cellular_raza_concepts/domain.rs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376
use std::collections::{BTreeMap, BTreeSet};
use crate::errors::{BoundaryError, DecomposeError};
/// Provides an abstraction of the physical total simulation domain.
///
/// [cellular_raza](https://github.com/jonaspleyer/cellular_raza) uses domain-decomposition
/// algorithms to split up the computational workload over multiple physical regions.
/// That's why the domain itself is mostly responsible for being deconstructed
/// into smaller [SubDomains](SubDomain) which can then be used to numerically solve our system.
///
/// This trait can be automatically implemented when the [SortCells], [DomainRngSeed],
/// and [DomainCreateSubDomains] are satisfied together with a small number of trait bounds to hash
/// and compare indices.
pub trait Domain<C, S, Ci = Vec<C>> {
/// Subdomains can be identified by their unique [SubDomainIndex](Domain::SubDomainIndex).
/// The backend uses this property to construct a mapping (graph) between subdomains.
type SubDomainIndex;
/// Similarly to the [SubDomainIndex](Domain::SubDomainIndex), voxels can be accessed by
/// their unique index. The backend will use this information to construct a mapping
/// (graph) between voxels inside their respective subdomains.
type VoxelIndex;
/// Deconstructs the [Domain] into its respective subdomains.
///
/// When using the blanket implementation of this function, the following steps are carried
/// out:
/// Its functionality consists of the following steps:
/// 1. Decompose the Domain into [Subdomains](SubDomain)
/// 2. Build a neighbor map between [SubDomains](SubDomain)
/// 3. Sort cells to their respective [SubDomain]
/// However, to increase performance or avoid trait bounds, one can also opt to implement this
/// trait directly.
fn decompose(
self,
n_subdomains: core::num::NonZeroUsize,
cells: Ci,
) -> Result<DecomposedDomain<Self::SubDomainIndex, S, C>, DecomposeError>;
}
/// Manage the current rng seed of a [Domain]
pub trait DomainRngSeed {
// fn set_rng_seed(&mut self, seed: u64);
/// Obtains the current rng seed
fn get_rng_seed(&self) -> u64;
}
/// Generate [SubDomains](SubDomain) from an existing [Domain]
pub trait DomainCreateSubDomains<S> {
/// This should always be identical to [Domain::SubDomainIndex].
type SubDomainIndex;
/// This should always be identical to [Domain::VoxelIndex].
type VoxelIndex;
/// Generates at most `n_subdomains`. This function can also return a lower amount of
/// subdomains but never less than 1.
fn create_subdomains(
&self,
n_subdomains: core::num::NonZeroUsize,
) -> Result<
impl IntoIterator<Item = (Self::SubDomainIndex, S, Vec<Self::VoxelIndex>)>,
DecomposeError,
>;
}
/// Generated by the [decompose](Domain::decompose) method. The backend will know how to
/// deal with this type and crate a working simulation from it.
pub struct DecomposedDomain<I, S, C> {
/// Number of spawned [SubDomains](SubDomain). This number is guaranteed to be
/// smaller or equal to the number may be different to the one given to the
/// [Domain::decompose] method.
/// Such behaviour can result from not being able to construct as many subdomains as desired.
/// Note that this function will attempt to construct more [SubDomains](SubDomain)
/// than available CPUs if given a larger number.
pub n_subdomains: core::num::NonZeroUsize,
/// Vector containing properties of individual [SubDomains](SubDomain).
/// Entries are [Domain::SubDomainIndex], [SubDomain], and a vector of cells.
// TODO can be use another iterator than Vec<(I, S, Vec<C>)>?
pub index_subdomain_cells: Vec<(I, S, Vec<C>)>,
/// Encapsulates how the subdomains are linked to each other.
/// Eg. two subdomains without any boundary will never appear in each others collection
/// of neighbors.
/// For the future, we might opt to change to an undirected graph rather than a [BTreeMap].
pub neighbor_map: BTreeMap<I, BTreeSet<I>>,
/// Initial seed of the simulation for random number generation.
pub rng_seed: u64,
}
/// Subdomains are produced by decomposing a [Domain] into multiple physical regions.
///
/// # Derivation
/// ```
/// # use cellular_raza_concepts::*;
/// struct MySubDomain {
/// x_min: f32,
/// x_max: f32,
/// n: usize,
/// }
///
/// impl SubDomain for MySubDomain {
/// type VoxelIndex = usize;
///
/// fn get_neighbor_voxel_indices(
/// &self,
/// voxel_index: &Self::VoxelIndex
/// ) -> Vec<Self::VoxelIndex> {
/// (voxel_index.saturating_sub(1)..voxel_index.saturating_add(1).min(self.n)+1)
/// .filter(|k| k!=voxel_index)
/// .collect()
/// }
///
/// fn get_all_indices(&self) -> Vec<Self::VoxelIndex> {
/// (0..self.n).collect()
/// }
/// }
///
/// #[derive(SubDomain)]
/// struct MyNewSubDomain {
/// #[Base]
/// base: MySubDomain,
/// }
/// # let _my_sdm = MyNewSubDomain {
/// # base: MySubDomain {
/// # x_min: -20.0,
/// # x_max: -11.0,
/// # n: 20,
/// # }
/// # };
/// # assert_eq!(_my_sdm.get_all_indices(), (0..20).collect::<Vec<_>>());
/// # assert_eq!(_my_sdm.get_neighbor_voxel_indices(&0), vec![1]);
/// # assert_eq!(_my_sdm.get_neighbor_voxel_indices(&3), vec![2,4]);
/// # assert_eq!(_my_sdm.get_neighbor_voxel_indices(&7), vec![6,8]);
/// ```
pub trait SubDomain {
/// Individual Voxels inside each subdomain can be accessed by this index.
type VoxelIndex;
/// Obtains the neighbor voxels of the specified voxel index. This function behaves similarly
/// to [SortCells::get_voxel_index_of] in that it also has to return
/// indices which are in other [SubDomains](SubDomain).
fn get_neighbor_voxel_indices(&self, voxel_index: &Self::VoxelIndex) -> Vec<Self::VoxelIndex>;
// fn apply_boundary(&self, cell: &mut C) -> Result<(), BoundaryError>;
/// Get all voxel indices of this [SubDomain].
fn get_all_indices(&self) -> Vec<Self::VoxelIndex>;
}
/// Assign an [VoxelIndex](SortCells::VoxelIndex) to a given cell.
///
/// This trait is used by the [Domain] and [SubDomain] trait to assign a [Domain::SubDomainIndex]
/// and [SubDomain::VoxelIndex] respectively.
///
/// # [SubDomain]
/// This trait is supposed to return the correct voxel index of the cell
/// even if this index is inside another [SubDomain].
/// This restriction might be lifted in the future but is still
/// required now.
pub trait SortCells<C> {
/// An index which determines to which next smaller unit the cell should be assigned.
type VoxelIndex;
/// If given a cell, we can sort this cell into the corresponding sub unit.
fn get_voxel_index_of(&self, cell: &C) -> Result<Self::VoxelIndex, BoundaryError>;
}
/// Apply boundary conditions to a cells position and velocity.
///
/// # Derivation
/// ```
/// # use cellular_raza_concepts::*;
/// # use cellular_raza_concepts::BoundaryError;
/// struct MyMechanics {
/// x_min: f64,
/// x_max: f64,
/// }
///
/// impl SubDomainMechanics<f64, f64> for MyMechanics {
/// fn apply_boundary(&self, pos: &mut f64, vel: &mut f64) -> Result<(), BoundaryError> {
/// if *pos < self.x_min {
/// *vel = vel.abs();
/// }
/// if *pos > self.x_max {
/// *vel = -vel.abs();
/// }
/// *pos = pos.clamp(self.x_min, self.x_max);
/// Ok(())
/// }
/// }
///
/// #[derive(SubDomain)]
/// struct MySubDomain {
/// #[Mechanics]
/// mechanics: MyMechanics,
/// }
/// # let _my_sdm = MySubDomain {
/// # mechanics: MyMechanics {
/// # x_min: 1.0,
/// # x_max: 33.0,
/// # }
/// # };
/// # let mut pos = 0.0;
/// # let mut vel = - 0.1;
/// # _my_sdm.apply_boundary(&mut pos, &mut vel).unwrap();
/// # assert_eq!(pos, 1.0);
/// # assert_eq!(vel, 0.1);
/// ```
pub trait SubDomainMechanics<Pos, Vel> {
/// If the subdomain has boundary conditions, this function will enforce them onto the cells.
/// For the future, we plan to replace this function to additionally obtain information
/// about the previous and current location of the cell.
fn apply_boundary(&self, pos: &mut Pos, vel: &mut Vel) -> Result<(), BoundaryError>;
}
/// Apply a force on a cell depending on its position and velocity.
///
/// # Derivation
/// ```
/// # use cellular_raza_concepts::*;
/// struct MyForce {
/// damping: f64,
/// }
///
/// impl SubDomainForce<f64, f64, f64> for MyForce {
/// fn calculate_custom_force(&self, pos: &f64, vel: &f64) -> Result<f64, CalcError> {
/// Ok(- self.damping * vel)
/// }
/// }
///
/// #[derive(SubDomain)]
/// struct MySubDomain {
/// #[Force]
/// force: MyForce,
/// }
/// # let _my_sdm = MySubDomain {
/// # force: MyForce {
/// # damping: 0.1,
/// # }
/// # };
/// # let calculated_force = _my_sdm.calculate_custom_force(&0.0, &1.0).unwrap();
/// # assert_eq!(calculated_force, -0.1);
/// ```
pub trait SubDomainForce<Pos, Vel, For> {
///
fn calculate_custom_force(&self, pos: &Pos, vel: &Vel) -> Result<For, crate::CalcError>;
}
/// Describes extracellular reactions and fluid dynamics
///
/// # Derivation
/// ```
/// # use cellular_raza_concepts::*;
///
/// #[derive(Clone, Debug)]
/// struct MyReactions<const N: usize> {
/// values: Vec<f32>,
/// pos: [f32; N],
/// }
///
/// impl<const N: usize> SubDomainReactions<[f32; N], Vec<f32>, f32> for MyReactions<N> {
/// type NeighborValue = Vec<f32>;
/// type BorderInfo = Self;
///
/// fn treat_increments<I, J>(
/// &mut self,
/// neighbors: I,
/// sources: J,
/// ) -> Result<(), CalcError>
/// where
/// I: IntoIterator<Item = Self::NeighborValue>,
/// J: IntoIterator<Item = ([f32; N], Vec<f32>)>,
/// {
/// Ok(())
/// }
///
/// fn update_fluid_dynamics(&mut self, dt: f32) -> Result<(), CalcError> {
/// Ok(())
/// }
///
/// fn get_extracellular_at_pos(&self, pos: &[f32; N]) -> Result<Vec<f32>, CalcError> {
/// Ok(self.values.clone())
/// }
///
/// fn get_neighbor_value(&self, border_info: Self::BorderInfo) -> Self::NeighborValue {
/// self.values.clone()
/// }
///
/// fn get_border_info(&self) -> Self::BorderInfo {
/// self.clone()
/// }
/// }
///
/// #[derive(SubDomain)]
/// struct DerivedSubDomain<const N: usize> {
/// #[Reactions]
/// reactions: MyReactions<N>,
/// }
/// ```
pub trait SubDomainReactions<Pos, Re, Float> {
/// Extracellular value of neighbor
type NeighborValue;
/// Exchanged information to locate neighboring subdomains.
type BorderInfo;
/// Combines increments which have been obtained by neighbors and cell-sources
fn treat_increments<I, J>(&mut self, neighbors: I, sources: J) -> Result<(), crate::CalcError>
where
I: IntoIterator<Item = Self::NeighborValue>,
J: IntoIterator<Item = (Pos, Re)>;
/// Main update function to calculate new values of extracellular concentrations.
fn update_fluid_dynamics(&mut self, dt: Float) -> Result<(), crate::CalcError>;
/// Obtain extracellular concentrations at given point.
fn get_extracellular_at_pos(&self, pos: &Pos) -> Result<Re, crate::CalcError>;
/// Obtains the [SubDomainReactions::NeighborValue] which should be sent to the neighbor which
/// has exposed the given [SubDomainReactions::BorderInfo].
fn get_neighbor_value(&self, border_info: Self::BorderInfo) -> Self::NeighborValue;
/// Obtains the [SubDomainReactions::BorderInfo] used to retrieve the
/// [SubDomainReactions::NeighborValue].
fn get_border_info(&self) -> Self::BorderInfo;
}
/// This trait derives the different aspects of a [SubDomain].
///
/// It serves similarly as the [cellular_raza_concepts_derive::CellAgent] trait to quickly
/// build new structures from already existing functionality.
///
/// | Attribute | Trait | Implemented |
/// | --- | --- |:---:|
/// | `Base` | [SubDomain] | ✅ |
/// | `SortCells` | [SortCells] | ✅ |
/// | `Mechanics` | [SubDomainMechanics] | ✅ |
/// | `Force` | [SubDomainForce] | ✅ |
/// | `Reactions` | [SubDomainReactions] | ❌ |
///
/// # Example Usage
/// ```
/// # use cellular_raza_concepts::*;
/// # struct MySubDomain;
/// # impl SubDomain for MySubDomain {
/// # type VoxelIndex = usize;
/// # fn get_neighbor_voxel_indices(&self, voxel_index: &Self::VoxelIndex) -> Vec<usize> {
/// # Vec::new()
/// # }
/// # fn get_all_indices(&self) -> Vec<Self::VoxelIndex> {
/// # Vec::new()
/// # }
/// # }
/// #[derive(SubDomain)]
/// struct MyDerivedSubDomain {
/// #[Base]
/// s: MySubDomain,
/// }
/// # let derived_subdomain = MyDerivedSubDomain {
/// # s: MySubDomain,
/// # };
/// # let all_indices = derived_subdomain.get_all_indices();
/// # assert_eq!(all_indices.len(), 0);
/// ```
#[doc(inline)]
pub use cellular_raza_concepts_derive::SubDomain;
/// Derives aspects related to the simulation [Domain]
///
/// | Attribute | Trait | Implemented |
/// | -- | --- |:---:|
/// | `Base` | [Domain] | ✅ |
/// | `DomainPartialDerive` | [Domain] | ✅ |
/// | `DomainRngSeed` | [DomainRngSeed] | ✅ |
/// | `DomainCreateSubDomains` | [DomainCreateSubDomains] | ✅ |
/// | `SortCells` | [SortCells] | ✅ |
#[doc(inline)]
pub use cellular_raza_concepts_derive::Domain;