scripts: update RTM iterator with touch-window timestamp picker

scripts/make_parquet_v2.py

@@ -189,8 +189,12 @@ def iter_tactile_tracking():
 
 def iter_real_tactile_mnist():
     """RTM seq-320x240: parquet with 'sensor_video' list-of-struct{bytes,path}.
-    Each row in the parquet is one 'round' = one digit object touched ~256 times.
-    We keep one frame per touch video (the middle frame, near peak contact).
+    Each row is one 'round' = one digit touched ~256 times. Each touch video
+    is 60-75 frames long but the actual contact window is only ~6 frames
+    (typically near the end of the video). We pick the frame at the *middle
+    of the touch window* using `touch_start_time_rel` / `touch_end_time_rel`
+    plus the per-frame `time_stamp_rel_seq`. As a fallback when timestamps
+    are missing or stale, we pick the frame with highest central std.
     """
     import cv2
     root = f"{BASE_DATA}/markerless/RealTactileMNIST"
@@ -206,6 +210,9 @@ def iter_real_tactile_mnist():
                 label = cols.get("label", [None]*n)[i]
                 obj_id = cols.get("object_id", [None]*n)[i]
                 videos = cols["sensor_video"][i] or []
+                ts_seq = cols.get("time_stamp_rel_seq", [None]*n)[i] or []
+                t_start = cols.get("touch_start_time_rel", [None]*n)[i] or []
+                t_end = cols.get("touch_end_time_rel", [None]*n)[i] or []
                 for tj, vid_struct in enumerate(videos):
                     if not vid_struct: continue
                     vid_bytes = vid_struct.get("bytes") if isinstance(vid_struct, dict) else None
@@ -222,9 +229,32 @@ def iter_real_tactile_mnist():
                     try: os.remove(tmpf)
                     except: pass
                     if not frames: continue
-                    # keep only the middle frame (near peak contact)
-                    mid = frames[len(frames)//2]
-                    yield mid, {
+
+                    pick_idx = None
+                    # Try timestamp-based pick first
+                    try:
+                        ts = ts_seq[tj] if tj < len(ts_seq) else None
+                        ts0 = t_start[tj] if tj < len(t_start) else None
+                        ts1 = t_end[tj] if tj < len(t_end) else None
+                        if ts is not None and ts0 is not None and ts1 is not None \
+                                and len(ts) == len(frames):
+                            t_mid = (ts0 + (ts1 - ts0) / 2)
+                            diffs = [abs((t - t_mid).total_seconds()) for t in ts]
+                            pick_idx = int(np.argmin(diffs))
+                    except Exception:
+                        pick_idx = None
+
+                    # Fallback: frame with max central-region std
+                    if pick_idx is None:
+                        stds = []
+                        for fr in frames:
+                            g = np.asarray(Image.fromarray(fr).convert("L"),
+                                           dtype=np.float32)
+                            h, w = g.shape
+                            stds.append(float(g[h//4:3*h//4, w//4:3*w//4].std()))
+                        pick_idx = int(np.argmax(stds))
+
+                    yield frames[pick_idx], {
                         "source": "real_tactile_mnist",
                         "markered": False,
                         "capture": f"r{round_id}_t{tj}",
@@ -232,7 +262,7 @@ def iter_real_tactile_mnist():
                         "obj_name": f"digit_{label}",
                         "digit_class": int(label) if label is not None else None,
                         "episode": str(obj_id) if obj_id is not None else None,
-                        "frame_idx": len(frames)//2,
+                        "frame_idx": pick_idx,
                     }