def testPyrosROSCtx():
# start ros system , before PyrosROS process and Client otherwise Client assumes there is problem ( discovery timeout )
# we might need to load pyros_setup here...
try:
import pyros_utils
except ImportError:
# TODO : find a proper way to log from a test like here...
try:
#_logger.warning("loading pyros_setup and configuring your ROS environment")
import pyros_setup
# This will load the pyros_setup configuration from the environment
pyros_setup.configurable_import().configure().activate()
import pyros_utils
except ImportError:
# This is expected when testing pyros by itself
raise nose.SkipTest(
"pyros_utils could not be imported, and trying to import pyros_setup for dynamic ros setup failed."
)
master, roscore_proc = pyros_utils.get_master(
spawn=True) # we start the master if needed
assert master.is_online()
with pyros_ctx(node_impl=pyros_interfaces_ros.PyrosROS) as ctx:
assert isinstance(ctx.client, PyrosClient)
# TODO : assert the context manager does his job ( HOW ? )
if roscore_proc is not None:
roscore_proc.terminate()
while roscore_proc.is_alive():
time.sleep(0.2) # waiting for roscore to die
def testPyrosROSCtx():
# start ros system , before PyrosROS process and Client otherwise Client assumes there is problem ( discovery timeout )
# we might need to load pyros_setup here...
try:
import pyros_utils
except ImportError:
# TODO : find a proper way to log from a test like here...
try :
#_logger.warning("loading pyros_setup and configuring your ROS environment")
import pyros_setup
# This will load the pyros_setup configuration from the environment
pyros_setup.configurable_import().configure().activate()
import pyros_utils
except ImportError:
# This is expected when testing pyros by itself
raise nose.SkipTest("pyros_utils could not be imported, and trying to import pyros_setup for dynamic ros setup failed.")
master, roscore_proc = pyros_utils.get_master(spawn=True) # we start the master if needed
assert master.is_online()
with pyros_ctx(node_impl=pyros_interfaces_ros.PyrosROS) as ctx:
assert isinstance(ctx.client, PyrosClient)
# TODO : assert the context manager does his job ( HOW ? )
if roscore_proc is not None:
roscore_proc.terminate()
while roscore_proc.is_alive():
time.sleep(0.2) # waiting for roscore to die
def test_rosnode_started():
master, roscore = pyros_utils.get_master()
try:
assert master.is_online()
# hack needed to wait for master until fix for https://github.com/ros/ros_comm/pull/711 is released
roslaunch.rlutil.get_or_generate_uuid(None, True)
launch = roslaunch.scriptapi.ROSLaunch()
launch.start()
assert launch.started
echo_node = roslaunch.core.Node('pyros_test', 'echo.py', name='echo')
echo_process = launch.launch(echo_node)
assert echo_process.is_alive()
node_api = None
with timeout(5) as t:
while not t.timed_out and node_api is None:
node_api = rosnode.get_api_uri(
master, 'echo'
) # would be good to find a way to do this without rosnode dependency
assert node_api is not None
launch.stop()
finally: # to make sure we always shutdown everything, even if test fails
if roscore is not None:
roscore.terminate()
rospy.signal_shutdown('test_rosnode_started done')
while roscore and roscore.is_alive():
time.sleep(0.2) # waiting for roscore to die
assert not (roscore and master.is_online())
def test_roscore_started():
master, roscore = pyros_utils.get_master()
try:
assert master.is_online()
finally: # to make sure we always shutdown everything, even if test fails
if roscore is not None:
roscore.terminate()
rospy.signal_shutdown('test_roscore_started done')
while roscore and roscore.is_alive():
time.sleep(0.2) # waiting for roscore to die
assert not (roscore and master.is_online())
def test_roslaunch_started():
master, roscore = pyros_utils.get_master()
try:
assert master.is_online()
# hack needed to wait for master until fix for https://github.com/ros/ros_comm/pull/711 is released
roslaunch.rlutil.get_or_generate_uuid(None, True)
launch = roslaunch.scriptapi.ROSLaunch()
launch.start()
assert launch.started
launch.stop()
finally: # to make sure we always shutdown everything, even if test fails
if roscore is not None:
roscore.terminate()
rospy.signal_shutdown('test_roslaunch_started done')
while roscore and roscore.is_alive():
time.sleep(0.2) # waiting for roscore to die
assert not (roscore and master.is_online())
def __init__(self,
node_name,
publishers=None,
subscribers=None,
services=None,
params=None,
enable_cache=False,
argv=None):
# This runs in a child process (managed by PyrosROS) and as a normal ros node)
# First thing to do : find the rosmaster...
# master has to be running here or we just wait for ever
# TODO : improve...
m, _ = pyros_utils.get_master(spawn=False)
while not m.is_online():
_logger.warning("ROSMASTER not found !!!...")
time.sleep(0.5)
# Second thing to do : initialize the ros node and disable signals to avoid overriding callers behavior
rospy.init_node(node_name, argv=argv, disable_signals=True)
rospy.loginfo(
'RosInterface {name} node started with args : {argv}'.format(
name=node_name, argv=argv))
# note on dynamic update (config reload, etc.) this is reinitialized
# rospy.init_node supports being reinitialized with the exact same arguments
self.enable_cache = enable_cache
if self.enable_cache and rocon_python_comms is None:
rospy.logerr(
"Connection Cache enabled for RosInterface, but rocon_python_comms not found. Disabling."
)
self.enable_cache = False
# only one queue to avoid sync issues
self.cb_ss = Queue.Queue()
# we add params from ROS environment, if we get something there (bwcompat behavior)
services = services or []
publishers = publishers or []
subscribers = subscribers or []
params = params or []
services += list(
set(ast.literal_eval(rospy.get_param('~services', "[]"))))
# bwcompat
publishers += list(
set(ast.literal_eval(rospy.get_param('~topics', "[]"))))
subscribers += list(
set(ast.literal_eval(rospy.get_param('~topics', "[]"))))
# new
publishers += list(
set(ast.literal_eval(rospy.get_param('~publishers', "[]"))))
subscribers += list(
set(ast.literal_eval(rospy.get_param('~subscribers', "[]"))))
params += list(set(ast.literal_eval(rospy.get_param('~params', "[]"))))
enable_cache = rospy.get_param('~enable_cache', enable_cache)
if enable_cache is not None:
self.enable_cache = enable_cache
# Note : None means no change ( different from [] )
rospy.loginfo("""[{name}] ROS Interface initialized with:
- services : {services}
- publishers : {publishers}
- subscribers : {subscribers}
- params : {params}
- enable_cache : {enable_cache}
""".format(name=__name__,
publishers="\n" + "- ".rjust(10) +
"\n\t- ".join(publishers) if publishers else [],
subscribers="\n" + "- ".rjust(10) +
"\n\t- ".join(subscribers) if subscribers else [],
services="\n" + "- ".rjust(10) +
"\n\t- ".join(services) if services else [],
params="\n" + "- ".rjust(10) +
"\n\t- ".join(params) if params else [],
enable_cache=enable_cache))
# This base constructor assumes the system to interface with is already available ( can do a get_svc_available() )
params_pool = RosParamIfPool(params)
services_pool = RosServiceIfPool(services)
subscribers_pool = RosSubscriberIfPool(subscribers)
publishers_pool = RosPublisherIfPool(publishers)
super(RosInterface, self).__init__(publishers_pool, subscribers_pool,
services_pool, params_pool)
# connecting to the master via proxy object
self._master = rospy.get_master()
#: If enabled, connection cache proxy will be setup in update() to allow dynamic update via config.
# TODO : double check : maybe useless now since we completely reinit the interface for dynamic update...
self.connection_cache = None
# Setup our debug log
# We need this debug log since rospy.logdebug does NOT store debug messages in the log.
# But it should Ref : http://wiki.ros.org/rospy/Overview/Logging#Reading_log_messages
# TODO : find why.
ros_home = os.environ.get(
'ROS_HOME', os.path.join(os.path.expanduser("~"), '.ros'))
self._debug_logger = logging.getLogger('pyros.ros_interface')
if not os.path.exists(os.path.join(ros_home, 'logdebug')):
os.makedirs(os.path.join(ros_home, 'logdebug'))
logfilename = os.path.join(
ros_home, 'logdebug',
rospy.get_name()[1:].replace(os.path.sep, "-") +
'_pyros_rosinterface.log')
file_handler = logging.handlers.RotatingFileHandler(logfilename,
maxBytes=100 *
131072,
backupCount=10)
self._debug_logger.propagate = False # to avoid propagating to root logger
self._debug_logger.setLevel(logging.DEBUG)
self._debug_logger.addHandler(file_handler)