Webots sim-to-real fixes, DAgger pipeline, 360° proto variant

Today's session worked across the full Webots delivery stack — found and
fixed a cluster of bugs blocking the BC/RL transfer, then explored
training-side mitigations for the residual perception gap.

Bug fixes:
- Makefile FP_RATE default 2.0 → 0.0: BC demos used fp_rate=0 but RL
  fine-tune defaulted to fp_rate=2, poisoning the BC obs distribution
  and stalling PPO at 0% success across 1.46M+ steps.
- controllers/{shepherd_dog,sheep}/runtime.ini: Webots was launching
  controllers under system python3 (no numpy) and they were crashing
  silently. Pinned to the conda tir env.
- herding/config.py HERDING_WEBOTS preset: pen_latch_depth 0.5 → 2.0,
  max_new_tracks_per_step 3 → 1, static_reject 0.8 → 1.2. Stops phantom
  FPs near the gate from latching as permanently-penned tracks.
- herding/perception/sheep_tracker.py: penned tracks now decay at
  forget_steps × 8 instead of living forever. Adds get_positions
  min_freshness filter for deploy-time use.

Training/eval matches deployment:
- training/bc/collect.py: --dagger-policy flag for DAgger rollouts
  (policy drives, teacher labels) + --use-webots-preset for matched
  140° tracker + DR config.
- controllers/shepherd_dog/shepherd_dog.py: scan-fallback (0, 0.6) when
  BC/RL sees empty sheep_positions — recovers from FOV gaps.

Tooling:
- tools/dagger_round.sh: one-shot DAgger round (collect + concat + bc).
- tools/webots_sweep_gt.sh: full sweep with HERDING_USE_GT=1 for the
  perception-gap diagnosis matrix.
- protos/ShepherdDog360.proto: 360° FOV variant for the FOV-ablation
  comparison. Canonical proto stays at 140° per project spec.

Artifacts: v1 BC/RL policies for all 4 (drive × world) combos trained
in clean gym (success: diff/field 90-100%, diff/round 58%, mec/field
60-100%, mec/round 50-100%). DAgger r1/r2 BCs for diff/field show
12%→38% progression on gym HERDING_WEBOTS proxy but did not close
to actual Webots LiDAR (0/5 throughout). Next: LSTM policy or
learned tracker per the project-state memory.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

herding/world/geometry.py

@@ -72,6 +72,36 @@ if FIELD_SHAPE == "field_round":
     GATE_Y = FIELD_ROUND_GATE_Y
 
 
+def configure_from_args(argv: list[str] | None = None) -> str:
+    """Parse ``--world`` from *argv* (or ``sys.argv[1:]``), call :func:`configure`,
+    and set ``HERDING_WORLD`` in the environment.
+
+    Returns the resolved world name (``"field"`` or ``"field_round"``).
+
+    Call this at the very top of any script that accepts a ``--world`` flag,
+    *before* importing anything from ``herding.*`` that depends on field
+    geometry.  This centralises the pre-parse logic that was previously
+    duplicated in ``bc/collect.py``, ``rl/train.py``, and ``eval.py``::
+
+        from herding.world.geometry import configure_from_args
+        configure_from_args()   # reads sys.argv
+    """
+    import os
+    import sys as _sys
+    args = argv if argv is not None else _sys.argv[1:]
+    world = "field"
+    for i, a in enumerate(args):
+        if a == "--world" and i + 1 < len(args):
+            world = args[i + 1]
+            break
+        if a.startswith("--world="):
+            world = a.split("=", 1)[1]
+            break
+    configure(world)
+    os.environ["HERDING_WORLD"] = world
+    return world
+
+
 def configure(shape: str) -> None:
     """Switch the active field geometry at runtime.