memory engine occlusion

.gitignore

@@ -172,4 +172,6 @@ cython_debug/
 *.prof
 
 # OS files
-.DS_Store
+.DS_Store
+
+.claude/

examples/base/config.py

@@ -3,10 +3,10 @@
 import math
 
 # Visualization
-vis_engine = gamms.visual.Engine.PYGAME
+vis_engine = gamms.visual.Engine.NO_VIS
 
 # The path to the graph file
-location = "West Point, New York, USA"
+location = "La Jolla, CA, USA"
 resolution = 10.0
 graph_path = 'graph.pkl'
 

examples/occlusion_game/ucsd_game.py

@@ -0,0 +1,106 @@
+import gamms
+from gamms import osm as gamms_osm
+
+import math
+
+LOCATION     = "University of California San Diego, La Jolla, CA, USA"
+RESOLUTION   = 10.0
+SENSOR_RANGE = 80.0
+SENSOR_FOV   = math.radians(360)   # 120° cone
+
+# ---------------------------------------------------------------------------
+# Load graph
+# ---------------------------------------------------------------------------
+print("Fetching UCSD walk graph from OSM...")
+G = gamms_osm.create_osm_graph(LOCATION, gamms_osm.OSMType.WALK, resolution=RESOLUTION)
+print(f"  {G.number_of_nodes()} nodes, {G.number_of_edges()} edges")
+
+# ---------------------------------------------------------------------------
+# Create context and attach graph
+# ---------------------------------------------------------------------------
+ctx = gamms.create_context(
+    graph_engine=gamms.graph.Engine.MEMORY,
+    vis_engine=gamms.visual.Engine.PYGAME,
+    vis_kwargs={"width": 1600, "height": 900},
+)
+ctx.graph.attach_networkx_graph(G)
+
+# ---------------------------------------------------------------------------
+# Load buildings
+# ---------------------------------------------------------------------------
+print("Fetching UCSD buildings from OSM (may take ~30s)...")
+face_count = 0
+for face in gamms_osm.obstacle_from_osm(LOCATION):
+    ctx.graph.add_obstacle_face(
+        face["face_id"],
+        tr=face["tr"],
+        tl=face["tl"],
+        br=face["br"],
+        bl=face["bl"],
+        type=face["type"],
+    )
+    face_count += 1
+print(f"  {face_count} building wall faces loaded")
+
+# ---------------------------------------------------------------------------
+# Sensors + agent
+# ---------------------------------------------------------------------------
+start_node = next(ctx.graph.graph.get_nodes())
+
+ctx.sensor.create_sensor("neighbor", gamms.sensor.SensorType.NEIGHBOR)
+ctx.sensor.create_sensor(
+    "arc",
+    gamms.sensor.SensorType.ARC,
+    sensor_range=SENSOR_RANGE,
+    fov=SENSOR_FOV,
+)
+ctx.sensor.create_sensor(
+    "occ_map",
+    gamms.sensor.SensorType.OCCLUDED_MAP,
+    sensor_range=SENSOR_RANGE,
+    fov=SENSOR_FOV,
+)
+
+ctx.agent.create_agent("player", sensors=["neighbor", "arc", "occ_map"], start_node_id=start_node)
+ctx.sensor.get_sensor("arc").set_owner("player")
+ctx.sensor.get_sensor("occ_map").set_owner("player")
+
+# ---------------------------------------------------------------------------
+# Visuals
+# ---------------------------------------------------------------------------
+ctx.visual.set_graph_visual()
+ctx.visual.set_obstacle_visual()
+ctx.visual.set_agent_visual("player", color="blue", size=10)
+ctx.visual.set_sensor_visual("arc",     node_color=(255, 200, 0),   edge_color=(200, 160, 0))
+ctx.visual.set_sensor_visual("occ_map", node_color=(0,   220, 220), edge_color=(0,   180, 180))
+
+# ---------------------------------------------------------------------------
+# Game loop
+# ---------------------------------------------------------------------------
+print("\n=== Game started ===\n")
+
+turn = 0
+while not ctx.is_terminated():
+    agent = ctx.agent.get_agent("player")
+    state = agent.get_state()
+
+    arc_data  = state["sensor"]["arc"][1]
+    occ       = state["sensor"]["occ_map"][1]
+    arc_nodes = arc_data.get("nodes", {})
+    vis_nodes = occ.get("nodes", {})
+    neighbors = state["sensor"]["neighbor"][1]
+    hidden    = len(arc_nodes) - len(vis_nodes)
+
+    turn += 1
+    print(f"Turn {turn:>3} | node {agent.current_node_id:>6} "
+          f"| arc: {len(arc_nodes):>4}  occluded: {len(vis_nodes):>4}  hidden by buildings: {hidden:>4} "
+          f"| neighbors: {sorted(neighbors)}")
+
+    next_node = ctx.visual.human_input("player", state)
+    state["action"] = next_node
+    agent.set_state()
+
+    ctx.visual.simulate()
+
+ctx.terminate()
+print("Game over.")