chore: Remove trailing whitespace #129

Open
glen wants to merge 10 commits from glen/dyna3:noTrailingWhitespace into main
20 changed files with 337 additions and 341 deletions

View file

@ -10,7 +10,7 @@ runs:
using: "composite"
steps:
- run: rustup target add wasm32-unknown-unknown
# install the Trunk binary to `ci-bin` within the workspace directory, which
# is determined by the `github.workspace` label and reflected in the
# `GITHUB_WORKSPACE` environment variable. then, make the `trunk` command

View file

@ -24,6 +24,6 @@ jobs:
# workspace directory (action variable `github.workspace`, environment
# variable `$GITHUB_WORKSPACE`):
- uses: https://code.forgejo.org/actions/checkout@v4
- uses: ./.forgejo/setup-trunk
- run: RUSTFLAGS='-D warnings' cargo test
- run: RUSTFLAGS='-D warnings' cargo test

View file

@ -52,20 +52,20 @@ The latest prototype is in the folder `app-proto`. It includes both a user inter
1. Use `sh` to run the script `tools/run-examples.sh`.
- The script is location-independent, so you can do this from anywhere in the dyna3 repository.
- The call from the top level of the repository is:
```bash
sh tools/run-examples.sh
```
- For each example problem, the engine will print the value of the loss function at each optimization step.
- The first example that prints is the same as the Irisawa hexlet example from the Julia version of the engine prototype. If you go into `engine-proto/gram-test`, launch Julia, and then execute
```julia
include("irisawa-hexlet.jl")
for (step, scaled_loss) in enumerate(history_alt.scaled_loss)
println(rpad(step-1, 4), " | ", scaled_loss)
end
```
you should see that it prints basically the same loss history until the last few steps, when the lower default precision of the Rust engine really starts to show.
### Run the automated tests

View file

@ -15,9 +15,9 @@ fn main() {
for k in 4..9 {
println!(" {} sun", 1.0 / config[(3, k)]);
}
// print the completed Gram matrix
print::gram_matrix(&config);
}
print::loss_history(&realization.history);
}
}

View file

@ -14,7 +14,7 @@ fn main() {
// print the completed Gram matrix and the realized configuration
print::gram_matrix(&config);
print::config(&config);
// find the kaleidocycle's twist motion by projecting onto the tangent
// space
const N_POINTS: usize = 12;
@ -29,4 +29,4 @@ fn main() {
let normalization = 5.0 / twist_motion[(2, 0)];
println!("\nTwist motion:{}", (normalization * twist_motion).to_string().trim_end());
}
}
}

View file

@ -6,7 +6,7 @@
<link data-trunk rel="css" href="main.css"/>
<link href="https://fonts.bunny.net/css?family=fira-sans:ital,wght@0,400;1,400&display=swap" rel="stylesheet">
<link href="https://fonts.bunny.net/css?family=noto-emoji:wght@400&text=%f0%9f%94%97%e2%9a%a0&display=swap" rel="stylesheet">
<!--
the Charming visualization crate, which we use to show engine diagnostics,
depends the ECharts JavaScript package

View file

@ -45,7 +45,7 @@ static NEXT_SERIAL: AtomicU64 = AtomicU64::new(0);
pub trait Serial {
// a serial number that uniquely identifies this element
fn serial(&self) -> u64;
// take the next serial number, panicking if that was the last one left
fn next_serial() -> u64 where Self: Sized {
// the technique we use to panic on overflow is taken from _Rust Atomics
@ -101,33 +101,33 @@ pub trait ProblemPoser {
pub trait Element: Serial + ProblemPoser + DisplayItem {
// the default identifier for an element of this type
fn default_id() -> String where Self: Sized;
// the default example of an element of this type
fn default(id: String, id_num: u64) -> Self where Self: Sized;
// the default regulators that come with this element
fn default_regulators(self: Rc<Self>) -> Vec<Rc<dyn Regulator>> {
Vec::new()
}
fn id(&self) -> &String;
fn label(&self) -> &String;
fn representation(&self) -> Signal<DVector<f64>>;
fn ghost(&self) -> Signal<bool>;
// the regulators the element is subject to. the assembly that owns the
// element is responsible for keeping this set up to date
fn regulators(&self) -> Signal<BTreeSet<Rc<dyn Regulator>>>;
// project a representation vector for this kind of element onto its
// normalization variety
fn project_to_normalized(&self, rep: &mut DVector<f64>);
// the configuration matrix column index that was assigned to the element
// last time the assembly was realized, or `None` if the element has never
// been through a realization
fn column_index(&self) -> Option<usize>;
// assign the element a configuration matrix column index. this method must
// be used carefully to preserve invariant (1), described in the comment on
// the `tangent` field of the `Assembly` structure
@ -179,7 +179,7 @@ pub struct Sphere {
impl Sphere {
const CURVATURE_COMPONENT: usize = 3;
pub fn new(
id: String,
label: String,
@ -203,7 +203,7 @@ impl Element for Sphere {
fn default_id() -> String {
"sphere".to_string()
}
fn default(id: String, id_num: u64) -> Self {
Self::new(
id,
@ -212,39 +212,39 @@ impl Element for Sphere {
sphere(0.0, 0.0, 0.0, 1.0),
)
}
fn default_regulators(self: Rc<Self>) -> Vec<Rc<dyn Regulator>> {
vec![Rc::new(HalfCurvatureRegulator::new(self))]
}
fn id(&self) -> &String {
&self.id
}
fn label(&self) -> &String {
&self.label
}
fn representation(&self) -> Signal<DVector<f64>> {
self.representation
}
fn ghost(&self) -> Signal<bool> {
self.ghost
}
fn regulators(&self) -> Signal<BTreeSet<Rc<dyn Regulator>>> {
self.regulators
}
fn project_to_normalized(&self, rep: &mut DVector<f64>) {
project_sphere_to_normalized(rep);
}
fn column_index(&self) -> Option<usize> {
self.column_index.get()
}
fn set_column_index(&self, index: usize) {
self.column_index.set(Some(index));
}
@ -279,7 +279,7 @@ pub struct Point {
impl Point {
const WEIGHT_COMPONENT: usize = 3;
const NORM_COMPONENT: usize = 4;
pub fn new(
id: String,
label: String,
@ -303,7 +303,7 @@ impl Element for Point {
fn default_id() -> String {
"point".to_string()
}
fn default(id: String, id_num: u64) -> Self {
Self::new(
id,
@ -321,35 +321,35 @@ impl Element for Point {
})
.collect()
}
fn id(&self) -> &String {
&self.id
}
fn label(&self) -> &String {
&self.label
}
fn representation(&self) -> Signal<DVector<f64>> {
self.representation
}
fn ghost(&self) -> Signal<bool> {
self.ghost
}
fn regulators(&self) -> Signal<BTreeSet<Rc<dyn Regulator>>> {
self.regulators
}
fn project_to_normalized(&self, rep: &mut DVector<f64>) {
project_point_to_normalized(rep);
}
fn column_index(&self) -> Option<usize> {
self.column_index.get()
}
fn set_column_index(&self, index: usize) {
self.column_index.set(Some(index));
}
@ -420,10 +420,10 @@ impl InversiveDistanceRegulator {
)
)
});
let set_point = create_signal(SpecifiedValue::from_empty_spec());
let serial = Self::next_serial();
Self { subjects, measurement, set_point, serial }
}
}
@ -432,11 +432,11 @@ impl Regulator for InversiveDistanceRegulator {
fn subjects(&self) -> Vec<Rc<dyn Element>> {
self.subjects.clone().into()
}
fn measurement(&self) -> ReadSignal<f64> {
self.measurement
}
fn set_point(&self) -> Signal<SpecifiedValue> {
self.set_point
}
@ -475,10 +475,10 @@ impl HalfCurvatureRegulator {
let measurement = subject.representation().map(
|rep| rep[Sphere::CURVATURE_COMPONENT]
);
let set_point = create_signal(SpecifiedValue::from_empty_spec());
let serial = Self::next_serial();
Self { subject, measurement, set_point, serial }
}
}
@ -487,11 +487,11 @@ impl Regulator for HalfCurvatureRegulator {
fn subjects(&self) -> Vec<Rc<dyn Element>> {
vec![self.subject.clone()]
}
fn measurement(&self) -> ReadSignal<f64> {
self.measurement
}
fn set_point(&self) -> Signal<SpecifiedValue> {
self.set_point
}
@ -518,7 +518,6 @@ impl ProblemPoser for HalfCurvatureRegulator {
#[derive(Clone, Copy, Sequence)]
pub enum Axis { X = 0, Y = 1, Z = 2 }
impl Axis {
fn name(&self) -> &'static str {
match self { Axis::X => "X", Axis::Y => "Y", Axis::Z => "Z" }
@ -601,7 +600,7 @@ pub struct Assembly {
// elements and regulators
pub elements: Signal<BTreeSet<Rc<dyn Element>>>,
pub regulators: Signal<BTreeSet<Rc<dyn Regulator>>>,
// solution variety tangent space. the basis vectors are stored in
// configuration matrix format, ordered according to the elements' column
// indices. when you realize the assembly, every element that's present
@ -613,13 +612,13 @@ pub struct Assembly {
// in that column of the tangent space basis matrices
//
pub tangent: Signal<ConfigSubspace>,
// indexing
pub elements_by_id: Signal<BTreeMap<String, Rc<dyn Element>>>,
// realization control
pub realization_trigger: Signal<()>,
// realization diagnostics
pub realization_status: Signal<Result<(), String>>,
pub descent_history: Signal<DescentHistory>,
@ -639,7 +638,7 @@ impl Assembly {
descent_history: create_signal(DescentHistory::new()),
step: create_signal(SpecifiedValue::from_empty_spec()),
};
// realize the assembly whenever the element list, the regulator list,
// a regulator's set point, or the realization trigger is updated
let assembly_for_realization = assembly.clone();
@ -653,7 +652,7 @@ impl Assembly {
assembly_for_realization.realization_trigger.track();
assembly_for_realization.realize();
});
// load a configuration from the descent history whenever the active
// step is updated
let assembly_for_step_selection = assembly.clone();
@ -665,12 +664,12 @@ impl Assembly {
assembly_for_step_selection.load_config(&config)
}
});
assembly
}
// --- inserting elements and regulators ---
// insert an element into the assembly without checking whether we already
// have an element with the same identifier. any element that does have the
// same identifier will get kicked out of the `elements_by_id` index
@ -680,13 +679,13 @@ impl Assembly {
let elt_rc = Rc::new(elt);
self.elements.update(|elts| elts.insert(elt_rc.clone()));
self.elements_by_id.update(|elts_by_id| elts_by_id.insert(id, elt_rc.clone()));
// create and insert the element's default regulators
for reg in elt_rc.default_regulators() {
self.insert_regulator(reg);
}
}
pub fn try_insert_element(&self, elt: impl Element + 'static) -> bool {
let can_insert = self.elements_by_id.with_untracked(
|elts_by_id| !elts_by_id.contains_key(elt.id())
@ -696,7 +695,7 @@ impl Assembly {
}
can_insert
}
pub fn insert_element_default<T: Element + 'static>(&self) {
// find the next unused identifier in the default sequence
let default_id = T::default_id();
@ -708,17 +707,17 @@ impl Assembly {
id_num += 1;
id = format!("{default_id}{id_num}");
}
// create and insert the default example of `T`
let _ = self.insert_element_unchecked(T::default(id, id_num));
}
pub fn insert_regulator(&self, regulator: Rc<dyn Regulator>) {
// add the regulator to the assembly's regulator list
self.regulators.update(
|regs| regs.insert(regulator.clone())
);
// add the regulator to each subject's regulator list
let subject_regulators: Vec<_> = regulator.subjects().into_iter().map(
|subj| subj.regulators()
@ -726,7 +725,7 @@ impl Assembly {
for regulators in subject_regulators {
regulators.update(|regs| regs.insert(regulator.clone()));
}
/* DEBUG */
// print an updated list of regulators
console_log!("Regulators:");
@ -749,9 +748,9 @@ impl Assembly {
}
});
}
// --- updating the configuration ---
pub fn load_config(&self, config: &DMatrix<f64>) {
for elt in self.elements.get_clone_untracked() {
elt.representation().update(
@ -759,9 +758,9 @@ impl Assembly {
);
}
}
// --- realization ---
pub fn realize(&self) {
// index the elements
self.elements.update_silent(|elts| {
@ -769,7 +768,7 @@ impl Assembly {
elt.set_column_index(index);
}
});
// set up the constraint problem
let problem = self.elements.with_untracked(|elts| {
let mut problem = ConstraintProblem::new(elts.len());
@ -783,21 +782,21 @@ impl Assembly {
});
problem
});
/* DEBUG */
// log the Gram matrix
console_log!("Gram matrix:\n{}", problem.gram);
console_log!("Frozen entries:\n{}", problem.frozen);
/* DEBUG */
// log the initial configuration matrix
console_log!("Old configuration:{:>8.3}", problem.guess);
// look for a configuration with the given Gram matrix
let Realization { result, history } = realize_gram(
&problem, 1.0e-12, 0.5, 0.9, 1.1, 200, 110
);
/* DEBUG */
// report the outcome of the search in the browser console
if let Err(ref message) = result {
@ -809,20 +808,20 @@ impl Assembly {
console_log!("Steps: {}", history.scaled_loss.len() - 1);
console_log!("Loss: {}", history.scaled_loss.last().unwrap());
}
// report the descent history
let step_cnt = history.config.len();
self.descent_history.set(history);
match result {
Ok(ConfigNeighborhood { nbhd: tangent, .. }) => {
/* DEBUG */
// report the tangent dimension
console_log!("Tangent dimension: {}", tangent.dim());
// report the realization status
self.realization_status.set(Ok(()));
// display the last realization step
self.step.set(
if step_cnt > 0 {
@ -832,7 +831,7 @@ impl Assembly {
SpecifiedValue::from_empty_spec()
}
);
// save the tangent space
self.tangent.set_silent(tangent);
},
@ -842,15 +841,15 @@ impl Assembly {
// `Err(message)` we received from the match: we're changing the
// `Ok` type from `Realization` to `()`
self.realization_status.set(Err(message));
// display the initial guess
self.step.set(SpecifiedValue::from(Some(0.0)));
},
}
}
// --- deformation ---
// project the given motion to the tangent space of the solution variety and
// move the assembly along it. the implementation is based on invariant (1)
// from above and the following additional invariant:
@ -867,7 +866,7 @@ impl Assembly {
if self.tangent.with(|tan| tan.dim() <= 0 && tan.assembly_dim() > 0) {
console::log_1(&JsValue::from("The assembly is rigid"));
}
// give a column index to each moving element that doesn't have one yet.
// this temporarily breaks invariant (1), but the invariant will be
// restored when we realize the assembly at the end of the deformation.
@ -885,7 +884,7 @@ impl Assembly {
}
next_column_index
};
// project the element motions onto the tangent space of the solution
// variety and sum them to get a deformation of the whole assembly. the
// matrix `motion_proj` that holds the deformation has extra columns for
@ -896,7 +895,7 @@ impl Assembly {
// we can unwrap the column index because we know that every moving
// element has one at this point
let column_index = elt_motion.element.column_index().unwrap();
if column_index < realized_dim {
// this element had a column index when we started, so by
// invariant (1), it's reflected in the tangent space
@ -914,7 +913,7 @@ impl Assembly {
target_column += unif_to_std * elt_motion.velocity;
}
}
// step the assembly along the deformation. this changes the elements'
// normalizations, so we restore those afterward
for elt in self.elements.get_clone_untracked() {
@ -932,7 +931,7 @@ impl Assembly {
};
});
}
// trigger a realization to bring the configuration back onto the
// solution variety. this also gets the elements' column indices and the
// saved tangent space back in sync
@ -943,9 +942,9 @@ impl Assembly {
#[cfg(test)]
mod tests {
use super::*;
use crate::engine;
#[test]
#[should_panic(expected =
"Sphere \"sphere\" must be indexed before it writes problem data")]
@ -955,7 +954,7 @@ mod tests {
elt.pose(&mut ConstraintProblem::new(1));
});
}
#[test]
#[should_panic(expected = "Subject \"sphere1\" must be indexed before \
inversive distance regulator writes problem data")]
@ -973,7 +972,7 @@ mod tests {
}.pose(&mut ConstraintProblem::new(2));
});
}
#[test]
fn curvature_drift_test() {
const INITIAL_RADIUS: f64 = 0.25;
@ -993,7 +992,7 @@ mod tests {
engine::sphere(0.0, 0.0, 0.0, INITIAL_RADIUS),
)
);
// nudge the sphere repeatedly along the `z` axis
const STEP_SIZE: f64 = 0.0025;
const STEP_CNT: usize = 400;
@ -1009,7 +1008,7 @@ mod tests {
]
);
}
// check how much the sphere's curvature has drifted
const INITIAL_HALF_CURV: f64 = 0.5 / INITIAL_RADIUS;
const DRIFT_TOL: f64 = 0.015;

View file

@ -54,7 +54,7 @@ fn StepInput() -> View {
// get the assembly
let state = use_context::<AppState>();
let assembly = state.assembly;
// the `last_step` signal holds the index of the last step
let last_step = assembly.descent_history.map(
|history| match history.config.len() {
@ -63,15 +63,15 @@ fn StepInput() -> View {
}
);
let input_max = last_step.map(|last| last.unwrap_or(0));
// these signals hold the entered step number
let value = create_signal(String::new());
let value_as_number = create_signal(0.0);
create_effect(move || {
value.set(assembly.step.with(|n| n.spec.clone()));
});
view! {
div(id = "step-input") {
label { "Step" }
@ -98,7 +98,7 @@ fn StepInput() -> View {
|val| val.clamp(0.0, input_max.get() as f64)
)
);
// set the input string and the assembly's active step
value.set(step.spec.clone());
assembly.step.set(step);
@ -124,7 +124,7 @@ fn LossHistory() -> View {
const CONTAINER_ID: &str = "loss-history";
let state = use_context::<AppState>();
let renderer = WasmRenderer::new_opt(None, Some(178));
on_mount(move || {
create_effect(move || {
// get the loss history
@ -136,13 +136,13 @@ fn LossHistory() -> View {
.map(into_log10_time_point)
.collect()
);
// initialize the chart axes
let step_axis = Axis::new()
.type_(AxisType::Category)
.boundary_gap(false);
let scaled_loss_axis = Axis::new();
// load the chart data. when there's no history, we load the data
// point (0, None) to clear the chart. it would feel more natural to
// load empty data vectors, but that turns out not to clear the
@ -164,7 +164,7 @@ fn LossHistory() -> View {
renderer.render(CONTAINER_ID, &chart).unwrap();
});
});
view! {
div(id = CONTAINER_ID, class = "diagnostics-chart")
}
@ -176,7 +176,7 @@ fn SpectrumHistory() -> View {
const CONTAINER_ID: &str = "spectrum-history";
let state = use_context::<AppState>();
let renderer = WasmRenderer::new(478, 178);
on_mount(move || {
create_effect(move || {
// get the spectrum of the Hessian at each step, split into its
@ -208,13 +208,13 @@ fn SpectrumHistory() -> View {
): (Vec<_>, Vec<_>) = hess_eigvals_nonzero
.into_iter()
.partition(|&(_, val)| val > 0.0);
// initialize the chart axes
let step_axis = Axis::new()
.type_(AxisType::Category)
.boundary_gap(false);
let eigval_axis = Axis::new();
// load the chart data. when there's no history, we load the data
// point (0, None) to clear the chart. it would feel more natural to
// load empty data vectors, but that turns out not to clear the
@ -270,7 +270,7 @@ fn SpectrumHistory() -> View {
renderer.render(CONTAINER_ID, &chart).unwrap();
});
});
view! {
div(id = CONTAINER_ID, class = "diagnostics-chart")
}
@ -302,7 +302,7 @@ pub fn Diagnostics() -> View {
let diagnostics_state = DiagnosticsState::new("loss".to_string());
let active_tab = diagnostics_state.active_tab.clone();
provide_context(diagnostics_state);
view! {
div(id = "diagnostics") {
div(id = "diagnostics-bar") {
@ -317,4 +317,4 @@ pub fn Diagnostics() -> View {
DiagnosticsPanel(name = "spectrum") { SpectrumHistory {} }
}
}
}
}

View file

@ -48,11 +48,11 @@ impl SceneSpheres {
highlights: Vec::new(),
}
}
fn len_i32(&self) -> i32 {
self.representations.len().try_into().expect("Number of spheres must fit in a 32-bit integer")
}
fn push(
&mut self, representation: DVector<f64>,
color: ElementColor, opacity: f32, highlight: f32,
@ -79,7 +79,7 @@ impl ScenePoints {
selections: Vec::new(),
}
}
fn push(
&mut self, representation: DVector<f64>,
color: ElementColor, opacity: f32, highlight: f32, selected: bool,
@ -107,7 +107,7 @@ impl Scene {
pub trait DisplayItem {
fn show(&self, scene: &mut Scene, selected: bool);
// the smallest positive depth, represented as a multiple of `dir`, where
// the line generated by `dir` hits the element. returns `None` if the line
// misses the element
@ -125,14 +125,14 @@ impl DisplayItem for Sphere {
const DEFAULT_OPACITY: f32 = 0.5;
const GHOST_OPACITY: f32 = 0.2;
const HIGHLIGHT: f32 = 0.2;
let representation = self.representation.get_clone_untracked();
let color = if selected { self.color.map(|channel| 0.2 + 0.8*channel) } else { self.color };
let opacity = if self.ghost.get() { GHOST_OPACITY } else { DEFAULT_OPACITY };
let highlight = if selected { 1.0 } else { HIGHLIGHT };
scene.spheres.push(representation, color, opacity, highlight);
}
// this method should be kept synchronized with `sphere_cast` in
// `spheres.frag`, which does essentially the same thing on the GPU side
fn cast(
@ -144,12 +144,12 @@ impl DisplayItem for Sphere {
// if `a/b` is less than this threshold, we approximate
// `a*u^2 + b*u + c` by the linear function `b*u + c`
const DEG_THRESHOLD: f64 = 1e-9;
let rep = self.representation.with_untracked(|rep| assembly_to_world * rep);
let a = -rep[3] * dir.norm_squared();
let b = rep.rows_range(..3).dot(&dir);
let c = -rep[4];
let adjust = 4.0*a*c/(b*b);
if adjust < 1.0 {
// as long as `b` is non-zero, the linear approximation of
@ -184,14 +184,14 @@ impl DisplayItem for Point {
/* SCAFFOLDING */
const GHOST_OPACITY: f32 = 0.4;
const HIGHLIGHT: f32 = 0.5;
let representation = self.representation.get_clone_untracked();
let color = if selected { self.color.map(|channel| 0.2 + 0.8*channel) } else { self.color };
let opacity = if self.ghost.get() { GHOST_OPACITY } else { 1.0 };
let highlight = if selected { 1.0 } else { HIGHLIGHT };
scene.points.push(representation, color, opacity, highlight, selected);
}
/* SCAFFOLDING */
fn cast(
&self,
@ -203,16 +203,16 @@ impl DisplayItem for Point {
if rep[2] < 0.0 {
// this constant should be kept synchronized with `point.frag`
const POINT_RADIUS_PX: f64 = 4.0;
// find the radius of the point in screen projection units
let point_radius_proj = POINT_RADIUS_PX * pixel_size;
// find the squared distance between the screen projections of the
// ray and the point
let dir_proj = -dir.fixed_rows::<2>(0) / dir[2];
let rep_proj = -rep.fixed_rows::<2>(0) / rep[2];
let dist_sq = (dir_proj - rep_proj).norm_squared();
// if the ray hits the point, return its depth
if dist_sq < point_radius_proj * point_radius_proj {
Some(rep[2] / dir[2])
@ -254,13 +254,13 @@ fn set_up_program(
WebGl2RenderingContext::FRAGMENT_SHADER,
fragment_shader_source,
);
// create the program and attach the shaders
let program = context.create_program().unwrap();
context.attach_shader(&program, &vertex_shader);
context.attach_shader(&program, &fragment_shader);
context.link_program(&program);
/* DEBUG */
// report whether linking succeeded
let link_status = context
@ -273,7 +273,7 @@ fn set_up_program(
"Linking failed"
};
console::log_1(&JsValue::from(link_msg));
program
}
@ -318,7 +318,7 @@ fn load_new_buffer(
// create a buffer and bind it to ARRAY_BUFFER
let buffer = context.create_buffer();
context.bind_buffer(WebGl2RenderingContext::ARRAY_BUFFER, buffer.as_ref());
// load the given data into the buffer. this block is unsafe because
// `Float32Array::view` creates a raw view into our module's
// `WebAssembly.Memory` buffer. allocating more memory will change the
@ -332,7 +332,7 @@ fn load_new_buffer(
WebGl2RenderingContext::STATIC_DRAW,
);
}
buffer
}
@ -353,11 +353,11 @@ fn event_dir(event: &MouseEvent) -> (Vector3<f64>, f64) {
let width = rect.width();
let height = rect.height();
let shortdim = width.min(height);
// this constant should be kept synchronized with `spheres.frag` and
// `point.vert`
const FOCAL_SLOPE: f64 = 0.3;
(
Vector3::new(
FOCAL_SLOPE * (2.0*(f64::from(event.client_x()) - rect.left()) - width) / shortdim,
@ -373,13 +373,13 @@ fn event_dir(event: &MouseEvent) -> (Vector3<f64>, f64) {
#[component]
pub fn Display() -> View {
let state = use_context::<AppState>();
// canvas
let display = create_node_ref();
// viewpoint
let assembly_to_world = create_signal(DMatrix::<f64>::identity(5, 5));
// navigation
let pitch_up = create_signal(0.0);
let pitch_down = create_signal(0.0);
@ -390,7 +390,7 @@ pub fn Display() -> View {
let zoom_in = create_signal(0.0);
let zoom_out = create_signal(0.0);
let turntable = create_signal(false); /* BENCHMARKING */
// manipulation
let translate_neg_x = create_signal(0.0);
let translate_pos_x = create_signal(0.0);
@ -400,7 +400,7 @@ pub fn Display() -> View {
let translate_pos_z = create_signal(0.0);
let shrink_neg = create_signal(0.0);
let shrink_pos = create_signal(0.0);
// change listener
let scene_changed = create_signal(true);
create_effect(move || {
@ -413,18 +413,18 @@ pub fn Display() -> View {
state.selection.track();
scene_changed.set(true);
});
/* INSTRUMENTS */
const SAMPLE_PERIOD: i32 = 60;
let mut last_sample_time = 0.0;
let mut frames_since_last_sample = 0;
let mean_frame_interval = create_signal(0.0);
let assembly_for_raf = state.assembly.clone();
on_mount(move || {
// timing
let mut last_time = 0.0;
// viewpoint
const ROT_SPEED: f64 = 0.4; // in radians per second
const ZOOM_SPEED: f64 = 0.15; // multiplicative rate per second
@ -432,18 +432,18 @@ pub fn Display() -> View {
let mut orientation = DMatrix::<f64>::identity(5, 5);
let mut rotation = DMatrix::<f64>::identity(5, 5);
let mut location_z: f64 = 5.0;
// manipulation
const TRANSLATION_SPEED: f64 = 0.15; // in length units per second
const SHRINKING_SPEED: f64 = 0.15; // in length units per second
// display parameters
const LAYER_THRESHOLD: i32 = 0; /* DEBUG */
const DEBUG_MODE: i32 = 0; /* DEBUG */
/* INSTRUMENTS */
let performance = window().unwrap().performance().unwrap();
// get the display canvas
let canvas = display.get().unchecked_into::<web_sys::HtmlCanvasElement>();
let ctx = canvas
@ -452,28 +452,28 @@ pub fn Display() -> View {
.unwrap()
.dyn_into::<WebGl2RenderingContext>()
.unwrap();
// disable depth testing
ctx.disable(WebGl2RenderingContext::DEPTH_TEST);
// set blend mode
ctx.enable(WebGl2RenderingContext::BLEND);
ctx.blend_func(WebGl2RenderingContext::SRC_ALPHA, WebGl2RenderingContext::ONE_MINUS_SRC_ALPHA);
// set up the sphere rendering program
let sphere_program = set_up_program(
&ctx,
include_str!("identity.vert"),
include_str!("spheres.frag"),
);
// set up the point rendering program
let point_program = set_up_program(
&ctx,
include_str!("point.vert"),
include_str!("point.frag"),
);
/* DEBUG */
// print the maximum number of vectors that can be passed as
// uniforms to a fragment shader. the OpenGL ES 3.0 standard
@ -490,10 +490,10 @@ pub fn Display() -> View {
&ctx.get_parameter(WebGl2RenderingContext::MAX_FRAGMENT_UNIFORM_VECTORS).unwrap(),
&JsValue::from("uniform vectors available"),
);
// find the sphere program's vertex attribute
let viewport_position_attr = ctx.get_attrib_location(&sphere_program, "position") as u32;
// find the sphere program's uniforms
const SPHERE_MAX: usize = 200;
let sphere_cnt_loc = ctx.get_uniform_location(&sphere_program, "sphere_cnt");
@ -513,7 +513,7 @@ pub fn Display() -> View {
let shortdim_loc = ctx.get_uniform_location(&sphere_program, "shortdim");
let layer_threshold_loc = ctx.get_uniform_location(&sphere_program, "layer_threshold");
let debug_mode_loc = ctx.get_uniform_location(&sphere_program, "debug_mode");
// load the viewport vertex positions into a new vertex buffer object
const VERTEX_CNT: usize = 6;
let viewport_positions: [f32; 3*VERTEX_CNT] = [
@ -527,20 +527,20 @@ pub fn Display() -> View {
1.0, -1.0, 0.0,
];
let viewport_position_buffer = load_new_buffer(&ctx, &viewport_positions);
// find the point program's vertex attributes
let point_position_attr = ctx.get_attrib_location(&point_program, "position") as u32;
let point_color_attr = ctx.get_attrib_location(&point_program, "color") as u32;
let point_highlight_attr = ctx.get_attrib_location(&point_program, "highlight") as u32;
let point_selection_attr = ctx.get_attrib_location(&point_program, "selected") as u32;
// set up a repainting routine
let (_, start_animation_loop, _) = create_raf(move || {
// get the time step
let time = performance.now();
let time_step = 0.001*(time - last_time);
last_time = time;
// get the navigation state
let pitch_up_val = pitch_up.get();
let pitch_down_val = pitch_down.get();
@ -551,7 +551,7 @@ pub fn Display() -> View {
let zoom_in_val = zoom_in.get();
let zoom_out_val = zoom_out.get();
let turntable_val = turntable.get(); /* BENCHMARKING */
// get the manipulation state
let translate_neg_x_val = translate_neg_x.get();
let translate_pos_x_val = translate_pos_x.get();
@ -561,7 +561,7 @@ pub fn Display() -> View {
let translate_pos_z_val = translate_pos_z.get();
let shrink_neg_val = shrink_neg.get();
let shrink_pos_val = shrink_pos.get();
// update the assembly's orientation
let ang_vel = {
let pitch = pitch_up_val - pitch_down_val;
@ -582,11 +582,11 @@ pub fn Display() -> View {
Rotation3::from_scaled_axis(time_step * ang_vel).matrix()
);
orientation = &rotation * &orientation;
// update the assembly's location
let zoom = zoom_out_val - zoom_in_val;
location_z *= (time_step * ZOOM_SPEED * zoom).exp();
// manipulate the assembly
/* KLUDGE */
// to avoid the complexity of making tangent space projection
@ -642,11 +642,11 @@ pub fn Display() -> View {
scene_changed.set(true);
}
}
if scene_changed.get() {
const SPACE_DIM: usize = 3;
const COLOR_SIZE: usize = 3;
/* INSTRUMENTS */
// measure mean frame interval
frames_since_last_sample += 1;
@ -655,11 +655,11 @@ pub fn Display() -> View {
last_sample_time = time;
frames_since_last_sample = 0;
}
// --- get the assembly ---
let mut scene = Scene::new();
// find the map from assembly space to world space
let location = {
let u = -location_z;
@ -672,7 +672,7 @@ pub fn Display() -> View {
])
};
let asm_to_world = &location * &orientation;
// set up the scene
state.assembly.elements.with_untracked(
|elts| for elt in elts {
@ -681,26 +681,26 @@ pub fn Display() -> View {
}
);
let sphere_cnt = scene.spheres.len_i32();
// --- draw the spheres ---
// use the sphere rendering program
ctx.use_program(Some(&sphere_program));
// enable the sphere program's vertex attribute
ctx.enable_vertex_attrib_array(viewport_position_attr);
// write the spheres in world coordinates
let sphere_reps_world: Vec<_> = scene.spheres.representations.into_iter().map(
|rep| (&asm_to_world * rep).cast::<f32>()
).collect();
// set the resolution
let width = canvas.width() as f32;
let height = canvas.height() as f32;
ctx.uniform2f(resolution_loc.as_ref(), width, height);
ctx.uniform1f(shortdim_loc.as_ref(), width.min(height));
// pass the scene data
ctx.uniform1i(sphere_cnt_loc.as_ref(), sphere_cnt);
for n in 0..sphere_reps_world.len() {
@ -722,33 +722,33 @@ pub fn Display() -> View {
scene.spheres.highlights[n],
);
}
// pass the display parameters
ctx.uniform1i(layer_threshold_loc.as_ref(), LAYER_THRESHOLD);
ctx.uniform1i(debug_mode_loc.as_ref(), DEBUG_MODE);
// bind the viewport vertex position buffer to the position
// attribute in the vertex shader
bind_to_attribute(&ctx, viewport_position_attr, SPACE_DIM as i32, &viewport_position_buffer);
// draw the scene
ctx.draw_arrays(WebGl2RenderingContext::TRIANGLES, 0, VERTEX_CNT as i32);
// disable the sphere program's vertex attribute
ctx.disable_vertex_attrib_array(viewport_position_attr);
// --- draw the points ---
if !scene.points.representations.is_empty() {
// use the point rendering program
ctx.use_program(Some(&point_program));
// enable the point program's vertex attributes
ctx.enable_vertex_attrib_array(point_position_attr);
ctx.enable_vertex_attrib_array(point_color_attr);
ctx.enable_vertex_attrib_array(point_highlight_attr);
ctx.enable_vertex_attrib_array(point_selection_attr);
// write the points in world coordinates
let asm_to_world_sp = asm_to_world.rows(0, SPACE_DIM);
let point_positions = DMatrix::from_columns(
@ -756,7 +756,7 @@ pub fn Display() -> View {
|rep| &asm_to_world_sp * rep
).collect::<Vec<_>>().as_slice()
).cast::<f32>();
// load the point positions and colors into new buffers and
// bind them to the corresponding attributes in the vertex
// shader
@ -764,22 +764,22 @@ pub fn Display() -> View {
bind_new_buffer_to_attribute(&ctx, point_color_attr, (COLOR_SIZE + 1) as i32, scene.points.colors_with_opacity.concat().as_slice());
bind_new_buffer_to_attribute(&ctx, point_highlight_attr, 1 as i32, scene.points.highlights.as_slice());
bind_new_buffer_to_attribute(&ctx, point_selection_attr, 1 as i32, scene.points.selections.as_slice());
// draw the scene
ctx.draw_arrays(WebGl2RenderingContext::POINTS, 0, point_positions.ncols() as i32);
// disable the point program's vertex attributes
ctx.disable_vertex_attrib_array(point_position_attr);
ctx.disable_vertex_attrib_array(point_color_attr);
ctx.disable_vertex_attrib_array(point_highlight_attr);
ctx.disable_vertex_attrib_array(point_selection_attr);
}
// --- update the display state ---
// update the viewpoint
assembly_to_world.set(asm_to_world);
// clear the scene change flag
scene_changed.set(
pitch_up_val != 0.0
@ -799,7 +799,7 @@ pub fn Display() -> View {
});
start_animation_loop();
});
let set_nav_signal = move |event: &KeyboardEvent, value: f64| {
let mut navigating = true;
let shift = event.shift_key();
@ -819,7 +819,7 @@ pub fn Display() -> View {
event.prevent_default();
}
};
let set_manip_signal = move |event: &KeyboardEvent, value: f64| {
let mut manipulating = true;
let shift = event.shift_key();
@ -838,7 +838,7 @@ pub fn Display() -> View {
event.prevent_default();
}
};
view! {
/* TO DO */
// switch back to integer-valued parameters when that becomes possible
@ -860,7 +860,7 @@ pub fn Display() -> View {
yaw_left.set(0.0);
pitch_up.set(0.0);
pitch_down.set(0.0);
// swap manipulation inputs
translate_pos_z.set(translate_neg_y.get());
translate_neg_z.set(translate_pos_y.get());
@ -886,7 +886,7 @@ pub fn Display() -> View {
roll_ccw.set(0.0);
zoom_in.set(0.0);
zoom_out.set(0.0);
// swap manipulation inputs
translate_pos_y.set(translate_neg_z.get());
translate_neg_y.set(translate_pos_z.get());
@ -927,7 +927,7 @@ pub fn Display() -> View {
None => (),
};
}
// if we clicked something, select it
match clicked {
Some((elt, _)) => state.select(&elt, event.shift_key()),
@ -936,4 +936,4 @@ pub fn Display() -> View {
},
)
}
}
}

View file

@ -21,16 +21,16 @@ fn RegulatorInput(regulator: Rc<dyn Regulator>) -> View {
// get the regulator's measurement and set point signals
let measurement = regulator.measurement();
let set_point = regulator.set_point();
// the `valid` signal tracks whether the last entered value is a valid set
// point specification
let valid = create_signal(true);
// the `value` signal holds the current set point specification
let value = create_signal(
set_point.with_untracked(|set_pt| set_pt.spec.clone())
);
// this `reset_value` closure resets the input value to the regulator's set
// point specification
let reset_value = move || {
@ -39,11 +39,11 @@ fn RegulatorInput(regulator: Rc<dyn Regulator>) -> View {
value.set(set_point.with(|set_pt| set_pt.spec.clone()));
})
};
// reset the input value whenever the regulator's set point specification
// is updated
create_effect(reset_value);
view! {
input(
r#type = "text",
@ -241,7 +241,7 @@ fn ElementOutlineItem(element: Rc<dyn Element>) -> View {
#[component]
pub fn Outline() -> View {
let state = use_context::<AppState>();
// list the elements alphabetically by ID
/* TO DO */
// this code is designed to generalize easily to other sort keys. if we only
@ -254,7 +254,7 @@ pub fn Outline() -> View {
.sorted_by_key(|elt| elt.id().clone())
.collect::<Vec<_>>()
);
view! {
ul(
id = "outline",
@ -272,4 +272,4 @@ pub fn Outline() -> View {
)
}
}
}
}

View file

@ -10,10 +10,10 @@ out vec4 outColor;
void main() {
float r = total_radius * length(2.*gl_PointCoord - vec2(1.));
const float POINT_RADIUS = 4.;
float border = smoothstep(POINT_RADIUS - 1., POINT_RADIUS, r);
float disk = 1. - smoothstep(total_radius - 1., total_radius, r);
vec4 color = mix(point_color, vec4(1.), border * point_highlight);
outColor = vec4(vec3(1.), disk) * color;
}
}

View file

@ -14,11 +14,11 @@ const float focal_slope = 0.3;
void main() {
total_radius = 5. + 0.5*selected;
float depth = -focal_slope * position.z;
gl_Position = vec4(position.xy / depth, 0., 1.);
gl_PointSize = 2.*total_radius;
point_color = color;
point_highlight = highlight;
}
}

View file

@ -75,7 +75,7 @@ Fragment sphere_shading(vecInv v, vec3 pt, vec4 base_color) {
// point. i calculated it in my head and decided that the result looked good
// enough for now
vec3 normal = normalize(-v.sp + 2.*v.lt.s*pt);
float incidence = dot(normal, light_dir);
float illum = mix(0.4, 1.0, max(incidence, 0.0));
return Fragment(pt, normal, vec4(illum * base_color.rgb, base_color.a));
@ -110,7 +110,7 @@ vec2 sphere_cast(vecInv v, vec3 dir) {
float a = -v.lt.s * dot(dir, dir);
float b = dot(v.sp, dir);
float c = -v.lt.t;
float adjust = 4.*a*c/(b*b);
if (adjust < 1.) {
// as long as `b` is non-zero, the linear approximation of
@ -136,7 +136,7 @@ vec2 sphere_cast(vecInv v, vec3 dir) {
void main() {
vec2 scr = (2.*gl_FragCoord.xy - resolution) / shortdim;
vec3 dir = vec3(focal_slope * scr, -1.);
// cast rays through the spheres
const int LAYER_MAX = 12;
TaggedDepth top_hits [LAYER_MAX];
@ -144,7 +144,7 @@ void main() {
for (int id = 0; id < sphere_cnt; ++id) {
// find out where the ray hits the sphere
vec2 hit_depths = sphere_cast(sphere_list[id], dir);
// insertion-sort the points we hit into the hit list
float dimming = 1.;
for (int side = 0; side < 2; ++side) {
@ -169,14 +169,14 @@ void main() {
}
}
}
/* DEBUG */
// in debug mode, show the layer count instead of the shaded image
if (debug_mode) {
// at the bottom of the screen, show the color scale instead of the
// layer count
if (gl_FragCoord.y < 10.) layer_cnt = int(16. * gl_FragCoord.x / resolution.x);
// convert number to color
ivec3 bits = layer_cnt / ivec3(1, 2, 4);
vec3 color = mod(vec3(bits), 2.);
@ -186,7 +186,7 @@ void main() {
outColor = vec4(color, 1.);
return;
}
// composite the sphere fragments
vec3 color = vec3(0.);
int layer = layer_cnt - 1;
@ -203,7 +203,7 @@ void main() {
// load the current fragment
Fragment frag = frag_next;
float highlight = highlight_next;
// shade the next fragment
hit = top_hits[layer];
sphere_color = color_list[hit.id];
@ -213,23 +213,23 @@ void main() {
vec4(hit.dimming * sphere_color.rgb, sphere_color.a)
);
highlight_next = highlight_list[hit.id];
// highlight intersections
float ixn_dist = intersection_dist(frag, frag_next);
float max_highlight = max(highlight, highlight_next);
float ixn_highlight = 0.5 * max_highlight * (1. - smoothstep(2./3.*ixn_threshold, 1.5*ixn_threshold, ixn_dist));
frag.color = mix(frag.color, vec4(1.), ixn_highlight);
frag_next.color = mix(frag_next.color, vec4(1.), ixn_highlight);
// highlight cusps
float cusp_cos = abs(dot(dir, frag.normal));
float cusp_threshold = 2.*sqrt(ixn_threshold * sphere_list[hit.id].lt.s);
float cusp_highlight = highlight * (1. - smoothstep(2./3.*cusp_threshold, 1.5*cusp_threshold, cusp_cos));
frag.color = mix(frag.color, vec4(1.), cusp_highlight);
// composite the current fragment
color = mix(color, frag.color.rgb, frag.color.a);
}
color = mix(color, frag_next.color.rgb, frag_next.color.a);
outColor = vec4(sRGB(color), 1.);
}
}

View file

@ -144,7 +144,7 @@ fn load_low_curvature(assembly: &Assembly) {
engine::sphere(2.0/3.0, 4.0/3.0 * a, 0.0, 1.0/3.0),
)
);
// impose the desired tangencies and make the sides planar
let index_range = 1..=3;
let [central, assemb_plane] = ["central", "assemb_plane"].map(
@ -217,7 +217,7 @@ fn load_pointed(assembly: &Assembly) {
for index_y in 0..=1 {
let x = index_x as f64 - 0.5;
let y = index_y as f64 - 0.5;
let _ = assembly.try_insert_element(
Sphere::new(
format!("sphere{index_x}{index_y}"),
@ -226,7 +226,7 @@ fn load_pointed(assembly: &Assembly) {
engine::sphere(x, y, 0.0, 1.0),
)
);
let _ = assembly.try_insert_element(
Point::new(
format!("point{index_x}{index_y}"),
@ -310,7 +310,7 @@ fn load_tridiminished_icosahedron(assembly: &Assembly) {
for vertex in vertices {
let _ = assembly.try_insert_element(vertex);
}
// create the faces
const COLOR_FACE: ElementColor = [0.75_f32, 0.75_f32, 0.75_f32];
let frac_1_sqrt_6 = 1.0 / 6.0_f64.sqrt();
@ -339,7 +339,7 @@ fn load_tridiminished_icosahedron(assembly: &Assembly) {
face.ghost().set(true);
let _ = assembly.try_insert_element(face);
}
let index_range = 1..=3;
for j in index_range.clone() {
// make each face planar
@ -352,7 +352,7 @@ fn load_tridiminished_icosahedron(assembly: &Assembly) {
curvature_regulator.set_point().set(
SpecifiedValue::try_from("0".to_string()).unwrap()
);
// put each A vertex on the face it belongs to
let vertex_a = assembly.elements_by_id.with_untracked(
|elts_by_id| elts_by_id[&format!("a{j}")].clone()
@ -360,7 +360,7 @@ fn load_tridiminished_icosahedron(assembly: &Assembly) {
let incidence_a = InversiveDistanceRegulator::new([face.clone(), vertex_a.clone()]);
incidence_a.set_point.set(SpecifiedValue::try_from("0".to_string()).unwrap());
assembly.insert_regulator(Rc::new(incidence_a));
// regulate the B-C vertex distances
let vertices_bc = ["b", "c"].map(
|series| assembly.elements_by_id.with_untracked(
@ -370,10 +370,10 @@ fn load_tridiminished_icosahedron(assembly: &Assembly) {
assembly.insert_regulator(
Rc::new(InversiveDistanceRegulator::new(vertices_bc))
);
// get the pair of indices adjacent to `j`
let adjacent_indices = [j % 3 + 1, (j + 1) % 3 + 1];
for k in adjacent_indices.clone() {
for series in ["b", "c"] {
// put each B and C vertex on the faces it belongs to
@ -383,14 +383,14 @@ fn load_tridiminished_icosahedron(assembly: &Assembly) {
let incidence = InversiveDistanceRegulator::new([face.clone(), vertex.clone()]);
incidence.set_point.set(SpecifiedValue::try_from("0".to_string()).unwrap());
assembly.insert_regulator(Rc::new(incidence));
// regulate the A-B and A-C vertex distances
assembly.insert_regulator(
Rc::new(InversiveDistanceRegulator::new([vertex_a.clone(), vertex]))
);
}
}
// regulate the A-A and C-C vertex distances
let adjacent_pairs = ["a", "c"].map(
|series| adjacent_indices.map(
@ -422,14 +422,14 @@ fn load_dodecahedral_packing(assembly: &Assembly) {
let substrate = assembly.elements_by_id.with_untracked(
|elts_by_id| elts_by_id["substrate"].clone()
);
// fix the substrate's curvature
substrate.regulators().with_untracked(
|regs| regs.first().unwrap().clone()
).set_point().set(
SpecifiedValue::try_from("0.5".to_string()).unwrap()
);
// add the circles to be packed
const COLOR_A: ElementColor = [1.00_f32, 0.25_f32, 0.00_f32];
const COLOR_B: ElementColor = [1.00_f32, 0.00_f32, 0.25_f32];
@ -445,10 +445,10 @@ fn load_dodecahedral_packing(assembly: &Assembly) {
for k in 0..2 {
let small_coord = face_scales[k] * (2.0*(j as f64) - 1.0);
let big_coord = face_scales[k] * (2.0*(k as f64) - 1.0) * phi;
let id_num = format!("{j}{k}");
let label_sub = format!("{}{}", subscripts[j], subscripts[k]);
// add the A face
let id_a = format!("a{id_num}");
let _ = assembly.try_insert_element(
@ -464,7 +464,7 @@ fn load_dodecahedral_packing(assembly: &Assembly) {
|elts_by_id| elts_by_id[&id_a].clone()
)
);
// add the B face
let id_b = format!("b{id_num}");
let _ = assembly.try_insert_element(
@ -480,7 +480,7 @@ fn load_dodecahedral_packing(assembly: &Assembly) {
|elts_by_id| elts_by_id[&id_b].clone()
)
);
// add the C face
let id_c = format!("c{id_num}");
let _ = assembly.try_insert_element(
@ -498,14 +498,14 @@ fn load_dodecahedral_packing(assembly: &Assembly) {
);
}
}
// make each face sphere perpendicular to the substrate
for face in faces {
let right_angle = InversiveDistanceRegulator::new([face, substrate.clone()]);
right_angle.set_point.set(SpecifiedValue::try_from("0".to_string()).unwrap());
assembly.insert_regulator(Rc::new(right_angle));
}
// set up the tangencies that define the packing
for [long_edge_plane, short_edge_plane] in [["a", "b"], ["b", "c"], ["c", "a"]] {
for k in 0..2 {
@ -524,14 +524,14 @@ fn load_dodecahedral_packing(assembly: &Assembly) {
)
)
);
// set up the short-edge tangency
let short_tangency = InversiveDistanceRegulator::new(short_edge.clone());
if k == 0 {
short_tangency.set_point.set(SpecifiedValue::try_from("-1".to_string()).unwrap());
}
assembly.insert_regulator(Rc::new(short_tangency));
// set up the side tangencies
for i in 0..2 {
for j in 0..2 {
@ -577,14 +577,14 @@ fn load_balanced(assembly: &Assembly) {
for sphere in spheres {
let _ = assembly.try_insert_element(sphere);
}
// get references to the spheres
let [outer, a, b] = ["outer", "a", "b"].map(
|id| assembly.elements_by_id.with_untracked(
|elts_by_id| elts_by_id[id].clone()
)
);
// fix the diameters of the outer, sun, and moon spheres
for (sphere, radius) in [
(outer.clone(), R_OUTER),
@ -599,7 +599,7 @@ fn load_balanced(assembly: &Assembly) {
SpecifiedValue::try_from(curvature.to_string()).unwrap()
);
}
// set the inversive distances between the spheres. as described above, the
// initial configuration deliberately violates these constraints
for inner in [a, b] {
@ -629,14 +629,14 @@ fn load_off_center(assembly: &Assembly) {
engine::sphere(0.0, 0.0, 0.0, 1.0),
),
);
// get references to the elements
let point_and_sphere = ["point", "sphere"].map(
|id| assembly.elements_by_id.with_untracked(
|elts_by_id| elts_by_id[id].clone()
)
);
// put the point on the sphere
let incidence = InversiveDistanceRegulator::new(point_and_sphere);
incidence.set_point.set(SpecifiedValue::try_from("0".to_string()).unwrap());
@ -650,7 +650,7 @@ fn load_off_center(assembly: &Assembly) {
// inversive distance of 0 between the circumsphere and each vertex
fn load_radius_ratio(assembly: &Assembly) {
let index_range = 1..=4;
// create the spheres
const GRAY: ElementColor = [0.75_f32, 0.75_f32, 0.75_f32];
let spheres = [
@ -670,7 +670,7 @@ fn load_radius_ratio(assembly: &Assembly) {
for sphere in spheres {
let _ = assembly.try_insert_element(sphere);
}
// create the vertices
let vertices = izip!(
index_range.clone(),
@ -699,7 +699,7 @@ fn load_radius_ratio(assembly: &Assembly) {
for vertex in vertices {
let _ = assembly.try_insert_element(vertex);
}
// create the faces
let base_dir = Vector3::new(1.0, 0.75, 1.0).normalize();
let offset = base_dir.dot(&Vector3::new(-0.6, 0.8, 0.6));
@ -731,7 +731,7 @@ fn load_radius_ratio(assembly: &Assembly) {
face.ghost().set(true);
let _ = assembly.try_insert_element(face);
}
// impose the constraints
for j in index_range.clone() {
let [face_j, vertex_j] = [
@ -742,7 +742,7 @@ fn load_radius_ratio(assembly: &Assembly) {
|elts_by_id| elts_by_id[&id].clone()
)
);
// make the faces planar
let curvature_regulator = face_j.regulators().with_untracked(
|regs| regs.first().unwrap().clone()
@ -750,12 +750,12 @@ fn load_radius_ratio(assembly: &Assembly) {
curvature_regulator.set_point().set(
SpecifiedValue::try_from("0".to_string()).unwrap()
);
for k in index_range.clone().filter(|&index| index != j) {
let vertex_k = assembly.elements_by_id.with_untracked(
|elts_by_id| elts_by_id[&format!("v{k}")].clone()
);
// fix the distances between the vertices
if j < k {
let distance_regulator = InversiveDistanceRegulator::new(
@ -763,7 +763,7 @@ fn load_radius_ratio(assembly: &Assembly) {
);
assembly.insert_regulator(Rc::new(distance_regulator));
}
// put the vertices on the faces
let incidence_regulator = InversiveDistanceRegulator::new([face_j.clone(), vertex_k.clone()]);
incidence_regulator.set_point.set(SpecifiedValue::try_from("0".to_string()).unwrap());
@ -799,7 +799,7 @@ fn load_irisawa_hexlet(assembly: &Assembly) {
[0.00_f32, 0.25_f32, 1.00_f32],
[0.25_f32, 0.00_f32, 1.00_f32],
].into_iter();
// create the spheres
let spheres = [
Sphere::new(
@ -836,7 +836,7 @@ fn load_irisawa_hexlet(assembly: &Assembly) {
for sphere in spheres {
let _ = assembly.try_insert_element(sphere);
}
// put the outer sphere in ghost mode and fix its curvature
let outer = assembly.elements_by_id.with_untracked(
|elts_by_id| elts_by_id["outer"].clone()
@ -848,7 +848,7 @@ fn load_irisawa_hexlet(assembly: &Assembly) {
outer_curvature_regulator.set_point().set(
SpecifiedValue::try_from((1.0 / 3.0).to_string()).unwrap()
);
// impose the desired tangencies
let [outer, sun, moon] = ["outer", "sun", "moon"].map(
|id| assembly.elements_by_id.with_untracked(
@ -872,11 +872,11 @@ fn load_irisawa_hexlet(assembly: &Assembly) {
assembly.insert_regulator(Rc::new(tangency));
}
}
let outer_sun_tangency = InversiveDistanceRegulator::new([outer.clone(), sun]);
outer_sun_tangency.set_point.set(SpecifiedValue::try_from("1".to_string()).unwrap());
assembly.insert_regulator(Rc::new(outer_sun_tangency));
let outer_moon_tangency = InversiveDistanceRegulator::new([outer.clone(), moon]);
outer_moon_tangency.set_point.set(SpecifiedValue::try_from("1".to_string()).unwrap());
assembly.insert_regulator(Rc::new(outer_moon_tangency));
@ -895,18 +895,18 @@ pub fn TestAssemblyChooser() -> View {
console::log_1(
&JsValue::from(format!("Showing assembly \"{}\"", name.clone()))
);
batch(|| {
let state = use_context::<AppState>();
let assembly = &state.assembly;
// clear state
assembly.regulators.update(|regs| regs.clear());
assembly.elements.update(|elts| elts.clear());
assembly.elements_by_id.update(|elts_by_id| elts_by_id.clear());
assembly.descent_history.set(DescentHistory::new());
state.selection.update(|sel| sel.clear());
// load assembly
match name.as_str() {
"general" => load_general(assembly),
@ -922,7 +922,7 @@ pub fn TestAssemblyChooser() -> View {
};
});
});
// build the chooser
view! {
select(bind:value = assembly_name) {
@ -938,4 +938,4 @@ pub fn TestAssemblyChooser() -> View {
option(value = "empty") { "Empty" }
}
}
}
}

View file

@ -62,19 +62,19 @@ impl PartialMatrix {
pub fn new() -> Self {
Self(Vec::<MatrixEntry>::new())
}
pub fn push(&mut self, row: usize, col: usize, value: f64) {
let Self(entries) = self;
entries.push(MatrixEntry { index: (row, col), value });
}
pub fn push_sym(&mut self, row: usize, col: usize, value: f64) {
self.push(row, col, value);
if row != col {
self.push(col, row, value);
}
}
fn freeze(&self, a: &DMatrix<f64>) -> DMatrix<f64> {
let mut result = a.clone();
for &MatrixEntry { index, value } in self {
@ -82,7 +82,7 @@ impl PartialMatrix {
}
result
}
fn proj(&self, a: &DMatrix<f64>) -> DMatrix<f64> {
let mut result = DMatrix::<f64>::zeros(a.nrows(), a.ncols());
for &MatrixEntry { index, .. } in self {
@ -90,7 +90,7 @@ impl PartialMatrix {
}
result
}
fn sub_proj(&self, rhs: &DMatrix<f64>) -> DMatrix<f64> {
let mut result = DMatrix::<f64>::zeros(rhs.nrows(), rhs.ncols());
for &MatrixEntry { index, value } in self {
@ -112,7 +112,7 @@ impl Display for PartialMatrix {
impl IntoIterator for PartialMatrix {
type Item = MatrixEntry;
type IntoIter = std::vec::IntoIter<Self::Item>;
fn into_iter(self) -> Self::IntoIter {
let Self(entries) = self;
entries.into_iter()
@ -122,7 +122,7 @@ impl IntoIterator for PartialMatrix {
impl<'a> IntoIterator for &'a PartialMatrix {
type Item = &'a MatrixEntry;
type IntoIter = std::slice::Iter<'a, MatrixEntry>;
fn into_iter(self) -> Self::IntoIter {
let PartialMatrix(entries) = self;
entries.into_iter()
@ -146,7 +146,7 @@ impl ConfigSubspace {
basis_std: Vec::new(),
}
}
// approximate the kernel of a symmetric endomorphism of the configuration
// space for `assembly_dim` elements. we consider an eigenvector to be part
// of the kernel if its eigenvalue is smaller than the constant `THRESHOLD`
@ -167,10 +167,10 @@ impl ConfigSubspace {
|(λ, v)| (λ.abs() < THRESHOLD).then_some(v)
).collect::<Vec<_>>().as_slice()
);
// express the basis in the standard coordinates
let basis_std = proj_to_std * &basis_proj;
const ELEMENT_DIM: usize = 5;
const UNIFORM_DIM: usize = 4;
Self {
@ -187,15 +187,15 @@ impl ConfigSubspace {
).collect(),
}
}
pub fn dim(&self) -> usize {
self.basis_std.len()
}
pub fn assembly_dim(&self) -> usize {
self.assembly_dim
}
// find the projection onto this subspace of the motion where the element
// with the given column index has velocity `v`. the velocity is given in
// projection coordinates, and the projection is done with respect to the
@ -253,7 +253,7 @@ impl ConstraintProblem {
guess: DMatrix::<f64>::zeros(ELEMENT_DIM, element_count),
}
}
#[cfg(feature = "dev")]
pub fn from_guess(guess_columns: &[DVector<f64>]) -> Self {
Self {
@ -377,10 +377,10 @@ pub fn realize_gram(
) -> Realization {
// destructure the problem data
let ConstraintProblem { gram, guess, frozen } = problem;
// start the descent history
let mut history = DescentHistory::new();
// handle the case where the assembly is empty. our general realization
// routine can't handle this case because it builds the Hessian using
// `DMatrix::from_columns`, which panics when the list of columns is empty
@ -394,20 +394,20 @@ pub fn realize_gram(
);
return Realization { result, history };
}
// find the dimension of the search space
let element_dim = guess.nrows();
let total_dim = element_dim * assembly_dim;
// scale the tolerance
let scale_adjustment = (gram.0.len() as f64).sqrt();
let tol = scale_adjustment * scaled_tol;
// convert the frozen indices to stacked format
let frozen_stacked: Vec<usize> = frozen.into_iter().map(
|MatrixEntry { index: (row, col), .. }| col*element_dim + row
).collect();
// use a regularized Newton's method with backtracking
let mut state = SearchState::from_config(gram, frozen.freeze(guess));
let mut hess = DMatrix::zeros(element_dim, assembly_dim);
@ -416,7 +416,7 @@ pub fn realize_gram(
let neg_grad = 4.0 * &*Q * &state.config * &state.err_proj;
let mut neg_grad_stacked = neg_grad.clone().reshape_generic(Dyn(total_dim), Const::<1>);
history.neg_grad.push(neg_grad.clone());
// find the negative Hessian of the loss function
let mut hess_cols = Vec::<DVector<f64>>::with_capacity(total_dim);
for col in 0..assembly_dim {
@ -435,7 +435,7 @@ pub fn realize_gram(
}
}
hess = DMatrix::from_columns(hess_cols.as_slice());
// regularize the Hessian
let hess_eigvals = hess.symmetric_eigenvalues();
let min_eigval = hess_eigvals.min();
@ -443,7 +443,7 @@ pub fn realize_gram(
hess -= reg_scale * min_eigval * DMatrix::identity(total_dim, total_dim);
}
history.hess_eigvals.push(hess_eigvals);
// project the negative gradient and negative Hessian onto the
// orthogonal complement of the frozen subspace
let zero_col = DVector::zeros(total_dim);
@ -454,12 +454,12 @@ pub fn realize_gram(
hess.set_column(k, &zero_col);
hess[(k, k)] = 1.0;
}
// stop if the loss is tolerably low
history.config.push(state.config.clone());
history.scaled_loss.push(state.loss / scale_adjustment);
if state.loss < tol { break; }
// compute the Newton step
/* TO DO */
/*
@ -479,7 +479,7 @@ pub fn realize_gram(
let base_step_stacked = hess_cholesky.solve(&neg_grad_stacked);
let base_step = base_step_stacked.reshape_generic(Dyn(element_dim), Dyn(assembly_dim));
history.base_step.push(base_step.clone());
// use backtracking line search to find a better configuration
if let Some((better_state, backoff_steps)) = seek_better_config(
gram, &state, &base_step, neg_grad.dot(&base_step),
@ -505,10 +505,10 @@ pub fn realize_gram(
.view_mut(block_start, (element_dim, UNIFORM_DIM))
.copy_from(&local_unif_to_std(state.config.column(n)));
}
// find the kernel of the Hessian. give it the uniform inner product
let tangent = ConfigSubspace::symmetric_kernel(hess, unif_to_std, assembly_dim);
Ok(ConfigNeighborhood { #[cfg(feature = "dev")] config: state.config, nbhd: tangent })
} else {
Err("Failed to reach target accuracy".to_string())
@ -521,9 +521,9 @@ pub fn realize_gram(
#[cfg(feature = "dev")]
pub mod examples {
use std::f64::consts::PI;
use super::*;
// this problem is from a sangaku by Irisawa Shintarō Hiroatsu. the article
// below includes a nice translation of the problem statement, which was
// recorded in Uchida Itsumi's book _Kokon sankan_ (_Mathematics, Past and
@ -547,40 +547,40 @@ pub mod examples {
)
).collect::<Vec<_>>().as_slice()
);
for s in 0..9 {
// each sphere is represented by a spacelike vector
problem.gram.push_sym(s, s, 1.0);
// the circumscribing sphere is tangent to all of the other
// spheres, with matching orientation
if s > 0 {
problem.gram.push_sym(0, s, 1.0);
}
if s > 2 {
// each chain sphere is tangent to the "sun" and "moon"
// spheres, with opposing orientation
for n in 1..3 {
problem.gram.push_sym(s, n, -1.0);
}
// each chain sphere is tangent to the next chain sphere,
// with opposing orientation
let s_next = 3 + (s-2) % 6;
problem.gram.push_sym(s, s_next, -1.0);
}
}
// the frozen entries fix the radii of the circumscribing sphere, the
// "sun" and "moon" spheres, and one of the chain spheres
for k in 0..4 {
problem.frozen.push(3, k, problem.guess[(3, k)]);
}
realize_gram(&problem, scaled_tol, 0.5, 0.9, 1.1, 200, 110)
}
// set up a kaleidocycle, made of points with fixed distances between them,
// and find its tangent space
pub fn realize_kaleidocycle(scaled_tol: f64) -> Realization {
@ -601,7 +601,7 @@ pub mod examples {
}
).collect::<Vec<_>>().as_slice()
);
const N_POINTS: usize = 2 * N_HINGES;
for block in (0..N_POINTS).step_by(2) {
let block_next = (block + 2) % N_POINTS;
@ -610,18 +610,18 @@ pub mod examples {
for k in j..2 {
problem.gram.push_sym(block + j, block + k, if j == k { 0.0 } else { -0.5 });
}
// non-hinge edges
for k in 0..2 {
problem.gram.push_sym(block + j, block_next + k, -0.625);
}
}
}
for k in 0..N_POINTS {
problem.frozen.push(3, k, problem.guess[(3, k)])
}
realize_gram(&problem, scaled_tol, 0.5, 0.9, 1.1, 200, 110)
}
}
@ -630,9 +630,9 @@ pub mod examples {
mod tests {
use nalgebra::Vector3;
use std::{f64::consts::{FRAC_1_SQRT_2, PI}, iter};
use super::{*, examples::*};
#[test]
fn freeze_test() {
let frozen = PartialMatrix(vec![
@ -651,7 +651,7 @@ mod tests {
]);
assert_eq!(frozen.freeze(&config), expected_result);
}
#[test]
fn sub_proj_test() {
let target = PartialMatrix(vec![
@ -670,7 +670,7 @@ mod tests {
]);
assert_eq!(target.sub_proj(&attempt), expected_result);
}
#[test]
fn zero_loss_test() {
let mut gram = PartialMatrix::new();
@ -690,7 +690,7 @@ mod tests {
let state = SearchState::from_config(&gram, config);
assert!(state.loss.abs() < f64::EPSILON);
}
/* TO DO */
// at the frozen indices, the optimization steps should have exact zeros,
// and the realized configuration should have the desired values
@ -720,13 +720,13 @@ mod tests {
assert_eq!(config[index], value);
}
}
#[test]
fn irisawa_hexlet_test() {
// solve Irisawa's problem
const SCALED_TOL: f64 = 1.0e-12;
let config = realize_irisawa_hexlet(SCALED_TOL).result.unwrap().config;
// check against Irisawa's solution
let entry_tol = SCALED_TOL.sqrt();
let solution_diams = [30.0, 10.0, 6.0, 5.0, 15.0, 10.0, 3.75, 2.5, 2.0 + 8.0/11.0];
@ -734,7 +734,7 @@ mod tests {
assert!((config[(3, k)] - 1.0 / diam).abs() < entry_tol);
}
}
#[test]
fn tangent_test_three_spheres() {
const SCALED_TOL: f64 = 1.0e-12;
@ -758,7 +758,7 @@ mod tests {
let ConfigNeighborhood { config, nbhd: tangent } = result.unwrap();
assert_eq!(config, problem.guess);
assert_eq!(history.scaled_loss.len(), 1);
// list some motions that should form a basis for the tangent space of
// the solution variety
const UNIFORM_DIM: usize = 4;
@ -786,11 +786,11 @@ mod tests {
0.0, 0.0, -1.0, 0.25, 1.0,
]),
];
// confirm that the dimension of the tangent space is no greater than
// expected
assert_eq!(tangent.basis_std.len(), tangent_motions_std.len());
// confirm that the tangent space contains all the motions we expect it
// to. since we've already bounded the dimension of the tangent space,
// this confirms that the tangent space is what we expect it to be
@ -802,13 +802,13 @@ mod tests {
assert!((motion_std - motion_proj).norm_squared() < tol_sq);
}
}
fn translation_motion_unif(vel: &Vector3<f64>, assembly_dim: usize) -> Vec<DVector<f64>> {
let mut elt_motion = DVector::zeros(4);
elt_motion.fixed_rows_mut::<3>(0).copy_from(vel);
iter::repeat(elt_motion).take(assembly_dim).collect()
}
fn rotation_motion_unif(ang_vel: &Vector3<f64>, points: Vec<DVectorView<f64>>) -> Vec<DVector<f64>> {
points.into_iter().map(
|pt| {
@ -819,7 +819,7 @@ mod tests {
}
).collect()
}
#[test]
fn tangent_test_kaleidocycle() {
// set up a kaleidocycle and find its tangent space
@ -827,7 +827,7 @@ mod tests {
let Realization { result, history } = realize_kaleidocycle(SCALED_TOL);
let ConfigNeighborhood { config, nbhd: tangent } = result.unwrap();
assert_eq!(history.scaled_loss.len(), 1);
// list some motions that should form a basis for the tangent space of
// the solution variety
const N_HINGES: usize = 6;
@ -838,12 +838,12 @@ mod tests {
translation_motion_unif(&Vector3::new(1.0, 0.0, 0.0), assembly_dim),
translation_motion_unif(&Vector3::new(0.0, 1.0, 0.0), assembly_dim),
translation_motion_unif(&Vector3::new(0.0, 0.0, 1.0), assembly_dim),
// the rotations about the coordinate axes
rotation_motion_unif(&Vector3::new(1.0, 0.0, 0.0), config.column_iter().collect()),
rotation_motion_unif(&Vector3::new(0.0, 1.0, 0.0), config.column_iter().collect()),
rotation_motion_unif(&Vector3::new(0.0, 0.0, 1.0), config.column_iter().collect()),
// the twist motion. more precisely: a motion that keeps the center
// of mass stationary and preserves the distances between the
// vertices to first order. this has to be the twist as long as:
@ -872,11 +872,11 @@ mod tests {
).collect::<Vec<_>>()
)
).collect::<Vec<_>>();
// confirm that the dimension of the tangent space is no greater than
// expected
assert_eq!(tangent.basis_std.len(), tangent_motions_unif.len());
// confirm that the tangent space contains all the motions we expect it
// to. since we've already bounded the dimension of the tangent space,
// this confirms that the tangent space is what we expect it to be
@ -888,7 +888,7 @@ mod tests {
assert!((motion_std - motion_proj).norm_squared() < tol_sq);
}
}
fn translation(dis: Vector3<f64>) -> DMatrix<f64> {
const ELEMENT_DIM: usize = 5;
DMatrix::from_column_slice(ELEMENT_DIM, ELEMENT_DIM, &[
@ -899,7 +899,7 @@ mod tests {
0.0, 0.0, 0.0, 0.0, 1.0,
])
}
// confirm that projection onto a configuration subspace is equivariant with
// respect to Euclidean motions
#[test]
@ -919,7 +919,7 @@ mod tests {
let ConfigNeighborhood { config: config_orig, nbhd: tangent_orig } = result_orig.unwrap();
assert_eq!(config_orig, problem_orig.guess);
assert_eq!(history_orig.scaled_loss.len(), 1);
// find another pair of spheres that meet at 120°. we'll think of this
// solution as a transformed version of the original one
let guess_tfm = {
@ -940,17 +940,17 @@ mod tests {
let ConfigNeighborhood { config: config_tfm, nbhd: tangent_tfm } = result_tfm.unwrap();
assert_eq!(config_tfm, problem_tfm.guess);
assert_eq!(history_tfm.scaled_loss.len(), 1);
// project a nudge to the tangent space of the solution variety at the
// original solution
let motion_orig = DVector::from_column_slice(&[0.0, 0.0, 1.0, 0.0]);
let motion_orig_proj = tangent_orig.proj(&motion_orig.as_view(), 0);
// project the equivalent nudge to the tangent space of the solution
// variety at the transformed solution
let motion_tfm = DVector::from_column_slice(&[FRAC_1_SQRT_2, 0.0, FRAC_1_SQRT_2, 0.0]);
let motion_tfm_proj = tangent_tfm.proj(&motion_tfm.as_view(), 0);
// take the transformation that sends the original solution to the
// transformed solution and apply it to the motion that the original
// solution makes in response to the nudge
@ -964,7 +964,7 @@ mod tests {
]);
let transl = translation(Vector3::new(0.0, 0.0, 7.0));
let motion_proj_tfm = transl * rot * motion_orig_proj;
// confirm that the projection of the nudge is equivariant. we loosen
// the comparison tolerance because the transformation seems to
// introduce some numerical error
@ -972,4 +972,4 @@ mod tests {
let tol_sq = ((problem_orig.guess.nrows() * problem_orig.guess.ncols()) as f64) * SCALED_TOL_TFM * SCALED_TOL_TFM;
assert!((motion_proj_tfm - motion_tfm_proj).norm_squared() < tol_sq);
}
}
}

View file

@ -30,7 +30,7 @@ impl AppState {
selection: create_signal(BTreeSet::default()),
}
}
// in single-selection mode, select the given element. in multiple-selection
// mode, toggle whether the given element is selected
fn select(&self, element: &Rc<dyn Element>, multi: bool) {
@ -53,10 +53,10 @@ fn main() {
// set the console error panic hook
#[cfg(feature = "console_error_panic_hook")]
console_error_panic_hook::set_once();
sycamore::render(|| {
provide_context(AppState::new());
view! {
div(id = "sidebar") {
AddRemove {}
@ -66,4 +66,4 @@ fn main() {
Display {}
}
});
}
}

View file

@ -20,7 +20,7 @@ impl SpecifiedValue {
pub fn from_empty_spec() -> Self {
Self { spec: String::new(), value: None }
}
pub fn is_present(&self) -> bool {
matches!(self.value, Some(_))
}
@ -42,7 +42,7 @@ impl From<Option<f64>> for SpecifiedValue {
// if the specification is properly formatted, and `Error` if not
impl TryFrom<String> for SpecifiedValue {
type Error = ParseFloatError;
fn try_from(spec: String) -> Result<Self, Self::Error> {
if spec.is_empty() {
Ok(Self::from_empty_spec())
@ -52,4 +52,4 @@ impl TryFrom<String> for SpecifiedValue {
)
}
}
}
}

View file

@ -33,7 +33,7 @@ The unification of spheres/planes is indeed attractive for a project like Dyna3.
Discussed coordinates with Alex Kontorovich. He was suggesting "inversive coordinates" -- for a sphere, that's 1/coradius, 1/radius, center/radius (where coradius is radius of sphere inverted in the unit sphere.) The advantage is tangent to and perpendicular to are linear in these coordinates (in the sense that if one is known, the condition of being tangent to or perpendicular to that one are linear). Planes have 1/radius = 0, and in fact, you can take the coordinates to be (2s, 0, x, y, z) where s is the distance to the origin and x,y,z are the normal direction. (Note the normal direction is only determined up to a scalar multiple. So could always scale so that the first non-zero coordinate is 1, or if you like only allow x, y to vary and let z be determined as sqrt(1-x^2^-y^2^). ) Points can be given by (r^2,1,x,y,z) where x,y,z are the coordinates and r is the distance to the origin. Quadratic form that tells you if something is a sphere/plane, or in the boundary, or up in the hyperbolic plane above. There are some details, but not quite explicit for modeling R^3, at http://sites.math.rutgers.edu/~alexk/files/LetterToDuke.pdf -- all this emphasize need to be agnostic with respect to geometric model so that we can experiment. Not really sure exactly how this relates or not to conformal geometric algebra, and whether it can be combined with geometric algebra. As formulated, there are clear-ish reps for planes/spheres and for points, but not as clear for lines. Have to see how to compute distance and/or specify a given distance. To combine inversive coordinates and geometric algebra, maybe think dually; there should be a lift from a normal vector and distance from origin to the five-vector; bivectors would rep circles/lines; trivectors would rep point pairs/points. What is the signature of this algebra, i.e. how many coordinates square to +1, -1, or 0? But it doesn't seem worth it for three dimensions, because there is a natural representation of points, as follows:
The signature of Q will be (1,4), and in fact Q(I1,I2) = 1/2(ab+ba) - E1\dot E2, where a is the "first" or "coradius" coordinate, "b" is the "second" or "radius" coordinate, and E is the Euclidean part (x,y,z). Then the inversive coordinates of a sphere with center (x,y,z) and radius r will be I = (1/\hat{r},1/r,x/r,y/r,z/r) where \hat{r} = r/(|E|^2 -r^2). These coordinates satisfy Q(I,I) = -1. For this to make sense, of course r > 0, but we get planes by letting the radius of a tangent sphere to the plane go to infinity, and we get I = (2s, 0, x0, y0, z0) where (x0,y0,z0) is the unit normal to the plane and s is the perpendicular distance from the plane to the origin. Still Q(I,I) = -1.
Since r>0, we can't represent individual points this way. Instead we will use some coordinates J for which Q(J,J) = 0. In particular, if you take for the Euclidean point E = (u,v,w) the coordinates J = (`|E|`^2,1,u,v,w) then Q(J,J) = 0 and moreover it comes out that Q(I,J) = 0
Since r>0, we can't represent individual points this way. Instead we will use some coordinates J for which Q(J,J) = 0. In particular, if you take for the Euclidean point E = (u,v,w) the coordinates J = (`|E|`^2,1,u,v,w) then Q(J,J) = 0 and moreover it comes out that Q(I,J) = 0
whenever E lies on the sphere or plane described by some I with Q(I,I) = -1.
The condition that two spheres I1 and I2 are tangent seems to be that Q(I1,I2) = 1. So given a fixed sphere, the condition that another sphere be tangent to it is linear in the coordinates of that other sphere.
This system does seem promising for encoding points, spheres, and planes, and doing basic computations with them. I guess I would just encode a circle as the intersection of the concentric sphere and the containing plane, and a line as either a pair of points or a pair of planes (modulo some equivalence relation, since I can't see any canonical choice of either two planes or two points). Or actually as described below, there is a more canonical choice.
@ -62,4 +62,4 @@ In the engine's coordinate conventions, a sphere with radius $r > 0$ centered on
$$I'_s = \left(\frac{P_x}{r}, \frac{P_y}{r}, \frac{P_z}{r}, \frac1{2r}, \frac{\|P\|^2 - r^2}{2r}\right),$$
which has the normalization $Q'(I'_s, I'_s) = 1$. The point $P$ is represented by the vector
$$I'_P = \left(P_x, P_y, P_z, \frac{1}{2}, \frac{\|P\|^2}{2}\right).$$
In the `engine` module, these formulas are encoded in the `sphere` and `point` functions.
In the `engine` module, these formulas are encoded in the `sphere` and `point` functions.

View file

@ -22,9 +22,10 @@ Jürgen also emphasized the need for an intuitive user interface. Notes on that
His final mathematical advice was reasonably encouraging, however:
"But still I would consider it all more or less doable. One should very precisely think about a doable scope.
I think three things are essential for the math no matter what you exactly plan.
I think three things are essential for the math no matter what you exactly
plan.
1. Think projectively,
1. Think projectively.
Use Projective Geometry, Homogeneous Coordinates (or to a certain extent Quaternions, and Clifford Algebras, which are more or less an elegant way to merge Complex numbers with projective concepts.)
2. Consider ambient complex spaces.
The true nature of the objects can only be understood if embedded into a complex ambient space.
@ -37,10 +38,8 @@ I think three things are essential for the math no matter what you exactly plan.
It would be nice to see how Jürgen handled some of these issues in a 2D system that he designed. Unfortunately, Cinderella was and remains closed-source; it was distributed for profit for some stretch of time. However, (a part of?) it was reimplemented in JavaScript as CindyJS, which is open source. I took a relatively quick look at that source code at one point, and these were my observations:
CindyJS uses very concrete basic objects: 2D points are represented via projective geometry as a list of three floating-point numbers, and everything is done numerically. There are no symbolic representations or exact algebraic numbers. (Not sure how a point on a circle or line is handled, that would take further investigation.)
CindyJS uses very concrete basic objects: 2D points are represented via projective geometry as a list of three floating-point numbers, and everything is done numerically. There are no symbolic representations or exact algebraic numbers. (Not sure how a point on a circle or line is handled, that would take further investigation.)
Lines are given by explicit coordinates as well (not sure of the internal details/exact coordinatization, or of how a "LineThrough" is represented).
Was unclear to me how the complex parametrization for preserving continuity was handled in the code, even though Jürgen harps on complex ambient spaces; where are the complex numbers? Perhaps that part of Cinderella was never re-implemented?

View file

@ -7,5 +7,3 @@
<body><script type="module" src="dyna3.js"></script>
</body>
</html>