app.py

import gradio as gr
import cv2
import numpy as np
from groq import Groq
import time
from PIL import Image as PILImage
import io
import os
import base64

def create_monitor_interface():
    api_key = os.getenv("GROQ_API_KEY")
    
    if not api_key:
        with gr.Blocks() as demo:
            gr.Markdown("""
            # ⚠️ Groq API Key Required
            
            ## Setup Instructions for Hugging Face Space:
            1. Go to your Space's Settings tab
            2. Scroll down to "Repository Secrets"
            3. Click "New Secret"
            4. Enter:
               - Secret name: `GROQ_API_KEY`
               - Secret value: Your Groq API key
            5. Click "Add secret"
            6. Rebuild the Space
            
            Once configured, the safety monitoring system will be available.
            """)
        return demo

    class SafetyMonitor:
        def __init__(self, model_name: str = "mixtral-8x7b-vision"):
            self.client = Groq(api_key=api_key)
            self.model_name = model_name

        def analyze_frame(self, frame: np.ndarray) -> str:
            if frame is None:
                return "No frame received"
                
            # Convert numpy array to PIL Image
            if len(frame.shape) == 2:  # If grayscale
                frame = cv2.cvtColor(frame, cv2.COLOR_GRAY2RGB)
            elif len(frame.shape) == 3 and frame.shape[2] == 4:  # If RGBA
                frame = cv2.cvtColor(frame, cv2.COLOR_RGBA2RGB)
                
            frame_pil = PILImage.fromarray(frame)
            
            # Convert to base64
            buffered = io.BytesIO()
            frame_pil.save(buffered, format="JPEG")
            img_base64 = base64.b64encode(buffered.getvalue()).decode('utf-8')
            
            prompt = """Analyze this image for workplace safety issues. Focus on:
            1. PPE usage (helmets, vests, etc.)
            2. Unsafe behaviors
            3. Equipment safety
            4. Environmental hazards
            Provide specific observations."""
            
            try:
                completion = self.client.chat.completions.create(
                    messages=[
                        {
                            "role": "user",
                            "content": [
                                {"type": "text", "text": prompt},
                                {"type": "image", "image_url": f"data:image/jpeg;base64,{img_base64}"}
                            ]
                        }
                    ],
                    model=self.model_name,
                    max_tokens=200,
                    temperature=0.2
                )
                return completion.choices[0].message.content
            except Exception as e:
                return f"Analysis Error: {str(e)}"

        def process_frame(self, frame: np.ndarray) -> tuple[np.ndarray, str]:
            if frame is None:
                return None, "No image provided"
                
            analysis = self.analyze_frame(frame)
            display_frame = frame.copy()
            
            # Add text overlay
            overlay = display_frame.copy()
            height, width = display_frame.shape[:2]
            cv2.rectangle(overlay, (5, 5), (width-5, 100), (0, 0, 0), -1)
            cv2.addWeighted(overlay, 0.3, display_frame, 0.7, 0, display_frame)
            
            # Add analysis text
            y_position = 30
            lines = analysis.split('\n')
            for line in lines:
                cv2.putText(display_frame, line[:80], (10, y_position),
                           cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
                y_position += 30
                if y_position >= 90:  # Prevent text from going outside the overlay
                    break

            return display_frame, analysis

    # Create the main interface
    monitor = SafetyMonitor()
    
    with gr.Blocks() as demo:
        gr.Markdown("""
        # Safety Monitoring System
        Upload an image to analyze workplace safety concerns.
        """)
        
        with gr.Row():
            input_image = gr.Image(label="Upload Image")
            output_image = gr.Image(label="Analysis Results")
        
        analysis_text = gr.Textbox(label="Detailed Safety Analysis", lines=5)
            
        def analyze_image(image):
            if image is None:
                return None, "No image provided"
            try:
                processed_frame, analysis = monitor.process_frame(image)
                return processed_frame, analysis
            except Exception as e:
                return None, f"Error processing image: {str(e)}"
            
        input_image.change(
            fn=analyze_image,
            inputs=input_image,
            outputs=[output_image, analysis_text]
        )
        
        gr.Markdown("""
        ## Instructions:
        1. Upload an image using the input panel
        2. The system will automatically analyze it for safety concerns
        3. View the analyzed image with overlay and detailed analysis below
        """)

    return demo

demo = create_monitor_interface()
demo.launch()