migration: scope subfolder download, keep normalized joints as-is

- download_subfolder: fetch only the target sub-dataset subtree instead of
  enumerating the whole community_dataset_v3 monorepo tree (fixes apparent hang)
- normalized SO gripper (RANGE_0_100) left in native 0..100 frame, matching
  degrees_new datasets, instead of remapping to +/-45deg
- uncalibrated normalized datasets: skip identity value rewrite, keep normalized
  units and flag them APPROXIMATE on the dataset card
- remove --allow-uncalibrated flag and its CANON_IS_CALIBRATED side effect

community_v3_migration/so_arm_frame.py

@@ -8,8 +8,9 @@ Two calibration-free branches + one that needs an assumed canonical range:
   * degrees_new  (`*_follower` recorded with use_degrees=True, not saturated): already
                  degrees. EXACT.
   * normalized   (`*_follower`, -100..100 joints / 0..100 gripper, saturates at bounds):
-                 already mid-range-zero; only the SCALE is missing (per-robot range_min/max
-                 is not stored) -> use assumed canonical per-joint spans below. APPROXIMATE.
+                 5 arm joints are mid-range-zero, only the SCALE is missing (per-robot
+                 range_min/max not stored) -> use assumed canonical spans below. APPROXIMATE.
+                 The gripper (0..100) is kept in its native frame, matching degrees_new.
   * radians      -> untouched.
 
 Joint order per arm: shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper.
@@ -23,13 +24,12 @@ JOINT_ORDER = ["shoulder_pan", "shoulder_lift", "elbow_flex", "wrist_flex", "wri
 SIGNS = np.array([1.0, -1.0, 1.0, 1.0, 1.0, 1.0], dtype=np.float64)
 OFFSETS_DEG = np.array([0.0, 90.0, 90.0, 0.0, 0.0, 0.0], dtype=np.float64)
 
-# --- Canonical per-joint spans (DEGREES) used ONLY to invert the -100..100 / 0..100
-#     normalization when per-robot calibration is unavailable. joints 0..4 (RANGE_M100_100):
-#     normalized +/-100 -> +/-HALF_RANGE. gripper (RANGE_0_100): 0..100 -> centered at 50,
-#     span FULL_RANGE. THESE ARE PLACEHOLDERS — run calibrate_canonical_ranges.py and paste
-#     the fitted values here before a production run. ---
+# --- Canonical per-joint spans (DEGREES) used ONLY to invert the -100..100 normalization of
+#     the 5 arm joints (RANGE_M100_100) when per-robot calibration is unavailable: normalized
+#     +/-100 -> +/-HALF_RANGE. The gripper (RANGE_0_100) is left in its native 0..100 frame in
+#     every SO dataset, so it needs no canonical span. THESE ARE PLACEHOLDERS — run
+#     calibrate_canonical_ranges.py and paste the fitted values here before a production run. ---
 CANON_HALF_RANGE_DEG = np.array([100.0, 100.0, 100.0, 100.0, 100.0], dtype=np.float64)  # 5 arm joints
-CANON_GRIPPER_FULL_RANGE_DEG = 90.0
 CANON_IS_CALIBRATED = False  # flipped to True once you paste fitted values
 
 
@@ -43,9 +43,10 @@ def _convert_arm(x: np.ndarray, encoding: str) -> np.ndarray:
     if encoding == "degrees_new":
         return x
     if encoding == "normalized":
-        new_deg = np.empty_like(x)
+        new_deg = np.array(x, dtype=np.float64)
         new_deg[..., :5] = (x[..., :5] / 100.0) * CANON_HALF_RANGE_DEG
-        new_deg[..., 5] = (x[..., 5] / 100.0 - 0.5) * CANON_GRIPPER_FULL_RANGE_DEG
+        # gripper is RANGE_0_100 in every SO dataset (including use_degrees=True / degrees_new),
+        # so it is already frame-consistent and must be left untouched, not remapped to +/-deg.
         return new_deg
     raise ValueError(f"unknown encoding: {encoding!r}")
 
@@ -59,7 +60,7 @@ def to_degrees(arr, encoding: str, n_joints_per_arm: int = 6) -> np.ndarray:
     if encoding == "normalized" and not CANON_IS_CALIBRATED:
         raise RuntimeError(
             "CANON ranges are placeholders. Run calibrate_canonical_ranges.py and set "
-            "CANON_* + CANON_IS_CALIBRATED=True, or pass --allow-uncalibrated to accept them."
+            "CANON_* + CANON_IS_CALIBRATED=True before converting 'normalized' datasets to degrees."
         )
     out = np.empty_like(arr)
     for a in range(d // n_joints_per_arm):