How to give path to carla server if build from source

[SExpert12]

CARLA version:0.9.14 Platform/OS:Ubuntu Problem you have experienced: Launch Carla server

Hi, I am reproducing results for RL algorithms from this repo -

https://github.com/yanlai00/RL-Carla/issues/new

But the problem is I am using "make launch" to start the carla server, so I don't understand how to write this line in my code.

server = subprocess.Popen(f'DISPLAY= ' + str(os.path.join(os.environ.get("home/Downloads/CARLA14"), "CarlaUE4.sh")) + f' -opengl '+ f' -carla-rpc-port={port}' + f" -quality-level=Epic ", stdout=None, stderr=subprocess.STDOUT, preexec_fn=os.setsid, env=env, shell=True)

I tried to give the path of CarlaGear server but it is not working the one which is suggested by @felipecode to collect the data.

/home/.local/lib/python3.8/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:572: DeprecationWarning: np.object is a deprecated alias for the builtin object. To silence this warning, use object by itself. Doing this will not modify any behavior and is safe. Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations (np.object, string), /home/.local/lib/python3.8/site-packages/tensorboard/compat/tensorflowstub/dtypes.py:573: DeprecationWarning: np.bool is a deprecated alias for the builtin bool. To silence this warning, use bool by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.boolhere. Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations (np.bool, bool), /home/.local/lib/python3.8/site-packages/tensorboard/util/tensor_util.py:111: DeprecationWarning:np.objectis a deprecated alias for the builtinobject. To silence this warning, useobjectby itself. Doing this will not modify any behavior and is safe. Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations np.object: SlowAppendObjectArrayToTensorProto, /home/.local/lib/python3.8/site-packages/tensorboard/util/tensor_util.py:112: DeprecationWarning:np.boolis a deprecated alias for the builtinbool. To silence this warning, useboolby itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, usenp.bool_here. Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations np.bool: SlowAppendBoolArrayToTensorProto, 2024-06-29 07:37:26.688240: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory 2024-06-29 07:37:26.688263: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine. /home/.local/lib/python3.8/site-packages/h5py/__init__.py:46: DeprecationWarning:np.typeDictis a deprecated alias fornp.sctypeDict. from ._conv import register_converters as _register_converters /home/.local/lib/python3.8/site-packages/scipy/fft/__init__.py:97: DeprecationWarning: The module numpy.dual is deprecated. Instead of using dual, use the functions directly from numpy or scipy. from numpy.dual import register_func /home/.local/lib/python3.8/site-packages/scipy/sparse/sputils.py:17: DeprecationWarning:np.typeDictis a deprecated alias fornp.sctypeDict. supported_dtypes = [np.typeDict[x] for x in supported_dtypes] /home/.local/lib/python3.8/site-packages/flatbuffers/compat.py:19: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses import imp /home/.local/lib/python3.8/site-packages/matplotlib/__init__.py:152: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead. if LooseVersion(module.__version__) < minver: /home/RL-Carla-main/Carla-RL/yanlai/lib/python3.8/site-packages/setuptools/_distutils/version.py:337: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead. other = LooseVersion(other) /home/.local/lib/python3.8/site-packages/pygame/pkgdata.py:25: DeprecationWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html from pkg_resources import resource_stream, resource_exists /home/RL-Carla-main/Carla-RL/yanlai/lib/python3.8/site-packages/pkg_resources/__init__.py:2976: DeprecationWarning: Deprecated call topkg_resources.declare_namespace('google'). Implementing implicit namespace packages (as specified in PEP 420) is preferred topkg_resources.declare_namespace. See https://setuptools.pypa.io/en/latest/references/keywords.html#keyword-namespace-packages declare_namespace(pkg) /home/RL-Carla-main/Carla-RL/yanlai/lib/python3.8/site-packages/pkg_resources/__init__.py:2976: DeprecationWarning: Deprecated call topkg_resources.declare_namespace('mpl_toolkits'). Implementing implicit namespace packages (as specified in PEP 420) is preferred topkg_resources.declare_namespace. See https://setuptools.pypa.io/en/latest/references/keywords.html#keyword-namespace-packages declare_namespace(pkg) /home/.local/lib/python3.8/site-packages/scipy/special/orthogonal.py:81: DeprecationWarning:np.intis a deprecated alias for the builtinint. To silence this warning, useintby itself. Doing this will not modify any behavior and is safe. When replacingnp.int, you may wish to use e.g.np.int64ornp.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information. Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations from numpy import (exp, inf, pi, sqrt, floor, sin, cos, around, int, Traceback (most recent call last): File "train_sac.py", line 79, in main(model_name, load_model, town, fps, im_width, im_height, repeat_action, start_transform_type, sensors, File "train_sac.py", line 15, in main env = CarlaEnv(town, fps, im_width, im_height, repeat_action, start_transform_type, sensors, File "/home/RL-Carla-main/Carla-RL/carla_env.py", line 29, in init self.client, self.world, self.frame, self.server = setup(town=town, fps=fps, client_timeout=timeout) File "/home/RL-Carla-main/Carla-RL/setup.py", line 66, in setup server = subprocess.Popen(f'DISPLAY= ' + str(os.path.join(os.environ.get("home/Downloads/CARLA14"), "CarlaUE4.sh")) + f' -opengl '+ f' -carla-rpc-port={port}' + f" -quality-level=Epic ", stdout=None, stderr=subprocess.STDOUT, preexec_fn=os.setsid, env=env, shell=True) File "/usr/lib/python3.8/posixpath.py", line 76, in join a = os.fspath(a) TypeError: expected str, bytes or os.PathLike object, not NoneType

How to solve this?

[GoodarzMehr]

If I understand correctly that line launches the server, if you are launching the server using make launch (and presumably hitting play in UE4) can't you just remove that line?

Alternatively, you could build a packaged version of CARLA using make package (it'll be built in the Dist folder) and then use that.

[SExpert12]

It is a variable taken as a argument in many file like this one file.

import carla import numpy as np import transforms3d.euler from absl import logging

logging.set_verbosity(logging.DEBUG)

def setup( town: str, fps: int = 20, server_timestop: float = 30.0, client_timeout: float = 20.0, num_max_restarts: int = 10, ): """Returns the CARLA server, client and world.

Args:
    town: The `CARLA` town identifier.
    fps: The frequency (in Hz) of the simulation.
    server_timestop: The time interval between spawing the server
    and resuming program.
    client_timeout: The time interval before stopping
    the search for the carla server.
    num_max_restarts: Number of attempts to connect to the server.

Returns:
    client: The `CARLA` client.
    world: The `CARLA` world.
    frame: The synchronous simulation time step ID.
    server: The `CARLA` server.
"""
assert town in ("Town01", "Town02", "Town03", "Town04", "Town05")

# The attempts counter.
attempts = 0

while attempts < num_max_restarts:
    logging.debug("{} out of {} attempts to setup the CARLA simulator".format(
        attempts + 1, num_max_restarts))

    # Random assignment of port.
    port = np.random.randint(2000, 3000)

    # Start CARLA server.
    env = os.environ.copy()
    env["SDL_VIDEODRIVER"] = "offscreen"
    env["SDL_HINT_CUDA_DEVICE"] = "0"
    logging.debug("Inits a CARLA server at port={}".format(port))
    server = subprocess.Popen(f'DISPLAY= ' + str(os.path.join(os.environ.get("CARLA_ROOT"), "CarlaUE4.sh")) + f' -opengl '+ f' -carla-rpc-port={port}' + f" -quality-level=Epic ", stdout=None, stderr=subprocess.STDOUT, preexec_fn=os.setsid, env=env, shell=True)
    atexit.register(os.killpg, server.pid, signal.SIGKILL)
    time.sleep(server_timestop)

    # Connect client.
    logging.debug("Connects a CARLA client at port={}".format(port))
    try:
        client = carla.Client("localhost", port)  # pylint: disable=no-member
        client.set_timeout(client_timeout)
        client.load_world(map_name=town)
        world = client.get_world()
        world.set_weather(carla.WeatherParameters.ClearNoon)  # pylint: disable=no-member
        frame = world.apply_settings(
            carla.WorldSettings(  # pylint: disable=no-member
                synchronous_mode=True,
                fixed_delta_seconds=1.0 / fps,
            ))
        logging.debug("Server version: {}".format(client.get_server_version()))
        logging.debug("Client version: {}".format(client.get_client_version()))
        return client, world, frame, server
    except RuntimeError as msg:
        logging.debug(msg)
        attempts += 1
        logging.debug("Stopping CARLA server at port={}".format(port))
        os.killpg(server.pid, signal.SIGKILL)
        atexit.unregister(lambda: os.killpg(server.pid, signal.SIGKILL))

logging.debug(
    "Failed to connect to CARLA after {} attempts".format(num_max_restarts))
sys.exit()

So as per my understanding I cannot remove that line.

Second option may work out. But would it launch the Carla server like CarlaUE4.sh command? Do you know that?

Thanks for your help.

[GoodarzMehr]

The server object returned by Popen is a system process. If you are using make launch I think you can change your code to get it from the system process list using something like

processes = [p for p in psutil.process_iter() if 'carla' in p.name().lower()]

The second option generates a packaged version of CARLA, similar to the ones available for download. It includes a CarlaUE4.sh file that launches the server.

[SExpert12]

Okay. Thanks. Let me try and I get back to you.

[SExpert12]

Hi, I use the packaged version of carla and from .sh file I am able to open a server. So I change line like this:

server = subprocess.Popen(f'DISPLAY= ' + str(os.path.join("/home/local/carla0914/","CarlaUE4.sh")) + f' -opengl '+ f' -carla-rpc-port={port}' + f" -quality-level=Epic ", stdout=None, stderr=subprocess.STDOUT, preexec_fn=os.setsid, env=env, shell=True)

It started like this: 4.26.2-0+++UE4+Release-4.26 522 0 Disabling core dumps.

But then I got this error.

Error in atexit._run_exitfuncs: ProcessLookupError: [Errno 3] No such process

Where I am making mistake? How to resolve this now?

[GoodarzMehr]

Hi,

I'm not really sure. At least the beginning (4.26.2-0+++UE4+Release-4.26 522 0 Disabling core dumps.) is what you'd expect. This is helpful in explaining what the error means, and you may need to do some more debugging around the atexit line(s) to figure out the root cause of the problem. I suspect for whatever reason the atexit after the server line cannot find server.pid, but I may be wrong.

[SExpert12]

Okay.

Thanks