uoe-agents · LukasSchaefer · Jul 22, 2024 · Jul 1, 2024 · Jul 2, 2024 · Jul 3, 2024
diff --git a/README.md b/README.md
diff --git a/env_requirements.txt b/env_requirements.txt
@@ -1,7 +1,9 @@
-lbforaging
-rware
-git+https://github.com/semitable/multiagent-particle-envs.git
-git+https://github.com/uoe-agents/matrix-games.git
+# lbforaging
+# rware
+pettingzoo
+# git+https://github.com/uoe-agents/matrix-games.git
 pysc2==3.0.0
 s2clientprotocol==4.10.1.75800.0
 git+https://github.com/oxwhirl/smac.git
+git+https://github.com/oxwhirl/smacv2.git
+# git+https://github.com/uoe-agents/smaclite.git
diff --git a/plot_results.py b/plot_results.py
@@ -82,13 +82,14 @@ def extract_alg_name_from_config(config):
 
 
 def extract_env_name_from_config(config):
-    if "map" in config["env_args"]:
-        env_name = config["env_args"]["map"]
+    env = config["env"]
+    if "map_name" in config["env_args"]:
+        env_name = config["env_args"]["map_name"]
     elif "key" in config["env_args"]:
         env_name = config["env_args"]["key"]
     else:
         env_name = None
-    return env_name
+    return f"{env}_{env_name}"
 
 
 def load_results(path, metric):
@@ -100,7 +101,11 @@ def load_results(path, metric):
     for file in metrics_files:
         # load json
         with open(file, "r") as f:
-            metrics = json.load(f)
+            try:
+                metrics = json.load(f)
+            except json.JSONDecodeError:
+                warnings.warn(f"Could not load metrics from {file} --> skipping")
+                continue
 
         # find corresponding config file
         config_file = file.parent / "config.json"

diff --git a/requirements.txt b/requirements.txt
@@ -1,7 +1,7 @@
 absl-py
 attrs
 click
-gym==0.21.0
+gymnasium
 idna==2.7
 imageio
 matplotlib
@@ -16,11 +16,12 @@ protobuf
 py==1.6.0
 pygame
 pyparsing==2.2.2
-pytest==4.3.1
+pytest
 python-dateutil==2.7.3
 PyYAML==5.3.1
 requests==2.19.1
 sacred
+seaborn
 scipy
 six==1.11.0
 sk-video==1.1.10

diff --git a/src/config/envs/sc2v2.yaml b/src/config/envs/sc2v2.yaml
@@ -0,0 +1,13 @@
+env: "sc2v2"
+
+env_args:
+  map_name: null
+  time_limit: 150
+
+test_greedy: True
+test_nepisode: 100
+test_interval: 50000
+log_interval: 50000
+runner_log_interval: 10000
+learner_log_interval: 10000
+t_max: 2050000
diff --git a/src/config/envs/smaclite.yaml b/src/config/envs/smaclite.yaml
@@ -0,0 +1,13 @@
+env: "smaclite"
+
+env_args:
+  map_name: null
+  time_limit: 150
+
+test_greedy: True
+test_nepisode: 100
+test_interval: 50000
+log_interval: 50000
+runner_log_interval: 10000
+learner_log_interval: 10000
+t_max: 2050000
diff --git a/src/config/envs/smacv2_configs/protoss_10_vs_10.yaml b/src/config/envs/smacv2_configs/protoss_10_vs_10.yaml
@@ -0,0 +1,67 @@
+env: sc2wrapped
+
+env_args:
+  continuing_episode: False
+  difficulty: "7"
+  game_version: null
+  map_name: "10gen_protoss"
+  move_amount: 2
+  obs_all_health: True
+  obs_instead_of_state: False
+  obs_last_action: False
+  obs_own_health: True
+  obs_pathing_grid: False
+  obs_terrain_height: False
+  obs_timestep_number: False
+  reward_death_value: 10
+  reward_defeat: 0
+  reward_negative_scale: 0.5
+  reward_only_positive: True
+  reward_scale: True
+  reward_scale_rate: 20
+  reward_sparse: False
+  reward_win: 200
+  replay_dir: ""
+  replay_prefix: ""
+  conic_fov: False
+  use_unit_ranges: True
+  min_attack_range: 2
+  obs_own_pos: True
+  num_fov_actions: 12
+  capability_config:
+    n_units: 10
+    n_enemies: 10
+    team_gen:
+      dist_type: "weighted_teams"
+      unit_types: 
+        - "stalker"
+        - "zealot"
+        - "colossus"
+      weights:
+        - 0.45
+        - 0.45
+        - 0.1
+      observe: True
+    start_positions:
+      dist_type: "surrounded_and_reflect"
+      p: 0.5
+      map_x: 32
+      map_y: 32
+
+    # enemy_mask:
+    #   dist_type: "mask"
+    #   mask_probability: 0.5
+    #   n_enemies: 5
+  state_last_action: True
+  state_timestep_number: False
+  step_mul: 8
+  heuristic_ai: False
+  # heuristic_rest: False
+  debug: False
+
+test_nepisode: 32
+test_interval: 10000
+log_interval: 2000
+runner_log_interval: 2000
+learner_log_interval: 2000
+t_max: 10050000
diff --git a/src/config/envs/smacv2_configs/protoss_10_vs_11.yaml b/src/config/envs/smacv2_configs/protoss_10_vs_11.yaml
@@ -0,0 +1,67 @@
+env: sc2wrapped
+
+env_args:
+  continuing_episode: False
+  difficulty: "7"
+  game_version: null
+  map_name: "10gen_protoss"
+  move_amount: 2
+  obs_all_health: True
+  obs_instead_of_state: False
+  obs_last_action: False
+  obs_own_health: True
+  obs_pathing_grid: False
+  obs_terrain_height: False
+  obs_timestep_number: False
+  reward_death_value: 10
+  reward_defeat: 0
+  reward_negative_scale: 0.5
+  reward_only_positive: True
+  reward_scale: True
+  reward_scale_rate: 20
+  reward_sparse: False
+  reward_win: 200
+  replay_dir: ""
+  replay_prefix: ""
+  conic_fov: False
+  use_unit_ranges: True
+  min_attack_range: 2
+  obs_own_pos: True
+  num_fov_actions: 12
+  capability_config:
+    n_units: 10
+    n_enemies: 11
+    team_gen:
+      dist_type: "weighted_teams"
+      unit_types: 
+        - "stalker"
+        - "zealot"
+        - "colossus"
+      weights:
+        - 0.45
+        - 0.45
+        - 0.1
+      observe: True
+    start_positions:
+      dist_type: "surrounded_and_reflect"
+      p: 0.5
+      map_x: 32
+      map_y: 32
+
+    # enemy_mask:
+    #   dist_type: "mask"
+    #   mask_probability: 0.5
+    #   n_enemies: 5
+  state_last_action: True
+  state_timestep_number: False
+  step_mul: 8
+  heuristic_ai: False
+  # heuristic_rest: False
+  debug: False
+
+test_nepisode: 32
+test_interval: 10000
+log_interval: 2000
+runner_log_interval: 2000
+learner_log_interval: 2000
+t_max: 10050000
diff --git a/src/config/envs/smacv2_configs/protoss_20_vs_20.yaml b/src/config/envs/smacv2_configs/protoss_20_vs_20.yaml
@@ -0,0 +1,67 @@
+env: sc2wrapped
+
+env_args:
+  continuing_episode: False
+  difficulty: "7"
+  game_version: null
+  map_name: "10gen_protoss"
+  move_amount: 2
+  obs_all_health: True
+  obs_instead_of_state: False
+  obs_last_action: False
+  obs_own_health: True
+  obs_pathing_grid: False
+  obs_terrain_height: False
+  obs_timestep_number: False
+  reward_death_value: 10
+  reward_defeat: 0
+  reward_negative_scale: 0.5
+  reward_only_positive: True
+  reward_scale: True
+  reward_scale_rate: 20
+  reward_sparse: False
+  reward_win: 200
+  replay_dir: ""
+  replay_prefix: ""
+  conic_fov: False
+  use_unit_ranges: True
+  min_attack_range: 2
+  obs_own_pos: True
+  num_fov_actions: 12
+  capability_config:
+    n_units: 20
+    n_enemies: 20
+    team_gen:
+      dist_type: "weighted_teams"
+      unit_types: 
+        - "stalker"
+        - "zealot"
+        - "colossus"
+      weights:
+        - 0.45
+        - 0.45
+        - 0.1
+      observe: True
+    start_positions:
+      dist_type: "surrounded_and_reflect"
+      p: 0.5
+      map_x: 32
+      map_y: 32
+
+    # enemy_mask:
+    #   dist_type: "mask"
+    #   mask_probability: 0.5
+    #   n_enemies: 5
+  state_last_action: True
+  state_timestep_number: False
+  step_mul: 8
+  heuristic_ai: False
+  # heuristic_rest: False
+  debug: False
+
+test_nepisode: 32
+test_interval: 10000
+log_interval: 2000
+runner_log_interval: 2000
+learner_log_interval: 2000
+t_max: 10050000
diff --git a/src/config/envs/smacv2_configs/protoss_20_vs_23.yaml b/src/config/envs/smacv2_configs/protoss_20_vs_23.yaml
@@ -0,0 +1,67 @@
+env: sc2wrapped
+
+env_args:
+  continuing_episode: False
+  difficulty: "7"
+  game_version: null
+  map_name: "10gen_protoss"
+  move_amount: 2
+  obs_all_health: True
+  obs_instead_of_state: False
+  obs_last_action: False
+  obs_own_health: True
+  obs_pathing_grid: False
+  obs_terrain_height: False
+  obs_timestep_number: False
+  reward_death_value: 10
+  reward_defeat: 0
+  reward_negative_scale: 0.5
+  reward_only_positive: True
+  reward_scale: True
+  reward_scale_rate: 20
+  reward_sparse: False
+  reward_win: 200
+  replay_dir: ""
+  replay_prefix: ""
+  conic_fov: False
+  use_unit_ranges: True
+  min_attack_range: 2
+  obs_own_pos: True
+  num_fov_actions: 12
+  capability_config:
+    n_units: 20
+    n_enemies: 23
+    team_gen:
+      dist_type: "weighted_teams"
+      unit_types: 
+        - "stalker"
+        - "zealot"
+        - "colossus"
+      weights:
+        - 0.45
+        - 0.45
+        - 0.1
+      observe: True
+    start_positions:
+      dist_type: "surrounded_and_reflect"
+      p: 0.5
+      map_x: 32
+      map_y: 32
+
+    # enemy_mask:
+    #   dist_type: "mask"
+    #   mask_probability: 0.5
+    #   n_enemies: 5
+  state_last_action: True
+  state_timestep_number: False
+  step_mul: 8
+  heuristic_ai: False
+  # heuristic_rest: False
+  debug: False
+
+test_nepisode: 32
+test_interval: 10000
+log_interval: 2000
+runner_log_interval: 2000
+learner_log_interval: 2000
+t_max: 10050000