            image = Image.open(io.BytesIO(response.content))
            return image, None
        elif response.status_code == 503:
            return None, "Model is loading. Please wait 20 seconds and try again."
        elif response.status_code == 401:
            return None, "Invalid API token. Please check your Hugging Face token."
        elif response.status_code == 402:
            return None, "You've exceeded your free tier limit. Consider upgrading to PRO or wait for next month."
        else:
            error_msg = response.json().get('error', 'Unknown error')
            return None, f"API Error ({response.status_code}): {error_msg}"
    
    except requests.Timeout:
        return None, "Request timed out. The model might be loading. Please try again."
    except Exception as e:
        return None, f"Error: {str(e)}"

def image_to_base64(image):
    """Convert PIL Image to base64 string"""
    buffered = io.BytesIO()
    image.save(buffered, format="PNG")
    return base64.b64encode(buffered.getvalue()).decode()

# Header
st.markdown('<h1 class="main-header">🎨 AdForge AI</h1>', unsafe_allow_html=True)
st.markdown('<p class="sub-header">Transform your concept into stunning advertisements with AI</p>', unsafe_allow_html=True)

# Sidebar Configuration
with st.sidebar:
    st.header("⚙️ Configuration")
    
    # API Token Input
    st.subheader("🔑 Hugging Face API Token")
    api_token_input = st.text_input(
        "Enter your HF token",
        type="password",
        value=st.session_state.api_token,
        help="Get your free token from https://huggingface.co/settings/tokens"
    )
    
    if api_token_input:
        st.session_state.api_token = api_token_input
        st.success("✅ Token configured!")
    else:
        st.warning("⚠️ Please enter your Hugging Face API token to generate images")
    
    st.markdown("---")
    
    # Advanced Settings
    st.subheader("🎛️ Generation Settings")
    
    num_variations = st.slider(
        "Number of Variations",
        min_value=1,
        max_value=4,
        value=1,
        help="Generate multiple variations of your ad"
    )
    
    image_size = st.selectbox(
        "Image Resolution",
        options=["1024x1024 (Square)", "1024x768 (Landscape)", "768x1024 (Portrait)"],
        index=0
    )
    
    st.markdown("---")
    
    # Model Info
    st.subheader("ℹ️ About")
    st.info("""
    **Model:** FLUX.1-schnell  
    **Speed:** 2-5 seconds  
    **Quality:** Professional-grade  
    **License:** Apache 2.0
    
    FLUX.1-schnell is optimized for rapid, high-quality image generation perfect for advertising.
    """)
    
    # Quick Guide
    with st.expander("📖 Quick Guide"):
        st.markdown("""
        **Steps to Generate:**
        1. Enter your HF API token above
        2. Fill in the JSON with your ad concept
        3. Click "Generate Advertisement"
        4. Download your results!
        
        **Get API Token:**
        1. Visit [Hugging Face](https://huggingface.co)
        2. Sign up/Login
        3. Go to Settings → Access Tokens
        4. Create new token with "Make calls to Inference API" permission
        """)

# Main Content Area
col1, col2 = st.columns([1, 1], gap="large")

with col1:
    st.subheader("📝 Advertisement Concept (JSON)")
    
    # Sample JSON templates
    sample_templates = {
        "Tech Product": {
            "product_name": "AirPods Pro Max",
            "description": "Premium wireless headphones with spatial audio and active noise cancellation",
            "style": "minimalist and modern",
            "mood": "sophisticated and premium",
            "target_audience": "tech enthusiasts and professionals",
            "colors": ["space gray", "silver", "white"],
            "setting": "modern studio with gradient background",
            "composition": "centered with dramatic angle",
            "lighting": "soft key light with rim lighting",
            "key_features": ["noise cancellation", "spatial audio", "premium build"]
        },
        "Food & Beverage": {
            "product_name": "Artisan Cold Brew Coffee",
            "description": "Smooth cold brew coffee with hints of chocolate and caramel",
            "style": "rustic and organic",
            "mood": "warm and inviting",
            "target_audience": "coffee lovers and millennials",
            "colors": ["rich brown", "cream", "amber"],
            "setting": "wooden table with natural elements",
            "composition": "rule of thirds with foreground elements",
            "lighting": "natural morning light with soft shadows",
            "key_features": ["organic", "small batch", "sustainably sourced"]
        },
        "Fashion": {
            "product_name": "Summer Collection Dress",
            "description": "Elegant flowing summer dress with floral patterns",
            "style": "editorial fashion",
            "mood": "elegant and breezy",
            "target_audience": "fashion-forward women 25-40",
            "colors": ["pastel pink", "soft white", "gold accents"],
            "setting": "outdoor garden at golden hour",
            "composition": "full body shot with movement",
            "lighting": "golden hour backlight with fill",
            "key_features": ["sustainable fabric", "handcrafted", "limited edition"]
        },
        "Automotive": {
            "product_name": "Electric Sports Car",
            "description": "Sleek electric vehicle with cutting-edge design and performance",
            "style": "dynamic and futuristic",
            "mood": "powerful and innovative",
            "target_audience": "luxury car enthusiasts",
            "colors": ["metallic blue", "chrome", "carbon black"],
            "setting": "modern urban environment at night",
            "composition": "low angle hero shot",
            "lighting": "dramatic city lights with reflections",
            "key_features": ["0-60 in 2.5s", "400 mile range", "autopilot"]
        }
    }
    
    template_choice = st.selectbox(
        "📋 Choose a template or start from scratch",
        options=["Custom"] + list(sample_templates.keys())
    )
    
    if template_choice != "Custom":
        default_json = json.dumps(sample_templates[template_choice], indent=2)
    else:
        default_json = json.dumps({
            "product_name": "",
            "description": "",
            "style": "professional",
            "mood": "energetic",
            "target_audience": "",
            "colors": [],
            "setting": "studio",
            "composition": "centered",
            "lighting": "soft studio lighting"
        }, indent=2)
    
    json_input = st.text_area(
        "Paste or edit your advertisement concept:",
        value=default_json,
        height=400,
        help="Provide details about your product and desired ad style"
    )
    
    # Generate Button
    generate_btn = st.button(
        "🚀 Generate Advertisement",
        type="primary",
        disabled=not st.session_state.api_token,
        use_container_width=True
    )
    
    if generate_btn:
        if not st.session_state.api_token:
            st.error("❌ Please enter your Hugging Face API token in the sidebar first!")
        else:
            # Validate JSON
            ad_data, error = validate_json(json_input)
            
            if error:
                st.error(f"❌ {error}")
            else:
                st.markdown('<div class="success-box">✅ JSON validated successfully!</div>', unsafe_allow_html=True)
                
                # Generate prompts
                main_prompt = generate_advanced_prompt(ad_data)
                negative_prompt = generate_negative_prompt(ad_data)
                
                # Display prompts
                with st.expander("🔍 View Generated Prompts"):
                    st.write("**Positive Prompt:**")
                    st.code(main_prompt, language=None)
                    st.write("**Negative Prompt:**")
                    st.code(negative_prompt, language=None)
                
                # Generate images
                st.markdown('<div class="info-box">⏳ Generating your advertisement... This takes 5-10 seconds.</div>', unsafe_allow_html=True)
                
                progress_bar = st.progress(0)
                generated_images = []
                
                for i in range(num_variations):
                    progress_bar.progress((i + 1) / num_variations)
                    
                    with st.spinner(f"Creating variation {i+1}/{num_variations}..."):
                        image, error = call_huggingface_api(
                            main_prompt,
                            negative_prompt,
                            st.session_state.api_token
                        )
                        
                        if error:
                            st.error(f"❌ Variation {i+1} failed: {error}")
                            if "loading" in error.lower():
                                st.info("💡 Tip: The model is warming up. Wait 20 seconds and try again.")
                            break
                        else:
                            generated_images.append({
                                'image': image,
                                'prompt': main_prompt,
                                'timestamp': datetime.now(),
                                'ad_data': ad_data
                            })
                            time.sleep(1)  # Small delay between requests
                
                progress_bar.progress(1.0)
                
                if generated_images:
                    st.session_state.generated_images = generated_images
                    st.session_state.generation_history.append({
                        'timestamp': datetime.now(),
                        'count': len(generated_images),
                        'product': ad_data.get('product_name', 'Unknown')
                    })
                    st.success(f"✨ Successfully generated {len(generated_images)} advertisement(s)!")
                    st.balloons()

with col2:
    st.subheader("🖼️ Generated Advertisements")
    
    if st.session_state.generated_images:
        for idx, item in enumerate(st.session_state.generated_images):
            st.image(
                item['image'],
                caption=f"Advertisement {idx + 1} - {item['ad_data'].get('product_name', 'Product')}",
                use_container_width=True
            )
            
            # Download and info buttons
            col_a, col_b = st.columns(2)
            
            with col_a:
                # Convert image to bytes for download
                img_buffer = io.BytesIO()
                item['image'].save(img_buffer, format='PNG')
                img_buffer.seek(0)
                
                st.download_button(
                    label=f"⬇️ Download Ad {idx + 1}",
                    data=img_buffer,
                    file_name=f"adforge_ad_{idx+1}_{item['timestamp'].strftime('%Y%m%d_%H%M%S')}.png",
                    mime="image/png",
                    key=f"download_{idx}",
                    use_container_width=True
                )
            
            with col_b:
                with st.expander("ℹ️ Details"):
                    st.json(item['ad_data'])
            
            st.markdown("---")
        
        # Clear history button
        if st.button("🗑️ Clear All", use_container_width=True):
            st.session_state.generated_images = []
            st.rerun()
    
    else:
        st.info("👈 Configure your advertisement concept and click 'Generate Advertisement' to see results here")
        
        # Show example
        st.markdown("### 🎯 Example Output")
        st.markdown("""
        Your generated advertisements will appear here with:
        - ✨ High-resolution professional images
        - 📥 One-click download buttons
        - 📊 Detailed generation information
        - 🎨 Multiple variations (if selected)
        """)

# JSON Schema Documentation
with st.expander("📚 JSON Schema Reference"):
    st.markdown("""
    ### Complete JSON Schema for Advertisement Generation
    
    ```json
    {
        "product_name": "Your Product Name",           // REQUIRED: Product identifier
        "description": "Detailed product description",  // REQUIRED: What the product is
        "style": "minimalist | professional | vibrant | editorial | rustic",
        "mood": "energetic | calm | luxurious | playful | sophisticated",
        "target_audience": "Description of target demographic",
        "colors": ["color1", "color2", "color3"],      // Preferred color palette
        "setting": "studio | outdoor | urban | natural",
        "composition": "centered | rule of thirds | dynamic",
        "lighting": "soft | dramatic | natural | studio",
        "key_features": ["feature1", "feature2"],       // Product highlights
        "avoid": ["element1", "element2"]               // Optional: Things to exclude
    }
    ```
    
    ### Field Descriptions
    
    **Required Fields:**
    - `product_name`: The name of your product
    - `description`: Brief but descriptive product overview
    
    **Style Options:**
    - `minimalist`: Clean, simple, modern aesthetics
    - `professional`: Corporate, polished look
    - `vibrant`: Colorful, energetic, eye-catching
    - `editorial`: Magazine-quality, artistic
    - `rustic`: Natural, organic, handcrafted feel
    
    **Mood Options:**
    - `energetic`: Dynamic, active, exciting
    - `calm`: Peaceful, serene, relaxing
    - `luxurious`: Premium, high-end, exclusive
    - `playful`: Fun, lighthearted, creative
    - `sophisticated`: Refined, elegant, classy
    
    **Pro Tips:**
    - Be specific with colors (e.g., "ocean blue" instead of just "blue")
    - Combine multiple style elements for unique results
    - Use the `avoid` field to exclude unwanted elements
    - Experiment with different composition and lighting combinations
    """)

# Footer with statistics
st.markdown("---")
col_foot1, col_foot2, col_foot3 = st.columns(3)

with col_foot1:
    st.metric("Ads Generated This Session", len(st.session_state.generation_history))

with col_foot2:
    total_images = sum(item['count'] for item in st.session_state.generation_history)
    st.metric("Total Images Created", total_images)

with col_foot3:
    if st.session_state.generation_history:
        last_gen = st.session_state.generation_history[-1]['product']
        st.metric("Last Product", last_gen)

st.markdown("""
<div style='text-align: center; color: gray; padding: 2rem 0;'>
    <p><strong>AdForge AI</strong> - Powered by FLUX.1-schnell & Hugging Face</p>
    <p>🌟 Create professional advertisements in seconds | 🚀 Free to use | 💎 Commercial license</p>
</div>
""", unsafe_allow_html=True)