How to Build and Test a ROS 2 Action Client in Python

ROS 2, Tutorials

In the previous article, we implemented a ROS 2 Action Server in Python that calculates the Fibonacci sequence. We also saw how to interact with it using the ROS 2 CLI (ros2 action send_goal).

Now, we’ll take a step further and implement a ROS 2 Action Client in Python.
This client node will connect directly to our Fibonacci Action Server, send a goal request, handle feedback messages in real time, and finally display the result when the computation is complete.

If you are following along in C++, you can skip ahead to the next article, where we will replicate the same action client in C++.

📖 A Quick Recap: What are ROS 2 Actions?

As we’ve already seen, ROS 2 Actions are designed for long-running tasks where:

A goal is sent by the client (e.g., “compute Fibonacci sequence up to order 10”).
The server works on that request, sending periodic feedback (e.g., partial sequence).
When complete, the server returns a result message.

This makes actions perfect for robotics tasks like navigation, arm manipulation, or—in our simplified case—computing the Fibonacci sequence.

Hands-On: Creating the Action Client Node

We’ll place our client inside the same package as before, arduinobot_py_examples, under a new script:

				
					arduinobot_py_examples/arduinobot_py_examples/simple_action_client.py

Here’s the complete code for the client:

				
					import rclpy
from rclpy.node import Node
from rclpy.action import ActionClient
from arduinobot_msgs.action import Fibonacci


class SimpleActionClient(Node):

    def __init__(self):
        super().__init__("simple_action_client")

        # Create action client to connect to the 'fibonacci' action server
        self.action_client = ActionClient(self, Fibonacci, "fibonacci")

        # Wait until the server becomes available
        self.action_client.wait_for_server()

        # Create a goal message
        goal = Fibonacci.Goal()
        goal.order = 10

        # Send the goal asynchronously, register callbacks
        self.future = self.action_client.send_goal_async(
            goal, feedback_callback=self.feedbackCallback
        )
        self.future.add_done_callback(self.responseCallback)

    def responseCallback(self, future):
        goal_handle = future.result()
        if not goal_handle.accepted:
            self.get_logger().info("Goal rejected :(")
            return

        self.get_logger().info("Goal accepted :)")

        # Wait for the result asynchronously
        result_future = goal_handle.get_result_async()
        result_future.add_done_callback(self.resultCallback)

    def feedbackCallback(self, feedback_msg):
        self.get_logger().info(
            f"Received feedback: {feedback_msg.feedback.partial_sequence}"
        )

    def resultCallback(self, future):
        result = future.result().result
        self.get_logger().info(f"Result: {result.sequence}")
        rclpy.shutdown()


def main():
    rclpy.init()
    action_client = SimpleActionClient()
    rclpy.spin(action_client)


if __name__ == "__main__":
    main()

🔎 Let’s break down the code

Let’s go line by line and understand what’s happening.

Imports

				
					import rclpy
from rclpy.node import Node
from rclpy.action import ActionClient
from arduinobot_msgs.action import Fibonacci

rclpy: ROS 2 Python client library.
Node: Base class for any ROS 2 node.
ActionClient: Class to create and manage clients for action servers.
Fibonacci: Our custom action definition (.action file).

Class Setup

				
					class SimpleActionClient(Node):
    def __init__(self):
        super().__init__("simple_action_client")

Defines a new node named "simple_action_client".
All ROS 2 interactions (logging, publishers, subscribers, actions) happen inside a node.

Creating the Action Client

				
					self.action_client = ActionClient(self, Fibonacci, "fibonacci")

Connects this client to an action server named "fibonacci".
Uses the Fibonacci action interface.
If the server isn’t running, this will wait until it becomes available.

Sending a Goal

				
					goal = Fibonacci.Goal()
goal.order = 10

Creates a goal message asking for Fibonacci sequence of order 10.

				
					self.future = self.action_client.send_goal_async(
    goal, feedback_callback=self.feedbackCallback
)
self.future.add_done_callback(self.responseCallback)

send_goal_async sends the request.
We also pass a feedback callback, executed whenever the server sends progress updates.
The returned future will tell us if the goal was accepted or rejected.

Handling Goal Response

				
					def responseCallback(self, future):
    goal_handle = future.result()
    if not goal_handle.accepted:
        self.get_logger().info("Goal rejected :(")
        return
    self.get_logger().info("Goal accepted :)")

The server may reject the request (e.g., invalid order).
Otherwise, we get a goal_handle, which lets us request the final result.

Feedback Handling

				
					def feedbackCallback(self, feedback_msg):
    self.get_logger().info(
        f"Received feedback: {feedback_msg.feedback.partial_sequence}"
    )

Triggered multiple times during execution.
Prints the partial Fibonacci sequence as it grows.

Receiving the Result

				
					def resultCallback(self, future):
    result = future.result().result
    self.get_logger().info(f"Result: {result.sequence}")
    rclpy.shutdown()

Once the server completes, this callback fires.
Prints the final sequence.
Shuts down the node since we’re done.

Main Function

				
					def main():
    rclpy.init()
    action_client = SimpleActionClient()
    rclpy.spin(action_client)

Initializes ROS 2.
Creates the action client node.
Spins until the client shuts down.

🛠️ Running the Client

Rebuild your workspace:

				
					cd ~/arduinobot_ws
colcon build

Source it:

				
					. install/setup.bash

Terminal 1: Run the action server (from previous article):

				
					ros2 run arduinobot_py_examples simple_action_server

Terminal 2: Run the client:

				
					ros2 run arduinobot_py_examples simple_action_client

We now have a working Python Action Client that:

Connects to the Fibonacci Action Server.
Sends a goal request.
Receives live feedback.
Prints the final result.

This is a foundation for more complex robotics tasks where you need long-running interactions between clients and servers.

Want to learn more?

Explore all the ROS 2 actions in the "Robotics and ROS 2 - Learn by Doing! Manipulators" course

Enroll Now

DISCOUNT