app.py

import gradio as gr
from core_pipeline import extract_frames, detect_trees, plot_detections

def process_video(video_file):
    import tempfile
    import numpy as np

    if video_file is None:
        return None

    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as tmp:
        tmp.write(video_file.read())
        video_path = tmp.name

    frames = extract_frames(video_path)
    results = []

    for i, frame in enumerate(frames[:3]):
        detected, bboxes, confs, labels = detect_trees(frame)
        annotated = plot_detections(detected, bboxes)
        results.append(annotated)

    if results:
        preview = np.hstack(results)
        return preview
    return None

gr.Interface(
    fn=process_video,
    inputs=gr.File(label="Upload Drone Video", file_types=[".mp4"]),  # ✅ FIXED
    outputs=gr.Image(label="Tree Detections (Sample Frames)"),
    title="🌳 Tree Height Detection from Drone Video",
    description="Upload drone video to detect trees. Sample frames will be shown with bounding boxes."
).launch()