Move grass regrowth time to PropertyLayer

Author: mgemaakbar

mesa/examples/advanced/wolf_sheep/agents.py

@@ -73,15 +73,6 @@ def move(self):
         if len(cells_without_wolves) == 0:
             return
 
-        target_cells_not_walls = cells_without_wolves.select(
-            lambda cell: not self.model.grid.wall.data[cell.coordinate[0]][
-                cell.coordinate[1]
-            ]
-        )
-
-        if len(target_cells_not_walls) == 0:
-            return
-
         # Among safe cells, prefer those with grown grass
         cells_with_grass = cells_without_wolves.select(
             lambda cell: any(
@@ -133,11 +124,15 @@ def fully_grown(self, value: bool) -> None:
         if not value:  # If grass was just eaten
             self.model.simulator.schedule_event_relative(
                 setattr,
-                self.grass_regrowth_time,
+                int(
+                    self.model.grid.grass_regrowth_time.data[self.cell.coordinate[0]][
+                        self.cell.coordinate[1]
+                    ]
+                ),
                 function_args=[self, "fully_grown", True],
             )
 
-    def __init__(self, model, countdown, grass_regrowth_time, cell):
+    def __init__(self, model, countdown, cell):
         """Create a new patch of grass.
 
         Args:
@@ -148,7 +143,6 @@ def __init__(self, model, countdown, grass_regrowth_time, cell):
         """
         super().__init__(model)
         self._fully_grown = countdown == 0
-        self.grass_regrowth_time = grass_regrowth_time
         self.cell = cell
 
         # Schedule initial growth if not fully grown

mesa/examples/advanced/wolf_sheep/model.py

@@ -10,9 +10,9 @@
 """
 
 import math
-import random
 
 import numpy as np
+from scipy.ndimage import gaussian_filter
 
 from mesa import Model
 from mesa.datacollection import DataCollector
@@ -96,38 +96,32 @@ def __init__(
 
         self.datacollector = DataCollector(model_reporters)
 
-        wall_arr = [[False] * self.width for i in range(self.height)]
+        def generate_grass_regrowth_time_array():  # Using Gaussian filter to make it look like spatial distribution
+            rows, cols = height, width
 
-        wall_coord = {
-            (random.randrange(self.height), random.randrange(self.width))
-            for i in range((width * height) // 10)
-        }  # set is used because the random number gen might return the same coordinate
-        for i, j in wall_coord:
-            wall_arr[i][j] = True
+            seeds = np.zeros((rows, cols))
+            num_seeds = grass_regrowth_time
 
-        wall_arr = np.array(wall_arr)
+            for _ in range(num_seeds):
+                x, y = np.random.randint(0, rows), np.random.randint(0, cols)
+                seeds[x, y] = np.random.randint(1, num_seeds)
 
-        self.grid.add_property_layer(PropertyLayer.from_data("wall", wall_arr))
+            # Smooth the array to create clustered patterns using SciPy's Gaussian filter
+            filtered_array = gaussian_filter(seeds, sigma=10)
 
-        def is_wall(row, col):
-            return (
-                True
-                if row < 0 or col < 0 or row >= height or col >= width  # corner case
-                else wall_arr[row][col]
-            )
+            # Normalize the array to the range [1, num_seeds]
+            filtered_array = (filtered_array - np.min(filtered_array)) / (
+                np.max(filtered_array) - np.min(filtered_array)
+            ) * (num_seeds - 1) + 1
+            filtered_array = filtered_array.astype(int)
 
-        def is_trapped_in_walls(row, col):
-            return (
-                is_wall(row + 1, col)
-                and is_wall(row - 1, col)
-                and is_wall(row, col + 1)
-                and is_wall(row, col - 1)
-            )
+            return filtered_array
 
-        possible_cells = self.grid.all_cells.select(
-            lambda cell: not wall_arr[cell.coordinate[0]][cell.coordinate[1]]
-            and not is_trapped_in_walls(cell.coordinate[0], cell.coordinate[1])
-        ).cells  # so we don't create an animal at wall cells. and make sure the animal is not trapped in walls
+        grass_regrowth_time_array = generate_grass_regrowth_time_array()
+
+        self.grid.add_property_layer(
+            PropertyLayer.from_data("grass_regrowth_time", grass_regrowth_time_array)
+        )
 
         # Create sheep:
         Sheep.create_agents(
@@ -136,7 +130,7 @@ def is_trapped_in_walls(row, col):
             energy=self.rng.random((initial_sheep,)) * 2 * sheep_gain_from_food,
             p_reproduce=sheep_reproduce,
             energy_from_food=sheep_gain_from_food,
-            cell=self.random.choices(possible_cells, k=initial_sheep),
+            cell=self.random.choices(self.grid.all_cells.cells, k=initial_sheep),
         )
         # Create Wolves:
         Wolf.create_agents(
@@ -145,7 +139,7 @@ def is_trapped_in_walls(row, col):
             energy=self.rng.random((initial_wolves,)) * 2 * wolf_gain_from_food,
             p_reproduce=wolf_reproduce,
             energy_from_food=wolf_gain_from_food,
-            cell=self.random.choices(possible_cells, k=initial_wolves),
+            cell=self.random.choices(self.grid.all_cells.cells, k=initial_wolves),
         )
 
         # Create grass patches if enabled
@@ -154,9 +148,16 @@ def is_trapped_in_walls(row, col):
             for cell in self.grid:
                 fully_grown = self.random.choice(possibly_fully_grown)
                 countdown = (
-                    0 if fully_grown else self.random.randrange(0, grass_regrowth_time)
+                    0
+                    if fully_grown
+                    else self.random.randrange(
+                        0,
+                        grass_regrowth_time_array[cell.coordinate[0]][
+                            cell.coordinate[1]
+                        ],
+                    )
                 )
-                GrassPatch(self, countdown, grass_regrowth_time, cell)
+                GrassPatch(self, countdown, cell)
 
         # Collect initial data
         self.running = True