app.py

import gradio as gr
from transformers import pipeline
import openai
import base64
import os
import json

# ---------------------------------------------------------------------------
# Car brands — must match the classes the ViT model was trained on
# ---------------------------------------------------------------------------
CAR_BRANDS = ['BMW', 'Dodge', 'Ferrari', 'Ford', 'Jeep', 'Lamborghini', 'Porsche', 'Rolls-Royce', 'Toyota']

# ---------------------------------------------------------------------------
# Load models (loaded once at startup)
# ---------------------------------------------------------------------------

# Custom fine-tuned ViT model (trained on Stanford Cars, 8 brand classes)
# Replace with your actual Hugging Face model ID after training and pushing
vit_classifier = pipeline(
    "image-classification",
    model="nenzilea/car-classification"
)

# CLIP zero-shot classifier
clip_classifier = pipeline(
    model="openai/clip-vit-large-patch14",
    task="zero-shot-image-classification"
)

# ---------------------------------------------------------------------------
# OpenAI helper
# ---------------------------------------------------------------------------

def encode_image_to_base64(image_path: str) -> str:
    with open(image_path, "rb") as f:
        return base64.b64encode(f.read()).decode("utf-8")


def classify_with_openai(image_path: str) -> dict:
    """Send image to GPT-4o and ask it to return confidence scores per brand."""
    try:
        api_key = os.environ.get("OPENAI_API_KEY")
        if not api_key:
            return {"Error: OPENAI_API_KEY not set": 1.0}
        client = openai.OpenAI(api_key=api_key)

        ext = os.path.splitext(image_path)[1].lower().lstrip(".")
        mime_type = "image/jpeg" if ext in ("jpg", "jpeg") else f"image/{ext}"
        base64_image = encode_image_to_base64(image_path)

        prompt = (
            f"You are a car classification expert. Classify the car brand shown in this image. "
            f"The possible classes are: {', '.join(CAR_BRANDS)}. "
            "Respond ONLY with a valid JSON object where each key is a brand name from the list and "
            "each value is a confidence score between 0.0 and 1.0. All scores must sum to 1.0. "
            'Example format: {"BMW": 0.05, "Ferrari": 0.85, "Ford": 0.02, ...}'
        )

        response = client.chat.completions.create(
            model="gpt-4o",
            messages=[
                {
                    "role": "user",
                    "content": [
                        {"type": "text", "text": prompt},
                        {
                            "type": "image_url",
                            "image_url": {"url": f"data:{mime_type};base64,{base64_image}"},
                        },
                    ],
                }
            ],
            max_tokens=300,
        )

        text = response.choices[0].message.content
        start = text.find("{")
        end = text.rfind("}") + 1
        scores = json.loads(text[start:end])
        return {brand: float(scores.get(brand, 0.0)) for brand in CAR_BRANDS}

    except openai.AuthenticationError:
        return {"Error: Invalid OpenAI API key": 1.0}
    except Exception as e:
        return {f"Error: {str(e)[:60]}": 1.0}


# ---------------------------------------------------------------------------
# Main classification function
# ---------------------------------------------------------------------------

def classify_car(image):
    if image is None:
        return {}, {}, {}

    # Custom ViT — fine-tuned on Stanford Cars
    vit_results = vit_classifier(image, top_k=len(CAR_BRANDS))
    vit_output = {r["label"]: round(r["score"], 4) for r in vit_results}

    # CLIP — zero-shot with brand names as candidate labels
    clip_results = clip_classifier(image, candidate_labels=CAR_BRANDS)
    clip_output = {r["label"]: round(r["score"], 4) for r in clip_results}

    # OpenAI GPT-4o Vision
    openai_output = classify_with_openai(image)

    return vit_output, clip_output, openai_output


# ---------------------------------------------------------------------------
# Example images (add representative car images to example_images/)
# ---------------------------------------------------------------------------

_img_dir = os.path.join(os.path.dirname(__file__), "example_images")
example_images = [
    [os.path.join(_img_dir, f)]
    for f in sorted(os.listdir(_img_dir))
    if f.lower().endswith((".jpg", ".jpeg", ".png", ".webp"))
]

# ---------------------------------------------------------------------------
# Gradio UI
# ---------------------------------------------------------------------------

css = """
.title { text-align: center; margin-bottom: 0.25rem; }
.subtitle { text-align: center; color: #6b7280; margin-bottom: 1.5rem; font-size: 0.95rem; }
.model-header { font-weight: 600; font-size: 1rem; margin-bottom: 0.25rem; padding: 0.4rem 0.75rem;
                border-radius: 6px; background: #f3f4f6; }
.classify-btn { max-width: 200px; margin: 0 auto; }
footer { display: none !important; }
"""

with gr.Blocks(title="Car Brand Classification", css=css, theme=gr.themes.Soft()) as demo:

    # Header
    gr.Markdown("# 🚗 Car Brand Classification", elem_classes="title")
    gr.Markdown(
        "Upload a car image and compare predictions from three models side by side.",
        elem_classes="subtitle"
    )

    # Input + button
    with gr.Row():
        with gr.Column(scale=1):
            input_image = gr.Image(
                type="filepath",
                label="Car Image",
                height=280,
            )
            classify_btn = gr.Button("Classify", variant="primary", elem_classes="classify-btn")

        # Results
        with gr.Column(scale=2):
            with gr.Row():
                with gr.Column():
                    gr.Markdown("**Custom ViT** — fine-tuned", elem_classes="model-header")
                    vit_output = gr.Label(num_top_classes=5, label="")
                with gr.Column():
                    gr.Markdown("**CLIP** — zero-shot", elem_classes="model-header")
                    clip_output = gr.Label(num_top_classes=5, label="")
                with gr.Column():
                    gr.Markdown("**GPT-4o** — vision LLM", elem_classes="model-header")
                    openai_output = gr.Label(num_top_classes=5, label="")

    # Examples
    gr.Examples(
        examples=example_images,
        inputs=input_image,
        label="Example Images",
    )

    classify_btn.click(
        fn=classify_car,
        inputs=[input_image],
        outputs=[vit_output, clip_output, openai_output],
    )

demo.launch()