ntprop 0.2.3

A way to create transparent NetworkTables variables

NTProp

A simple mechanism for using Network Tables entries as semi-transparent member variables

Installation

Using your Python package manager of choice, just install the ntprop package from PyPI.

pip install ntprop

Usage

To start using these, you just need to import NTPropertyHost from ntprop and have your class inherit from it. From there, any member varaible that is of the types NumberProperty, BooleanProperty, or StringProperty (also from ntprop) will act like a simple, strictly-typed member variable that transparently updates and is updated by the relevant NT entry.

Example

from ntprop import NTPropertyHost, NumberProperty

class Foo(NTPropertyHost):
    # for purposes of type-checking here, we could just call this a float, since there's no difference in 
    # usage at runtime. This is only valid if the owning class is an NTPropertyHost.
    bar: NumberProperty

    def __init__(self):
        self.bar = NumberProperty("ntbar", default=1, readonly=False) #defaults are 0.0 and False, respectively

    def do_something(self):
        # += may work?
        self.bar = self.bar + 1 
        print(self.bar)

It's also possible to use an NTProperty without it being a member of ah NTPropertyHost. To do that, you just need to use the get and set methods of the type. NOTE: These methods are not available in NTPropertyHosts since that type already does the necessary calls to get and set behind the scenes and resolves them prior to returning a value.

Bindings

It's possible to bind to changes to a property using the NTProperty.bind method, and providing a callback compatible with Callable[[Any], None]. The callback will be called any time the value changes, both from the local program and NetworkTables.

To bind a callback for a property in an NTPropertyHost, call the NTPropertyHost.bind method, passing the name of the NT entry with the callback using kwargs syntax. If your entry name is incompatible with Python name syntax, you can use a dict with the entry name as the key and the callback as the associated value. Either argument method allows multiple key-value pairs, so you can bind multiple callbacks at once.

Note: Be aware that local changes may cause the callback to be called multiple times. It is recommended to only use idempotent (same every time) callbacks as well as not make them any more expensive to call than strictly necessary.

Warning: There is no guarantee of the callback being called in the same thread it was created it (e.g. the NT listener that does background updates runs in a different non-Python thread), so any interactions with the outside world must be thread safe. If you don't know what that means, you're probably fine, and if you do know what it means, but not how to fix it, visit the Python docs on the threading module.