backports 1.2.2

Project.toml

@@ -1,7 +1,7 @@
 name = "Oscar"
 uuid = "f1435218-dba5-11e9-1e4d-f1a5fab5fc13"
 authors = ["The OSCAR Team <oscar@oscar-system.org>"]
-version = "1.2.1"
+version = "1.2.2"
 
 [deps]
 AbstractAlgebra = "c3fe647b-3220-5bb0-a1ea-a7954cac585d"

README.md

@@ -37,7 +37,7 @@ julia> using Oscar
  / _ \ / ___| / ___|  / \  |  _ \   |  Combining ANTIC, GAP, Polymake, Singular
 | | | |\___ \| |     / _ \ | |_) |  |  Type "?Oscar" for more information
 | |_| | ___) | |___ / ___ \|  _ <   |  Manual: https://docs.oscar-system.org
- \___/ |____/ \____/_/   \_\_| \_\  |  Version 1.2.1
+ \___/ |____/ \____/_/   \_\_| \_\  |  Version 1.2.2
 
 julia> k, a = quadratic_field(-5)
 (Imaginary quadratic field defined by x^2 + 5, sqrt(-5))
@@ -114,7 +114,7 @@ pm::Array<topaz::HomologyGroup<pm::Integer> >
 If you have used OSCAR in the preparation of a paper please cite it as described below:
 
     [OSCAR]
-        OSCAR -- Open Source Computer Algebra Research system, Version 1.2.1,
+        OSCAR -- Open Source Computer Algebra Research system, Version 1.2.2,
         The OSCAR Team, 2024. (https://www.oscar-system.org)
     [OSCAR-book]
         Wolfram Decker, Christian Eder, Claus Fieker, Max Horn, Michael Joswig, eds.
@@ -127,7 +127,7 @@ If you are using BibTeX, you can use the following BibTeX entries:
       key          = {OSCAR},
       organization = {The OSCAR Team},
       title        = {OSCAR -- Open Source Computer Algebra Research system,
-                      Version 1.2.1},
+                      Version 1.2.2},
       year         = {2024},
       url          = {https://www.oscar-system.org},
       }

docs/src/AlgebraicGeometry/ToricVarieties/AlgebraicCycles.md

@@ -30,26 +30,30 @@ the variety in question.
 For complete and simplicial toric varieties, many things are
 known about the Chow ring and algebraic cycles (cf. section 12.5
 in [CLS11](@cite):
-* By therorem 12.5.3 of [CLS11](@cite), there is an isomorphism
-among the Chow ring and the cohomology ring. Note that the
-cohomology ring is naturally graded (cf. last paragraph
-on page 593 in [CLS11](@cite)). However, the Chow ring
-is usually considered as a non-graded ring. To match this general
-convention, and in particular the implementation of the Chow ring
-for matroids in OSCAR, the toric Chow ring is constructed as a
-non-graded ring.
-* By therorem 12.5.3 of [CLS11](@cite), the Chow ring is isomorphic
-to the quotient of the non-graded Cox ring and a certain ideal.
-Specifically, the ideal in question is the sum of the ideal of
-linear relations and the Stanley-Reisner ideal.
-* It is worth noting that the ideal of linear relations is not
-homogeneous with respect to the class group grading of the Cox ring.
-In order to construct the cohomology ring, one can introduce a
-$\mathbb{Z}$-grading on the Cox ring such that the ideal of linear
-relations and the Stanley-Reißner ideal are homogeneous.
-* Finally, by lemma 12.5.1 of [CLS11](@cite), generators of the
-rational equivalence classes of algebraic cycles are one-to-one to
-the cones in the fan of the toric variety.
+
+  * By therorem 12.5.3 of [CLS11](@cite), there is an isomorphism
+    among the Chow ring and the cohomology ring. Note that the
+    cohomology ring is naturally graded (cf. last paragraph
+    on page 593 in [CLS11](@cite)). However, the Chow ring
+    is usually considered as a non-graded ring. To match this general
+    convention, and in particular the implementation of the Chow ring
+    for matroids in OSCAR, the toric Chow ring is constructed as a
+    non-graded ring.
+
+  * By therorem 12.5.3 of [CLS11](@cite), the Chow ring is isomorphic
+    to the quotient of the non-graded Cox ring and a certain ideal.
+    Specifically, the ideal in question is the sum of the ideal of
+    linear relations and the Stanley-Reisner ideal.
+
+  * It is worth noting that the ideal of linear relations is not
+    homogeneous with respect to the class group grading of the Cox ring.
+    In order to construct the cohomology ring, one can introduce a
+    $\mathbb{Z}$-grading on the Cox ring such that the ideal of linear
+    relations and the Stanley-Reißner ideal are homogeneous.
+
+  * Finally, by lemma 12.5.1 of [CLS11](@cite), generators of the
+    rational equivalence classes of algebraic cycles are one-to-one to
+    the cones in the fan of the toric variety.
 
 
 ## Constructors

docs/src/DeveloperDocumentation/documentation.md

@@ -131,6 +131,13 @@ changes in the bibliography. If so, this test fails and indicates that the
 is not required that this test is passed. Still, please feel encouraged to fix
 this failure by running `bibtool` locally as explained above.
 
+!!! note "bibtool produces changes in unrelated parts of oscar_references.bib"
+    Sometimes `bibtool` will produce many changes when run locally. This can be
+    caused by a version difference. The version used in our github actions is
+    2.68. Check your version by running `bibtool -V`. When running this
+    command, please also pay attention whether any "Special configuration
+    options" are set.
+
 Please follow the additional guidelines below, that are not checked by bibtool:
 
 - Do not escape special characters like umlauts or accented characters. Instead, use the unicode character directly.

gap/OscarInterface/PackageInfo.g

@@ -10,8 +10,8 @@ SetPackageInfo( rec(
 
 PackageName := "OscarInterface",
 Subtitle := "GAP interface to OSCAR",
-Version := "1.2.1",
-Date := "20/11/2024", # dd/mm/yyyy format
+Version := "1.2.2",
+Date := "11/12/2024", # dd/mm/yyyy format
 License := "GPL-2.0-or-later",
 
 Persons := [

src/AlgebraicGeometry/ToricVarieties/NormalToricVarieties/constructors.jl

@@ -66,8 +66,7 @@ end
     normal_toric_variety(max_cones::IncidenceMatrix, rays::AbstractCollection[RayVector]; non_redundant::Bool = false)
 
 Construct a normal toric variety $X$ by providing the rays and maximal cones
-as vector of vectors. By default, this method assumes that the input is not
-non-redundant (e.g. that a ray was entered twice by accident). If the user
+as vector of vectors. By default, this method allows redundancies in the input, e.g. duplicate rays and non-maximal cones. If the user
 is certain that no redundancy exists in the entered information, one can
 pass `non_redundant = true` as third argument. This will bypass these consistency
 checks. In addition, this will ensure that the order of the rays is not

src/NumberTheory/GaloisGrp/GaloisGrp.jl

@@ -2667,6 +2667,10 @@ function Hecke.absolute_minpoly(a::Oscar.NfNSGenElem{QQFieldElem, QQMPolyRingEle
 end
 
 function blow_up(G::PermGroup, C::GaloisCtx, lf::Vector, con::PermGroupElem=one(G))
+  #TODO: currently con is useless here, the cluster detection and the reduce
+  #      tree is re-arranging the factors in lf randomly
+  #      one would need to trace the re-arranged lf as well
+  #      Since we don't, we have to compute roots and evaluate
 
   if all(x->x[2] == 1, lf)
     return G, C
@@ -2678,28 +2682,53 @@ function blow_up(G::PermGroup, C::GaloisCtx, lf::Vector, con::PermGroupElem=one(
 
   icon = inv(con)
 
-  gs = map(Vector{Int}, gens(G))
+  rr = roots(C, 2; raw = true)
+  renum = Int[]
   for (g, k) = lf
+    l = findall(iszero, map(g, rr))
+    append!(renum, l)
+  end
+  re = inv(symmetric_group(degree(G))(renum))
+
+  gs = map(Vector{Int}, gens(G))
+
+  # G is a sub group of prod G_i <= prod sym(n_i) the galois groups
+  # of the factors
+  # con describes the current ordering of the roots in so G^con <= prod G_i
+  #for the factors with multiplicity > 1, we need to append more factors
+  H = [G]
+  mps = [gens(G)]
+  n = degree(G)
+  for (g,k) = lf
+    if k == 1
+      st += degree(g)
+      continue
+    end
+    S = symmetric_group(degree(g))
+    #need proj G -> G_i, so need a subset of the points (st+1:st+deg(g))
+    # moved by re, then mapped to 1:deg(g)
+    h = [S([(i+st)^(gg^re)-st for i=1:degree(g)]) for gg = gens(G)]
     for j=2:k
-      for i=1:degree(g)
-        mp[n+i] = (st+i)^con
-      end
-      for h = gs
-        for i=1:degree(g)
-          push!(h, h[(st+i)^con]^icon-st+n)
-        end
+      S = symmetric_group(degree(g))
+      push!(H, S)
+      push!(mps, h)
+      for i = 1:degree(g)
+        mp[n+i] = (st+i)^icon
       end
       n += degree(g)
     end
     st += degree(g)
   end
 
+  D, emb, pro = inner_direct_product(H; morphisms = true)
+  h = hom(G, D, [prod(emb[i](mps[i][j]) for i=1:length(H)) for j=1:ngens(G)])
+
   C.rt_num = mp
-  S = symmetric_group(n)
-  GG, _ = sub(S, map(S, gs))
+  GG, _ = image(h)
 
   h = hom(G, GG, gens(G), gens(GG))
   @assert is_injective(h) && is_surjective(h)
+  C.G = GG
   return GG, C
 end
 
@@ -2825,6 +2854,7 @@ function galois_group(f::PolyRingElem{<:FieldElem}; prime=0, pStart::Int = 2*deg
   @vprint :GaloisGroup 1 "found $(length(cl)) connected components\n"
 
   res = Vector{Tuple{typeof(C[1]), PermGroupElem}}()
+  llf = Int[]
   function setup(C::Vector{<:GaloisCtx})
     G, emb, pro = inner_direct_product([x.G for x = C], morphisms = true)
     g = prod(x.f for x = C)

src/PolyhedralGeometry/PolyhedralFan/properties.jl

@@ -212,8 +212,21 @@ julia> cones(PF, 2)
 """
 function cones(PF::_FanLikeType, cone_dim::Int)
   l = cone_dim - length(lineality_space(PF))
-  l < 1 && return nothing
-  return SubObjectIterator{Cone{_get_scalar_type(PF)}}(
+  t = Cone{_get_scalar_type(PF)}
+  (l < 0 || dim(PF) == -1) && return _empty_subobjectiterator(t, PF)
+
+  if l == 0
+    return SubObjectIterator{t}(
+      PF,
+      (_, _, _) -> positive_hull(
+        coefficient_field(PF), zeros(Int, ambient_dim(PF)), lineality_space(PF)
+      ),
+      1,
+      NamedTuple(),
+    )
+  end
+
+  return SubObjectIterator{t}(
     PF, _cone_of_dim, size(Polymake.fan.cones_of_dim(pm_object(PF), l), 1), (c_dim=l,)
   )
 end

src/PolyhedralGeometry/iterators.jl

@@ -106,6 +106,10 @@ Base.:(==)(::PointVector, ::RayVector) =
 Base.:(==)(::RayVector, ::PointVector) =
   throw(ArgumentError("Cannot compare PointVector to RayVector"))
 
+Base.isequal(x::RayVector, y::RayVector) = x == y
+
+Base.hash(x::RayVector, h::UInt) = hash(collect(sign.(x)), hash(coefficient_field(x), h))
+
 ################################################################################
 ######## Halfspaces and Hyperplanes
 ################################################################################

src/Rings/localization_interface.jl

@@ -399,7 +399,15 @@
 
 canonical_unit(f::LocRingElemType) where {LocRingElemType<:AbsLocalizedRingElem} = one(parent(f))
 
-characteristic(W::AbsLocalizedRing) = characteristic(base_ring(W))
+function characteristic(W::AbsLocalizedRing)
+  # It might happen that localization collapses the ring.
+  is_zero(one(W)) && return 1
+  # If the characteristic is prime, then it is preserved.
+  c = characteristic(base_ring(W))
+  (is_zero(c) || is_prime(c)) && return c
+  # All other cases have to be overwritten manually.
+  error("computation of characteristic not implemented")
+end
 
 function Base.show(io::IO, ::MIME"text/plain", W::AbsLocalizedRing)
   io = pretty(io)

test/AlgebraicGeometry/ToricVarieties/toric_blowups.jl

@@ -68,7 +68,8 @@
     @test !issubset(pbJ_str, ideal_sheaf(E))
   end
 
-  I = ideal(S, [S[1], S[2]])
+  # Select variables corresponding to rays [1,0] and [0,1]
+  I = ideal(S, gens(S)[findall(v->v[1]>=0&&v[2]>=0, rays(P2))])
   bl2 = blow_up(P2, I)
   II = IdealSheaf(P2, I)
 

test/NumberTheory/galthy.jl

@@ -10,7 +10,7 @@
   @test degree(L) == order(G)
   @test length(roots(L, k.pol)) == 5
 
-  R, x = polynomial_ring(QQ, :x)
+  R, x = polynomial_ring(QQ, :x; cached = false)
   pol = x^6 - 366*x^4 - 878*x^3 + 4329*x^2 + 14874*x + 10471
   g, C = galois_group(pol)
   @test order(g) == 18
@@ -24,7 +24,7 @@
   G, C = galois_group((1//13)*x^2+2)
   @test order(G) == 2
 
-  K, a = number_field(x^4-2)
+  K, a = number_field(x^4-2; cached = false)
   G, C = galois_group(K)
   Gc, Cc = galois_group(K, algorithm = :Complex)
   Gs, Cs = galois_group(K, algorithm = :Symbolic)
@@ -46,6 +46,17 @@
   G, C = galois_group((2*x+1)^2)
   @test order(G) == 1
   @test degree(G) == 2
+
+  #from errors:
+  G, C = galois_group((x^4 + 1)^3 * x^2 * (x^2 - 4*x + 1)^5)
+  @test order(G) == 8
+  @test degree(G) == 24
+  k = fixed_field(C, sub(G, [one(G)])[1])
+  @test degree(k) == 8
+
+  G, C = galois_group((x^3-2)^2*(x^3-5)^2*(x^2-6))
+  @test order(G) == 36
+  @test degree(G) == 14
 end
 
 import Oscar.GaloisGrp: primitive_by_shape, an_sn_by_shape, cycle_structures

test/PolyhedralGeometry/cone.jl

@@ -28,7 +28,7 @@
     if T == QQFieldElem
       @test hilbert_basis(Cone1) isa SubObjectIterator{PointVector{ZZRingElem}}
       @test length(hilbert_basis(Cone1)) == 2
-      @test issetequal(hilbert_basis(Cone1), ray_vector.(Ref(ZZ), [[1, 0], [0, 1]]))
+      @test issetequal(hilbert_basis(Cone1), point_vector.(Ref(ZZ), [[1, 0], [0, 1]]))
       @test generator_matrix(hilbert_basis(Cone1)) == _oscar_matrix_from_property(ZZ, hilbert_basis(Cone1))
     end
     @test n_rays(Cone1) == 2

test/PolyhedralGeometry/linear_program.jl

@@ -7,12 +7,7 @@
   C1 = cube(f, 2, 0, 1)
   Pos = polyhedron(f, [-1 0 0; 0 -1 0; 0 0 -1], [0, 0, 0])
   L = polyhedron(f, [-1 0 0; 0 -1 0], [0, 0])
-  point = convex_hull(f, [0 1 0])
   empty = polyhedron(f, [1 0 0; -1 0 0], [0, -1])
-  # this is to make sure the order of some matrices below doesn't change
-  Polymake.prefer("beneath_beyond") do
-    affine_hull(point)
-  end
   s = simplex(f, 2)
   rsquare = cube(f, 2, QQFieldElem(-3, 2), QQFieldElem(3, 2))
 

test/PolyhedralGeometry/polyhedral_fan.jl

@@ -58,11 +58,18 @@
     @test size(cones(F2, 2)) == (2,)
     @test lineality_space(cones(F2, 2)[1]) == [[0, 1, 0]]
     @test rays.(cones(F2, 2)) == [[], []]
-    @test isnothing(cones(F2, 1))
     @test _check_im_perm_rows(ray_indices(cones(F1, 2)), incidence1)
     @test _check_im_perm_rows(incidence_matrix(cones(F1, 2)), incidence1)
     @test _check_im_perm_rows(cones(IncidenceMatrix, F1, 2), incidence1)
 
+    A3 = affine_space(NormalToricVariety, 3)
+    @test length(cones(A3, 1)) == 3
+    @test length(cones(A3, 0)) == 1
+    @test length(cones(A3, -1)) == 0
+    @test cones(A3, 1) isa SubObjectIterator{Cone{QQFieldElem}}
+    @test cones(A3, 0) isa SubObjectIterator{Cone{QQFieldElem}}
+    @test cones(A3, -1) isa SubObjectIterator{Cone{QQFieldElem}}
+
     II = ray_indices(maximal_cones(NFsquare))
     NF0 = polyhedral_fan(II, rays(NFsquare))
     @test n_rays(NF0) == 4

test/PolyhedralGeometry/polyhedron.jl

@@ -13,10 +13,6 @@
   L = polyhedron(f, [-1 0 0; 0 -1 0], [0, 0])
   full = polyhedron(f, zero_matrix(f, 0, 3), [])
   point = convex_hull(f, [0 1 0])
-  # this is to make sure the order of some matrices below doesn't change
-  Polymake.prefer("beneath_beyond") do
-    affine_hull(point)
-  end
   s = simplex(f, 2)
   R, x = polynomial_ring(QQ, :x)
   v = T[f(1), f(1)]

test/PolyhedralGeometry/types.jl

@@ -109,6 +109,9 @@
           end
         end
       end
+      if T == RayVector
+        @test length(Set(ray_vector.(Ref(f), [[1,2,3],[2,4,6],[-1,-2,-3]]))) == 2
+      end
     end
   end