Running PrimeHub more than once

[bcfletcher]

Does anyone know if there are ramifications of running PrimeHub() more than once in a single program.

The rationale is the following.

Currently to configure top_side and front_side parameters for PrimeHub() you have do that during initialization instead of post initialization. This means we have to know which hub we are connecting to PRIOR to running PrimeHub(). This means we have to pass the hubNumber (it's in an index into an array of configuration objects) when we call our initializeRobot() function. This means we need to know which robot we are configuring. ideally we would like to determine programmatically which hub is connected then run PrimeHub to configure it.

currently we run PrimeHub()
the get the hub name while running through the list of configuration objects to find the object with the right name. then we rerun PrimeHub() passing in the correct values for top_side and front_side but we are not sure if there are potential unknown side effects by doing this.

ideally the top_side and front_side parameters should be moved to a function like

axis.orientation(topside=Axis.?, front_size=Axis.?)

similar to display.orientation()

[dlech]

Does anyone know if there are ramifications of running PrimeHub() more than once in a single program.

Undefined behavior. Pybricks wasn't designed to do this so it hasn't really been tested and wasn't programmed to support it. In the future, initializing it multiple times may raise an exception.

---

As an alternative, you could make a common file and separate main entry points for each different hub, then have the hub-specific stuff in one file and pass that to the common main function as args.

[bcfletcher]

We are essentially doing that suggestion but we have different robot designs so the hub initializations parameters are different and we want our robot menu to pick which robot configuration to use based on name of robot which can’t be done until after PrimeHub() is called . Thus the chicken and egg problem.

[laurensvalk]

I like the suggestion to move (or also have) the IMU configuration to a method you can call at any time.

[BertLindeman]

For multihub models I instanciate the current hub to get it's name and then import the program for that hub.
This further defines the hub.
Like the top and bottom hub of the Liebherr crane.

Made a smallish program that gets HUB.system.name() and then defines another instance of the same hub depending on the name.
This would not be a redefinition of the hub in the same object.
But would it be future-safe?

issue_2130_no_redef_one_program.py

# issue_2130_no_redef_one_program.py

from pybricks.hubs import TechnicHub
from pybricks import version
from pybricks.parameters import Color
from pybricks.tools import wait

# Clear the log window
print("\x1b[H\x1b[2J", end="")

# Define hub roles
HUBS = {
    "top": {"huba 3798", "hubc beta"},
    "bottom": {"hubb beta", "hubd beta"}
}

BCAST = 26  # Broadcast channel


def init_hub(role, **kwargs):
    """ Instantiate a new TechnicHub with role-specific parameters """
    print(f"Instantiating {role}hub with arguments: {kwargs}")
    return TechnicHub(**kwargs)


try:
    # temporary hub instance to get the name
    temp_hub = TechnicHub()
    HNAM = str(temp_hub.system.name())
    print(f"\nRunning on HUB '{HNAM}'\n\tVersion: {version}")

    # No longer need this object
    HUB = None

    # Instantiate the hub properly using `init_hub`
    if HNAM in HUBS["top"]:
        print("Initializing as tophub")
        HUB = init_hub("top", broadcast_channel=BCAST)
        HUB.light.on(Color.GREEN)
        print(f"Tophub broadcasting on channel: {BCAST}")
        HUB.ble.broadcast(True)
        print("End of tophub program")
        HUB.light.on(Color.CYAN)
        wait(500)  # allow some time for broadcast

    elif HNAM in HUBS["bottom"]:
        print("Initializing as bottomhub")
        HUB = init_hub("bottom", observe_channels=[BCAST])
        HUB.light.animate([Color.GREEN, Color.NONE, Color.CYAN, Color.NONE], 128)
        print(f"Bottomhub waiting for a message on channel {BCAST}")
        while not (msg := HUB.ble.observe(BCAST)):
            wait(100)

        print(f"Bottomhub received message: {msg}")
        HUB.light.on(Color.GREEN)
        wait(1000)
        print("\nBottomhub done\n")

    else:
        print("\n\tNo matching hub name found\nDone\n")

except Exception as e:
    print(f"\nFailed to initialize hub: {e}\n")

print("Main program ended")

[bcfletcher]

any chance the change could get implemented in 3.6?