app.py

import io
import gradio as gr
import matplotlib.pyplot as plt
import requests, validators
import torch
import pathlib
from PIL import Image
from transformers import AutoFeatureExtractor, YolosForObjectDetection, DetrForObjectDetection
import os
import warnings

warnings.filterwarnings("ignore", category=FutureWarning)
os.environ["KMP_DUPLICATE_LIB_OK"]="TRUE"

# colors for visualization
COLORS = [
    [0.000, 0.447, 0.741],
    [0.850, 0.325, 0.098],
    [0.929, 0.694, 0.125],
    [0.494, 0.184, 0.556],
    [0.466, 0.674, 0.188],
    [0.301, 0.745, 0.933]
]

def make_prediction(img, feature_extractor, model):
    inputs = feature_extractor(img, return_tensors="pt")
    outputs = model(**inputs)
    img_size = torch.tensor([tuple(reversed(img.size))])
    processed_outputs = feature_extractor.post_process(outputs, img_size)
    return processed_outputs[0]

def fig2img(fig):
    buf = io.BytesIO()
    fig.savefig(buf)
    buf.seek(0)
    pil_img = Image.open(buf)
    basewidth = 750
    wpercent = (basewidth/float(pil_img.size[0]))
    hsize = int((float(pil_img.size[1])*float(wpercent)))
    img = pil_img.resize((basewidth,hsize), Image.Resampling.LANCZOS) 
    return img

def visualize_prediction(img, output_dict, threshold=0.5, id2label=None):
    keep = output_dict["scores"] > threshold
    boxes = output_dict["boxes"][keep].tolist()
    scores = output_dict["scores"][keep].tolist()
    labels = output_dict["labels"][keep].tolist()
    
    if id2label is not None:
        labels = [id2label[x] for x in labels]

    plt.figure(figsize=(50, 50))
    plt.imshow(img)
    ax = plt.gca()
    colors = COLORS * 100
    for score, (xmin, ymin, xmax, ymax), label, color in zip(scores, boxes, labels, colors):
        if label == 'license-plates':
            ax.add_patch(plt.Rectangle((xmin, ymin), xmax - xmin, ymax - ymin, fill=False, color=color, linewidth=10))
            ax.text(xmin, ymin, f"{label}: {score:0.2f}", fontsize=60, bbox=dict(facecolor="yellow", alpha=0.8))
    plt.axis("off")
    return fig2img(plt.gcf())

def get_original_image(url_input):
    if validators.url(url_input):
        try:
            response = requests.get(url_input, stream=True)
            response.raise_for_status()
            image = Image.open(response.raw)
            return image
        except Exception as e:
            print(f"Error loading image from URL: {e}")
            return None
    return None

def detect_objects(model_name, url_input, image_input, webcam_input, threshold):
    # Handle case where no image is provided
    image = None
    
    if validators.url(url_input) and url_input.strip():
        image = get_original_image(url_input)
    elif image_input is not None:
        image = image_input
    elif webcam_input is not None:
        image = webcam_input
    
    if image is None:
        raise gr.Error("Please provide an image via URL, file upload, or webcam")
    
    # Extract model and feature extractor
    feature_extractor = AutoFeatureExtractor.from_pretrained(model_name)
    
    if "yolos" in model_name:
        model = YolosForObjectDetection.from_pretrained(model_name)
    elif "detr" in model_name:
        model = DetrForObjectDetection.from_pretrained(model_name)
    
    # Make prediction
    processed_outputs = make_prediction(image, feature_extractor, model)
    
    # Visualize prediction
    viz_img = visualize_prediction(image, processed_outputs, threshold, model.config.id2label)
    
    return viz_img

def set_example_image(example: list) -> dict:
    return gr.Image.update(value=example[0])

def set_example_url(example: list) -> dict:
    image = get_original_image(example[0])
    return gr.Textbox.update(value=example[0]), gr.Image.update(value=image)

title = """<h1 id="title">License Plate Detection with YOLOS</h1>"""

description = """
# 🚗✨ Customize Your Biblical Porsche Scene Showcase ✨🚗

**YOLOS: When a Vision Transformer Gets Divine Revelation**

Behold! YOLOS is a Vision Transformer (ViT) that achieved 42 AP on COCO - not just a number, but *the answer to everything* (including which disciple gets shotgun in your biblical Porsche). 

**The Scripture According to YOLOS:**
- "In the beginning was the Sequence, and the Sequence was One" - YOLOS 1:1
- Trained on 118k sacred images from the COCO testament
- Performs miracles at detecting heavenly vehicles and license plates
- Fine-tuned on the "Book of Car Plates" (443 verses of automotive divinity)

**Biblical Porsche Detection Capabilities:**
- ✅ Finds Peter's Porsche at the Gates of Heaven
- ✅ Spots Moses' license plate ("LET-M-PPL-GO")
- ✅ Detects David's sports car facing Goliath's SUV
- ✅ Locates the Holy Ghost's invisible convertible

*"And lo, the model saith: thou shalt look at only one sequence, and it shall be enough to find thy Porsche in the Red Sea of data."*

**Warning:** May occasionally confuse manna with hubcaps. Results not guaranteed in actual biblical times (camels not detected).
Links to HuggingFace Models:
- [nickmuchi/yolos-small-rego-plates-detection](https://huggingface.co/nickmuchi/yolos-small-rego-plates-detection) 
"""

models = ["nickmuchi/yolos-small-finetuned-license-plate-detection","nickmuchi/detr-resnet50-license-plate-detection"]

# FIXED: Use "resolve/main" URLs instead of "blob/main" for raw images
urls = [
    "https://huggingface.co/spaces/TroglodyteDerivations/Customize_your_biblical_Porsche_scene_Showcase/resolve/main/images/flux_krea_00005_.png",
    "https://huggingface.co/spaces/TroglodyteDerivations/Customize_your_biblical_Porsche_scene_Showcase/resolve/main/images/flux_krea_00007_.png"
]

images = [[path.as_posix()] for path in sorted(pathlib.Path('images').rglob('*.*')) if path.suffix.lower() in ['.webp', '.jpg', '.jpeg', '.png']]

tik_tok_link = """
[![](https://img.shields.io/badge/TikTok-@porsche-000000?style=flat&logo=tiktok&logoColor=white)](https://www.tiktok.com/@porsche)
"""

css = '''
h1#title {
  text-align: center;
}
'''
demo = gr.Blocks()

with demo:
    gr.Markdown(title)
    gr.Markdown(description)
    gr.Markdown(tik_tok_link)
    options = gr.Dropdown(choices=models,label='Object Detection Model',value=models[0],show_label=True)
    slider_input = gr.Slider(minimum=0.2,maximum=1,value=0.5,step=0.1,label='Prediction Threshold')
    
    with gr.Tabs():
        with gr.TabItem('Image URL'):
            with gr.Row():
                with gr.Column():
                    url_input = gr.Textbox(lines=2,label='Enter valid image URL here..')
                    original_image = gr.Image(height=750, width=750)
                    # Update the change event to handle errors
                    url_input.change(
                        get_original_image, 
                        inputs=[url_input], 
                        outputs=[original_image],
                        show_progress=True
                    )
                with gr.Column():
                    img_output_from_url = gr.Image(height=750, width=750)
                
            with gr.Row():
                example_url = gr.Examples(
                    examples=urls,
                    inputs=[url_input],
                    outputs=[original_image],
                    fn=set_example_url,
                    cache_examples=False
                )
                
            url_but = gr.Button('Detect')
     
        with gr.TabItem('Image Upload'):
            with gr.Row():
                img_input = gr.Image(type='pil', height=750, width=750)
                img_output_from_upload= gr.Image(height=750, width=750)
                
            with gr.Row(): 
                example_images = gr.Examples(examples=images,inputs=[img_input])                            
                                                   
                
            img_but = gr.Button('Detect')
            
        with gr.TabItem('WebCam'):
            with gr.Row():
                web_input = gr.Image(sources=['webcam'], type='pil', height=750, width=750, streaming=True)
                img_output_from_webcam= gr.Image(height=750, width=750)

            cam_but = gr.Button('Detect')
            
    url_but.click(detect_objects,inputs=[options,url_input,img_input,web_input,slider_input],outputs=[img_output_from_url],queue=True)
    img_but.click(detect_objects,inputs=[options,url_input,img_input,web_input,slider_input],outputs=[img_output_from_upload],queue=True)
    cam_but.click(detect_objects,inputs=[options,url_input,img_input,web_input,slider_input],outputs=[img_output_from_webcam],queue=True)

    gr.Markdown("[![](https://img.shields.io/badge/TikTok-Follow%20@porsche-000000?style=social&logo=tiktok)](https://www.tiktok.com/@porsche)")

demo.launch(debug=True, css=css)