Forgetting the module structure of a sheaf of modules is exact #
Let R be a sheaf of rings on a site (C, J). Mathlib's SheafOfModules.toSheaf R sends a
sheaf of R-modules to its underlying abelian sheaf, and knows that this functor preserves and
reflects finite limits. This file supplies the missing half: it preserves finite colimits as
well, hence carries short exact sequences of sheaves of modules to short exact sequences of
abelian sheaves.
Main declarations #
TauCeti.SheafOfModules.preservesFiniteColimits_toSheaf:SheafOfModules.toSheafpreserves finite colimits;TauCeti.SheafOfModules.shortExact_map_toSheaf: a short exact sequence of sheaves of modules stays short exact after forgetting the module structures.
The proof is a transfer along the sheafification of presheaves of modules. That functor is a
left adjoint whose counit is an isomorphism, and composing it with SheafOfModules.toSheaf
gives PresheafOfModules.toPresheaf ⋙ presheafToSheaf, which preserves finite colimits because
colimits of presheaves of modules are computed sectionwise and presheafToSheaf is a left
adjoint.
Exactness of SheafOfModules.toSheaf is what makes the cohomology of a sheaf of modules, which
is defined through the underlying abelian sheaf, fit into a long exact sequence; see
TauCeti/AlgebraicGeometry/Cohomology/LongExactSequence.lean. It is therefore Layer B
infrastructure for TauCetiRoadmap/JacobianChallenge/README.md. No formalization is vendored:
the ingredients are Mathlib's PresheafOfModules.sheafificationAdjunction,
PresheafOfModules.sheafificationCompToSheaf and ShortComplex.ShortExact.map_of_exact.
Forgetting the module structures of a short exact sequence of sheaves of modules leaves a short exact sequence of abelian sheaves.