sw-mc-builder 0.0.4

Framework for building Stormworks microcontrollers

Stormworks Microcontroller framework

A framework for building microcontrollers in Stormworks: Build and Rescue using python scripts.

Features

• Modular design: Easily create and manage multiple microcontroller scripts.
• Latency optimization: Arithmetic functions are collapsed into function blocks in order to reduce latency.
• Auto layout: Automatically arrange function blocks for better readability.
• Integration with tumfl for lua verification and minification.
• Implicit constant number creation.

Getting started

After installing, use the sw_mc_builder module to create microcontrollers.

sw_mc_builder init test_mc.py

from sw_mc_builder import *

input1 = comp.input(SignalType.Number, "Input 1")
input2 = comp.input(SignalType.Number, "Input 2")

added = input1 + input2 * 2
highest = input1.max(input2)

mc = Microcontroller("Example MC")
mc.place_input(input1, 0, 0)
mc.place_input(input2, 0, 1)
mc.place_output(added, "Added", x=1, y=0)
mc.place_output(highest, "Highest", x=1, y=1)

handle_mcs(mc)

Design

There are two predominant design patterns used in this framework:

Python operators

from sw_mc_builder import *

input1 = comp.input(SignalType.Number, "Input 1")
input2 = comp.input(SignalType.Number, "Input 2")

added = input1 + input2
maxed = (added * input2).max(1)

Each operation is represented as a python expression. This makes it very easy to write and read code, but latency is harder to reason about, as some operations can result in multiple components.

[!WARNING] There are some operands which are impossible to implement as python expressions, see below.

Component functions

from sw_mc_builder import *

input1 = comp.input(SignalType.Number, "Input 1")
input2 = comp.input(SignalType.Number, "Input 2")

added = comp.add(input1, input2)
multiplied = comp.mul(added, input2)
maxed = comp.function("max(x,y)", multiplied, 1)

Each function represents exactly one component. This makes latency very easy to reason about, as each function has a fixed latency of one tick.

[!WARNING] The optimizer might collapse multiple arithmetic operations into a single function block, which can affect latency.

[!INFO] It is discouraged to use numerical or boolean function blocks. In part due to the poor readability, but also because some parts of the optimizer work better, if there are no function blocks.

Caveats of using python operators

• and: Use & instead. Keep in mind that & has a different operator precedence than and.
• or: Use | instead. Keep in mind that | has a different operator precedence than or.
• not: Use ~ instead. Keep in mind that ~ has a different operator precedence than not.
• int: Use <Wire>.int() instead. Will convert a boolean to a number (0 or 1).
• __getindex__: (<Wire>[index]): will always return a number. If you want to get a boolean, use <Wire>.get_bool(index) instead.
• __setindex__: (<Wire>[index] = value): This will create a new composite write for each write.

Functions other than python operators

The Wire class has a few functions that are possible with function blocks.

• max(other): Maximum of two wires.
• min(other): Minimum of two wires.
• clamp(min, max): Clamp the wire between min and max.
• lerp(start, end): Linear interpolation between two wires.
• sin(): Sine of the wire (in radians).
• cos(): Cosine of the wire (in radians).
• tan(): Tangent of the wire (in radians).
• asin(): Arcsine of the wire (in radians).
• acos(): Arccosine of the wire (in radians).
• atan(): Arctangent of the wire (in radians).
• atan2(other): Arctangent of the wire and another wire (in radians).
• sqrt(): Square root of the wire.
• ceil(): Ceiling of the wire.
• floor(): Floor of the wire.
• round(ndigits): Round the wire.
• sgn(): Sign of the wire.
• switch(on_wire, off_wire): Switch between two wires based on the boolean value of the wire.

Optimizer

All arithmetic operations are collapsed into a single function block in order to reduce latency. This can be disabled per node or per mc.

from sw_mc_builder import *

input1 = comp.input(SignalType.Number, "Input 1")

# Delay signal by one tick, disabling optimization for this node
delayed = comp.function("x", input1).stop_optimization()

# Disable optimization for the entire microcontroller
mc = Microcontroller("name")
mc.stop_optimization()

Working principles

Microcontroller handler

The handle_mcs function is the entity that will bring your microcontrollers to life. It manages the actual generation (including optimization and layout) and placement of microcontrollers. It can replace existing microcontrollers in vehicles, or simply make them available ingame. It can handle multiple microcontrollers at once.

Command line arguments

• -m or --microcontroller: Export generated microcontrollers to the games microcontrollers directory.
• -v or --vehicle: Replace microcontrollers in vehicles in the games vehicles directory. Will reserve parameter settings.
• -s or --select: Only process microcontrollers with the given, comma separated, names.

Recursive definitions

There is a special component, a Placeholder, which can be used to create recursive definitions. A Placeholder must be resolved before submitting the microcontroller to handle_mcs, otherwise an error will be raised.

from sw_mc_builder import *

counter = comp.placeholder(SignalType.Number)
counter.replace_producer(counter + 1)

Input and output management

Inputs will usually be the first thing to be defined in a microcontroller. They produce Wire objects that can be used in the microcontroller. If an input is used in any calculation that is present in an output, wi has to be placed in the microcontroller.

Outputs are simply defined by placing them in a microcontroller.

If inputs or outputs are outside of the boundaries of the microcontroller, a warning will be generated, and the microcontroller expanded.

from sw_mc_builder import *

input1 = comp.input(SignalType.Number, "Input 1")
input2 = comp.input(SignalType.Number, "Input 2")

# Program logic
added = input1 + input2 * 2

mc = Microcontroller("Example MC")
mc.place_input(input1, 0, 0)
mc.place_input(input2, 0, 1)
mc.place_output(added, "Added", x=1, y=0)

handle_mcs(mc)