cs = flipCandidates Tcs = flipCandidates(A, tri)Two kinds of $(d{+}1)$-sets are considered:
$\bullet$ codim-2 walls of $T$ -- $(d{+}1)$-sets formed as the union of two adjacent maximal simplices of $T$. These support the standard "fine" flips that preserve the vertex set.
$\bullet$ $\sigma \cup \{v\}$ for each maximal simplex $\sigma$ of $T$ and each column $v$ not used in $T$. These are not walls of $T$, but they are the supports of bistellar flips that insert $v$; in non-acyclic configurations (e.g., complete simplicial fans) such insertion flips can be the only non-fine neighbors of $T$.
For each $(d{+}1)$-set $c$, the function returns the signed kernel partition of those columns as a pair $\{neg, pos\}$ of column indices.
Each returned circuit is a candidate input for bistellarFlip, but not all candidates yield a valid flip in $T$; see bistellarFlip. In contrast, orientedCircuits returns all circuits of the underlying point configuration (most of which are not supported on walls of $T$), and flips -- a wrapper around topcomFlips -- returns only those circuits that are actually flippable in $T$.
|
|
|
The object flipCandidates is a method function.
The source of this document is in Triangulations.m2:2161:0.