app.py

from ultralytics import YOLO
from PIL import Image
import gradio as gr
from huggingface_hub import snapshot_download
import os
import tempfile
import cv2
import zipfile
import shutil

# === Load model ===
def load_model(repo_id):
    download_dir = snapshot_download(repo_id)
    path = os.path.join(download_dir, "best_int8_openvino_model")
    return YOLO(path, task='detect')

REPO_ID = "sensura/belisha-beacon-zebra-crossing-yoloV8"
detection_model = load_model(REPO_ID)

# === Image prediction ===
def predict_image(image, conf_threshold, iou_threshold):
    result = detection_model.predict(image, conf=conf_threshold, iou=iou_threshold)
    img_bgr = result[0].plot()
    out_img = Image.fromarray(img_bgr[..., ::-1])  # Convert BGR to RGB
    return out_img  # Return as PIL

# === Video prediction ===
def predict_video(video, conf_threshold, iou_threshold):
    cap = cv2.VideoCapture(video)
    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    fps = cap.get(cv2.CAP_PROP_FPS)
    out_path = tempfile.NamedTemporaryFile(suffix=".mp4", delete=False).name
    out_writer = cv2.VideoWriter(out_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (width, height))

    while cap.isOpened():
        ret, frame = cap.read()
        if not ret:
            break
        result = detection_model.predict(frame, conf=conf_threshold, iou=iou_threshold)
        annotated = result[0].plot()
        out_writer.write(annotated)

    cap.release()
    out_writer.release()
    return out_path

# === Multiple images prediction ===
def predict_multiple(files, conf_threshold, iou_threshold):
    if not files:
        return None, None

    output_dir = tempfile.mkdtemp()
    annotated_images = []

    for file in files:
        try:
            img = Image.open(file).convert("RGB")
            result = detection_model.predict(img, conf=conf_threshold, iou=iou_threshold)
            img_bgr = result[0].plot()
            out_img = Image.fromarray(img_bgr[..., ::-1])  # Convert BGR to RGB for PIL
            out_path = os.path.join(output_dir, os.path.basename(file.name))
            out_img.save(out_path)
            annotated_images.append(out_img)
        except Exception as e:
            print(f"Failed to process {file.name}: {e}")

    zip_path = shutil.make_archive(output_dir, 'zip', output_dir)
    return annotated_images, zip_path

# === Gradio Interfaces ===

image_tab = gr.Interface(
    fn=predict_image,
    inputs=[
        gr.Image(type="pil", label="Upload Image"),
        gr.Slider(0.1, 1.0, value=0.5, step=0.05, label="Confidence Threshold"),
        gr.Slider(0.1, 1.0, value=0.6, step=0.05, label="IoU Threshold"),
    ],
    outputs=gr.Image(type="pil", label="Detected Image"),  # PIL to avoid color bugs
    title="Single Image Detection"
)

video_tab = gr.Interface(
    fn=predict_video,
    inputs=[
        gr.Video(label="Upload Video"),
        gr.Slider(0.1, 1.0, value=0.5, step=0.05, label="Confidence Threshold"),
        gr.Slider(0.1, 1.0, value=0.6, step=0.05, label="IoU Threshold"),
    ],
    outputs=gr.Video(label="Detected Video"),
    title="Single Video Detection"
)

gallery_tab = gr.Interface(
    fn=predict_multiple,
    inputs=[
        gr.File(file_types=["image"], file_count="multiple", label="Upload Multiple Images"),
        gr.Slider(0.1, 1.0, value=0.5, step=0.05, label="Confidence Threshold"),
        gr.Slider(0.1, 1.0, value=0.6, step=0.05, label="IoU Threshold"),
    ],
    outputs=[
        gr.Gallery(label="Detected Gallery", columns=3, height="auto"),
        gr.File(label="Download Annotated ZIP")
    ],
    title="Batch Image Detection"
)

# === Tabbed UI Launch ===
gr.TabbedInterface(
    [image_tab, video_tab, gallery_tab],
    tab_names=["Image", "Video", "Gallery"]
).launch(share=True)