Ensemble: disable redundant heavy features in classical pipeline

The previous ensemble commit (9cbefb9) timed out because the classical
sklearn pipeline calls predict_wireframe_tracks via USE_TRACK_ENSEMBLE
and USE_TRACKS_AS_VERTICES — which is the same heavy DLT triangulation
already called by script.py's hybrid_merge. Doubling that step blows
the time budget.

Before calling predict_wireframe_sklearn, monkey-patch its module-level
flags to disable:
- USE_TRACK_ENSEMBLE (duplicate of hybrid_merge)
- USE_TRACKS_AS_VERTICES (duplicate of hybrid_merge)
- USE_WINNER_CANDIDATES (expensive iterative dilation)
- USE_LINE_EDGES (3D line clustering)

What we keep: per-view 2D semantic vertex detection, RANSAC affine
depth fit, 3D unprojection, multi-view merge, sklearn edge classifier.
This is the cheap-but-distinct signal — independent from anything the
learned pipeline produces. Flags are restored after the call.

script.py

@@ -468,12 +468,34 @@ if __name__ == "__main__":
                         print(f"  Track ensemble failed for {order_id}: {track_e_err}")
 
                     # Pipeline ensemble: also run the classical sklearn pipeline
-                    # and union with the learned-pipeline output. Independent
-                    # error modes (semantic 2D detection + line cloud + sklearn
-                    # edge classifier) complement the learned 3D segment regression.
+                    # but disable its heaviest features. predict_wireframe_tracks
+                    # is already called above by hybrid_merge — calling it again
+                    # via USE_TRACK_ENSEMBLE/USE_TRACKS_AS_VERTICES doubles the
+                    # slowest step and causes timeout. Winner candidates and line
+                    # cloud are also expensive. We keep just the core classical
+                    # signal: 2D semantic vertex detection + RANSAC depth fit +
+                    # 3D unprojection + sklearn edge classifier.
                     try:
-                        from sklearn_submission import predict_wireframe_sklearn
-                        skl_v, skl_e = predict_wireframe_sklearn(sample)
+                        import sklearn_submission as _skl
+                        _saved_flags = (
+                            _skl.USE_TRACK_ENSEMBLE,
+                            _skl.USE_TRACKS_AS_VERTICES,
+                            _skl.USE_WINNER_CANDIDATES,
+                            _skl.USE_LINE_EDGES,
+                        )
+                        _skl.USE_TRACK_ENSEMBLE = False
+                        _skl.USE_TRACKS_AS_VERTICES = False
+                        _skl.USE_WINNER_CANDIDATES = False
+                        _skl.USE_LINE_EDGES = False
+                        try:
+                            skl_v, skl_e = _skl.predict_wireframe_sklearn(sample)
+                        finally:
+                            (
+                                _skl.USE_TRACK_ENSEMBLE,
+                                _skl.USE_TRACKS_AS_VERTICES,
+                                _skl.USE_WINNER_CANDIDATES,
+                                _skl.USE_LINE_EDGES,
+                            ) = _saved_flags
                         if isinstance(pred_v, np.ndarray) and len(pred_v) >= 1 and \
                                 skl_v is not None and len(skl_v) >= 1:
                             pred_v, pred_e = ensemble_merge(