app.py

import gradio as gr
from src.agents.text_agent import TextAgent
from src.agents.image_agent import ImageAgent
from src.agents.coordinator_agent import CoordinatorAgent
from PIL import Image

async def process_content(text: str = None, image: Image.Image = None):
    text_result = None
    image_result = None
    
    # Process text if provided
    if text and text.strip():
        try:
            text_agent = TextAgent(confidence_threshold=0.8)
            text_result = await text_agent.process(text)
        except Exception as e:
            return f"Error processing text: {str(e)}"
    
    # Process image if provided
    if image is not None:
        try:
            image_agent = ImageAgent(confidence_threshold=0.8)
            image_result = await image_agent.process(image)
        except Exception as e:
            return f"Error processing image: {str(e)}"
    
    # Coordinate results
    try:
        coordinator = CoordinatorAgent()
        #final_analysis = await coordinator.analyze_content(text_result, image_result)
        final_analysis = await coordinator.process({
            'text_result': text_result,
            'image_result': image_result
            })
        # Format the output
        output_parts = []
        
        # Add original results if available
        if text_result:
            output_parts.append(f"""
Text Analysis Results:
---------------------
Text: {text_result['text']}
Model Used: {text_result['model_used']}
Label: {text_result['label']}
Confidence: {text_result['confidence']:.2%}
""")
        
#         if image_result:
#             output_parts.append(f"""
# Image Analysis Results:
# ----------------------
# Caption: {image_result['caption']}
# Model Used: {image_result['model_used']}
# Confidence: {image_result['confidence']:.2%}
# """)
        
        if image_result:
            output_parts.append(f"""
Image Analysis Results:
----------------------
Caption: {image_result['caption']}
Model Used: {image_result['model_used']}
Safety: {"Safe" if image_result['is_safe'] else "Potentially Unsafe"}
Confidence: {image_result['confidence']:.2%}
""")

        
        # Add correlation analysis if both text and image were processed
        if final_analysis["correlation"]:
            output_parts.append(f"""
Correlation Analysis:
-------------------
Correlation Level: {final_analysis["correlation"]["level"]}
Correlation Score: {final_analysis["correlation"]["score"]:.2f}
""")
        
        # Add final analysis
        output_parts.append("""
Combined Analysis:
----------------""")
        for analysis_point in final_analysis["analysis"]:
            output_parts.append(f"- {analysis_point}")
        
        return "\n".join(output_parts)
        
    except Exception as e:
        return f"Error in coordination: {str(e)}"

def create_interface():
    with gr.Blocks() as demo:
        gr.Markdown("# Content Analysis System")
        gr.Markdown("Upload an image and/or enter text for analysis. The system will analyze both individually and provide a combined analysis.")
        
        with gr.Row():
            # Left column for text
            with gr.Column():
                text_input = gr.Textbox(
                    label="Enter text (max 50 words)",
                    placeholder="Type your text here...",
                    max_lines=3
                )
            
            # Right column for image
            with gr.Column():
                image_input = gr.Image(
                    type="pil", 
                    label="Upload Image"
                )
        
        # Submit button
        submit_btn = gr.Button("Analyze Content")
        
        # Output area with increased height for detailed analysis
        output = gr.Textbox(
            label="Analysis Results",
            lines=10
        )
        
        # Click handler
        submit_btn.click(
            fn=process_content,
            inputs=[text_input, image_input],
            outputs=output
        )
    
    return demo

# Create and launch the interface
demo = create_interface()

# For HuggingFace Spaces deployment
if __name__ == "__main__":
    demo.launch()