Unity 机器学习代理工具包 (ML-Agents) 是一个开源项目,它使游戏和模拟能够作为训练智能代理的环境。
您最多选择25个主题 主题必须以中文或者字母或数字开头,可以包含连字符 (-),并且长度不得超过35个字符
 
 
 
 
 

235 行
8.1 KiB

import os
import shutil
import subprocess
import yaml
from sys import platform
from typing import List, Optional
def get_unity_executable_path():
if platform == "darwin":
downloader_install_path = "./.Editor/Unity.app/Contents/MacOS/Unity"
else: # if platform == "linux":
downloader_install_path = "./.Editor/Unity"
if os.path.exists(downloader_install_path):
return downloader_install_path
raise FileNotFoundError("Can't find executable from unity-downloader-cli")
def get_base_path():
# We might need to do some more work here if the working directory ever changes
# E.g. take the full path and back out the main module main.
# But for now, this should work
return os.getcwd()
def get_base_output_path():
""""
Returns the artifact folder to use for yamato jobs.
"""
return os.path.join(get_base_path(), "artifacts")
def run_standalone_build(
base_path: str,
verbose: bool = False,
output_path: str = None,
scene_path: str = None,
build_target: str = None,
log_output_path: str = f"{get_base_output_path()}/standalone_build.txt",
) -> int:
"""
Run BuildStandalonePlayerOSX test to produce a player. The location defaults to
artifacts/standalone_build/testPlayer.
"""
unity_exe = get_unity_executable_path()
print(f"Running BuildStandalonePlayer via {unity_exe}")
# enum values from https://docs.unity3d.com/2019.4/Documentation/ScriptReference/BuildTarget.html
build_target_to_enum = {
"mac": "StandaloneOSX",
"osx": "StandaloneOSX",
"linux": "StandaloneLinux64",
}
test_args = [
unity_exe,
"-projectPath",
f"{base_path}/Project",
"-batchmode",
"-executeMethod",
"Unity.MLAgents.StandaloneBuildTest.BuildStandalonePlayerOSX",
]
os.makedirs(os.path.dirname(log_output_path), exist_ok=True)
subprocess.run(["touch", log_output_path])
test_args += ["-logfile", log_output_path]
if output_path is not None:
output_path = os.path.join(get_base_output_path(), output_path)
test_args += ["--mlagents-build-output-path", output_path]
os.makedirs(os.path.dirname(output_path), exist_ok=True)
if scene_path is not None:
test_args += ["--mlagents-build-scene-path", scene_path]
if build_target is not None:
test_args += ["--mlagents-build-target", build_target_to_enum[build_target]]
print(f"{' '.join(test_args)} ...")
timeout = 30 * 60 # 30 minutes, just in case
res: subprocess.CompletedProcess = subprocess.run(test_args, timeout=timeout)
# Copy the default build name into the artifacts folder.
if output_path is None and res.returncode == 0:
exe_name = "testPlayer.app" if platform == "darwin" else "testPlayer"
shutil.move(
os.path.join(base_path, "Project", exe_name),
os.path.join(get_base_output_path(), exe_name),
)
# Print if we fail or want verbosity.
if verbose or res.returncode != 0:
subprocess.run(["cat", log_output_path])
return res.returncode
def find_executables(root_dir: str) -> List[str]:
"""
Try to find the player executable. This seems to vary between Unity versions.
"""
ignored_extension = frozenset([".dll", ".dylib", ".bundle"])
ignored_files = frozenset(["macblas"])
exes = []
for root, _, files in os.walk(root_dir):
for filename in files:
file_root, ext = os.path.splitext(filename)
if ext in ignored_extension or filename in ignored_files:
continue
file_path = os.path.join(root, filename)
if os.access(file_path, os.X_OK):
exes.append(file_path)
return exes
def init_venv(
mlagents_python_version: str = None, extra_packages: Optional[List[str]] = None
) -> str:
"""
Set up the virtual environment, and return the venv path.
:param mlagents_python_version: The version of mlagents python packcage to install.
If None, will do a local install, otherwise will install from pypi
:return:
"""
# Use a different venv path for different versions
venv_path = "venv"
if mlagents_python_version:
venv_path += "_" + mlagents_python_version
# Set up the venv and install mlagents
# if platform == "linux":
# subprocess.check_call(
# "sudo apt-get update && sudo apt-get install -y python3-venv", shell=True
# )
if platform == "darwin":
subprocess.check_call(f"python3 -m venv {venv_path}", shell=True)
pip_commands = ["--upgrade pip", "--upgrade setuptools"]
if mlagents_python_version:
# install from pypi
if platform != "darwin":
raise RuntimeError("Yamato can only run tensorflow on mac platforms!")
pip_commands += [
f"mlagents=={mlagents_python_version}",
f"gym-unity=={mlagents_python_version}",
# TODO build these and publish to internal pypi
"~/tensorflow_pkg/tensorflow-2.0.0-cp37-cp37m-macosx_10_14_x86_64.whl",
"tf2onnx==1.6.1",
]
else:
# Local install
pip_commands += ["-e ./ml-agents-envs", "-e ./ml-agents", "-e ./gym-unity"]
if extra_packages:
pip_commands += extra_packages
for cmd in pip_commands:
pip_index_url = "--index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple"
print(f'Running "python3 -m pip install -q {cmd} {pip_index_url}"')
subprocess.check_call(
f"python3 -m pip install -q {cmd} {pip_index_url}", shell=True
)
return venv_path
def checkout_csharp_version(csharp_version):
"""
Checks out the specific git revision (usually a tag) for the C# package and Project.
If csharp_version is None, no changes are made.
:param csharp_version:
:return:
"""
if csharp_version is None:
return
csharp_tag = f"com.unity.ml-agents_{csharp_version}"
csharp_dirs = ["com.unity.ml-agents", "com.unity.ml-agents.extensions", "Project"]
for csharp_dir in csharp_dirs:
subprocess.check_call(f"rm -rf {csharp_dir}", shell=True)
# Allow the checkout to fail, since the extensions folder isn't availabe in 1.0.0
subprocess.call(f"git checkout {csharp_tag} -- {csharp_dir}", shell=True)
def undo_git_checkout():
"""
Clean up the git working directory.
"""
subprocess.check_call("git reset HEAD .", shell=True)
subprocess.check_call("git checkout -- .", shell=True)
# Ensure the cache isn't polluted with old compiled assemblies.
subprocess.check_call("rm -rf Project/Library", shell=True)
def override_config_file(src_path, dest_path, overrides):
"""
Override settings in a trainer config file. For example,
override_config_file(src_path, dest_path, max_steps=42)
will copy the config file at src_path to dest_path, but override the max_steps field to 42 for all brains.
"""
with open(src_path) as f:
configs = yaml.safe_load(f)
behavior_configs = configs["behaviors"]
for config in behavior_configs.values():
_override_config_dict(config, overrides)
with open(dest_path, "w") as f:
yaml.dump(configs, f)
def _override_config_dict(config, overrides):
for key, val in overrides.items():
if isinstance(val, dict):
_override_config_dict(config[key], val)
else:
config[key] = val
def override_legacy_config_file(python_version, src_path, dest_path, **kwargs):
"""
Override settings in a trainer config file, using an old version of the src_path. For example,
override_config_file("0.16.0", src_path, dest_path, max_steps=42)
will sync the file at src_path from version 0.16.0, copy it to dest_path, and override the
max_steps field to 42 for all brains.
"""
# Sync the old version of the file
python_tag = f"python-packages_{python_version}"
subprocess.check_call(f"git checkout {python_tag} -- {src_path}", shell=True)
with open(src_path) as f:
configs = yaml.safe_load(f)
for config in configs.values():
config.update(**kwargs)
with open(dest_path, "w") as f:
yaml.dump(configs, f)