2026-05-18-17-55-29 增加运行确认和演示录制

scripts/record_demo.py

@@ -0,0 +1,189 @@
+from __future__ import annotations
+
+import json
+import shutil
+import subprocess
+import textwrap
+import time
+import urllib.request
+from pathlib import Path
+
+import cv2
+import numpy as np
+
+
+ROOT = Path(__file__).resolve().parents[1]
+DEMO_DIR = ROOT / "storage" / "demos"
+SAMPLE_VIDEO = ROOT / "storage" / "samples" / "synthetic_guidewire.mp4"
+HOME_SCREENSHOT = DEMO_DIR / "01_home.png"
+RESPONSE_JSON = DEMO_DIR / "latest_demo_response.json"
+OUTPUT_VIDEO = DEMO_DIR / "isiseg_usage_demo.mp4"
+BASE_URL = "http://127.0.0.1:8001"
+
+
+def run(command: list[str], timeout: int = 120) -> subprocess.CompletedProcess[str]:
+    return subprocess.run(
+        command,
+        cwd=ROOT,
+        check=True,
+        text=True,
+        capture_output=True,
+        timeout=timeout,
+    )
+
+
+def wait_for_service() -> None:
+    deadline = time.time() + 20
+    while time.time() < deadline:
+        try:
+            with urllib.request.urlopen(f"{BASE_URL}/api/health", timeout=2) as response:
+                if response.status == 200:
+                    return
+        except OSError:
+            time.sleep(0.5)
+    raise RuntimeError(f"ISISeg service is not reachable at {BASE_URL}")
+
+
+def ensure_sample() -> None:
+    if not SAMPLE_VIDEO.exists():
+        run(["bash", "scripts/generate_sample.sh"])
+
+
+def capture_homepage() -> None:
+    chrome = shutil.which("google-chrome") or shutil.which("chromium") or shutil.which("chromium-browser")
+    if not chrome:
+        raise RuntimeError("Chrome/Chromium is required to capture the UI screenshot")
+    run(
+        [
+            chrome,
+            "--headless=new",
+            "--no-sandbox",
+            "--disable-gpu",
+            "--hide-scrollbars",
+            "--window-size=1440,1000",
+            f"--screenshot={HOME_SCREENSHOT}",
+            BASE_URL,
+        ],
+        timeout=60,
+    )
+
+
+def call_segmentation() -> dict:
+    command = [
+        "curl",
+        "-s",
+        "-X",
+        "POST",
+        "-F",
+        f"file=@{SAMPLE_VIDEO}",
+        "-F",
+        "method=fusion",
+        "-F",
+        "sensitivity=0.68",
+        "-F",
+        "frame_stride=12",
+        "-F",
+        "max_frames=4",
+        f"{BASE_URL}/api/segment",
+    ]
+    completed = run(command, timeout=180)
+    payload = json.loads(completed.stdout)
+    RESPONSE_JSON.write_text(json.dumps(payload, ensure_ascii=False, indent=2), encoding="utf-8")
+    return payload
+
+
+def fit_image(image: np.ndarray, box_w: int, box_h: int) -> np.ndarray:
+    h, w = image.shape[:2]
+    scale = min(box_w / w, box_h / h)
+    resized = cv2.resize(image, (int(w * scale), int(h * scale)), interpolation=cv2.INTER_AREA)
+    canvas = np.full((box_h, box_w, 3), 18, dtype=np.uint8)
+    y = (box_h - resized.shape[0]) // 2
+    x = (box_w - resized.shape[1]) // 2
+    canvas[y : y + resized.shape[0], x : x + resized.shape[1]] = resized
+    return canvas
+
+
+def put_lines(frame: np.ndarray, lines: list[str], x: int, y: int, size: float = 0.82, color=(238, 245, 239)) -> None:
+    for line in lines:
+        cv2.putText(frame, line, (x, y), cv2.FONT_HERSHEY_SIMPLEX, size, color, 2, cv2.LINE_AA)
+        y += int(34 * size / 0.82)
+
+
+def wrap(text: str, width: int = 44) -> list[str]:
+    return textwrap.wrap(text, width=width, break_long_words=False)
+
+
+def slate(title: str, subtitle: str, width: int = 1280, height: int = 720) -> np.ndarray:
+    frame = np.zeros((height, width, 3), dtype=np.uint8)
+    frame[:, :] = (18, 20, 18)
+    cv2.rectangle(frame, (0, 0), (width, height), (48, 64, 56), 18)
+    cv2.line(frame, (70, 90), (1210, 90), (102, 209, 185), 3)
+    cv2.putText(frame, title, (70, 185), cv2.FONT_HERSHEY_SIMPLEX, 1.35, (255, 209, 102), 3, cv2.LINE_AA)
+    put_lines(frame, wrap(subtitle, 58), 74, 260, size=0.82, color=(226, 234, 226))
+    cv2.putText(frame, BASE_URL, (74, 640), cv2.FONT_HERSHEY_SIMPLEX, 0.78, (102, 209, 185), 2, cv2.LINE_AA)
+    return frame
+
+
+def screenshot_slide() -> np.ndarray:
+    image = cv2.imread(str(HOME_SCREENSHOT), cv2.IMREAD_COLOR)
+    frame = slate("1. Open Web Console", "Browser opens the ISISeg guidewire segmentation workspace.")
+    fitted = fit_image(image, 1120, 500)
+    frame[160:660, 80:1200] = fitted
+    return frame
+
+
+def result_slide(payload: dict, frame_payload: dict, index: int) -> np.ndarray:
+    frame = slate(
+        f"2.{index} Segmentation Result",
+        "The demo uploads the synthetic guidewire video and renders real fusion masks returned by the API.",
+    )
+    overlay = cv2.imread(str(ROOT / frame_payload["overlay_url"].lstrip("/")), cv2.IMREAD_COLOR)
+    mask = cv2.imread(str(ROOT / frame_payload["mask_url"].lstrip("/")), cv2.IMREAD_GRAYSCALE)
+    mask_color = cv2.cvtColor(mask, cv2.COLOR_GRAY2BGR)
+    frame[178:598, 75:635] = fit_image(overlay, 560, 420)
+    frame[178:598, 645:1205] = fit_image(mask_color, 560, 420)
+    metrics = frame_payload["metrics"]
+    lines = [
+        f"job: {payload['job_id']}  frame: {frame_payload['frame_index']}",
+        f"coverage: {metrics['coverage'] * 100:.3f}%    skeleton: {metrics['skeleton_length']}    components: {metrics['components']}",
+    ]
+    put_lines(frame, lines, 75, 650, size=0.62, color=(238, 245, 239))
+    return frame
+
+
+def write_video(slides: list[np.ndarray]) -> None:
+    writer = cv2.VideoWriter(str(OUTPUT_VIDEO), cv2.VideoWriter_fourcc(*"mp4v"), 24.0, (1280, 720))
+    for slide in slides:
+        for _ in range(72):
+            writer.write(slide)
+    writer.release()
+
+
+def main() -> None:
+    DEMO_DIR.mkdir(parents=True, exist_ok=True)
+    wait_for_service()
+    ensure_sample()
+    capture_homepage()
+    payload = call_segmentation()
+    slides = [
+        slate(
+            "ISISeg Usage Demo",
+            "A runnable Web system for guidewire segmentation: open the console, upload a fluoroscopy-like video, choose fusion mode, and inspect overlay/mask outputs.",
+        ),
+        screenshot_slide(),
+    ]
+    for index, frame_payload in enumerate(payload["frames"][:4], start=1):
+        slides.append(result_slide(payload, frame_payload, index))
+    slides.append(
+        slate(
+            "Ready to Use",
+            f"Demo output video and API artifacts are stored under storage/demos and storage/jobs. Overlay video URL: {payload.get('video_url')}",
+        )
+    )
+    write_video(slides)
+    print(OUTPUT_VIDEO)
+    print(RESPONSE_JSON)
+
+
+if __name__ == "__main__":
+    main()