WitnessCofaceSimplexStream

org.appliedtopology.tda4j.streams.WitnessCofaceSimplexStream
See theWitnessCofaceSimplexStream companion object
class WitnessCofaceSimplexStream(val geometry: WitnessGeometry, maxFiltrationValue: Double = ..., keepCriterion: PartialFunction[Simplex[Int], Boolean] = ...) extends RipserCofaceSimplexStream

The general witness complex (De Silva & Carlsson 2004; JavaPlex's plain WitnessStream): unlike the lazy variant above, NOT a flag complex -- a higher simplex's own witness condition uses a DIMENSION-SPECIFIC threshold m_k (the (k+1)-th nearest landmark, k = the simplex's own dimension) that need not be monotone facet-to-coface on its own, so every simplex's recorded filtration value is max(own_k(sigma), max over its own facets' filtration values) (JavaPlex's own addCofaces_: filtrationIndex = max(filtrationIndex, ...) over the boundary, then maxed again with the simplex's own witness value). This recursive max is what makes "the complex at threshold R" automatically downward-closed for every R -- the same way VR's own "max pairwise distance" does -- and it also makes JavaPlex's separate containsElement(face) gate redundant here (any facet whose own value exceeds a threshold forces its coface's value above that threshold too, via the max): see .claude/WORKLOG-witness-complex.md for the proof. So, unlike the eager reference implementation, RipserCofaceSimplexStream's plain "generate from the canonical (min-vertex-removed) facet, filter by filtrationValue <= threshold" shape is already correct once fed this recursive filtration value -- no extra "are all my facets already accepted" check needed.

nu plays no role here (each dimension has its own fixed m_k, not a caller-chosen parameter) -- the WitnessMetricSpace handed to the superclass exists only to satisfy RipserCofaceSimplexStream's constructor; its distance is never actually read (see that class's own "lazy, not eager" note), since filtrationValueOverride replaces filtrationValue for every dimension including edges, and maxFiltrationValue is always supplied explicitly (default +Infinity, i.e. untruncated) rather than left None -- None would fall back to minimumEnclosingRadius, NOT a valid truncation for a non-flag complex like this one (see .claude/WORKLOG-witness-complex.md).

'''Fact used to cross-validate against the lazy stream''' (WitnessStreamSpec): this class's own 1-skeleton is IDENTICAL to LazyWitnessSimplexStream(..., nu = 2)'s -- both use the 2nd-nearest-landmark threshold for edges (m_1 here, m_nu there with its sentinel-shifted index), reached via different code paths.

Built via the companion apply (below), not new, so the shared WitnessGeometry (O(L*N) to build) is computed exactly once and reused both for the superclass's WitnessMetricSpace and for recursiveFiltrationValue -- constructing it twice from raw (ambientMetricSpace, landmarks) would silently duplicate that work.

'''Performance hazard, standing for any unbounded coface stream, not specific to this one''': iterating this stream at its default maxFiltrationValue = +Infinity enumerates EVERY dimension up to landmarks.size - 1 -- the full 2^L power set for L landmarks, since nothing about the recursive filtration value ever prunes a candidate at an unbounded threshold. matlab.TDA4j's own facade avoids this by always wrapping in LimitedCofaceSimplexStream (its maxDimension option defaults to reporting H_0..H_2, i.e. simplices up to 4 vertices); a caller driving this class directly should do the same, or pass a finite maxFiltrationValue -- see .claude/WORKLOG-witness-complex.md for tutorial-scale timing measurements (machine-specific, kept there rather than here).

Attributes

Companion
object
Experimental
true
Graph
Supertypes
trait CofaceSimplexStream[Int, Double]
trait StratifiedSimplexStream[Int, Double]
trait StratifiedCellStream[Simplex[Int], Double]
trait CellStream[Simplex[Int], Double]
trait IterableOnce[Simplex[Int]]
trait Filtration[Simplex[Int], Double]
trait Filterable[Double]
class Object
trait Matchable
class Any
Show all

Members list

Value members

Concrete methods

def landmarks: IndexedSeq[Int]

Inherited methods

override def iterateDimension: PartialFunction[Int, Iterator[Simplex[Int]]]

Contract .iterator below relies on: the domain must be contiguous starting at 0 -- defined for 0, 1, ..., k for some k (or empty, or all of the non-negative integers), never with a gap. .iterator stops at the first dimension this is undefined for, so a non-contiguous domain (defined at d but not at d - 1) would silently truncate iteration instead of skipping the gap. Every implementation in this codebase already satisfies this (a simplicial complex can't have a d-simplex without its (d-1)-dimensional faces, so "no cells at d" implies "no cells at any dimension beyond d" too); a new implementation must preserve it.

Contract .iterator below relies on: the domain must be contiguous starting at 0 -- defined for 0, 1, ..., k for some k (or empty, or all of the non-negative integers), never with a gap. .iterator stops at the first dimension this is undefined for, so a non-contiguous domain (defined at d but not at d - 1) would silently truncate iteration instead of skipping the gap. Every implementation in this codebase already satisfies this (a simplicial complex can't have a d-simplex without its (d-1)-dimensional faces, so "no cells at d" implies "no cells at any dimension beyond d" too); a new implementation must preserve it.

Attributes

Definition Classes
Inherited from:
RipserCofaceSimplexStream
override def iterator: Iterator[Simplex[Int]]

Dimension-major: all of dimension d before any of dimension d + 1.

Dimension-major: all of dimension d before any of dimension d + 1.

MUST NOT be implemented as Iterator.from(0).filter(iterateDimension.isDefinedAt)....fold(...) (a real, confirmed bug this replaced -- see .claude/WORKLOG-cohomology.md): Iterator.filter on an infinite source can never prove "no more matches ahead", so once past the last dimension iterateDimension is defined for, it spins forever searching for a d that will never come -- and Int silently wrapping from Int.MaxValue to Int.MinValue after ~2^31 iterations can eventually feed a huge negative d straight to iterateDimension instead, surfacing as a BinomialCoefficient range exception rather than a hang. .takeWhile instead stops at the first d this is undefined for and never asks about any d beyond it, relying on exactly the contiguous-domain contract documented on iterateDimension above.

Attributes

Definition Classes
StratifiedCellStream -> IterableOnce
Inherited from:
StratifiedCellStream
def knownSize: Int

The number of elements in this collection, if it can be cheaply computed, -1 otherwise. Cheaply usually means: Not requiring a collection traversal.

The number of elements in this collection, if it can be cheaply computed, -1 otherwise. Cheaply usually means: Not requiring a collection traversal.

Attributes

Inherited from:
IterableOnce
def stepper[S <: Stepper[_]](implicit shape: StepperShape[Simplex[Int], S]): S

Returns a scala.collection.Stepper for the elements of this collection.

Returns a scala.collection.Stepper for the elements of this collection.

The Stepper enables creating a Java stream to operate on the collection, see scala.jdk.StreamConverters. For collections holding primitive values, the Stepper can be used as an iterator which doesn't box the elements.

The implicit scala.collection.StepperShape parameter defines the resulting Stepper type according to the element type of this collection.

  • For collections of Int, Short, Byte or Char, an scala.collection.IntStepper is returned
  • For collections of Double or Float, a scala.collection.DoubleStepper is returned
  • For collections of Long a scala.collection.LongStepper is returned
  • For any other element type, an scala.collection.AnyStepper is returned

Note that this method is overridden in subclasses and the return type is refined to S with EfficientSplit, for example scala.collection.IndexedSeqOps.stepper. For Steppers marked with scala.collection.Stepper.EfficientSplit, the converters in scala.jdk.StreamConverters allow creating parallel streams, whereas bare Steppers can be converted only to sequential streams.

Type parameters

S

the type of the returned Stepper, determined by the implicit StepperShape

Attributes

Inherited from:
IterableOnce

Concrete fields

Inherited fields

var currentDimension: Int

Attributes

Inherited from:
EnumeratingCofaceSimplexStream
var currentDimensionCache: Queue[Simplex[Int]]

Attributes

Inherited from:
EnumeratingCofaceSimplexStream
lazy val edges: Iterable[Simplex[Int]]

Attributes

Inherited from:
EnumeratingCofaceSimplexStream
override val filtrationOrdering: Ordering[Simplex[Int]]

Filtration value, reversed (so smaller-under-this-ordering means YOUNGER, matching SimplexStream's own established convention), then dimension, then COLEXICOGRAPHIC order on the vertex set (via simplexIndexing's own combinatorial-number-system index) -- the "lexicographically refined" tie-break Ripser's own apparent-pairs machinery (Definition 3.2/Proposition 3.9, see RipserCohomologyContext) is defined in terms of, so using it here keeps this stream's ordering consistent with every other Ripser-flavored piece of this codebase, not just internally self-consistent -- deliberately not the plain lexicographic tie-break FilteredSimplexOrdering uses.

Filtration value, reversed (so smaller-under-this-ordering means YOUNGER, matching SimplexStream's own established convention), then dimension, then COLEXICOGRAPHIC order on the vertex set (via simplexIndexing's own combinatorial-number-system index) -- the "lexicographically refined" tie-break Ripser's own apparent-pairs machinery (Definition 3.2/Proposition 3.9, see RipserCohomologyContext) is defined in terms of, so using it here keeps this stream's ordering consistent with every other Ripser-flavored piece of this codebase, not just internally self-consistent -- deliberately not the plain lexicographic tie-break FilteredSimplexOrdering uses.

Fixes a real, previously-confirmed bug (.claude/WORKLOG-cohomology.md): a bare Ordering.by(filtrationValue) has no tie-break at all, so two DIFFERENT simplices tied at the same filtration value compare as equal -- not a total order. This happens by construction on any Vietoris-Rips complex with a triangle, since a triangle's filtration value always equals that of its own longest edge; CellularHomologyContext bakes a stream's filtrationOrdering into Chain.reduceBy's SortedMap, so two cells that compare equal collide as a single map key and the reduction silently garbles pairings for that complex.

iterateDimension sorts each dimension's bucket by filtrationOrdering.reverse -- deliberately .reverse on this SAME Ordering object, not an independently-built "oldest first" comparator: two individually-valid orderings that disagree on tie-break direction let a coface sort before its own tied facet, corrupting Chain.reduceBy's pivot table the same way the no-tie-break bug did. A stream's iteration order and its filtrationOrdering (pivot order) must be THE SAME total order, one the consistent reverse of the other.

Attributes

Inherited from:
EnumeratingCofaceSimplexStream
override val filtrationValue: PartialFunction[Simplex[Int], Double]

Attributes

Inherited from:
EnumeratingCofaceSimplexStream
val largest: Double

Attributes

Inherited from:
DoubleFiltration
var lastDimensionCache: IndexedSeq[Simplex[Int]]

Attributes

Inherited from:
EnumeratingCofaceSimplexStream

Attributes

Inherited from:
EnumeratingCofaceSimplexStream

Attributes

Inherited from:
EnumeratingCofaceSimplexStream
val smallest: Double

Attributes

Inherited from:
DoubleFiltration