Add controller manager introspection topic system

[christophfroehlich]

This is useful to replay bags and inject state interfaces from a hardware component into the ros2_control stack.

I'm open for any name suggestions ;)

[codecov-commenter]

Codecov Report

❌ Patch coverage is 41.66667% with 28 lines in your changes missing coverage. Please review.
✅ Project coverage is 39.37%. Comparing base (007cff1) to head (036b594).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
...ware_component/src/cm_topic_hardware_component.cpp	16.66%	24 Missing and 1 partial ⚠️
...omponent/test/cm_topic_hardware_component_test.cpp	83.33%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #33      +/-   ##
==========================================
+ Coverage   38.62%   39.37%   +0.75%     
==========================================
  Files           2        4       +2     
  Lines         145      193      +48     
  Branches       34       40       +6     
==========================================
+ Hits           56       76      +20     
- Misses         80      107      +27     
- Partials        9       10       +1     

Flag	Coverage Δ
unittests	39.37% <41.66%> (+0.75%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...omponent/test/cm_topic_hardware_component_test.cpp	83.33% <83.33%> (ø)
...ware_component/src/cm_topic_hardware_component.cpp	16.66% <16.66%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[saikishor]

I think it is better to check if both the data are of same size before continuing, because I'm not sure if the variables are appended or reordered for every instance

[christophfroehlich]

I just copied this from plotjuggler
https://github.com/facontidavide/PlotJuggler/blob/75f9885826a7aca8c337057d368fe9e4ccfd89b2/plotjuggler_plugins/ParserROS/ros_parser.cpp#L585-L590

not sure why this std::min makes sense here? maybe for robustness for different data sources publishing the same names_version?

[saikishor]

Do we need to worry about concurrent access?

[christophfroehlich]

that's what my comment in L30 is about. But as there is no such thing in the other component here, I thought that it might not be worth it

topic_based_hardware_interfaces/joint_state_topic_hardware_interface/src/joint_state_topic_hardware_interface.cpp

Lines 132 to 134 in c179773

	topic_based_joint_states_subscriber_ = get_node()->create_subscription<sensor_msgs::msg::JointState>(
	get_hardware_parameter("joint_states_topic", "/robot_joint_states"), rclcpp::SensorDataQoS(),
	[this](const sensor_msgs::msg::JointState::SharedPtr joint_state) { latest_joint_state_ = *joint_state; });

[MarqRazz]

Should we add a warning if nothing is publishing on the configured topics? That or add a throttled print when // no data received yet in read().

[christophfroehlich]

in my case it was obvious because no joint_states/tf was published then. But sure, a throttled warning does not hurt

[MarqRazz]

What do you think about adding a test that publishes on the topics this system is listening to and verifying the state matches?

[christophfroehlich]

would be great :D

[MarqRazz]

Awesome! Maybe have some data in there that is should nicely ignore 😄.

        if (joint_state_interfaces_.find(name) != joint_state_interfaces_.end() ||
            sensor_state_interfaces_.find(name) != sensor_state_interfaces_.end() ||
            gpio_state_interfaces_.find(name) != gpio_state_interfaces_.end() ||
            unlisted_state_interfaces_.find(name) != unlisted_state_interfaces_.end())