The common datum of two adjacent levels and its summand inclusions #
The labels of two adjacent levels n, n + 1 of the exhaustion are compared through ONE run —
the functor's value on the coproduct of ALL the pieces of both levels, padded to a common
dimension n₁ + n₂ (PieceEmbedding.padLeft, padRight: the two embeddings (i₁, 0), (0, i₂)
into 𝔸ⁿ¹⁺ⁿ² of [Kol07, Lemma 39]), over the union of the reading opens — Kollár's single run on
the disjoint union of a cover [Kol07, Proposition 37, proof]. This module builds that datum and
its two summand inclusions:
padLeftData D m,padRightData D m— the same pieces, every embedding padded (dimensionD.n + m,m + D.n);sumData D D' h— the pieces of both data, indexed byD.ι ⊕ D'.ι, overU ∪ U'(the embeddings ofD'transported alongh : D.n = D'.n);sumPadData D₁ D₂— THE COMMON DATUM (a cast-free literal, equal field by field tosumData (D₁.padLeftData D₂.n) (D₂.padRightData D₁.n) rfl:sumPadData_eq_sumData), whose single run (sigmaRun) is the common run and whose label set (sigmaIndex) is the common label set;sumInlAmbient D D' h : D.sigmaAmbient → (sumData D D' h).sigmaAmbient— the inclusion of the first summand's coproduct ambient, the descent (sigmaDescMap) of the summand inclusionssigmaMk (Sum.inl i); it is an analytic open embedding (isAnalyticOpenEmbedding_sumInlAmbient) along which the first datum's coproduct triple is the pull-back of the sum's (isPullbackOf_sigmaTriple_sum, byIdealSheaf.ext_of_comap_sigmaMksummand by summand); at the common datum the two summand inclusionssumPadInlAmbient D₁ D₂,sumInrAmbient D₁ D₂likewise;- the reading opens of the common run from those of the two padded data (
sumPadOpens), with the first datum's ambient image carried into the sum's (image_sumInlAmbient_ambImage_subset); - the label maps —
exists_orderEmbedding_comap_sumInlResIn_sigmaMembersand itsinrcounterpart: the GENERAL restriction identity ofRunFamilyRestrict.lean([Kol07, 34.1]) at the summand inclusion, read on the datum'ssigmaMembers: an order embedding of the padded level's label set into the common label set, the common run's member of labellab kpulling back along the open immersionsumInlResInof the local resolutions to the level's memberk, the members off the range to⊤.
The padding identity — the padded level's labels and members against the level's own
(exists_isIso_localResolution_padAlong at the coproduct slice) — is the remaining link of the
comparison of the two levels; it is CoproductPadIdentity.lean. hbed is the only hypothesis.
Not in the sources beyond Kollár's single run; bookkeeping.
The left-padded datum ([Kol07, Lemma 39]) — the same pieces, each embedding padded by m
trailing coordinates (PieceEmbedding.padLeft); dimension D.n + m.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The right-padded datum — each embedding padded by m leading coordinates
(PieceEmbedding.padRight); dimension m + D.n.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The sum of two data of the same dimension over the union of their opens
([Kol07, Proposition 37, proof]) — the pieces of both, indexed by the sum type, the second datum's
embeddings transported along h.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The common padded datum of two adjacent levels — D₁ padded on the right by D₂.n
coordinates and D₂ padded on the left by D₁.n, both of dimension D₁.n + D₂.n, summed; its
single run is the common run. Written as a literal (not through sumData … rfl): the two padded
dimensions are the same term D₁.n + D₂.n, so no transport of the second summand's embeddings is
needed — sumData's h ▸ on the inr summand made every identification of the sum's inr
pieces with the padded pieces a whnf timeout at the level of morphisms of local resolutions (the
comparison of adjacent levels); sumPadData_eq_sumData records that the literal is
sumData … rfl — field by field by rfl, the match auxiliaries by cases.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The literal IS the generic sum of the two padded data along rfl — every field agrees by
rfl (the transport along rfl reduces); the two match auxiliaries on the index are identified
summand by summand.
The inclusion of D's coproduct ambient into the sum's — the descent of the summand
inclusions sigmaMk (Sum.inl i).
Equations
- One or more equations did not get rendered due to their size.
Instances For
The inclusion after a summand inclusion of D is the summand inclusion of the sum.
The inclusion is an analytic open embedding (the argument of
isCoprodOfOpenEmbeddings_sigmaDescMap) — a local isomorphism (each summand inclusion is one) and
injective (Sigma.mk.inj_iff, the summands' ranges disjoint).
D's coproduct triple is the pull-back of the sum's along the inclusion (the pull-back of
[Kol07, 34.1]) — summand by summand (IdealSheaf.ext_of_comap_sigmaMk,
isPullbackOf_ambientTriple_sigmaTriple on both sides); the divisors are empty.
The inclusion of the first padded level's coproduct ambient into the common datum's — the
descent of the summand inclusions sigmaMk (Sum.inl i) (sumInlAmbient at the literal
sumPadData).
Equations
- One or more equations did not get rendered due to their size.
Instances For
The inl inclusion after a summand inclusion.
The inl inclusion is an analytic open embedding.
The first padded level's coproduct triple is the pull-back of the common datum's along the
inl inclusion.
The inclusion of the second padded level's coproduct ambient into the common datum's — the
descent of the summand inclusions sigmaMk (Sum.inr j).
Equations
- One or more equations did not get rendered due to their size.
Instances For
The inr inclusion after a summand inclusion.
The inr inclusion is an analytic open embedding.
The second padded level's coproduct triple is the pull-back of the common datum's along the
inr inclusion.
The reading opens of the common run from those of the two padded levels.
Equations
- D₁.sumPadOpens D₂ W₁ W₂ (Sum.inl i) = W₁ i
- D₁.sumPadOpens D₂ W₁ W₂ (Sum.inr j) = W₂ j
Instances For
Relatively compact reading opens sum to relatively compact ones.
The first padded level's ambient image lands in the common run's (coe_sigmaCoordImage: the
images are unions of summand images).
The second padded level's ambient image lands in the common run's.
The open immersion of the first padded level's local resolution into the common run's —
localResolutionHomOn (LocalResolutionOn.lean) at the inl inclusion.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The open immersion of the second padded level's local resolution into the common run's —
localResolutionHomOn at the inr inclusion.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The label map of the first padded level into the common label set ([Kol07, 34.1]) — the
general restriction identity
(BEDanFamStar.exists_orderEmbedding_comap_localResolutionHomOn_runMembers) at the inl
inclusion, read on the data's sigmaMembers: the common run's member of label lab k pulls back
along sumInlResIn to the level's member k, the members off the range to ⊤.
The label map of the second padded level into the common label set ([Kol07, 34.1]) — the
general identity at the inr inclusion.