feat: add limo speaker package

[cooperj]

What type of PR is this? (check all applicable)

• Refactor
• Feature
• Bug Fix
• Optimization
• Documentation Update

Description

Adds a car horn and text to speech support for the LIMO platform. This allows for simple use of the speakers for student projects etc.

Related Tickets & Documents

• https://github.com/LCAS/limo_ros2/tree/josh/car-horn

QA Instructions, Screenshots, Recordings

Please replace this line with instructions on how to test your changes.

Limo should launch the nodes /speaker/play and /speaker/tts with the launch file

ros2 launch limo_speaker limo_speaker.launch.py

You can then use the following lines in the code-server instance to publish the messages.

ros2 topic pub -1 /speaker/tts std_msgs/msg/String "data: 'Hello, LCAS'"

ros2 topic pub -1 /speaker/play std_msgs/msg/String "data: 'car'"
ros2 topic pub -1 /speaker/play std_msgs/msg/String "data: 'clown'"
ros2 topic pub -1 /speaker/play std_msgs/msg/String "data: 'truck'"

and also

ros2 run limo_speaker console_tts
ros2 run limo_speaker console_horn

[optional] Are there any post deployment tasks we need to perform?

Before you merge this branch, a PR needs to be raised against the limo_ros2 repo to merge in josh/car-horn and then this PR updated to have the correct submodule ref.

[cooperj]

limo_drivers-1  | [ERROR] [launch]: Caught exception in launch (see debug for traceback): [Errno 2] No such file or directory: '/opt/ros/lcas/install/limo_speaker/share/limo_speaker/launch/limo_start.launch.py'

Oops!

[cooperj]

Didn't work, sounddevice module not installed... code-server not openable, have i broken something badly?

limo_drivers-1  | [play_tts-7] Traceback (most recent call last):
limo_drivers-1  | [play_tts-7]   File "/opt/ros/lcas/install/limo_speaker/lib/limo_speaker/play_tts", line 33, in <module>
limo_drivers-1  | [play_tts-7]     sys.exit(load_entry_point('limo-speaker==0.0.0', 'console_scripts', 'play_tts')())
limo_drivers-1  | [play_tts-7]   File "/opt/ros/lcas/install/limo_speaker/lib/limo_speaker/play_tts", line 25, in importlib_load_entry_point
limo_drivers-1  | [play_tts-7]     return next(matches).load()
limo_drivers-1  | [play_tts-7]   File "/usr/lib/python3.10/importlib/metadata/__init__.py", line 171, in load
limo_drivers-1  | [play_tts-7]     module = import_module(match.group('module'))
limo_drivers-1  | [play_tts-7]   File "/usr/lib/python3.10/importlib/__init__.py", line 126, in import_module
limo_drivers-1  | [play_tts-7]     return _bootstrap._gcd_import(name[level:], package, level)
limo_drivers-1  | [play_tts-7]   File "<frozen importlib._bootstrap>", line 1050, in _gcd_import
limo_drivers-1  | [play_tts-7]   File "<frozen importlib._bootstrap>", line 1027, in _find_and_load
limo_drivers-1  | [play_tts-7]   File "<frozen importlib._bootstrap>", line 1006, in _find_and_load_unlocked
limo_drivers-1  | [play_tts-7]   File "<frozen importlib._bootstrap>", line 688, in _load_unlocked
limo_drivers-1  | [play_tts-7]   File "<frozen importlib._bootstrap_external>", line 883, in exec_module
limo_drivers-1  | [play_tts-7]   File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
limo_drivers-1  | [play_tts-7]   File "/opt/ros/lcas/install/limo_speaker/lib/python3.10/site-packages/limo_speaker/playTTS.py", line 3, in <module>
limo_drivers-1  | [play_tts-7]     import sounddevice as sd
limo_drivers-1  | [play_tts-7] ModuleNotFoundError: No module named 'sounddevice'
limo_drivers-1  | [ERROR] [play_tts-7]: process has died [pid 369, exit code 1, cmd '/opt/ros/lcas/install/limo_speaker/lib/limo_speaker/play_tts --ros-args -r __node:=play_tts'].

and

limo_drivers-1  | [play_audio-6] Traceback (most recent call last):
limo_drivers-1  | [play_audio-6]   File "/opt/ros/lcas/install/limo_speaker/lib/limo_speaker/play_audio", line 33, in <module>
limo_drivers-1  | [play_audio-6]     sys.exit(load_entry_point('limo-speaker==0.0.0', 'console_scripts', 'play_audio')())
limo_drivers-1  | [play_audio-6]   File "/opt/ros/lcas/install/limo_speaker/lib/limo_speaker/play_audio", line 25, in importlib_load_entry_point
limo_drivers-1  | [play_audio-6]     return next(matches).load()
limo_drivers-1  | [play_audio-6]   File "/usr/lib/python3.10/importlib/metadata/__init__.py", line 171, in load
limo_drivers-1  | [play_audio-6]     module = import_module(match.group('module'))
limo_drivers-1  | [play_audio-6]   File "/usr/lib/python3.10/importlib/__init__.py", line 126, in import_module
limo_drivers-1  | [play_audio-6]     return _bootstrap._gcd_import(name[level:], package, level)
limo_drivers-1  | [play_audio-6]   File "<frozen importlib._bootstrap>", line 1050, in _gcd_import
limo_drivers-1  | [play_audio-6]   File "<frozen importlib._bootstrap>", line 1027, in _find_and_load
limo_drivers-1  | [play_audio-6]   File "<frozen importlib._bootstrap>", line 1006, in _find_and_load_unlocked
limo_drivers-1  | [play_audio-6]   File "<frozen importlib._bootstrap>", line 688, in _load_unlocked
limo_drivers-1  | [play_audio-6]   File "<frozen importlib._bootstrap_external>", line 883, in exec_module
limo_drivers-1  | [play_audio-6]   File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
limo_drivers-1  | [play_audio-6]   File "/opt/ros/lcas/install/limo_speaker/lib/python3.10/site-packages/limo_speaker/playAudio.py", line 3, in <module>
limo_drivers-1  | [play_audio-6]     import sounddevice as sd
limo_drivers-1  | [play_audio-6] ModuleNotFoundError: No module named 'sounddevice'

running sudo pip3 install sounddevice once in the container works as expected? is this a PATH issue?

[cooperj]

Currently facing an issue where code-server will not work on anything other than the LIMO, this is an upstream issue from code-server themselves, see issue coder/code-server#7219

This is working fine on the LIMOs back screen and if I open an ssh tunnel to the limo, so the page is served over a secure context, then its fine?! ie:

ssh -L 9999:localhost:9999 [email protected]

[marc-hanheide]

I conditionally approve, assuming you have tested it works.

[marc-hanheide]

This is not quite how it should be used, I understand, see https://docs.python.org/3/library/tempfile.html#tempfile.NamedTemporaryFile

I think in your case https://docs.python.org/3/library/tempfile.html#tempfile.mkstemp may be better used, but if it works for you and you have tested this, I'm happy to approve.

[cooperj]

Still working on it, currently working on getting the .wav files copied over into the share directory

[cooperj]

limo_speaker doesn't currently automatically launch (do we want that?) but otherwise just works as expected.

To get sound you have to bring it up with

ros2 launch limo_speaker limo_speaker.launch.py 

[marc-hanheide]

If there are no side effect, I can't see why we wouldn't want to launch it.

[marc-hanheide]

why has this been removed? This changes the final image as it leaves it with root user. Is it intentional?

[cooperj]

Ooops, I'll add that back… Was modifying it when I was debugging the issues with packages installed into the path.

This line, adds it to the compose hence it was missed.

limo_platform/configs/docker-compose.yaml

Line 42 in c8b93e3

user: "ros"