One Edge Endpoint Germ Positive Ray Distinct
ProvedOneEdgeEndpointGermPositiveRayDistinctLet be written as
for a finite vertex set and a finite set of nondegenerate closed straight edges. Assume distinct old edge relative interiors are disjoint. Let be a nondegenerate new segment whose relative interior is disjoint from .
At the endpoint , index the endpoint germs by one new germ in the direction , together with one old germ for each old edge of incident with , directed from toward its other endpoint. Then all these germ vectors are nonzero, and no two distinct indexed germs determine the same positive ray from .
Formalization Note This node is ported from the Trellis formalization of the crossing lemma and its consequences (wpegden/crossing-consequences, commit 8769d142, W. Pegden, Apache-2.0); the statement is the Trellis node OneEdgeEndpointGermPositiveRayDistinct.
import Mathlib.Tactic import Mathlib.Analysis.InnerProductSpace.PiL2 import Mathlib.Analysis.SpecialFunctions.Pow.Real import Mathlib.Combinatorics.SimpleGraph.Finite import Mathlib.Data.Finset.Prod import Mathlib.Data.Set.Finite.Basic import Mathlib.LinearAlgebra.AffineSpace.AffineSubspace.Basic open Classical noncomputable section
lemma OneEdgeEndpointGermPositiveRayDistinct
(A : Set (EuclideanSpace ℝ (Fin 2)))
(V : Finset (EuclideanSpace ℝ (Fin 2)))
(E : Finset (EuclideanSpace ℝ (Fin 2) × EuclideanSpace ℝ (Fin 2)))
(p q : EuclideanSpace ℝ (Fin 2))
(hA :
A =
(V : Set (EuclideanSpace ℝ (Fin 2))) ∪
⋃ e : {e // e ∈ E}, segment ℝ e.1.1 e.1.2)
(hEdgeNondegenerate :
∀ e : EuclideanSpace ℝ (Fin 2) × EuclideanSpace ℝ (Fin 2),
e ∈ E → e.1 ≠ e.2)
(hEdgeOpenInteriorsDisjoint :
∀ e f : EuclideanSpace ℝ (Fin 2) × EuclideanSpace ℝ (Fin 2),
e ∈ E → f ∈ E → e ≠ f →
Disjoint (openSegment ℝ e.1 e.2) (openSegment ℝ f.1 f.2))
(hpq : p ≠ q)
(hNewInteriorDisjoint : Disjoint (openSegment ℝ p q) A) :
let Incident :=
{e : {e // e ∈ E} // e.1.1 = p ∨ e.1.2 = p}
let u : Option Incident → EuclideanSpace ℝ (Fin 2) :=
fun i =>
match i with
| none => q - p
| some e =>
if e.1.1.1 = p then e.1.1.2 - p else e.1.1.1 - p
(∀ i : Option Incident, u i ≠ 0) ∧
(∀ {i j : Option Incident},
(∃ t : ℝ, 0 < t ∧ u j = t • u i) → i = j) := by sorry