Develop

.gitignore

@@ -3,3 +3,6 @@ __pycache__*
 *.xml
 *.iml
 *.ipynb_checkpoints/
+./blockference/.ipynb_checkpoints
+.DS_Store
+**checkpoint.py

blockference/.ipynb_checkpoints/gridference-checkpoint.py

@@ -345,7 +345,7 @@ def __init__(self, grid, planning_length: int = 2, env_state: tuple = (0, 0)) ->
 
     def get_A(self):
         """
-        State Matrix (identity matrix)
+        State Matrix (identity matrix for the single agent gridworld)
         Params:
             - n_observations: int: number of possible observations
             - n_states: int: number of possible states

blockference/agent.py

@@ -48,6 +48,33 @@ def __init__(
         lr_pD=1.0,
         use_BMA=True,
         policy_sep_prior=False,
-        save_belief_hist=False
+        save_belief_hist=False,
     ):
-        super().__init__()
+        super().__init__(A,
+                         B,
+                         C=None,
+                         D=None,
+                         E=None,
+                         pA=None,
+                         pB=None,
+                         pD=None,
+                         num_controls=None,
+                         policy_len=1,
+                         inference_horizon=1,
+                         control_fac_idx=None,
+                         policies=None,
+                         gamma=16.0,
+                         use_utility=True,
+                         use_states_info_gain=True,
+                         use_param_info_gain=False,
+                         action_selection="deterministic",
+                         inference_algo="VANILLA",
+                         inference_params=None,
+                         modalities_to_learn="all",
+                         lr_pA=1.0,
+                         factors_to_learn="all",
+                         lr_pB=1.0,
+                         lr_pD=1.0,
+                         use_BMA=True,
+                         policy_sep_prior=False,
+                         save_belief_hist=False,)

blockference/envs/grid_env.py

@@ -6,12 +6,13 @@ def __init__(self, grid_len, num_agents, grid_dim=2) -> None:
         self.grid = self.get_grid(grid_len, grid_dim)
         self.grid_dim = grid_dim
         self.no_actions = 2 * grid_dim + 1
-        self.agents = self.init_agents(num_agents)
+        self.n_observations = grid_len ** 2
+        self.n_states = grid_len ** 2
+        self.border = np.sqrt(self.n_states) - 1
+        # self.agents = self.init_agents(num_agents)
 
     def get_grid(self, grid_len, grid_dim):
         g = list(itertools.product(range(grid_len), repeat=grid_dim))
-        for i, p in enumerate(g):
-            g[i] += (0,)
         return g
 
     def move_grid(self, agent, chosen_action):
@@ -29,7 +30,7 @@ def move_grid(self, agent, chosen_action):
                 new_state[index] = state[index] - 1 if state[index] > 0 else state[index]
             elif chosen_action % 2 == 0:
                 index = chosen_action / 2
-                new_state[index] = state[index] + 1 if state[index] < agent.border else state[index]
+                new_state[index] = state[index] + 1 if state[index] < self.border else state[index]
         return new_state
 
     def init_agents(self, no_agents):