Merge branch 'release/1.2.1'

.github/ISSUE_TEMPLATE/bug_report.md

@@ -24,7 +24,7 @@ If applicable, add screenshots to help explain your problem.
 
 **Desktop (please complete the following information):**
  - OS: [e.g. mac/linux/windows]
- - Version [e.g. 1.2.0]
+ - Version [e.g. 1.2.1]
 
 **Additional context**
 Add any other context about the problem here.

readthedocs.yml → .readthedocs.yaml

@@ -11,14 +11,14 @@ sphinx:
 
 # Build documentation with MkDocs
 #mkdocs:
-#  configuration: mkdocs.yml
+#  configuration: mkdocs.yml 
 
 # Optionally build your docs in additional formats such as PDF
 formats:
   - pdf
 
 # Optionally set the version of Python and requirements required to build your docs
 python:
-  version: 3.7
+  version: 3.8
   install:
     - requirements: docs/requirements.txt

CMakeLists.txt

@@ -1,5 +1,5 @@
 cmake_minimum_required(VERSION 3.10.0)
-project(Griddly VERSION 1.2.0)
+project(Griddly VERSION 1.2.1)
 
 set(BINARY ${CMAKE_PROJECT_NAME})
 

bindings/python.cpp

@@ -12,7 +12,7 @@ namespace griddly {
 
 PYBIND11_MODULE(python_griddly, m) {
   m.doc() = "Griddly python bindings";
-  m.attr("version") = "1.2.0";
+  m.attr("version") = "1.2.1";
 
 #ifndef NDEBUG
   spdlog::set_level(spdlog::level::debug);
@@ -107,6 +107,7 @@ PYBIND11_MODULE(python_griddly, m) {
   observer_type.value("BLOCK_2D", ObserverType::BLOCK_2D);
   observer_type.value("ISOMETRIC", ObserverType::ISOMETRIC);
   observer_type.value("VECTOR", ObserverType::VECTOR);
+  observer_type.value("ASCII", ObserverType::ASCII);
 
   py::class_<NumpyWrapper<uint8_t>, std::shared_ptr<NumpyWrapper<uint8_t>>>(m, "Observation", py::buffer_protocol())
       .def_buffer([](NumpyWrapper<uint8_t> &m) -> py::buffer_info {

docs/conf.py

@@ -22,7 +22,7 @@
 author = 'Chris Bamford'
 
 # The full version, including alpha/beta/rc tags
-release = '1.2.0'
+release = '1.2.1'
 
 
 # -- General configuration ---------------------------------------------------

docs/getting-started/action spaces/index.rst

@@ -94,7 +94,7 @@ In order to easily support games with large action spaces such as RTS games, sev
 
   If ``mask_type == 'full'`` then a mask of dimensions (grid_height, grid_width) is returned. This mask can be used in the case where a one-hot representation of the entire grid is used for location selection. 
 
-  If ``mask_type == 'reduced'`` then two masks are returned. One for ``grid_height`` and one for ``grid_width``. This mask can be used when two seperate one-hot representations are used for ``x`` and ``y`` selection.
+  If ``mask_type == 'reduced'`` then two masks are returned. One for ``grid_height`` and one for ``grid_width``. This mask can be used when two separate one-hot representations are used for ``x`` and ``y`` selection.
 
 .. warning:: player_id=0 is reserved for NPCs and internal actions
 

docs/getting-started/procedural content generation/index.rst

@@ -41,14 +41,15 @@ This is a basic example and generating levels for this environment might not be
 Clusters Level Generator
 ************************
 
-A much more complicated example would be to use the `Clusters<doc_clusters>` game and generate new levels. The aim of the Clusters game is for the agent to push coloured blocks together to form "clusters", whilst avoiding spikes.
+A much more complicated example would be to use the :ref:`Clusters <doc_clusters>` game and generate new levels. The aim of the Clusters game is for the agent to push coloured blocks together to form "clusters", whilst avoiding spikes.
 The game is fully deterministic and there are only 5 levels supplied in the original GDY file. This makes it a perfect candidate for building new levels and testing if Reinforcement Learning can still solve these levels!
 
 
 Level Generator Class
 =====================
 
-Here's an example of a level generator for the cluster's game. 
+Here's an example of a level generator for the cluster's game. Levels are generated with simple configurable heuristics such as maximum number of each coloured boxes and maximum numbers of spikes. 
+The boxes and spikes are randomly placed in the grid to create the initial game layout. The agent is also added to the grid in a random position.
 
 The ``LevelGenerator`` class can be used as a base class. Only the ``generate`` function needs to be implemented.
 

docs/requirements.txt

@@ -1,5 +1,5 @@
-sphinx==3.1.0
+sphinx==4.0.2
 recommonmark==0.6.0
-sphinx_rtd_theme==0.4.3
-sphinxcontrib-images==0.9.2
+sphinx_rtd_theme==0.5.2
+sphinxcontrib-images==0.9.3
 sphinx-copybutton==0.3.1

docs/rllib/rts/index.rst

@@ -26,7 +26,7 @@ Full Example
     <div class="figure align-center" id="vid1">
         <video onloadeddata="this.play();" playsinline loop muted width="50%">
 
-            <source src="/_static/video/griddly_rts.mp4"
+            <source src="../../_static/video/griddly_rts.mp4"
                     type="video/mp4">
 
             Sorry, your browser doesn't support embedded videos.

python/examples/snippet.py

@@ -8,9 +8,9 @@
 
 def make_env(name):
     wrapper = GymWrapperFactory()
-    wrapper.build_gym_from_yaml(name, 'Single-Player/Mini-Grid/minigrid-spiders.yaml',
+    wrapper.build_gym_from_yaml(name, 'Single-Player/GVGAI/spider-nest.yaml',
                                 player_observer_type=gd.ObserverType.SPRITE_2D,
-                                global_observer_type=gd.ObserverType.BLOCK_2D,
+                                global_observer_type=gd.ObserverType.ASCII,
                                 level=0,
                                 max_steps=200)
 
@@ -38,8 +38,8 @@ def make_env(name):
 
         frames += 1
         obs, reward, done, info = env.step(action)
-        #env.render()
-        env.render(observer='global')
+        env.render()
+        print(env.render(observer='global'))
 
         if frames % 1000 == 0:
             end = timer()

python/griddly/GymWrapper.py

@@ -11,7 +11,7 @@
 class GymWrapper(gym.Env):
     metadata = {'render.modes': ['human', 'rgb_array']}
 
-    def __init__(self, yaml_file=None, level=0, global_observer_type=gd.ObserverType.VECTOR,
+    def __init__(self, yaml_file=None, yaml_string=None, level=0, global_observer_type=gd.ObserverType.VECTOR,
                  player_observer_type=gd.ObserverType.VECTOR, max_steps=None, gdy_path=None, image_path=None,
                  shader_path=None,
                  gdy=None, game=None, **kwargs):
@@ -29,10 +29,14 @@ def __init__(self, yaml_file=None, level=0, global_observer_type=gd.ObserverType
         self._renderWindow = {}
 
         # If we are loading a yaml file
-        if yaml_file is not None:
+        if yaml_file is not None or yaml_string is not None:
             self._is_clone = False
             loader = GriddlyLoader(gdy_path, image_path, shader_path)
-            self.gdy = loader.load(yaml_file)
+            if yaml_file is not None:
+                self.gdy = loader.load(yaml_file)
+            else:
+                self.gdy = loader.load_string(yaml_string)
+
             self.game = self.gdy.create_game(global_observer_type)
 
             if max_steps is not None:
@@ -204,10 +208,29 @@ def render(self, mode='human', observer=0):
             observation = np.array(self.game.observe(), copy=False)
             if self._global_observer_type == gd.ObserverType.VECTOR:
                 observation = self._vector2rgb.convert(observation)
+            if self._global_observer_type == gd.ObserverType.ASCII:
+                observation = observation \
+                    .swapaxes(2, 0) \
+                    .reshape(-1, observation.shape[0] * observation.shape[1]) \
+                    .view('c')
+                ascii_string = ''.join(np.column_stack(
+                    (observation, np.repeat(['\n'], observation.shape[0]))
+                ).flatten().tolist())
+                return ascii_string
+
         else:
             observation = self._player_last_observation[observer]
             if self._player_observer_type[observer] == gd.ObserverType.VECTOR:
                 observation = self._vector2rgb.convert(observation)
+            if self._player_observer_type[observer] == gd.ObserverType.ASCII:
+                observation = observation \
+                    .swapaxes(2, 0) \
+                    .reshape(-1, observation.shape[0] * observation.shape[1]) \
+                    .view('c')
+                ascii_string = ''.join(np.column_stack(
+                    (observation, np.repeat(['\n'], observation.shape[0]))
+                ).flatten().tolist())
+                return ascii_string
 
         if mode == 'human':
             if self._renderWindow.get(observer) is None:
@@ -314,3 +337,17 @@ def build_gym_from_yaml(self, environment_name, yaml_file, global_observer_type=
                 'player_observer_type': player_observer_type
             }
         )
+
+    def build_gym_from_yaml_string(self, environment_name, yaml_string, global_observer_type=gd.ObserverType.SPRITE_2D,
+                                   player_observer_type=gd.ObserverType.SPRITE_2D, level=None, max_steps=None):
+        register(
+            id=f'GDY-{environment_name}-v0',
+            entry_point='griddly:GymWrapper',
+            kwargs={
+                'yaml_string': yaml_string,
+                'level': level,
+                'max_steps': max_steps,
+                'global_observer_type': global_observer_type,
+                'player_observer_type': player_observer_type
+            }
+        )

python/griddly/util/environment_generator_generator.py

@@ -0,0 +1,139 @@
+import os
+
+import gym
+import numpy as np
+import yaml
+
+from griddly import GymWrapper, gd, GymWrapperFactory
+from griddly.RenderTools import VideoRecorder
+from griddly.util.wrappers import ValidActionSpaceWrapper
+
+
+class EnvironmentGeneratorGenerator():
+
+    def __init__(self, gdy_path=None, yaml_file=None):
+        module_path = os.path.dirname(os.path.realpath(__file__))
+        self._gdy_path = os.path.realpath(
+            os.path.join(module_path, '../', 'resources', 'games')) if gdy_path is None else gdy_path
+        self._input_yaml_file = self._get_full_path(yaml_file)
+
+    def _get_full_path(self, gdy_path):
+        # Assume the file is relative first and if not, try to find it in the pre-defined games
+        fullpath = gdy_path if os.path.exists(gdy_path) else os.path.join(self._gdy_path, gdy_path)
+        # (for debugging only) look in parent directory resources because we might not have built the latest version
+        fullpath = fullpath if os.path.exists(fullpath) else os.path.realpath(
+            os.path.join(self._gdy_path + '../../../../resources/games', gdy_path))
+        return fullpath
+
+    def generate_env_yaml(self, level_shape):
+        level_generator_gdy = {}
+        with open(self._input_yaml_file, 'r') as fs:
+            self._gdy = yaml.load(fs, Loader=yaml.FullLoader)
+
+        objects = [o for o in self._gdy['Objects'] if 'MapCharacter' in o]
+        environment = self._gdy['Environment']
+
+        # Create the placement actions
+        actions = []
+        for obj in objects:
+            object_name = obj["Name"]
+            place_action = {
+                'InputMapping': {
+                    'Inputs': {
+                        '1': {'Description': f'Places objects of type \"{object_name}\"'}
+                    }
+                },
+                'Name': f'place_{object_name.lower()}',
+                'Behaviours': [{
+                    'Src': {
+                        'Object': '_empty',
+                        'Commands': [
+                            {'spawn': object_name}
+                        ]
+                    }
+                }]
+
+            }
+            actions.append(place_action)
+
+        level_generator_gdy['Actions'] = actions
+
+        # Copy the Objects
+        level_generator_gdy['Objects'] = [{
+            'Name': o['Name'],
+            'MapCharacter': o['MapCharacter'],
+            'Observers': o['Observers']
+        } for o in objects]
+
+        # Generate a default empty level
+        empty_level = np.empty(level_shape, dtype='str')
+        empty_level[:] = '.'
+
+        level_0_string = '\n'.join(['   '.join(list(r)) for r in empty_level])
+
+        # Create the environment template
+        level_generator_gdy['Environment'] = {
+            'Name': f'{environment["Name"]} Generator',
+            'Description': f'Level Generator environment for {environment["Name"]}',
+            'Observers': {k: v for k, v in environment['Observers'].items() if k in ['Sprite2D', 'Isometric']},
+            'Player': {
+                'Observer': {
+                    'TrackAvatar': False,
+                    'Height': level_shape[1],
+                    'Width': level_shape[0],
+                    'OffsetX': 0,
+                    'OffsetY': 0,
+                }
+            },
+            'Levels': [level_0_string],
+        }
+
+        return yaml.dump(level_generator_gdy)
+
+    def generate_env(self, size, **env_kwargs):
+        env_yaml = self.generate_env_yaml(size)
+
+        env_args = {
+            **env_kwargs,
+            'yaml_string': env_yaml,
+        }
+
+        return GymWrapper(*env_args)
+
+
+if __name__ == '__main__':
+    wrapper_factory = GymWrapperFactory()
+    yaml_file = 'Single-Player/GVGAI/sokoban.yaml'
+
+    egg = EnvironmentGeneratorGenerator(yaml_file=yaml_file)
+
+    for i in range(100):
+        generator_yaml = egg.generate_env_yaml((10, 10))
+
+        env_name = f'test_{i}'
+        wrapper_factory.build_gym_from_yaml_string(
+            env_name,
+            yaml_string=generator_yaml,
+            # TODO: Change this to ASCII observer when its ready
+            global_observer_type=gd.ObserverType.VECTOR,
+            player_observer_type=gd.ObserverType.VECTOR,
+        )
+
+        env = gym.make(f'GDY-{env_name}-v0')
+        env.reset()
+        #env = ValidActionSpaceWrapper(env)
+
+        # visualization = env.render(observer=0, mode='rgb_array')
+        # video_recorder = VideoRecorder()
+        # video_recorder.start('generator_video_test.mp4', visualization.shape)
+
+        # Place 10 Random Objects
+        for i in range(0, 100):
+            action = env.action_space.sample()
+            obs, reward, done, info = env.step(action)
+
+            #state = env.get_state()
+
+            #visual = env.render(observer=0, mode='rgb_array')
+            # video_recorder.add_frame(visual)
+

python/setup.py

@@ -71,7 +71,7 @@ def griddly_package_data(config='Debug'):
 
 setup(
     name='griddly',
-    version="1.2.0",
+    version="1.2.1",
     author_email="[email protected]",
     description="Griddly Python Libraries",
     long_description=long_description,
@@ -82,7 +82,7 @@ def griddly_package_data(config='Debug'):
     install_requires=[
         "numpy>=1.20.3",
         "gym>=0.17.3",
-        "pyyaml>-5.3.1",
+        "pyyaml>=5.3.1",
         "imageio>=2.9.0"
     ],
     cmdclass={