Preservation of pretriangulated dg structure #
Every dg functor preserves the representability witnesses that define shifts
and cones. DGFunctor.preservesShifts and
DGFunctor.preservesChosenCones supply the two dg-level capabilities
automatically. For cones, the inverse to the image splitting map is built
from the images of the two projections extracted from the source cone.
Transport to the ordinary shift functors and cone triangles on H⁰ belongs
to the dg-enhancement layer.
A dg functor preserves shifts when it sends every dg shift witness to a shift witness on the image objects, with the defining homogeneous element given by the dg functor's map.
No pretriangulated instance is required to state the capability: callers may
provide individual witnesses, while H⁰ adapters add existence assumptions
only when they choose shifts for every object.
The target shift witness associated to a source shift witness.
The associated shift element is the image of the source element.
Instances For
The identity dg functor preserves shift witnesses.
Equations
- CategoryTheory.DGFunctor.PreservesShifts.id C = { mapShift := fun {X Y : C} {n : ℤ} (s : CategoryTheory.IsShiftBy X n Y) => s, mapShift_hom := ⋯ }
Instances For
Shift preservation is closed under composition of dg functors.
Equations
Instances For
Every dg functor preserves shifts.
A shift element is a closed, two-sided invertible element of degree -n:
IsShiftBy.inv extracts the inverse from surjectivity of right composition, and
hom_inv and inv_hom are the two identities. A dg functor preserves the
graded composition and the identities on the nose, so it carries an invertible
element to an invertible element, and right composition with the image is
bijective with right composition with the image of the inverse.
So PreservesShifts is not a hypothesis anybody has to discharge; it is
supplied here for every dg functor. The structure stays because it names the
capability and because mapShift is the useful accessor, but a caller that used
to take a PreservesShifts argument can now call this instead.
Equations
Instances For
Apply a dg functor to a chosen homotopy-commutative square. The homotopy
itself is mapped, so this construction retains the witness needed by later
cone maps rather than merely proving that the image square commutes in H⁰.
Equations
- F.mapHomotopySquare s = CategoryTheory.DGCategory.HomotopySquare.ofBoundary ⋯ ⋯ ((F.map (-1)) s.homotopy) ⋯
Instances For
Strong dg-level preservation of cone witnesses.
For every supplied cone witness, the image object is equipped with a cone
witness whose two inclusions are exactly the images of the source inclusions.
This chosen, composable capability is intentionally stronger than the later
H⁰ proposition that an image triangle is merely isomorphic to some cone
triangle.
- mapCone {X Y Z : C} {f : ↑((DGCategoryStruct.dgHom X Y).X 0)} (hc : IsConeOf f Z) : IsConeOf ((F.map 0) f) (F.obj Z)
The target cone witness carried by the image object.
- mapCone_inr {X Y Z : C} {f : ↑((DGCategoryStruct.dgHom X Y).X 0)} (hc : IsConeOf f Z) : (self.mapCone hc).inr = (F.map 0) hc.inr
The target cone inclusion is the image of the source cone inclusion.
- mapCone_inl {X Y Z : C} {f : ↑((DGCategoryStruct.dgHom X Y).X 0)} (hc : IsConeOf f Z) : (self.mapCone hc).inl = (F.map (-1)) hc.inl
The shifted source inclusion is the image of the source cone inclusion.
Instances For
The identity dg functor preserves chosen cone witnesses.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Strong cone preservation is closed under composition of dg functors.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Strong preservation of the chosen cone witnesses is invariant under an isomorphism of dg functors in the closed degree-zero category.
Equations
Instances For
The projection from an image cone to its source is the image of the source
cone projection. This is forced by uniqueness of the cone splitting; it is
not an additional field of PreservesChosenCones.
The projection from an image cone to its target is the image of the source cone projection, again derived from uniqueness of the cone splitting.
A cone-preserving, shift-preserving dg functor carries the connecting map of a supplied cone to the connecting map built from the corresponding image witnesses.
Every dg functor preserves the chosen cone witnesses.
The cone universal property is a split representability statement. A source
cone supplies projections fst and snd satisfying the usual five matrix
identities with inl and inr. After applying a dg functor, those identities
still hold. For an arbitrary target object W, they exhibit
c ↦ (c ≫ F(fst), c ≫ F(snd))
as the inverse to the image cone's splitting map. Thus neither fullness nor a
pretriangulated existence assumption is needed. The structure remains useful
as a named package consumed by the H⁰ exactness adapters.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The inverse of the image shift witness is forced by the image of the
source inverse. This removes the Classical.choice hidden in IsShiftBy.inv
from every later transport calculation.
F commutes with the dg action on morphisms induced by a shift witness.
F commutes with comparisons between two shifts of the same object.