app.py

import gradio as gr
from transformers import pipeline
import requests
from PIL import Image, ImageDraw
import io

# Load model from Hugging Face Hub
MODEL_ID = "sandbox338/wild-life-model"

# Function to detect wildlife in images
def detect_wildlife(image):
    # Convert image to bytes for API call
    img_byte_arr = io.BytesIO()
    image.save(img_byte_arr, format='JPEG')
    img_byte_arr = img_byte_arr.getvalue()
    
    # API endpoint for your model
    API_URL = f"https://huggingface.co/sandbox338/wild-life-model"
    headers = {"Authorization": f"Bearer {HF_API_TOKEN}"}  
    
    # Make prediction request
    response = requests.post(API_URL, headers=headers, data=img_byte_arr)
    
    if response.status_code != 200:
        return None, f"Error: {response.status_code} - {response.text}"
    
    # Parse the response
    result = response.json()
    
    # Draw predictions on image
    image_with_boxes = draw_predictions(image, result)
    
    # Create text summary of results
    summary = create_summary(result)
    
    return image_with_boxes, summary

# Function to draw bounding boxes on image
def draw_predictions(image, predictions):
    img = image.copy()
    draw = ImageDraw.Draw(img)
    
    colors = {
        "animal": (255, 0, 0),    # Red
        "bird": (0, 255, 0),      # Green
        "mammal": (0, 0, 255),    # Blue
        "reptile": (255, 255, 0)  # Yellow
    }
    
    for pred in predictions:
        box = pred.get("box", {})
        label = pred.get("label", "unknown")
        score = pred.get("score", 0)
        
        # Get color based on label (default to purple if not in colors dict)
        color = colors.get(label.lower(), (255, 0, 255))
        
        # Draw rectangle
        xmin, ymin = box.get("xmin", 0), box.get("ymin", 0)
        xmax, ymax = box.get("xmax", 0), box.get("ymax", 0)
        draw.rectangle([xmin, ymin, xmax, ymax], outline=color, width=3)
        
        # Draw label
        draw.text((xmin, ymin-15), f"{label}: {score:.2f}", fill=color)
    
    return img

# Function to create text summary of detections
def create_summary(predictions):
    if not predictions:
        return "No wildlife detected in this image."
    
    summary = f"Detected {len(predictions)} wildlife instances:\n\n"
    
    # Count occurrences of each species
    species_count = {}
    for pred in predictions:
        label = pred.get("label", "unknown")
        score = pred.get("score", 0)
        
        if label not in species_count:
            species_count[label] = 0
        species_count[label] += 1
    
    # Add to summary
    for species, count in species_count.items():
        summary += f"- {species}: {count} {'instance' if count == 1 else 'instances'}\n"
    
    return summary

## inputs
examples = [
    ["https://www.google.com/url?sa=i&url=https%3A%2F%2Fwww.ifaw.org%2Fanimals%2Fantelopes&psig=AOvVaw1S3fshLDMiIZ1r9YOa8NIg&ust=1747246824703000&source=images&cd=vfe&opi=89978449&ved=0CBQQjRxqFwoTCMCKlOWHoY0DFQAAAAAdAAAAABAE"],
    ["example_images/wildlife2.jpg"],
    ["example_images/wildlife3.jpg"]
]

# Create Gradio interface
interface = gr.Interface(
    fn=detect_wildlife,
    inputs=gr.Image(type="pil"),
    outputs=[
        gr.Image(label="Detection Results"),
        gr.Textbox(label="Detection Summary")
    ],
    title="Wildlife Detection Model",
    description="Upload an image to detect wildlife using our trained model",
    examples=examples
)

if __name__ == "__main__":
    interface.launch()