app.py

import json
import gradio as gr
import time
import modal
import threading
import tempfile
import os
import re
import asyncio
from concurrent.futures import ThreadPoolExecutor

# Note: TTS and visualization features are disabled for HF Spaces compatibility
# These require system-level dependencies not available in the cloud environment

print("🌍 Starting Agentic Ecosystem Simulator for Hugging Face Spaces...")

# Initialize Modal app for simplified functions
app = modal.App("agentic-ecosystem-simple")

def simulate_ecosystem_step(environment):
    """Main orchestration function - runs the text-based ecosystem simulation using Modal"""
    
    if not environment or not environment.strip():
        return None, [], "Please provide an environment description.", None
    
    try:
        print(f"🌱 Simulating ecosystem for environment: {environment}")
        
        # Call deployed Modal function using the new API
        f = modal.Function.from_name("agentic-ecosystem-simple", "simulate_ecosystem")
        results = f.remote(environment)
            
        # Check for errors
        if 'error' in results and results['error']:
            print(f"❌ Error in simulation: {results['error']}")
            return None, [], f"Simulation error: {results['error']}", None
        
        # Process species data
        species_list = results['species']
        print(f"✓ Generated {len(species_list)} species")
        
        # Process events and summary
        events = results['events']
        summary = results['summary']
        print("✓ Generated events and summary")
        
        print("🎉 Ecosystem simulation completed!")
        
        return (
            species_list,
            events,
            summary,
            "🔇 Audio narration not available in HF Spaces environment"
        )
        
    except Exception as e:
        print(f"Error in ecosystem simulation: {e}")
        return None, [], f"Simulation error: {str(e)}", None

def check_modal_deployment():
    """Check if Modal is properly configured"""
    try:
        # Try to lookup the deployed function using new API
        f = modal.Function.from_name("agentic-ecosystem-simple", "simulate_ecosystem")
        return "✅ Modal is properly configured and deployed!\n\nThe ecosystem simulator is ready to use."
    except Exception as e:
        return f"❌ Modal error: {str(e)}\nPlease ensure Modal functions are deployed and accessible."

# Create custom nature-themed color scheme
nature_theme = gr.themes.Soft().set(
    body_background_fill="#d4edda",  # Light green background
    background_fill_primary="#e8f5e8",  # Medium light green
    background_fill_secondary="#d6ebf5",  # Light blue
    block_background_fill="#f8f9fa",  # Off-white for content blocks
    button_primary_background_fill="#1e7e34",  # Forest green for primary buttons
    button_primary_background_fill_hover="#28a745",  # Medium green on hover
    button_primary_text_color="#ffffff",
    button_secondary_background_fill="#c3e6cb",  # Light green for secondary buttons
    button_secondary_background_fill_hover="#b3d7e6",  # Light blue on hover
    button_secondary_text_color="#155724"  # Dark green text
)

with gr.Blocks(title="Agentic Ecosystem Simulator", theme=nature_theme, css="""

/* Additional custom CSS for nature theme */

.gradio-container {

    background: linear-gradient(135deg, #c8e6c8 0%, #b8d4e8 50%, #d0e2f0 100%);

    min-height: 100vh;

}



/* Card-like styling for content blocks */

.block {

    background: rgba(248, 249, 250, 0.95);

    border-radius: 8px;

    border: 1px solid #6c9a6c;

    box-shadow: 0 2px 8px rgba(28, 126, 52, 0.1);

}



/* Nature-themed headers */

h1, h2, h3 {

    color: #155724;

    text-shadow: 1px 1px 2px rgba(28, 126, 52, 0.15);

}

""") as demo:
    gr.Markdown("""

    # 🌍 Agentic Ecosystem Simulator

    ## Powered by Modal Cloud Functions

    

    Welcome to the Cloud-Based Ecosystem Simulator! This application uses multiple AI models working together 

    through Modal cloud functions to create and simulate a living ecosystem through text.

    

    **How it works:**

    1. **🧬 Species Generation** - AI creates plants and animals specifically adapted to your environment

    2. **🔄 Interaction Simulation** - AI simulates ecological interactions specific to your environment  

    3. **📋 Ecosystem Summary** - Key events are summarized with environmental context

    4. **🌿 Environment Analysis** - Your input is analyzed for environment type, climate, time, season & more

    

    **The more detailed your environment description, the more tailored your ecosystem will be!**

    

    🚀 **Status**: Running on Hugging Face Spaces with Modal cloud backend

    

    ⚠️ **Note**: Audio narration and visualization features are disabled in this cloud environment

    """)
    
    with gr.Row():
        with gr.Column(scale=2):
            environment_input = gr.Textbox(
                label="🌿 Environment Description (be specific and detailed!)",
                placeholder="e.g., 'tropical rainforest with morning mist and flowering vines', 'desert oasis with palm trees and rocky outcrops at sunset'",
                lines=3,
                value="temperate meadow in spring morning with wildflowers and gentle breeze"
            )
            
            simulate_btn = gr.Button("🚀 Simulate Ecosystem Day", variant="primary", size="lg")
            
            modal_status = gr.Textbox(
                label="🖥️ Modal Status",
                value="Click 'Check Modal' to see connection status",
                lines=4,
                interactive=False
            )
            check_modal_btn = gr.Button("🔍 Check Modal Connection", variant="secondary")
            
        with gr.Column(scale=3):
            species_output = gr.JSON(
                label="🧬 Generated Species",
                show_label=True
            )
    
    with gr.Row():
        with gr.Column():
            events_output = gr.Textbox(
                label="📅 Day's Events", 
                lines=8,
                show_label=True
            )
        
        with gr.Column():
            summary_output = gr.Textbox(
                label="📋 Ecosystem Summary",
                lines=8,
                show_label=True
            )
            
            audio_status = gr.Textbox(
                label="🎵 Audio Status",
                value="🔇 Audio narration not available in HF Spaces environment",
                lines=2,
                interactive=False
            )
    
    # Event handlers
    simulate_btn.click(
        fn=simulate_ecosystem_step,
        inputs=[environment_input],
        outputs=[species_output, events_output, summary_output, audio_status]
    )
    
    check_modal_btn.click(
        fn=check_modal_deployment,
        outputs=[modal_status]
    )
    
    # Add example environments
    gr.Markdown("""

    ### 💡 Detailed Environments to Try:

    

    **Try adding environmental details like:**

    - **Time of day**: morning, afternoon, evening, night

    - **Season**: spring, summer, autumn, winter  

    - **Weather**: misty, sunny, rainy, cloudy, stormy

    - **Features**: rocky, sandy, flowering, foggy, lush, sparse

    

    **Example environments with rich details:**

    - `tropical rainforest with morning mist and flowering plants`

    - `desert oasis with palm trees and rocky outcrops at sunset`

    - `temperate forest clearing in autumn with golden leaves`

    - `alpine meadow with wildflowers on a spring morning`

    - `wetland marsh with reeds and shallow water during summer evening`

    - `volcanic island with unique flora adapting to the harsh terrain`

    - `underground cave system with phosphorescent fungi and still pools`

    - `arctic tundra with hardy vegetation under the midnight sun`

    

    ### 🔧 Technical Details

    

    This simulator demonstrates **agentic AI orchestration** by chaining together multiple specialized AI models:

    - Environment analysis and parsing

    - Species generation based on habitat conditions

    - Ecological interaction simulation

    - Narrative summary generation

    

    All processing happens in the cloud via Modal, making this a true **serverless AI application**.

    """)

if __name__ == "__main__":
    print("\n" + "="*60)
    print("🌍 AGENTIC ECOSYSTEM SIMULATOR - HUGGING FACE SPACES")
    print("="*60)
    print("\n🚀 Starting Hugging Face Spaces interface...")
    print("\n📋 SETUP STATUS:")
    print("✅ Modal cloud functions ready")
    print("✅ Hugging Face Spaces compatible version")
    print("⚠️ Audio and visualization features disabled for cloud compatibility")
    print("\n✨ Ready to simulate ecosystems!")
    print("="*60)
    
    demo.launch()