snappy 3.1.1__cp38-cp38-win_amd64.whl → 3.2__cp38-cp38-win_amd64.whl

snappy/drilling/geometric_structure.py

@@ -1,366 +0,0 @@
-from .line import R13LineWithMatrix
-
-from ..verify.shapes import compute_hyperbolic_shapes # type: ignore
-from ..snap.fundamental_polyhedron import FundamentalPolyhedronEngine # type: ignore
-from ..snap.kernel_structures import TransferKernelStructuresEngine # type: ignore
-from ..snap.t3mlite import simplex, Mcomplex, Tetrahedron, Vertex # type: ignore
-from ..SnapPy import word_as_list # type: ignore
-
-from ..hyperboloid import (o13_inverse,  # type: ignore
-                           space_r13_normalise,
-                           r13_dot,
-                           unnormalised_plane_eqn_from_r13_points)
-from ..upper_halfspace import sl2c_inverse, psl2c_to_o13 # type: ignore
-from ..upper_halfspace.ideal_point import ideal_point_to_r13 # type: ignore
-from ..matrix import vector, matrix, mat_solve # type: ignore
-from ..math_basics import prod, xgcd # type: ignore
-
-from collections import deque
-
-from typing import Tuple, Sequence, Optional, Any
-
-Filling = Tuple[int, int]
-FillingMatrix = Tuple[Filling, Filling]
-
-
-def compute_r13_planes_for_tet(tet : Tetrahedron):
-    """
-    Computes outward facing normals/plane equations from the vertices of
-    positively oriented tetrahedra - all in the hyperboloid model.
-    """
-
-    tet.R13_unnormalised_planes = {
-        f: unnormalised_plane_eqn_from_r13_points(
-            [ tet.R13_vertices[v] for v in verts ])
-        for f, verts in simplex.VerticesOfFaceCounterclockwise.items() }
-    tet.R13_planes = {
-        f : space_r13_normalise(plane)
-        for f, plane in tet.R13_unnormalised_planes.items() }
-
-
-def word_to_psl2c_matrix(mcomplex : Mcomplex, word : str):
-    """
-    Given a triangulation with a R13 geometric structure (that is
-    the structure attached by calling add_r13_geometry) and a word
-    in the simplified fundamental group (given as string), returns
-    the corresponding PSL(2,C)-matrix.
-    """
-
-    return word_list_to_psl2c_matrix(
-        mcomplex, word_as_list(word, mcomplex.num_generators))
-
-
-def word_list_to_psl2c_matrix(mcomplex : Mcomplex, word_list : Sequence[int]):
-    """
-    Like word_to_psl2c_matrix, but taking the word as a sequence of
-    non-zero integers with positive integers corresponding to generators and
-    negative integers corresponding to their inverses.
-    """
-
-    return prod([mcomplex.GeneratorMatrices[g]
-                 for g in word_list])
-
-
-def add_r13_geometry(
-        mcomplex : Mcomplex,
-        manifold,
-        verified : bool = False,
-        bits_prec : Optional[int] = None):
-    """
-    Given the same triangulation once as Mcomplex and once as SnapPy Manifold,
-    develops the vertices of the tetrahedra (using the same fundamental
-    polyhedron as the SnapPea kernel), computes the face-pairing matrices and
-    the matrices corresponding to the generators of the unsimplified
-    fundamental group, computes the incenter of the base tetrahedron and
-    the core curve for each vertex of each tetrahedron corresponding to a
-    filled cusp.
-
-    The precision can be given by bits_prec (if not given, the precision of
-    the Manifold type is used, i.e., 53 for Manifold and 212 for ManifoldHP).
-
-    If verified is True, intervals will be computed for all the above
-    information.
-    """
-
-    shapes = compute_hyperbolic_shapes(
-        manifold, verified=verified, bits_prec=bits_prec)
-    z = shapes[0]
-    RF = z.real().parent()
-
-    # Develop the vertices in the upper half space model - we will
-    # convert them to the hyperboloid model later.
-    poly = FundamentalPolyhedronEngine.from_manifold_and_shapes(
-        manifold, shapes, normalize_matrices=True)
-
-    # Match the order of the mcomplex.Vertices to the one the SnapPea
-    # kernel sees and copy meridians and longitudes to tet.PeripheralCurves.
-    TransferKernelStructuresEngine(
-        mcomplex, manifold).reindex_cusps_and_transfer_peripheral_curves()
-    mcomplex.verified = verified
-    mcomplex.RF = RF
-    # PSL(2,C)-matrices corresponding to generators of fundamental group.
-    # Positive integers map to the generators, negative integrs to their
-    # inverses and 0 to the identity.
-    mcomplex.GeneratorMatrices = {
-        g : _to_matrix(m)
-        for g, m in poly.mcomplex.GeneratorMatrices.items() }
-    # Number of generators of the fundamental group.
-    mcomplex.num_generators = len(mcomplex.GeneratorMatrices) // 2
-
-    for tet, developed_tet in zip(mcomplex.Tetrahedra, poly.mcomplex):
-        # Shape for each edge, keys are simplex.OneSubsimplices
-        tet.ShapeParameters = developed_tet.ShapeParameters
-        # Vertices in C union infinity on the boundary of
-        # upper halfspace model
-        tet.ideal_vertices = {
-            V: developed_tet.Class[V].IdealPoint
-            for V in simplex.ZeroSubsimplices }
-        # Vertices in hyperboloid model
-        tet.R13_vertices = {
-            V: ideal_point_to_r13(z, RF)
-            for V, z in tet.ideal_vertices.items() }
-        # Add plane equations for faces
-        compute_r13_planes_for_tet(tet)
-        # Compute face-pairing matrices for hyperboloid model
-        tet.O13_matrices = {
-            F : psl2c_to_o13(mcomplex.GeneratorMatrices.get(-g))
-            for F, g in developed_tet.GeneratorsInfo.items() }
-        # Dict, keys are a subset of simplex.ZeroSubsimplices
-        #
-        # If a vertex of a tet corresponds to a filled cusp, this dictionary
-        # will contain the appropriate lift of the core curve in the
-        # hyperboloid model.
-        tet.core_curves = { }
-
-    # Set base tetrahedron and compute its in-radius and center.
-    mcomplex.baseTet = mcomplex.Tetrahedra[
-        poly.mcomplex.ChooseGenInitialTet.Index]
-    mcomplex.baseTetInRadius, mcomplex.R13_baseTetInCenter = (
-        _compute_inradius_and_incenter_from_planes(
-            [ mcomplex.baseTet.R13_planes[f]
-              for f in simplex.TwoSubsimplices]))
-
-    # For each cusp, a pair of words for the meridian and longitude as
-    # sequence of non-zero integers.
-    #
-    # Only computed when needed.
-    all_peripheral_words : Optional[Sequence[Sequence[Sequence[int]]]] = None
-
-    # For each cusp
-    for v, info in zip(mcomplex.Vertices, manifold.cusp_info()):
-        # v.filling_matrix is a matrix of integers (as list of lists) such that
-        # v.filling_matrix[0] contains the filling coefficients
-        # (e.g., [3,4] for m004(3,4)) and the determinant is 1 if the cusp is
-        # filled. That is, v.filling_matrix[1] determines a curve intersecting
-        # the filling curve once (as sum of a multiple of meridian and
-        # longitude) and that is thus parallel to the core curve.
-        # For an unfilled cusp, v.filling_matrix is ((0,0), (0,0))
-
-        v.filling_matrix = _filling_matrix(info)
-        if v.filling_matrix[0] != (0,0):
-            if all_peripheral_words is None:
-                # Make the SnapPea kernel compute peripheral curves the first
-                # time when we need them.
-                G = manifold.fundamental_group(False)
-                all_peripheral_words = G.peripheral_curves(as_int_list=True)
-            # Note that a cusp only determines the words for the meridian
-            # and longitude only up to conjugacy, we need to pick a lift of the
-            # cusp and a path from the basepoint to the lift.
-            #
-            # Similarly, the lift of the core curve of a filled cusp to the
-            # hyperboloid model depends on a lift of a cusp a path from the
-            # basepoint to the lift.
-            #
-            # We compute the lift for each vertex of each tetrahedron in the
-            # fundamental domain corresponding to the cusp (with the path
-            # connecting the basepoint to the vertex being the one contained
-            # in the fundamental domain).
-            #
-            # Starting with the one choice the SnapPea kernel did and computing
-            # the resulting lift of the core curve, we need to transfer it
-            # to the other choices of vertices of tetrahedra corresponding to
-            # the cusp by "developing" the cusp.
-            #
-            _develop_core_curve_cusp(
-                mcomplex,
-                v,
-                _compute_core_curve(
-                    mcomplex,
-                    all_peripheral_words[v.Index],
-                    v.filling_matrix[1]))
-
-    return mcomplex
-
-###############################################################################
-# Helpers
-
-
-def _to_matrix(m):
-    """
-    Necesssary conversion when not SageMath.
-
-    This is needed because we have two matrix types outside of Sage:
-    SimpleMatrix and Matrix2x2. Convert to the former.
-    """
-    return matrix([[m[0,0],m[0,1]],
-                   [m[1,0],m[1,1]]])
-
-
-def _compute_core_curve(
-        mcomplex : Mcomplex,
-        peripheral_words : Sequence[Sequence[int]],
-        core_curve_coefficients : Filling) -> R13LineWithMatrix:
-    """
-    Compute core curve given words for meridian and longitude and
-    the integers determining a curve (as sum of a multiple of meridian
-    and longitude) that is parallel to the core curve.
-    """
-
-    result = mcomplex.GeneratorMatrices[0]
-
-    for word, f in zip(peripheral_words, core_curve_coefficients):
-        if f != 0:
-            m = word_list_to_psl2c_matrix(mcomplex, word)
-            if f < 0:
-                m = sl2c_inverse(m)
-            for i in range(abs(f)):
-                result = result * m
-
-    return R13LineWithMatrix.from_psl2c_matrix(result)
-
-
-def _find_standard_basepoint(mcomplex : Mcomplex,
-                             vertex : Vertex) -> Tuple[Tetrahedron, int]:
-    """
-    Reimplements find_standard_basepoint in fundamental_group.c.
-
-    That is, it finds the same tetrahedron and vertex of that tetrahedron
-    in the fundamental domain that the SnapPea kernel used to compute the
-    words for the meridian and longitude of the given cusp.
-
-    The SnapPea kernel picks the first vertex it finds where the meridian
-    and longitude intersect.
-    """
-
-    # Traverse tets and their vertices in the same order the SnapPea kernel
-    # does
-    for tet in mcomplex.Tetrahedra:
-        for v in simplex.ZeroSubsimplices:
-            # Only consider vertices corresponding to the given cusp
-            if tet.Class[v] is vertex:
-                for f in simplex.TwoSubsimplices:
-                    # Check that the meridian and longitude both
-                    # go through the same leg of the spine of the cusp
-                    # triangle.
-                    #
-                    # Note that we only support orientable manifolds,
-                    # so we only consider the 0-sheet of the orientation
-                    # double-cover of the cusp triangulation.
-                    if (tet.PeripheralCurves[0][0][v][f] != 0 and
-                        tet.PeripheralCurves[1][0][v][f] != 0):
-                        return tet, v
-
-    raise Exception("Could not find basepoint for cusp. This is a bug.")
-
-
-def _develop_core_curve_cusp(
-        mcomplex : Mcomplex,
-        v : Vertex,
-        core_curve : R13LineWithMatrix) -> None:
-    """
-    Given the core curve computed from the SnapPea kernel's given
-    words for the meridian and longitude for the given cusp,
-    compute the lift of the core curve for all vertices of the
-    tetrahedra corresponding to the given cusp.
-    """
-
-    # Start with the tet and vertex that the SnapPea kernel used
-    # to compute the words.
-    tet, vertex = _find_standard_basepoint(mcomplex, v)
-
-    tet.core_curves[vertex] = core_curve
-    pending_tet_verts = deque([ (tet, vertex, core_curve) ])
-
-    # Breadth-first traversal of cusp triangles to compute appropriate
-    # transform of core curve.
-    while pending_tet_verts:
-        tet, vertex, core_curve = pending_tet_verts.popleft()
-        for f in simplex.FacesAroundVertexCounterclockwise[vertex]:
-            new_tet = tet.Neighbor[f]
-            new_vertex = tet.Gluing[f].image(vertex)
-            if new_vertex in new_tet.core_curves:
-                continue
-            new_core_curve = core_curve.transformed(tet.O13_matrices[f])
-            new_tet.core_curves[new_vertex] = new_core_curve
-            pending_tet_verts.append(
-                (new_tet, new_vertex, new_core_curve))
-
-# Depending on whether we are using SnapPy inside SageMath or not, we
-# use different python classes to represent numbers, vectors and matrices.
-# Thus, using Any as type annotation for now :(
-
-
-def _compute_inradius_and_incenter_from_planes(planes) -> Tuple[Any, Any]:
-    """
-    Given outside-facing normals for the four faces of a
-    tetrahedron, compute the hyperbolic inradius and the
-    incenter (as unit time vector) of the tetrahedron (in the
-    hyperboloid model).
-    """
-
-    # We need to c and r such that
-    #  * r13_dot(c, c) = -1 and
-    #  * r13_dot(plane, c) = -sinh(r) for every plane
-    #
-    # We instead solve for the following system of linear equations:
-    #  * r13_dot(plane, pt) = -1 for every plane
-
-    RF = planes[0][0].parent()
-    m = matrix([[-plane[0], plane[1], plane[2], plane[3]]
-                for plane in planes])
-    v = vector([RF(-1), RF(-1), RF(-1), RF(-1)])
-
-    pt = mat_solve(m, v)
-
-    # And then use the inverse length of pt to scale pt to be
-    # a unit time vector and to compute the r.
-    scale = 1 / (-r13_dot(pt, pt)).sqrt()
-
-    return scale.arcsinh(), scale * pt
-
-
-def _filling_matrix(cusp_info : dict) -> FillingMatrix:
-    """
-    Given one of the dictionaries returned by Manifold.cusp_info(),
-    returns the "filling matrix" filling_matrix.
-
-    filling_matrix is a matrix of integers (as list of lists) such that
-    filling_matrix[0] contains the filling coefficients
-    (e.g., [3,4] for m004(3,4)) and the determinant is 1 if the cusp is
-    filled. That is, filling_matrix[1] determines a curve intersecting
-    the filling curve once (as sum of a multiple of meridian and
-    longitude) and that is thus parallel to the core curve.
-
-    For an unfilled cusp, filling_matrix is ((0,0), (0,0))
-
-    Raises an exception if the filling coefficients are non-integral or
-    not coprime.
-    """
-
-    float_m, float_l = cusp_info['filling']
-    m = int(float_m)
-    l = int(float_l)
-    if float_m != m or float_l != l:
-        raise ValueError("Filling coefficients (%r,%r) are not integral." % (
-            float_m, float_l))
-    if (m, l) == (0,0):
-        return ((0,0),
-                (0,0))
-
-    n, a, b = xgcd(m, l)
-    if n != 1:
-        raise ValueError("Filling coefficients (%d,%d) are not co-prime." % (
-            m, l))
-
-    return(( m, l),
-           (-b, a))

snappy/drilling/line.py

@@ -1,122 +0,0 @@
-from .fixed_points import r13_fixed_points_of_psl2c_matrix # type: ignore
-
-from ..hyperboloid import r13_dot, o13_inverse # type: ignore
-from ..upper_halfspace import psl2c_to_o13 # type: ignore
-from ..math_basics import is_RealIntervalFieldElement # type: ignore
-from ..sage_helper import _within_sage # type: ignore
-
-if _within_sage:
-    import sage.all # type: ignore
-
-__all__ = [ 'R13Line',
-            'distance_r13_lines',
-            'R13LineWithMatrix' ]
-
-def _safe_sqrt(p):
-    """
-    Compute the sqrt of a number that is known to be non-negative
-    though might not be non-negative because of floating point
-    issues. When using interval arithmetic, this means that
-    while the upper bound will be non-negative, the lower bound
-    we computed might be negative because it is too conservative.
-
-    Example of a quantity that can be given to this function:
-    negative inner product of two vectors in the positive
-    light cone. This is because we know that the inner product
-    of two such vectors is always non-positive.
-    """
-
-    if is_RealIntervalFieldElement(p):
-        RIF = p.parent()
-        p = p.intersection(RIF(0, sage.all.Infinity))
-    else:
-        if p < 0:
-            RF = p.parent()
-            return RF(0)
-    return p.sqrt()
-
-class R13Line:
-    """
-    A line in the hyperboloid model - represented by two
-    like-like vectors spanning the line.
-
-    For distance computations, the inner product between the two
-    vectors is stored as well.
-    """
-
-    def __init__(self,
-                 points, # Two light-like vectors
-                 inner_product=None): # Optional: their inner product
-        """
-        inner_product can be given if known, otherwise, will be computed.
-        """
-        self.points = points
-        if inner_product is None:
-            self.inner_product = r13_dot(points[0], points[1])
-        else:
-            self.inner_product = inner_product
-
-    def transformed(self,
-                    m): # O13-matrix
-        """
-        Returns image of the line under given O13-matrix m.
-        """
-
-        return R13Line(
-            [ m * point for point in self.points],
-            self.inner_product)
-
-
-def distance_r13_lines(line0 : R13Line, line1 : R13Line):
-    """
-    Computes distance between two hyperbolic lines.
-    """
-
-    p00 = r13_dot(line0.points[0], line1.points[0])
-    p01 = r13_dot(line0.points[0], line1.points[1])
-    p10 = r13_dot(line0.points[1], line1.points[0])
-    p11 = r13_dot(line0.points[1], line1.points[1])
-
-    pp = line0.inner_product * line1.inner_product
-
-    t0 = _safe_sqrt((p00 * p11) / pp)
-    t1 = _safe_sqrt((p01 * p10) / pp)
-
-    p = (t0 + t1 - 1) / 2
-
-    return 2 * _safe_sqrt(p).arcsinh()
-
-
-class R13LineWithMatrix:
-    """
-    A line in the hyperboloid model together with a O(1,3)-matrix fixing
-    the line (set-wise).
-    """
-    def __init__(self,
-                 r13_line : R13Line,
-                 o13_matrix):
-        self.r13_line = r13_line
-        self.o13_matrix = o13_matrix
-
-    @staticmethod
-    def from_psl2c_matrix(m):
-        """
-        Given a loxodromic PSL(2,C)-matrix m, returns the line (together
-        with the O(1,3)-matrix corresponding to m) fixed by m in
-        the hyperboloid model.
-        """
-
-        return R13LineWithMatrix(
-            R13Line(r13_fixed_points_of_psl2c_matrix(m)),
-            psl2c_to_o13(m))
-
-    def transformed(self, m):
-        """
-        Returns image of line with matrix under given O13-matrix m.
-
-        That is, the matrix will be conjugated by m so that the new
-        matrix will fix the image of the line (set-wise).
-        """
-        return R13LineWithMatrix(
-            self.r13_line.transformed(m),
-            m * self.o13_matrix * o13_inverse(m))

snappy/drilling/quotient_space.py

@@ -1,94 +0,0 @@
-from .spatial_dict import SpatialDict, floor_as_integers
-from .line import R13Line, R13LineWithMatrix
-from .geodesic_info import LiftedTetrahedron
-
-from ..hyperboloid import ( # type: ignore
-    r13_dot, o13_inverse, distance_unit_time_r13_points)
-from ..snap.t3mlite import Mcomplex # type: ignore
-from ..matrix import matrix # type: ignore
-
-
-def balance_end_points_of_line(line_with_matrix : R13LineWithMatrix,
-                               point) -> R13LineWithMatrix:
-    return R13LineWithMatrix(
-        R13Line(
-            [ endpoint / -r13_dot(point, endpoint)
-              for endpoint in line_with_matrix.r13_line.points]),
-        line_with_matrix.o13_matrix)
-
-
-class ZQuotientLiftedTetrahedronSet:
-    def __init__(self,
-                 mcomplex : Mcomplex,
-                 line_with_matrix : R13LineWithMatrix):
-        self._dict = _ZQuotientDict(mcomplex, line_with_matrix)
-        self._mcomplex = mcomplex
-
-    def add(self, lifted_tetrahedron : LiftedTetrahedron) -> bool:
-        tets = self._dict.setdefault(
-            lifted_tetrahedron.o13_matrix * self._mcomplex.R13_baseTetInCenter,
-            set())
-        if lifted_tetrahedron.tet in tets:
-            return False
-        tets.add(lifted_tetrahedron.tet)
-        return True
-
-
-class _ZQuotientDict(SpatialDict):
-    def __init__(self,
-                 mcomplex : Mcomplex,
-                 line_with_matrix : R13LineWithMatrix):
-        super().__init__(mcomplex.baseTetInRadius, mcomplex.verified)
-
-        self._line = line_with_matrix.r13_line
-        self._power_cache = _O13MatrixPowerCache(line_with_matrix.o13_matrix)
-
-        a = self._line.points[0][0]
-        b = (line_with_matrix.o13_matrix * self._line.points[0])[0]
-
-        self._log_scale_factor = 2 * (b / a).log()
-
-        RF = a.parent()
-        self._weights = [ RF(1.2003), RF(0.94553), RF(1.431112)]
-
-    def distance(self, point_0, point_1):
-        return distance_unit_time_r13_points(point_0, point_1)
-
-    def representatives(self, point):
-
-        a = r13_dot(point, self._line.points[0])
-        b = r13_dot(point, self._line.points[1])
-
-        r = (a / b).log() / self._log_scale_factor
-
-        return [ self._power_cache.power(i) * point
-                 for i in floor_as_integers(r) ]
-
-    def float_hash(self, pt):
-        return (pt[0] * self._weights[0] +
-                pt[1] * self._weights[1] +
-                pt[2] * self._weights[2])
-
-
-class _O13MatrixPowerCache:
-    def __init__(self, m):
-        self._positive_cache = _MatrixNonNegativePowerCache(m)
-        self._negative_cache = _MatrixNonNegativePowerCache(o13_inverse(m))
-
-    def power(self, i):
-        if i >= 0:
-            return self._positive_cache.power( i)
-        else:
-            return self._negative_cache.power(-i)
-
-
-class _MatrixNonNegativePowerCache:
-    def __init__(self, m):
-        self._m = m
-        self._powers = [ matrix.identity(ring=m.base_ring(),
-                                         n=m.dimensions()[0]) ]
-
-    def power(self, i):
-        while not i < len(self._powers):
-            self._powers.append(self._m * self._powers[-1])
-        return self._powers[i]