8203d55702 | ||
---|---|---|
.github/workflows | ||
include | ||
messaging | ||
site_scons/site_tools | ||
visionipc | ||
.dockerignore | ||
.gitignore | ||
.pre-commit-config.yaml | ||
Dockerfile | ||
LICENSE | ||
README.md | ||
SConscript | ||
SConstruct | ||
__init__.py | ||
car.capnp | ||
codecov.yml | ||
generate_javascript.sh | ||
install_capnp.sh | ||
legacy.capnp | ||
log.capnp | ||
maptile.capnp | ||
services.py |
README.md
What is cereal?
cereal is both a messaging spec for robotics systems as well as generic high performance IPC pub sub messaging with a single publisher and multiple subscribers.
Imagine this use case:
- A sensor process reads gyro measurements directly from an IMU and publishes a
sensorEvents
packet - A calibration process subscribes to the
sensorEvents
packet to use the IMU - A localization process subscribes to the
sensorEvents
packet to use the IMU also
Messaging Spec
You'll find the message types in log.capnp. It uses Cap'n proto and defines one struct called Event.
All Events have a logMonoTime
and a valid
. Then a big union defines the packet type.
Message definition Best Practices
-
All fields must describe quantities in SI units, unless otherwise specified in the field name.
-
In the context of the message they are in, field names should be completely unambigious.
-
All values should be easy to plot and be human-readable with minimal parsing.
Pub Sub Backends
cereal supports two backends, one based on zmq and another called msgq, a custom pub sub based on shared memory that doesn't require the bytes to pass through the kernel.
Example
import cereal.messaging as messaging
# in subscriber
sm = messaging.SubMaster(['sensorEvents'])
while 1:
sm.update()
print(sm['sensorEvents'])
# in publisher
pm = messaging.PubMaster(['sensorEvents'])
dat = messaging.new_message('sensorEvents', size=1)
dat.sensorEvents[0] = {"gyro": {"v": [0.1, -0.1, 0.1]}}
pm.send('sensorEvents', dat)