Differential graded categories from explicit Hom-complex data #
A differential graded category is defined as a category enriched in cochain complexes of
R-modules, and TauCeti/CategoryTheory/DG/Basic.lean unpacks that enrichment into homogeneous
morphisms, their differential, and their composition. This file goes the other way. Explicit
Hom-complex data, TauCeti.DGCategoryData, consists of a Hom complex for every pair of objects,
bilinear composition maps on homogeneous degrees, and identities, subject to the graded Leibniz
rule, associativity, and the unit laws stated on elements. Such data determines a differential
graded category, TauCeti.DGCategoryData.toDGCategory, whose homogeneous morphisms, differential,
identities and composition are the given ones.
The two presentations are equivalent: extracting the explicit data from a differential graded category and rebuilding the enrichment recovers the enrichment, and rebuilding the data from the enrichment of explicit data recovers the data. Hence a differential graded category may be specified by elementwise formulas, and a differential graded category built from homogeneous morphisms, such as the one-object category of a differential graded algebra or the category of differential graded modules, needs no monoidal plumbing of its own.
Composition in TauCeti.DGCategoryData is in Mathlib's enriched factor order, so comp f g is
f followed by g and the Leibniz sign is carried by the degree of f, exactly as for
TauCeti.dgComp. Keller's convention, in which g ∘ f carries the Leibniz sign of g, differs
by the Koszul sign (-1) ^ (|f| * |g|); TauCeti.DGCategoryData.ofKeller converts data given in
that convention.
Main definitions #
TauCeti.DGCategoryData: explicit Hom-complex data for a differential graded category.TauCeti.DGCategoryData.toDGCategory: the differential graded category it determines.TauCeti.DGCategoryData.ofKeller: the same data from Keller-ordered composition.TauCeti.dgCategoryData: the explicit Hom-complex data of a differential graded category.TauCeti.DGCategoryData.equivDGCategory: the two presentations are equivalent.
Main results #
TauCeti.DGCategoryData.dgComp_toDGCategory,TauCeti.DGCategoryData.dgId_toDGCategory,TauCeti.DGCategoryData.dgDifferential_toDGCategory: the homogeneous calculus of the constructed category is the given data.TauCeti.DGCategoryData.toDGCategory_dgCategoryDataandTauCeti.DGCategoryData.dgCategoryData_toDGCategory: the two round trips.
References #
- B. Keller, Deriving DG categories, Section 1.
- B. Keller, Introduction to A-infinity algebras and modules, Section 3.1, for the sign conventions and the comparison sign between the two factor orders.
TauCeti/Algebra/Homology/LinearHomComplex/Composition.lean.TauCeti/Algebra/Homology/LinearHomComplex/Enrichment.lean.
Explicit Hom-complex data for a differential graded category over R on the objects
C: a Hom complex hom X Y for every pair of objects, bilinear composition
comp p q n h : (hom X Y).X p →ₗ[R] (hom Y Z).X q →ₗ[R] (hom X Z).X n for p + q = n, in
Mathlib's enriched factor order, and identities id X of degree zero, satisfying the graded
Leibniz rule, associativity, and the unit laws on homogeneous elements. Such data determines the
differential graded category TauCeti.DGCategoryData.toDGCategory.
- hom : C → C → CochainComplex (ModuleCat R) ℤ
The Hom complex from
XtoY. - comp {X Y Z : C} (p q n : ℤ) (h : p + q = n) : ↑((self.hom X Y).X p) →ₗ[R] ↑((self.hom Y Z).X q) →ₗ[R] ↑((self.hom X Z).X n)
Composition of a degree-
pmorphismX ⟶ Ywith a degree-qmorphismY ⟶ Z, in Mathlib's enriched factor order:comp p q n h f gisffollowed byg. The identity of
X, a morphism of degree zero.- d_comp {X Y Z : C} {p q n : ℤ} (h : p + q = n) (f : ↑((self.hom X Y).X p)) (g : ↑((self.hom Y Z).X q)) : (ModuleCat.Hom.hom ((self.hom X Z).d n (n + 1))) (((self.comp p q n h) f) g) = ((self.comp (p + 1) q (n + 1) ⋯) ((ModuleCat.Hom.hom ((self.hom X Y).d p (p + 1))) f)) g + p.negOnePow • ((self.comp p (q + 1) (n + 1) ⋯) f) ((ModuleCat.Hom.hom ((self.hom Y Z).d q (q + 1))) g)
The graded Leibniz rule, with the Koszul sign carried by the degree of the first factor.
- comp_assoc {W X Y Z : C} {p q r pq qr n : ℤ} (hpq : p + q = pq) (hqr : q + r = qr) (h : p + q + r = n) (f : ↑((self.hom W X).X p)) (g : ↑((self.hom X Y).X q)) (k : ↑((self.hom Y Z).X r)) : ((self.comp pq r n ⋯) (((self.comp p q pq hpq) f) g)) k = ((self.comp p qr n ⋯) f) (((self.comp q r qr hqr) g) k)
Composition is associative.
The identity is a left unit.
The identity is a right unit.
Instances For
The identity is a cycle.
The enriched composition and identity #
The bidegree-(p, q) component of composition, as a morphism out of the tensor product of
the two homogeneous modules.
Equations
- D.compMap X Y Z p q n h = ModuleCat.ofHom (TensorProduct.lift (D.comp p q n h))
Instances For
Composition as a morphism of cochain complexes out of the totalized tensor product of the two Hom complexes. The Leibniz rule makes it a chain map.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The degree-n component of the enriched composition is assembled from the bidegree
components.
On the bidegree-(p, q) summand, the enriched composition is the bilinear composition of the
data.
On the bidegree-(p, q) summand, the enriched composition is the bilinear composition of the
data.
The identity of X, as a morphism from the tensor unit to the Hom complex of X with
itself.
Equations
- D.enrichedId X = HomologicalComplex.mkHomFromSingle (ModuleCat.ofHom (LinearMap.toSpanSingleton R (↑((D.hom X X).1 0)) (D.id X))) ⋯
Instances For
The degree-zero component of the enriched identity.
The degree-zero component of the enriched identity sends a scalar to that multiple of the identity.
The differential graded category #
The enriched identity is a left unit for the enriched composition.
The enriched identity is a left unit for the enriched composition.
The enriched identity is a right unit for the enriched composition.
The enriched identity is a right unit for the enriched composition.
The enriched composition is associative, with the tensor products identified by the monoidal associator.
The enriched composition is associative, with the tensor products identified by the monoidal associator.
The differential graded category determined by explicit Hom-complex data. Its body is exposed
so that its homogeneous morphisms TauCeti.DGHom remain definitionally the given
(D.hom X Y).X n.
Equations
- D.toDGCategory = { Hom := D.hom, id := D.enrichedId, comp := D.enrichedComp, id_comp := ⋯, comp_id := ⋯, assoc := ⋯ }
Instances For
The Hom complex of the constructed category is the given one.
The enriched identity of the constructed category is the given one.
The enriched composition of the constructed category is the given one.
The differential of the constructed category is the differential of the given Hom complexes.
The bidegree components of composition in the constructed category are the given ones.
The identity of the constructed category is the given identity.
Keller-ordered data #
Explicit Hom-complex data from Keller-ordered composition. Here comp q p n h g f is
g ∘ f for g : Y ⟶ Z of degree q and f : X ⟶ Y of degree p, the Leibniz rule
d (g ∘ f) = d g ∘ f + (-1) ^ q • g ∘ d f carries the sign of g, and associativity and the unit
laws are the usual ones. The enriched-order composition of the resulting data is
(-1) ^ (p * q) • g ∘ f, the Koszul sign converting between the two factor orders. The body is
exposed so that the homogeneous morphisms of the resulting data remain definitionally the given
(hom X Y).X n.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The enriched-order composition of Keller-ordered data is Keller composition in reversed order,
with the Koszul sign (-1) ^ (p * q).
The explicit data of a differential graded category #
The explicit Hom-complex data of a differential graded category: its Hom complexes, homogeneous composition, and identities. Its body is exposed so that its Hom complexes remain definitionally those of the category.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The enriched identity of a differential graded category is the enriched identity of its explicit data.
The enriched composition of a differential graded category is the enriched composition of its explicit data.
Rebuilding a differential graded category from its explicit Hom-complex data recovers the category.
Extracting the explicit Hom-complex data of the category built from explicit data recovers the data.
The two presentations of a differential graded category agree: explicit Hom-complex data
on C is the same as an enrichment of C in cochain complexes of R-modules.
Equations
- One or more equations did not get rendered due to their size.