initial

watcher.py

@@ -0,0 +1,130 @@
+import os
+import time
+import json
+import re
+from ultralytics import YOLO
+
+SNAPSHOT_DIR = "/snapshots"
+
+# only process images from last N minutes
+MAX_AGE_MINUTES = 3
+MAX_AGE_MS = MAX_AGE_MINUTES * 60 * 1000
+
+# remember processed files (for runtime)
+PROCESSED = set()
+
+# strict filename pattern
+PATTERN = re.compile(r"^snapshot_video\d+_(\d+)\.jpg$")
+
+model = YOLO("yolov8m.pt")
+#model = YOLO("yolov8s-pose.pt")
+
+
+def extract_timestamp(filename):
+    match = PATTERN.match(filename)
+    if match:
+        return int(match.group(1))
+    return None
+
+
+def get_recent_files():
+    now_ms = int(time.time() * 1000)
+    files = []
+
+    for f in os.listdir(SNAPSHOT_DIR):
+        ts = extract_timestamp(f)
+        if ts is None:
+            continue
+
+        # filter by age
+        if now_ms - ts > MAX_AGE_MS:
+            continue
+
+        files.append((f, ts))
+
+    # newest first
+    files.sort(key=lambda x: x[1], reverse=True)
+
+    return [f[0] for f in files]
+
+
+def is_file_stable(path):
+    """Avoid processing partially written files"""
+    try:
+        size1 = os.path.getsize(path)
+        time.sleep(0.1)
+        size2 = os.path.getsize(path)
+        return size1 == size2
+    except:
+        return False
+
+
+def detect_and_save(file):
+    full_path = os.path.join(SNAPSHOT_DIR, file)
+
+    if not is_file_stable(full_path):
+        return
+
+    results = model(full_path, imgsz=960)
+
+    detections = []
+
+    for r in results:
+        boxes = r.boxes
+        if boxes is None:
+            continue
+
+        for b in boxes:
+            detections.append({
+                "class": model.names[int(b.cls)],
+                "confidence": float(b.conf),
+                "bbox_xyxy": b.xyxy[0].tolist()
+            })
+        
+        # ✅ --- NEW: create annotated image ---
+        annotated_img = r.plot()  # returns image with boxes drawn
+
+        out_img_file = file.replace(".jpg", "_yolo.jpg")
+        out_img_path = os.path.join(SNAPSHOT_DIR, out_img_file)
+
+        # save using OpenCV
+        import cv2
+        cv2.imwrite(out_img_path, annotated_img)
+
+
+    out_file = file.replace(".jpg", "_yolo.json")
+    out_path = os.path.join(SNAPSHOT_DIR, out_file)
+
+    with open(out_path, "w") as f:
+        json.dump(detections, f, indent=2)
+
+    print(f"Processed {file} → {out_file}")
+
+
+while True:
+    try:
+        files = get_recent_files()
+
+        for file in files[:10]:  # limit per loop
+            if file in PROCESSED:
+                continue
+
+            json_file = file.replace(".jpg", "_yolo.json")
+            json_path = os.path.join(SNAPSHOT_DIR, json_file)
+
+            # skip if already processed (persistent)
+            if os.path.exists(json_path):
+                PROCESSED.add(file)
+                continue
+
+            detect_and_save(file)
+            PROCESSED.add(file)
+
+        # cleanup memory
+        if len(PROCESSED) > 1000:
+            PROCESSED.clear()
+
+    except Exception as e:
+        print("Error:", e)
+
+    time.sleep(1)