Services

To find the current list of supported services, one can do:

rosLogic = slicer.util.getModuleLogic('ROS2')
rosNode = rosLogic.GetDefaultROS2Node()
rosNode.RegisteredROS2ServiceClientNodes()

The method RegisteredROS2xxxxxNodes returns a long string with all the services client classes available.

Clients

To create a new service client, one should use the MRML ROS 2 node method vtkMRMLROS2NodeNode::CreateAndAddServiceClientNode. This method takes two parameters:

the class (type) of service to be used (full or short name). We provide clients for the most common service types (from Slicer to ROS messages).
the service name (std::string).

Service requests are sent asynchronously with SendAsyncRequest. After sending a request, use the ROS 2 logic’s WaitForServiceResponse helper before reading GetLastResponse. The helper keeps Slicer’s Qt event loop active so the module-level ROS 2 spin timer can process the response; user scripts do not need to call Spin directly.

For non-blocking workflows, observe vtkMRMLROS2ServiceClientNode.ResponseReceivedEvent on the service client node. This event is emitted when a valid response has been received and GetLastResponse is ready.

The example below calls turtlesim’s Spawn service. The request creates a new turtle, and the response contains the name assigned to that turtle.

# make sure service server is started
# for this example we use the turtle sim package
ros2 launch turtlesim multisim.launch.py

# create clients
rosLogic = slicer.util.getModuleLogic('ROS2')
rosNode = rosLogic.GetDefaultROS2Node()
# optional, shows which services are available
rosNode.RegisteredROS2ServiceClientNodes()
# example with full class name
spawn1 = rosNode.CreateAndAddServiceClientNode('vtkMRMLROS2ServiceClientSpawnNode', '/turtlesim1/spawn')
# to check if the client is started, in a shell: ros2 service info /turtlesim1/spawn

# now create a blank request, a new turtle will be created
req = spawn1.CreateBlankRequest()
req.SetX(4.0)
req.SetY(4.0)
# send the request
spawn1.SendAsyncRequest(req)
# wait for the asynchronous response
if not rosLogic.WaitForServiceResponse(spawn1, 5.0):
    raise TimeoutError(f"Timed out waiting for response from {spawn1.GetService()}")
res = spawn1.GetLastResponse()
# Spawn request returns the name given to the new turtle
print(res.GetName())

# example with short class name, Spawn will be expanded to vtkMRMLROS2ServiceClientSpawnNode
spawn2 = rosNode.CreateAndAddServiceClientNode('Spawn', '/turtlesim2/spawn')
req = spawn2.CreateBlankRequest()
req.SetX(6.0)
req.SetY(4.0)
spawn2.SendAsyncRequest(req)
if not rosLogic.WaitForServiceResponse(spawn2, 5.0):
    raise TimeoutError(f"Timed out waiting for response from {spawn2.GetService()}")
res = spawn2.GetLastResponse()
print(res.GetName())

# Optional non-blocking pattern: attach the observer before sending.
req = spawn2.CreateBlankRequest()
req.SetX(8.0)
req.SetY(4.0)

def onSpawnResponse(caller, event):
    print(caller.GetLastResponse().GetName())

observerId = spawn2.AddObserver(
    slicer.vtkMRMLROS2ServiceClientNode.ResponseReceivedEvent,
    onSpawnResponse)
spawn2.SendAsyncRequest(req)

// example with full class name
auto spawn1 = rosNode->CreateAndAddServiceClientNode("vtkMRMLROS2ServiceClientSpawnNode", "/turtlesim1/spawn");

// example with short class name, Spawn will be expanded to vtkMRMLROS2ServiceClientSpawnNode
auto spawn2 = rosNode->CreateAndAddServiceClientNode("Spawn", "/turtlesim2/spawn");

To remove the service client node, use the method vtkMRMLROS2NodeNode::RemoveAndDeleteServiceClientNode. This method takes one parameter:

the service name (std::string)

Servers

Not yet implemented.