gr-libs 0.1.4__py3-none-any.whl → 0.1.6.post1__py3-none-any.whl

CI/README.md

@@ -0,0 +1,12 @@
+## How to build a new docker image including new trained agents:
+1. Install docker
+2. Make sure you have a dataset.zip at your repo root
+3. Make sure you have a classic token in github: https://github.com/settings/tokens . If you don't, create one with package write, read and delete permissions and copy it somewhere safe.
+4. Authenticate to ghcr with docker by running:
+```sh
+echo ghp_REST_OF_TOKEN | docker login ghcr.io -u MatanShamir1 --password-stdin
+```
+3. docker build -t ghcr.io/<your-username>/gr_test_base:latest -f CI/Dockerfile .
+(the -f Dockerfile tells docker which Dockerfile to use and the '.' tells docker what's the build context, or where the dataset.zip should live)
+4. docker push ghcr.io/<your-username>/gr_test_base:latest
+docker push ghcr.io/MatanShamir1/gr_test_base:latest

CI/docker_build_context/Dockerfile

@@ -0,0 +1,15 @@
+FROM python:3.11-slim
+
+# Set workdir
+WORKDIR /app
+
+# Install unzip
+RUN apt-get update && apt-get install -y unzip && rm -rf /var/lib/apt/lists/*
+
+# Copy and unzip the dataset
+COPY dataset.zip .
+RUN unzip dataset.zip && rm dataset.zip
+RUN mv dataset_new dataset
+
+# Just start with bash by default
+CMD [ "bash" ]

gr_libs/_version.py

@@ -0,0 +1,21 @@
+# file generated by setuptools-scm
+# don't change, don't track in version control
+
+__all__ = ["__version__", "__version_tuple__", "version", "version_tuple"]
+
+TYPE_CHECKING = False
+if TYPE_CHECKING:
+    from typing import Tuple
+    from typing import Union
+
+    VERSION_TUPLE = Tuple[Union[int, str], ...]
+else:
+    VERSION_TUPLE = object
+
+version: str
+__version__: str
+__version_tuple__: VERSION_TUPLE
+version_tuple: VERSION_TUPLE
+
+__version__ = version = '0.1.6.post1'
+__version_tuple__ = version_tuple = (0, 1, 6)

gr_libs/environment/__init__.py

@@ -0,0 +1,22 @@
+import importlib.metadata
+import warnings
+
+def is_extra_installed(package: str, extra: str) -> bool:
+    """Check if an extra was installed for a given package."""
+    try:
+        # Get metadata for the installed package
+        dist = importlib.metadata.metadata(package)
+        requires = dist.get_all("Requires-Dist", [])  # Dependencies listed in the package metadata
+        return any(extra in req for req in requires)
+    except importlib.metadata.PackageNotFoundError:
+        return False  # The package is not installed
+
+# Check if `gr_libs[minigrid]` was installed
+for env in ["minigrid", "panda", "highway", "point_maze"]:
+	if is_extra_installed("gr_libs", f"gr_envs[{env}]"):
+		try:
+			importlib.import_module(f"gr_envs.{env}_scripts.envs")
+		except ImportError:
+			raise ImportError(f"gr_envs[{env}] was not installed, but gr_libs[{env}] requires it! if you messed with gr_envs installation, you can reinstall gr_libs.")
+	else:
+		warnings.warn(f"gr_libs[{env}] was not installed, skipping {env} imports.", RuntimeWarning)

gr_libs/environment/environment.py

@@ -105,7 +105,7 @@ class MinigridProperty(EnvProperty):
         env_id = problem_name.split("-DynamicGoal-")[0] + "-DynamicGoal-" + problem_name.split("-DynamicGoal-")[1]
         result = register(
             id=env_id,
-            entry_point="gr_libss.minigrid_scripts.envs:CustomColorEnv",
+            entry_point="gr_envs.minigrid_scripts.envs:CustomColorEnv",
             kwargs={"size": 13 if 'Simple' in problem_name else 9,
                     "num_crossings": 4 if 'Simple' in problem_name else 3,
                     "goal_pos": self.str_to_goal(problem_name),
@@ -168,8 +168,6 @@ class PandaProperty(GCEnvProperty):
 
         
 class ParkingProperty(GCEnvProperty):
-    # def str_to_goal(self): # TODO not use it, goal is not a part of the env property anymore.
-    #     return self.name.split("-")[-2]
 
     def __init__(self, name):
         super().__init__(name)

gr_libs/metrics/metrics.py

@@ -5,7 +5,6 @@ import numpy as np
 
 from typing import Callable, Generator, List, Dict, Tuple, Any
 from math import log2
-from numpy.core.fromnumeric import mean
 from scipy.stats import wasserstein_distance
 from gymnasium.spaces.discrete import Discrete
 # import torch
@@ -43,7 +42,7 @@ def kl_divergence_norm_softmax(observations: List[Tuple[State, Any]], agent, act
         qp2_flatten_distribution_list: List[float] = agent.get_actions_probabilities(
             observation=(observation, agent_pos))
         distances.append(kl_divergence(qp1, qp2_flatten_distribution_list))
-    return mean(distances)
+    return np.mean(distances)
 
 
 def amplify(values, alpha=1.0):

gr_libs/ml/neural/deep_rl_learner.py

@@ -13,11 +13,6 @@ if __name__ != "__main__":
 	from gr_libs.ml.utils.format import random_subset_with_order
 from stable_baselines3 import SAC, PPO
 from stable_baselines3.common.vec_env import DummyVecEnv
-from gr_envs.custom_env_wrappers.flat_obs_wrapper import CombineAchievedGoalAndObservationWrapper
-
-# important for registration of envs! do not remove lad
-import gr_envs.maze_scripts.envs.maze
-import gr_envs.highway_env_scripts.envs.parking_env
 from gr_libs.ml.utils import device
 
 # built-in python modules
@@ -32,13 +27,15 @@ def create_vec_env(kwargs):
 	return DummyVecEnv([lambda: env])
 
 def change_goal_to_specific_desired(obs, desired):
-	try:
-		if desired!=None: obs['desired_goal'] = desired
-	except Exception as e:
-		try:
-			if all(desired!=None): obs['desired_goal'] = desired
-		except Exception as e:
-			if all([desiredy!=None for desiredish in desired for desiredy in desiredish]): obs['desired_goal'] = desired
+	if desired is not None:
+		obs['desired_goal'] = desired
+	# try:
+	# 	if desired!=None: obs['desired_goal'] = desired
+	# except Exception as e:
+	# 	try:
+	# 		if all(desired!=None): obs['desired_goal'] = desired
+	# 	except Exception as e:
+	# 		if all([desiredy!=None for desiredish in desired for desiredy in desiredish]): obs['desired_goal'] = desired
 
 
 NETWORK_SETUP = {
@@ -265,6 +262,7 @@ class DeepRLAgent():
 			assert fig_path == None, "You can't specify a vid path when you don't even save the figure."
 		else:
 			assert fig_path != None, "You need to specify a vid path when you save the figure."
+		# The try-except is a bug fix for the env not being reset properly in panda. If someone wants to check why and provide a robust solution they're welcome.
 		try:
 			obs = self.env.reset()
 			change_goal_to_specific_desired(obs, desired)