get rendering context

Scripting in Blender with Python, and working on the API

Moderators: jesterKing, stiv

Post Reply
donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

get rendering context

Post by donato » Fri May 18, 2012 7:30 pm

Hello everyone.
I tryed google for this, but no luck.
I need to get the current render context in a Python script, what's the right syntax?

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

Re: get rendering context

Post by donato » Sat May 19, 2012 1:01 pm

donato wrote:Hello everyone.
I tryed google for this, but no luck.
I need to get the current render context in a Python script, what's the right syntax?
up

CoDEmanX
Posts: 894
Joined: Sun Apr 05, 2009 7:42 pm
Location: Germany

Post by CoDEmanX » Sat May 19, 2012 3:58 pm

what would this context be for example?
I'm sitting, waiting, wishing, building Blender in superstition...

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

Post by donato » Sun May 20, 2012 3:16 pm

i'm trying to port vector rendering method to python 3 and at a certain line it needs to get the render context

CoDEmanX
Posts: 894
Joined: Sun Apr 05, 2009 7:42 pm
Location: Germany

Post by CoDEmanX » Sun May 20, 2012 8:24 pm

question remains
I'm sitting, waiting, wishing, building Blender in superstition...

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

Post by donato » Sun May 20, 2012 8:30 pm

i'm unsure, the old code says just to get the current rendering context and i'm new to both python and blender scripting

CoDEmanX
Posts: 894
Joined: Sun Apr 05, 2009 7:42 pm
Location: Germany

Post by CoDEmanX » Sun May 20, 2012 11:25 pm

after reading
http://wiki.blender.org/index.php/Dev:R ... r#Concepts
and
http://en.wikibooks.org/wiki/Blender_3D ... er_Context

i'd say, the context is bpy.context and render context = current settings for rendering, stored here:

bpy.context.scene.render

also relevant is bpy.context.scene, as it holds the frame_start, _current and _end values
I'm sitting, waiting, wishing, building Blender in superstition...

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

Post by donato » Sun May 20, 2012 11:58 pm

thanks so much, i'll try asap

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

almost done

Post by donato » Fri May 25, 2012 8:49 am

I'm almost done with syntax, I just need to know where to find the base filename:

Code: Select all

basename = Blender.sys.basename(Blender.Get('filename'))
What's this in Blender 2.6x and Python 3?

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

Re: almost done

Post by donato » Sat May 26, 2012 3:06 pm

donato wrote:I'm almost done with syntax, I just need to know where to find the base filename:

Code: Select all

basename = Blender.sys.basename(Blender.Get('filename'))
What's this in Blender 2.6x and Python 3?
up, does anyone know?

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

Re: almost done

Post by donato » Sun May 27, 2012 10:00 am

donato wrote:
donato wrote:I'm almost done with syntax, I just need to know where to find the base filename:

Code: Select all

basename = Blender.sys.basename(Blender.Get('filename'))
What's this in Blender 2.6x and Python 3?
up, does anyone know?
up

CoDEmanX
Posts: 894
Joined: Sun Apr 05, 2009 7:42 pm
Location: Germany

Post by CoDEmanX » Mon May 28, 2012 7:58 pm

Code: Select all

bpy.path.basename(bpy.data.filepath)
:)
I'm sitting, waiting, wishing, building Blender in superstition...

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

Thanks

Post by donato » Mon May 28, 2012 8:09 pm

Thanks, now I'm done with syntax. But the script wont show the gui, here's the code as I can't upload anywhere, hoping someone will help me.

Code: Select all

#!BPY
"""
Name: 'VRM'
Blender: 245
Group: 'Render'
Tooltip: 'Vector Rendering Method script'
"""

__author__ = "Antonio Ospite"
__url__ = ["http://vrm.ao2.it"]
__version__ = "0.3"

__bpydoc__ = """\
    Render the scene and save the result in vector format.
"""

# ---------------------------------------------------------------------
#    Copyright (c) 2006, 2007, 2008, 2009, 2012 Antonio Ospite
#
#    This program is free software; you can redistribute it and/or modify
#    it under the terms of the GNU General Public License as published by
#    the Free Software Foundation; either version 2 of the License, or
#    (at your option) any later version.
#
#    This program is distributed in the hope that it will be useful,
#    but WITHOUT ANY WARRANTY; without even the implied warranty of
#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#    GNU General Public License for more details.
#
#    You should have received a copy of the GNU General Public License
#    along with this program; if not, write to the Free Software
#    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
#
# ---------------------------------------------------------------------
#
# Additional credits:
#   Thanks to Emilio Aguirre for S2flender from which I took inspirations :)
#   Thanks to Nikola Radovanovic, the author of the original VRM script,
#       the code you read here has been rewritten _almost_ entirely
#       from scratch but Nikola gave me the idea, so I thank him publicly.
#
# ---------------------------------------------------------------------
#
# Things TODO for a next release:
#   - Shadeless shader
#   - FIX the issue with negative scales in object tranformations!
#   - Use a better depth sorting algorithm
#   - Review how selections are made (this script uses selection states of
#     primitives to represent visibility infos)
#   - Use a data structure other than Mesh to represent the 2D image?
#     Think to a way to merge (adjacent) polygons that have the same color.
#     Or a way to use paths for silhouettes and contours.
#   - Consider SMIL for animation handling instead of ECMA Script? (Firefox do
#     not support SMIL for animations)
#   - Switch to the Mesh structure, should be considerably faster
#     (partially done, but with Mesh we cannot sort faces, yet)
#   - Implement Edge Styles (silhouettes, contours, etc.) (partially done).
#   - Implement Shading Styles? (partially done, to make more flexible).
#   - Add Vector Writers other than SVG.
#   - set the background color!
#   - Check memory use!!
#
# ---------------------------------------------------------------------

import bpy
from bpy import *
from mathutils import *
from math import *
import sys
import time

try:
    set()
except NameError:
    from sets import Set as set


def uniq(alist):
    tmpdict = dict()
    return [tmpdict.setdefault(e, e) for e in alist if e not in tmpdict]
    # in python > 2.4 we ca use the following
    #return [ u for u in alist if u not in locals()['_[1]'] ]


# Constants
EPS = 10e-5

# We use a global progress Indicator Object
progresses = None


# Config class for global settings

class config:
    polygons = dict()
    polygons['SHOW'] = True
    polygons['SHADING'] = 'FLAT'  # FLAT or TOON
    polygons['HSR'] = 'PAINTER'  # PAINTER or NEWELL
    # Hidden to the user for now
    polygons['EXPANSION_TRICK'] = True

    polygons['TOON_LEVELS'] = 2

    edges = dict()
    edges['SHOW'] = False
    edges['SHOW_HIDDEN'] = False
    edges['STYLE'] = 'MESH'  # MESH or SILHOUETTE
    edges['WIDTH'] = 2
    edges['COLOR'] = [0, 0, 0]

    output = dict()
    output['FORMAT'] = 'SVG'
    output['ANIMATION'] = False
    output['JOIN_OBJECTS'] = True

    def saveToRegistry():
        registry = {}

        for k, v in config.__dict__.iteritems():

            # config class store settings in dictionaries
            if v.__class__ == dict().__class__:

                regkey_prefix = k.upper() + "_"

                for opt_k, opt_v in v.iteritems():
                    regkey = regkey_prefix + opt_k

                    registry[regkey] = opt_v

        Blender.Registry.SetKey('VRM', registry, True)

    saveToRegistry = staticmethod(saveToRegistry)


# Utility functions
print_debug = False


def dumpfaces(flist, filename):
    """Dump a single face to a file.
    """
    if not print_debug:
        return

    class tmpmesh:
        pass

    m = tmpmesh()
    m.faces = flist

    writerobj = SVGVectorWriter(filename)

    writerobj.open()
    writerobj._printPolygons(m)

    writerobj.close()


def debug(msg):
    if print_debug:
        sys.stderr.write(msg)


def EQ(v1, v2):
    return (abs(v1[0] - v2[0]) < EPS and
            abs(v1[1] - v2[1]) < EPS)
by_furthest_z = (lambda f1, f2:
    cmp(max([v.co[2] for v in f1]), max([v.co[2] for v in f2]) + EPS)
    )


def sign(x):

    if x < -EPS:
    #if x < 0:
        return -1
    elif x > EPS:
    #elif x > 0:
        return 1
    else:
        return 0


# ---------------------------------------------------------------------
#
## HSR Utility class
#
# ---------------------------------------------------------------------

EPS = 10e-5
INF = 10e5


class HSR:
    """A utility class for HSR processing.
    """

    def is_nonplanar_quad(face):
        """Determine if a quad is non-planar.

        From: http://mathworld.wolfram.com/Coplanar.html

        Geometric objects lying in a common plane are said to be coplanar.
        Three noncollinear points determine a plane and so are trivially
        coplanar. Four points are coplanar iff the volume of the tetrahedron
        defined by them is 0,

            | x_1 y_1 z_1 1 |
            | x_2 y_2 z_2 1 |
            | x_3 y_3 z_3 1 |
            | x_4 y_4 z_4 1 | == 0

        Coplanarity is equivalent to the statement that the pair of lines
        determined by the four points are not skew, and can be equivalently
        stated in vector form as (x_3-x_1).[(x_2-x_1)x(x_4-x_3)]==0.

        An arbitrary number of n points x_1, ..., x_n can be tested for
        coplanarity by finding the point-plane distances of the points
        x_4, ..., x_n from the plane determined by (x_1,x_2,x_3)
        and checking if they are all zero.
        If so, the points are all coplanar.

        We here check only for 4-point complanarity.
        """
        n = len(face)

        # assert(n>4)
        if n < 3 or n > 4:
            print ("ERROR a mesh in Blender can't have more than 4 vertices or less than 3")
            raise AssertionError

        elif n == 3:
            # three points must be complanar
            return False
        else:  # n == 4
            x1 = Vector(face[0].co)
            x2 = Vector(face[1].co)
            x3 = Vector(face[2].co)
            x4 = Vector(face[3].co)

            v = (x3 - x1) * CrossVecs((x2 - x1), (x4 - x3))
            if v != 0:
                return True

        return False

    is_nonplanar_quad = staticmethod(is_nonplanar_quad)

    def pointInPolygon(poly, v):
        return False

    pointInPolygon = staticmethod(pointInPolygon)

    def edgeIntersection(s1, s2, do_perturbate=False):

        (x1, y1) = s1[0].co[0], s1[0].co[1]
        (x2, y2) = s1[1].co[0], s1[1].co[1]

        (x3, y3) = s2[0].co[0], s2[0].co[1]
        (x4, y4) = s2[1].co[0], s2[1].co[1]

        #z1 = s1[0].co[2]
        #z2 = s1[1].co[2]
        #z3 = s2[0].co[2]
        #z4 = s2[1].co[2]

        # calculate delta values (vector components)
        dx1 = x2 - x1
        dx2 = x4 - x3
        dy1 = y2 - y1
        dy2 = y4 - y3

        #dz1 = z2 - z1
        #dz2 = z4 - z3

        C = dy2 * dx1 - dx2 * dy1  # cross product
        if C == 0:  # parallel
            return None

        dx3 = x1 - x3  # combined origin offset vector
        dy3 = y1 - y3

        a1 = (dy3 * dx2 - dx3 * dy2) / C
        a2 = (dy3 * dx1 - dx3 * dy1) / C

        # check for degeneracies
        #print_debug("\n")
        #print_debug(str(a1)+"\n")
        #print_debug(str(a2)+"\n\n")

        if (a1 == 0 or a1 == 1 or a2 == 0 or a2 == 1):
            # Intersection on boundaries, we consider the point external?
            return None

        elif (a1 > 0.0 and a1 < 1.0 and a2 > 0.0 and a2 < 1.0):  # lines cross
            x = x1 + a1 * dx1
            y = y1 + a1 * dy1

            #z = z1 + a1 * dz1
            z = 0
            return (NMesh.Vert(x, y, z), a1, a2)

        else:
            # lines have intersections but not those segments
            return None

    edgeIntersection = staticmethod(edgeIntersection)

    def isVertInside(self, v):
        winding_number = 0
        coincidence = False

        # Create point at infinity
        point_at_infinity = NMesh.Vert(-INF, v.co[1], -INF)

        for i in range(len(self.v)):
            s1 = (point_at_infinity, v)
            s2 = (self.v[i - 1], self.v[i])

            if EQ(v.co, s2[0].co) or EQ(v.co, s2[1].co):
                coincidence = True

            if HSR.edgeIntersection(s1, s2, do_perturbate=False):
                winding_number += 1

        # Check even or odd
        if (winding_number % 2) == 0:
            return False
        else:
            if coincidence:
                return False
            return True

    isVertInside = staticmethod(isVertInside)

    def det(a, b, c):
        return ((b[0] - a[0]) * (c[1] - a[1]) -
                (b[1] - a[1]) * (c[0] - a[0]))

    det = staticmethod(det)

    def pointInPolygon(q, P):
        is_in = False

        point_at_infinity = NMesh.Vert(-INF, q.co[1], -INF)

        det = HSR.det

        for i in range(len(P.v)):
            p0 = P.v[i - 1]
            p1 = P.v[i]
            if (det(q.co, point_at_infinity.co, p0.co) < 0) != (det(q.co, point_at_infinity.co, p1.co) < 0):
                if det(p0.co, p1.co, q.co) == 0:
                    #print "On Boundary"
                    return False
                elif (det(p0.co, p1.co, q.co) < 0) != (det(p0.co, p1.co, point_at_infinity.co) < 0):
                    is_in = not is_in

        return is_in

    pointInPolygon = staticmethod(pointInPolygon)

    def projectionsOverlap(f1, f2):
        """ If you have nonconvex, but still simple polygons, an acceptable method
        is to iterate over all vertices and perform the Point-in-polygon test[1].
        The advantage of this method is that you can compute the exact
        intersection point and collision normal that you will need to simulate
        collision. When you have the point that lies inside the other polygon, you
        just iterate over all edges of the second polygon again and look for edge
        intersections. Note that this method detects collsion when it already
        happens. This algorithm is fast enough to perform it hundreds of times per
        sec.  """

        for i in range(len(f1.v)):

            # If a point of f1 in inside f2, there is an overlap!
            v1 = f1.v[i]
            #if HSR.isVertInside(f2, v1):
            if HSR.pointInPolygon(v1, f2):
                return True

            # If not the polygon can be ovelap as well, so we check for
            # intersection between an edge of f1 and all the edges of f2

            v0 = f1.v[i - 1]

            for j in range(len(f2.v)):
                v2 = f2.v[j - 1]
                v3 = f2.v[j]

                e1 = v0, v1
                e2 = v2, v3

                intrs = HSR.edgeIntersection(e1, e2)
                if intrs:
                    #print_debug(str(v0.co) + " " + str(v1.co) + " " +
                    #        str(v2.co) + " " + str(v3.co) )
                    #print_debug("\nIntersection\n")

                    return True

        return False

    projectionsOverlap = staticmethod(projectionsOverlap)

    def midpoint(p1, p2):
        """Return the midpoint of two vertices.
        """
        m = MidpointVecs(Vector(p1), Vector(p2))
        mv = NMesh.Vert(m[0], m[1], m[2])

        return mv

    midpoint = staticmethod(midpoint)

    def facesplit(P, Q, facelist, nmesh):
        """Split P or Q according to the strategy illustrated in the Newell's
        paper.
        """

        by_furthest_z = (lambda f1, f2:
                cmp(max([v.co[2] for v in f1]), max([v.co[2] for v in f2]) + EPS)
                )

        # Choose if split P on Q plane or vice-versa

        n = 0
        for Pi in P:
            d = HSR.Distance(Vector(Pi), Q)
            if d <= EPS:
                n += 1
        pIntersectQ = (n != len(P))

        n = 0
        for Qi in Q:
            d = HSR.Distance(Vector(Qi), P)
            if d >= -EPS:
                n += 1
        qIntersectP = (n != len(Q))

        newfaces = []

        # 1. If parts of P lie in both half-spaces of Q
        # then splice P in two with the plane of Q
        if pIntersectQ:
            #print "We split P"
            f = P
            plane = Q

            newfaces = HSR.splitOn(plane, f)

        # 2. Else if parts of Q lie in both half-space of P
        # then splice Q in two with the plane of P
        if qIntersectP and newfaces == None:
            #print "We split Q"
            f = Q
            plane = P

            newfaces = HSR.splitOn(plane, f)
            #print "After"

        # 3. Else slice P in half through the mid-point of
        # the longest pair of opposite sides
        if newfaces == None:

            print("We ignore P...")
            facelist.remove(P)
            return facelist

            #f = P

            #if len(P)==3:
            #    v1 = midpoint(f[0], f[1])
            #    v2 = midpoint(f[1], f[2])
            #if len(P)==4:
            #    v1 = midpoint(f[0], f[1])
            #    v2 = midpoint(f[2], f[3])
            #vec3 = (Vector(v2)+10*Vector(f.normal))
            #
            #v3 = NMesh.Vert(vec3[0], vec3[1], vec3[2])

            #plane = NMesh.Face([v1, v2, v3])
            #
            #newfaces = splitOn(plane, f)

        if newfaces == None:
            print("Big FAT problem, we weren't able to split POLYGONS!")
            raise AssertionError

        #print newfaces
        if newfaces:
            #for v in f:
            #    if v not in plane and v in nmesh.verts:
            #        nmesh.verts.remove(v)
            for nf in newfaces:

                nf.mat = f.mat
                nf.sel = f.sel
                nf.col = [f.col[0]] * len(nf.v)

                nf.smooth = 0

                for v in nf:
                    nmesh.verts.append(v)
                # insert pieces in the list
                facelist.append(nf)

            facelist.remove(f)

        # and resort the faces
        facelist.sort(by_furthest_z)
        facelist.sort(lambda f1, f2: cmp(f1.smooth, f2.smooth))
        facelist.reverse()

        #print [ f.smooth for f in facelist ]

        return facelist

    facesplit = staticmethod(facesplit)

    def isOnSegment(v1, v2, p, extremes_internal=False):
        """Check if point p is in segment v1v2.
        """

        l1 = (v1 - p).length
        l2 = (v2 - p).length

        # Should we consider extreme points as internal ?
        # The test:
        # if p == v1 or p == v2:
        if l1 < EPS or l2 < EPS:
            return extremes_internal

        l = (v1 - v2).length

        # if the sum of l1 and l2 is circa l, then the point is on segment,
        if abs(l - (l1 + l2)) < EPS:
            return True
        else:
            return False

    isOnSegment = staticmethod(isOnSegment)

    def Distance(point, face):
        """ Calculate the distance between a point and a face.

        An alternative but more expensive method can be:

            ip = Intersect(Vector(face[0]), Vector(face[1]), Vector(face[2]),
                    Vector(face.no), Vector(point), 0)

            d = Vector(ip - point).length

        See: http://mathworld.wolfram.com/Point-PlaneDistance.html
        """

        p = Vector(point)
        plNormal = Vector(face.no)
        plVert0 = Vector(face.v[0])

        d = (plVert0 * plNormal) - (p * plNormal)

        #d = plNormal * (plVert0 - p)

        #print "\nd: %.10f - sel: %d, %s\n" % (d, face.sel, str(point))

        return d

    Distance = staticmethod(Distance)

    def makeFaces(vl):
        #
        # make one or two new faces based on a list of vertex-indices
        #
        newfaces = []

        if len(vl) <= 4:
            nf = NMesh.Face()

            for v in vl:
                nf.v.append(v)

            newfaces.append(nf)

        else:
            nf = NMesh.Face()

            nf.v.append(vl[0])
            nf.v.append(vl[1])
            nf.v.append(vl[2])
            nf.v.append(vl[3])
            newfaces.append(nf)

            nf = NMesh.Face()
            nf.v.append(vl[3])
            nf.v.append(vl[4])
            nf.v.append(vl[0])
            newfaces.append(nf)

        return newfaces

    makeFaces = staticmethod(makeFaces)

    def splitOn(Q, P, return_positive_faces=True, return_negative_faces=True):
        """Split P using the plane of Q.
        Logic taken from the knife.py python script
        """

        # Check if P and Q are parallel
        u = CrossVecs(Vector(Q.no), Vector(P.no))
        ax = abs(u[0])
        ay = abs(u[1])
        az = abs(u[2])

        if (ax + ay + az) < EPS:
            print("PARALLEL planes!!")
            return

        # The final aim is to find the intersection line between P
        # and the plane of Q, and split P along this line

        nP = len(P.v)

        # Calculate point-plane Distance between vertices of P and plane Q
        d = []
        for i in range(0, nP):
            d.append(HSR.Distance(P.v[i], Q))

        newVertList = []

        posVertList = []
        negVertList = []
        for i in range(nP):
            d0 = d[i - 1]
            V0 = P.v[i - 1]

            d1 = d[i]
            V1 = P.v[i]

            #print "d0:", d0, "d1:", d1

            # if the vertex lies in the cutplane
            if abs(d1) < EPS:
                #print "d1 On cutplane"
                posVertList.append(V1)
                negVertList.append(V1)
            else:
                # if the previous vertex lies in cutplane
                if abs(d0) < EPS:
                    #print "d0 on Cutplane"
                    if d1 > 0:
                        #print "d1 on positive Halfspace"
                        posVertList.append(V1)
                    else:
                        #print "d1 on negative Halfspace"
                        negVertList.append(V1)
                else:
                    # if they are on the same side of the plane
                    if (d1 * d0) > 0:
                        #print "On the same half-space"
                        if d1 > 0:
                            #print "d1 on positive Halfspace"
                            posVertList.append(V1)
                        else:
                            #print "d1 on negative Halfspace"
                            negVertList.append(V1)

                    # the vertices are not on the same side of the plane, so we have an intersection
                    else:
                        #print "Intersection"

                        e = Vector(V0), Vector(V1)
                        tri = Vector(Q[0]), Vector(Q[1]), Vector(Q[2])

                        inters = Intersect(tri[0], tri[1], tri[2], e[1] - e[0], e[0], 0)
                        if inters == None:
                            print("Split Break")
                            break

                        #print "Intersection", inters

                        nv = NMesh.Vert(inters[0], inters[1], inters[2])
                        newVertList.append(nv)

                        posVertList.append(nv)
                        negVertList.append(nv)

                        if d1 > 0:
                            posVertList.append(V1)
                        else:
                            negVertList.append(V1)

        # uniq for python > 2.4
        #posVertList = [ u for u in posVertList if u not in locals()['_[1]'] ]
        #negVertList = [ u for u in negVertList if u not in locals()['_[1]'] ]

        # a more portable way
        posVertList = uniq(posVertList)
        negVertList = uniq(negVertList)

        # If vertex are all on the same half-space, return
        #if len(posVertList) < 3:
        #    print "Problem, we created a face with less that 3 vertices??"
        #    posVertList = []
        #if len(negVertList) < 3:
        #    print "Problem, we created a face with less that 3 vertices??"
        #    negVertList = []

        if len(posVertList) < 3 or len(negVertList) < 3:
            #print "RETURN NONE, SURE???"
            return None

        if not return_positive_faces:
            posVertList = []
        if not return_negative_faces:
            negVertList = []

        newfaces = HSR.addNewFaces(posVertList, negVertList)

        return newfaces

    splitOn = staticmethod(splitOn)

    def addNewFaces(posVertList, negVertList):
        # Create new faces resulting from the split
        outfaces = []
        if len(posVertList) or len(negVertList):

            #newfaces = [posVertList] + [negVertList]
            newfaces = ([[NMesh.Vert(v[0], v[1], v[2]) for v in posVertList]] +
                    [[NMesh.Vert(v[0], v[1], v[2]) for v in negVertList]])

            for nf in newfaces:
                if nf and len(nf) > 2:
                    outfaces += HSR.makeFaces(nf)

        return outfaces

    addNewFaces = staticmethod(addNewFaces)


# ---------------------------------------------------------------------
#
## Mesh Utility class
#
# ---------------------------------------------------------------------

class MeshUtils:

    def buildEdgeFaceUsersCache(me):
        '''
        Takes a mesh and returns a list aligned with the meshes edges.
        Each item is a list of the faces that use the edge
        would be the equiv for having ed.face_users as a property

        Taken from .blender/scripts/bpymodules/BPyMesh.py,
        thanks to ideasman_42.
        '''

        def sorted_edge_indicies(ed):
            i1 = ed.v1.index
            i2 = ed.v2.index
            if i1 > i2:
                i1, i2 = i2, i1
            return i1, i2

        face_edges_dict = dict([(sorted_edge_indicies(ed), (ed.index, [])) for ed in me.edges])
        for f in me.faces:
            fvi = [v.index for v in f.v]  # face vert idx's
            for i in xrange(len(f)):
                i1 = fvi[i]
                i2 = fvi[i - 1]

                if i1 > i2:
                    i1, i2 = i2, i1

                face_edges_dict[i1, i2][1].append(f)

        face_edges = [None] * len(me.edges)
        for ed_index, ed_faces in face_edges_dict.itervalues():
            face_edges[ed_index] = ed_faces

        return face_edges

    def isMeshEdge(adjacent_faces):
        """Mesh edge rule.

        A mesh edge is visible if _at_least_one_ of its adjacent faces is selected.
        Note: if the edge has no adjacent faces we want to show it as well,
        useful for "edge only" portion of objects.
        """

        if len(adjacent_faces) == 0:
            return True

        selected_faces = [f for f in adjacent_faces if f.sel]

        if len(selected_faces) != 0:
            return True
        else:
            return False

    def isSilhouetteEdge(adjacent_faces):
        """Silhuette selection rule.

        An edge is a silhuette edge if it is shared by two faces with
        different selection status or if it is a boundary edge of a selected
        face.
        """

        if ((len(adjacent_faces) == 1 and adjacent_faces[0].sel == 1) or
            (len(adjacent_faces) == 2 and
                adjacent_faces[0].sel != adjacent_faces[1].sel)
            ):
            return True
        else:
            return False

    buildEdgeFaceUsersCache = staticmethod(buildEdgeFaceUsersCache)
    isMeshEdge = staticmethod(isMeshEdge)
    isSilhouetteEdge = staticmethod(isSilhouetteEdge)


# ---------------------------------------------------------------------
#
## Shading Utility class
#
# ---------------------------------------------------------------------

class ShadingUtils:

    shademap = None

    def toonShadingMapSetup():
        levels = config.polygons['TOON_LEVELS']

        texels = 2 * levels - 1
        tmp_shademap = [0.0] + [(i) / float(texels - 1) for i in xrange(1, texels - 1)] + [1.0]

        return tmp_shademap

    def toonShading(u):

        shademap = ShadingUtils.shademap

        if not shademap:
            shademap = ShadingUtils.toonShadingMapSetup()

        v = 1.0
        for i in xrange(0, len(shademap) - 1):
            pivot = (shademap[i] + shademap[i + 1]) / 2.0
            j = int(u > pivot)

            v = shademap[i + j]

            if v < shademap[i + 1]:
                return v

        return v

    toonShadingMapSetup = staticmethod(toonShadingMapSetup)
    toonShading = staticmethod(toonShading)


# ---------------------------------------------------------------------
#
## Projections classes
#
# ---------------------------------------------------------------------

class Projector:
    """Calculate the projection of an object given the camera.

    A projector is useful to so some per-object transformation to obtain the
    projection of an object given the camera.

    The main method is #doProjection# see the method description for the
    parameter list.
    """

    def __init__(self, cameraObj, canvasRatio):
        """Calculate the projection matrix.

        The projection matrix depends, in this case, on the camera settings.
        TAKE CARE: This projector expects vertices in World Coordinates!
        """

        camera = cameraObj.getData()

        aspect = float(canvasRatio[0]) / float(canvasRatio[1])
        near = camera.clipStart
        far = camera.clipEnd

        scale = float(camera.scale)

        fovy = atan(0.5 / aspect / (camera.lens / 32))
        fovy = fovy * 360.0 / pi

        if Blender.Get('version') < 243:
            camPersp = 0
            camOrtho = 1
        else:
            camPersp = 'persp'
            camOrtho = 'ortho'

        # What projection do we want?
        if camera.type == camPersp:
            mP = self._calcPerspectiveMatrix(fovy, aspect, near, far)
        elif camera.type == camOrtho:
            mP = self._calcOrthoMatrix(fovy, aspect, near, far, scale)

        # View transformation
        cam = Matrix(cameraObj.getInverseMatrix())
        cam.transpose()

        mP = mP * cam

        self.projectionMatrix = mP

    ##
    # Public methods
    #

    def doProjection(self, v):
        """Project the point on the view plane.

        Given a vertex calculate the projection using the current projection
        matrix.
        """

        # Note that we have to work on the vertex using homogeneous coordinates
        # From blender 2.42+ we don't need to resize the vector to be 4d
        # when applying a 4x4 matrix, but we do that anyway since we need the
        # 4th coordinate later
        p = self.projectionMatrix * Vector(v).resize4D()

        # Perspective division
        if p[3] != 0:
            p[0] = p[0] / p[3]
            p[1] = p[1] / p[3]
            p[2] = p[2] / p[3]

        # restore the size
        p[3] = 1.0
        p.resize3D()

        return p

    ##
    # Private methods
    #

    def _calcPerspectiveMatrix(self, fovy, aspect, near, far):
        """Return a perspective projection matrix.
        """

        top = near * tan(fovy * pi / 360.0)
        bottom = -top
        left = bottom * aspect
        right = top * aspect
        x = (2.0 * near) / (right - left)
        y = (2.0 * near) / (top - bottom)
        a = (right + left) / (right - left)
        b = (top + bottom) / (top - bottom)
        c = - ((far + near) / (far - near))
        d = - ((2 * far * near) / (far - near))

        m = Matrix(
                [x,   0.0,    a,    0.0],
                [0.0,   y,    b,    0.0],
                [0.0, 0.0,    c,      d],
                [0.0, 0.0, -1.0,    0.0])

        return m

    def _calcOrthoMatrix(self, fovy, aspect, near, far, scale):
        """Return an orthogonal projection matrix.
        """

        # The 11 in the formula was found emiprically
        top = near * tan(fovy * pi / 360.0) * (scale * 11)
        bottom = -top
        left = bottom * aspect
        right = top * aspect
        rl = right - left
        tb = top - bottom
        fn = near - far
        tx = -((right + left) / rl)
        ty = -((top + bottom) / tb)
        tz = ((far + near) / fn)

        m = Matrix(
                [2.0 / rl, 0.0,      0.0,       tx],
                [0.0,      2.0 / tb, 0.0,       ty],
                [0.0,      0.0,      2.0 / fn,  tz],
                [0.0,      0.0,      0.0,      1.0])

        return m


# ---------------------------------------------------------------------
#
## Progress Indicator
#
# ---------------------------------------------------------------------

class Progress:
    """A model for a progress indicator.

    Do the progress calculation calculation and
    the view independent stuff of a progress indicator.
    """
    def __init__(self, steps=0):
        self.name = ""
        self.steps = steps
        self.completed = 0
        self.progress = 0

    def setSteps(self, steps):
        """Set the number of steps of the activity wich we want to track.
        """
        self.steps = steps

    def getSteps(self):
        return self.steps

    def setName(self, name):
        """Set the name of the activity wich we want to track.
        """
        self.name = name

    def getName(self):
        return self.name

    def getProgress(self):
        return self.progress

    def reset(self):
        self.completed = 0
        self.progress = 0

    def update(self):
        """Update the model, call this method when one step is completed.
        """
        if self.progress == 100:
            return False

        self.completed += 1
        self.progress = (float(self.completed) / float(self.steps)) * 100
        self.progress = int(self.progress)

        return True


class ProgressIndicator:
    """An abstraction of a View for the Progress Model
    """
    def __init__(self):

        # Use a refresh rate so we do not show the progress at
        # every update, but every 'self.refresh_rate' times.
        self.refresh_rate = 10
        self.shows_counter = 0

        self.quiet = False

        self.progressModel = None

    def setQuiet(self, value):
        self.quiet = value

    def setActivity(self, name, steps):
        """Initialize the Model.

        In a future version (with subactivities-progress support) this method
        could only set the current activity.
        """
        self.progressModel = Progress()
        self.progressModel.setName(name)
        self.progressModel.setSteps(steps)

    def getActivity(self):
        return self.progressModel

    def update(self):
        """Update the model and show the actual progress.
        """
        assert(self.progressModel)

        if self.progressModel.update():
            if self.quiet:
                return

            self.show(self.progressModel.getProgress(),
                    self.progressModel.getName())

        # We return always True here so we can call the update() method also
        # from lambda funcs (putting the call in logical AND with other ops)
        return True

    def show(self, progress, name=""):
        self.shows_counter = (self.shows_counter + 1) % self.refresh_rate
        if self.shows_counter != 0:
            return

        if progress == 100:
            self.shows_counter = -1


class ConsoleProgressIndicator(ProgressIndicator):
    """Show a progress bar on stderr, a la wget.
    """
    def __init__(self):
        ProgressIndicator.__init__(self)

        self.swirl_chars = ["-", "\\", "|", "/"]
        self.swirl_count = -1

    def show(self, progress, name):
        ProgressIndicator.show(self, progress, name)

        bar_length = 70
        bar_progress = int((progress / 100.0) * bar_length)
        bar = ("=" * bar_progress).ljust(bar_length)

        self.swirl_count = (self.swirl_count + 1) % len(self.swirl_chars)
        swirl_char = self.swirl_chars[self.swirl_count]

        progress_bar = "%s |%s| %c %3d%%" % (name, bar, swirl_char, progress)

        sys.stderr.write(progress_bar + "\r")
        if progress == 100:
            sys.stderr.write("\n")


class GraphicalProgressIndicator(ProgressIndicator):
    """Interface to the Blender.Window.DrawProgressBar() method.
    """
    def __init__(self):
        ProgressIndicator.__init__(self)

        #self.swirl_chars = ["-", "\\", "|", "/"]
        # We have to use letters with the same width, for now!
        # Blender progress bar considers the font widths when
        # calculating the progress bar width.
        self.swirl_chars = ["\\", "/"]
        self.swirl_count = -1

    def show(self, progress, name):
        ProgressIndicator.show(self, progress)

        self.swirl_count = (self.swirl_count + 1) % len(self.swirl_chars)
        swirl_char = self.swirl_chars[self.swirl_count]

        progress_text = "%s - %c %3d%%" % (name, swirl_char, progress)

        # Finally draw  the Progress Bar
        Window.WaitCursor(1)  # Maybe we can move that call in the constructor?
        Window.DrawProgressBar(progress / 100.0, progress_text)

        if progress == 100:
            Window.DrawProgressBar(1, progress_text)
            Window.WaitCursor(0)

# ---------------------------------------------------------------------
#
## 2D Object representation class
#
# ---------------------------------------------------------------------

# TODO: a class to represent the needed properties of a 2D vector image
# For now just using a [N]Mesh structure.


# ---------------------------------------------------------------------
#
## Vector Drawing Classes
#
# ---------------------------------------------------------------------

## A generic Writer

class VectorWriter:
    """
    A class for printing output in a vectorial format.

    Given a 2D representation of the 3D scene the class is responsible to
    write it is a vector format.

    Every subclasses of VectorWriter must have at last the following public
    methods:
        - open(self)
        - close(self)
        - printCanvas(self, scene,
            doPrintPolygons=True, doPrintEdges=False, showHiddenEdges=False):
    """

    def __init__(self, fileName):
        """Set the output file name and other properties"""

        try:
            config.writer
        except:
            config.writer = dict()
            config.writer['SETTING'] = True

        self.outputFileName = fileName

        context = bpy.context.scene.render
        self.canvasSize = (context.resolution_x, context.resolution_y)

        self.fps = context.fps

        self.startFrame = 1
        self.endFrame = 1
        self.animation = False

    ##
    # Public Methods
    #

    def open(self, startFrame=1, endFrame=1):
        if startFrame != endFrame:
            self.startFrame = startFrame
            self.endFrame = endFrame
            self.animation = True

        print("Outputting to: ", self.outputFileName)

        return

    def close(self):
        return

    def printCanvas(self, scene, doPrintPolygons=True, doPrintEdges=False,
            showHiddenEdges=False):
        """This is the interface for the needed printing routine.
        """
        return


## SVG Writer

class SVGVectorWriter(VectorWriter):
    """A concrete class for writing SVG output.
    """

    def __init__(self, fileName):
        """Simply call the parent Contructor.
        """
        VectorWriter.__init__(self, fileName)

        self.file = None

    ##
    # Public Methods
    #

    def open(self, startFrame=1, endFrame=1):
        """Do some initialization operations.
        """
        VectorWriter.open(self, startFrame, endFrame)

        self.file = open(self.outputFileName, "w")

        self._printHeader()

    def close(self):
        """Do some finalization operation.
        """
        self._printFooter()

        if self.file:
            self.file.close()

        # remember to call the close method of the parent as last
        VectorWriter.close(self)

    def printCanvas(self, scene, doPrintPolygons=True, doPrintEdges=False,
            showHiddenEdges=False):
        """Convert the scene representation to SVG.
        """

        Objects = scene.objects

        context = scene.getRenderingContext()
        framenumber = context.currentFrame()

        if self.animation:
            framestyle = "display:none"
        else:
            framestyle = "display:block"

        # Assign an id to this group so we can set properties on it using DOM
        self.file.write("<g id=\"frame%d\" style=\"%s\">\n" %
                (framenumber, framestyle))

        for obj in Objects:

            if obj.getType() != 'Mesh':
                continue

            self.file.write("<g id=\"%s\">\n" % obj.getName())

            mesh = obj.getData(mesh=1)

            if doPrintPolygons:
                self._printPolygons(mesh)

            if doPrintEdges:
                self._printEdges(mesh, showHiddenEdges)

            self.file.write("</g>\n")

        self.file.write("</g>\n")

    ##
    # Private Methods
    #

    def _calcCanvasCoord(self, v):
        """Convert vertex in scene coordinates to canvas coordinates.
        """

        pt = Vector([0, 0, 0])

        mW = float(self.canvasSize[0]) / 2.0
        mH = float(self.canvasSize[1]) / 2.0

        # rescale to canvas size
        pt[0] = v.co[0] * mW + mW
        pt[1] = v.co[1] * mH + mH
        pt[2] = v.co[2]

        # For now we want (0,0) in the top-left corner of the canvas.
        # Mirror and translate along y
        pt[1] *= -1
        pt[1] += self.canvasSize[1]

        return pt

    def _printHeader(self):
        """Print SVG header."""

        self.file.write("<?xml version=\"1.0\"?>\n")
        self.file.write("<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.0//EN\"\n")
        self.file.write("\t\"http://www.w3.org/TR/2001/REC-SVG-20010904/DTD/svg10.dtd\">\n")
        self.file.write("<svg version=\"1.0\"\n")
        self.file.write("\txmlns=\"http://www.w3.org/2000/svg\"\n")
        self.file.write("\twidth=\"%d\" height=\"%d\">\n\n" %
                self.canvasSize)

        if self.animation:
            delay = 1000 / self.fps

            self.file.write("""\n<script type="text/javascript"><![CDATA[
            globalStartFrame=%d;
            globalEndFrame=%d;

            timerID = setInterval("NextFrame()", %d);
            globalFrameCounter=%d;
            \n""" % (self.startFrame, self.endFrame, delay, self.startFrame))

            self.file.write("""\n
            function NextFrame()
            {
              currentElement  = document.getElementById('frame'+globalFrameCounter)
              previousElement = document.getElementById('frame'+(globalFrameCounter-1))

              if (!currentElement)
              {
                return;
              }

              if (globalFrameCounter > globalEndFrame)
              {
                clearInterval(timerID)
              }
              else
              {
                if(previousElement)
                {
                    previousElement.style.display="none";
                }
                currentElement.style.display="block";
                globalFrameCounter++;
              }
            }
            \n]]></script>\n
            \n""")

    def _printFooter(self):
        """Print the SVG footer."""

        self.file.write("\n</svg>\n")

    def _printPolygons(self, mesh):
        """Print the selected (visible) polygons.
        """

        if len(mesh.faces) == 0:
            return

        self.file.write("<g>\n")

        for face in mesh.faces:
            if not face.sel:
                continue

            self.file.write("<path d=\"")

            #p = self._calcCanvasCoord(face.verts[0])
            p = self._calcCanvasCoord(face.v[0])
            self.file.write("M %g,%g L " % (p[0], p[1]))

            for v in face.v[1:]:
                p = self._calcCanvasCoord(v)
                self.file.write("%g,%g " % (p[0], p[1]))

            # get rid of the last blank space, just cosmetics here.
            self.file.seek(-1, 1)
            self.file.write(" z\"\n")

            # take as face color the first vertex color
            if face.col:
                fcol = face.col[0]
                color = [fcol.r, fcol.g, fcol.b, fcol.a]
            else:
                color = [255, 255, 255, 255]

            # Convert the color to the #RRGGBB form
            str_col = "#%02X%02X%02X" % (color[0], color[1], color[2])

            # Handle transparent polygons
            opacity_string = ""
            if color[3] != 255:
                opacity = float(color[3]) / 255.0
                opacity_string = " fill-opacity: %g; stroke-opacity: %g; opacity: 1;" % (opacity, opacity)
                #opacity_string = "opacity: %g;" % (opacity)

            self.file.write("\tstyle=\"fill:" + str_col + ";")
            self.file.write(opacity_string)

            # use the stroke property to alleviate the "adjacent edges" problem,
            # we simulate polygon expansion using borders,
            # see http://www.antigrain.com/svg/index.html for more info
            stroke_width = 1.0

            # EXPANSION TRICK is not that useful where there is transparency
            if config.polygons['EXPANSION_TRICK'] and color[3] == 255:
                # str_col = "#000000" # For debug
                self.file.write(" stroke:%s;\n" % str_col)
                self.file.write(" stroke-width:" + str(stroke_width) + ";\n")
                self.file.write(" stroke-linecap:round;stroke-linejoin:round")

            self.file.write("\"/>\n")

        self.file.write("</g>\n")

    def _printEdges(self, mesh, showHiddenEdges=False):
        """Print the wireframe using mesh edges.
        """

        stroke_width = config.edges['WIDTH']
        stroke_col = config.edges['COLOR']

        self.file.write("<g>\n")

        for e in mesh.edges:

            hidden_stroke_style = ""

            if e.sel == 0:
                if showHiddenEdges == False:
                    continue
                else:
                    hidden_stroke_style = ";\n stroke-dasharray:3, 3"

            p1 = self._calcCanvasCoord(e.v1)
            p2 = self._calcCanvasCoord(e.v2)

            self.file.write("<line x1=\"%g\" y1=\"%g\" x2=\"%g\" y2=\"%g\"\n"
                    % (p1[0], p1[1], p2[0], p2[1]))
            self.file.write(" style=\"stroke:rgb(" + str(stroke_col[0]) + "," + str(stroke_col[1]) + "," + str(stroke_col[2]) + ");")
            self.file.write(" stroke-width:" + str(stroke_width) + ";\n")
            self.file.write(" stroke-linecap:round;stroke-linejoin:round")
            self.file.write(hidden_stroke_style)
            self.file.write("\"/>\n")

        self.file.write("</g>\n")


## SWF Writer

try:
    from ming import *
    SWFSupported = True
except:
    SWFSupported = False


class SWFVectorWriter(VectorWriter):
    """A concrete class for writing SWF output.
    """

    def __init__(self, fileName):
        """Simply call the parent Contructor.
        """
        VectorWriter.__init__(self, fileName)

        self.movie = None
        self.sprite = None

    ##
    # Public Methods
    #

    def open(self, startFrame=1, endFrame=1):
        """Do some initialization operations.
        """
        VectorWriter.open(self, startFrame, endFrame)
        self.movie = SWFMovie()
        self.movie.setDimension(self.canvasSize[0], self.canvasSize[1])
        if self.animation:
            self.movie.setRate(self.fps)
            numframes = endFrame - startFrame + 1
            self.movie.setFrames(numframes)

    def close(self):
        """Do some finalization operation.
        """
        self.movie.save(self.outputFileName)

        # remember to call the close method of the parent
        VectorWriter.close(self)

    def printCanvas(self, scene, doPrintPolygons=True, doPrintEdges=False,
            showHiddenEdges=False):
        """Convert the scene representation to SVG.
        """
        context = scene.getRenderingContext()
        framenumber = context.currentFrame()

        Objects = scene.objects

        if self.sprite:
            self.movie.remove(self.sprite)

        sprite = SWFSprite()

        for obj in Objects:

            if(obj.getType() != 'Mesh'):
                continue

            mesh = obj.getData(mesh=1)

            if doPrintPolygons:
                self._printPolygons(mesh, sprite)

            if doPrintEdges:
                self._printEdges(mesh, sprite, showHiddenEdges)

        sprite.nextFrame()
        i = self.movie.add(sprite)
        # Remove the instance the next time
        self.sprite = i
        if self.animation:
            self.movie.nextFrame()

    ##
    # Private Methods
    #

    def _calcCanvasCoord(self, v):
        """Convert vertex in scene coordinates to canvas coordinates.
        """

        pt = Vector([0, 0, 0])

        mW = float(self.canvasSize[0]) / 2.0
        mH = float(self.canvasSize[1]) / 2.0

        # rescale to canvas size
        pt[0] = v.co[0] * mW + mW
        pt[1] = v.co[1] * mH + mH
        pt[2] = v.co[2]

        # For now we want (0,0) in the top-left corner of the canvas.
        # Mirror and translate along y
        pt[1] *= -1
        pt[1] += self.canvasSize[1]

        return pt

    def _printPolygons(self, mesh, sprite):
        """Print the selected (visible) polygons.
        """

        if len(mesh.faces) == 0:
            return

        for face in mesh.faces:
            if not face.sel:
                continue

            if face.col:
                fcol = face.col[0]
                color = [fcol.r, fcol.g, fcol.b, fcol.a]
            else:
                color = [255, 255, 255, 255]

            s = SWFShape()
            f = s.addFill(color[0], color[1], color[2], color[3])
            s.setRightFill(f)

            # The starting point of the shape
            p0 = self._calcCanvasCoord(face.verts[0])
            s.movePenTo(p0[0], p0[1])

            for v in face.verts[1:]:
                p = self._calcCanvasCoord(v)
                s.drawLineTo(p[0], p[1])

            # Closing the shape
            s.drawLineTo(p0[0], p0[1])

            s.end()
            sprite.add(s)

    def _printEdges(self, mesh, sprite, showHiddenEdges=False):
        """Print the wireframe using mesh edges.
        """

        stroke_width = config.edges['WIDTH']
        stroke_col = config.edges['COLOR']

        s = SWFShape()

        for e in mesh.edges:

            # Next, we set the line width and color for our shape.
            s.setLine(stroke_width, stroke_col[0], stroke_col[1], stroke_col[2],
            255)

            if e.sel == 0:
                if showHiddenEdges == False:
                    continue
                else:
                    # SWF does not support dashed lines natively, so -for now-
                    # draw hidden lines thinner and half-trasparent
                    s.setLine(stroke_width / 2, stroke_col[0], stroke_col[1],
                            stroke_col[2], 128)

            p1 = self._calcCanvasCoord(e.v1)
            p2 = self._calcCanvasCoord(e.v2)

            s.movePenTo(p1[0], p1[1])
            s.drawLineTo(p2[0], p2[1])

        s.end()
        sprite.add(s)


## PDF Writer

try:
    from reportlab.pdfgen import canvas
    PDFSupported = True
except:
    PDFSupported = False


class PDFVectorWriter(VectorWriter):
    """A concrete class for writing PDF output.
    """

    def __init__(self, fileName):
        """Simply call the parent Contructor.
        """
        VectorWriter.__init__(self, fileName)

        self.canvas = None

    ##
    # Public Methods
    #

    def open(self, startFrame=1, endFrame=1):
        """Do some initialization operations.
        """
        VectorWriter.open(self, startFrame, endFrame)
        size = (self.canvasSize[0], self.canvasSize[1])
        self.canvas = canvas.Canvas(self.outputFileName, pagesize=size, bottomup=0)

    def close(self):
        """Do some finalization operation.
        """
        self.canvas.save()

        # remember to call the close method of the parent
        VectorWriter.close(self)

    def printCanvas(self, scene, doPrintPolygons=True, doPrintEdges=False,
            showHiddenEdges=False):
        """Convert the scene representation to SVG.
        """
        context = scene.getRenderingContext()
        framenumber = context.currentFrame()

        Objects = scene.objects

        for obj in Objects:

            if(obj.getType() != 'Mesh'):
                continue

            mesh = obj.getData(mesh=1)

            if doPrintPolygons:
                self._printPolygons(mesh)

            if doPrintEdges:
                self._printEdges(mesh, showHiddenEdges)

        self.canvas.showPage()

    ##
    # Private Methods
    #

    def _calcCanvasCoord(self, v):
        """Convert vertex in scene coordinates to canvas coordinates.
        """

        pt = Vector([0, 0, 0])

        mW = float(self.canvasSize[0]) / 2.0
        mH = float(self.canvasSize[1]) / 2.0

        # rescale to canvas size
        pt[0] = v.co[0] * mW + mW
        pt[1] = v.co[1] * mH + mH
        pt[2] = v.co[2]

        # For now we want (0,0) in the top-left corner of the canvas.
        # Mirror and translate along y
        pt[1] *= -1
        pt[1] += self.canvasSize[1]

        return pt

    def _printPolygons(self, mesh):
        """Print the selected (visible) polygons.
        """

        if len(mesh.faces) == 0:
            return

        for face in mesh.faces:
            if not face.sel:
                continue

            if face.col:
                fcol = face.col[0]
                color = [fcol.r / 255.0, fcol.g / 255.0, fcol.b / 255.0,
                        fcol.a / 255.0]
            else:
                color = [1, 1, 1, 1]

            self.canvas.setFillColorRGB(color[0], color[1], color[2])
            # For debug
            self.canvas.setStrokeColorRGB(0, 0, 0)

            path = self.canvas.beginPath()

            # The starting point of the path
            p0 = self._calcCanvasCoord(face.verts[0])
            path.moveTo(p0[0], p0[1])

            for v in face.verts[1:]:
                p = self._calcCanvasCoord(v)
                path.lineTo(p[0], p[1])

            # Closing the shape
            path.close()

            self.canvas.drawPath(path, stroke=0, fill=1)

    def _printEdges(self, mesh, showHiddenEdges=False):
        """Print the wireframe using mesh edges.
        """

        stroke_width = config.edges['WIDTH']
        stroke_col = config.edges['COLOR']

        self.canvas.setLineCap(1)
        self.canvas.setLineJoin(1)
        self.canvas.setLineWidth(stroke_width)
        self.canvas.setStrokeColorRGB(stroke_col[0] / 255.0, stroke_col[1] / 255.0,
            stroke_col[2] / 255)

        for e in mesh.edges:

            self.canvas.setLineWidth(stroke_width)

            if e.sel == 0:
                if showHiddenEdges == False:
                    continue
                else:
                    #

donato
Posts: 26
Joined: Tue May 15, 2012 12:59 pm

code, part II

Post by donato » Mon May 28, 2012 8:14 pm

Code: Select all

                else:
                    # PDF does not support dashed lines natively, so -for now-
                    # draw hidden lines thinner
                    self.canvas.setLineWidth(stroke_width / 2.0)

            p1 = self._calcCanvasCoord(e.v1)
            p2 = self._calcCanvasCoord(e.v2)

            self.canvas.line(p1[0], p1[1], p2[0], p2[1])

# ---------------------------------------------------------------------
#
## Rendering Classes
#
# ---------------------------------------------------------------------

# A dictionary to collect different shading style methods
shadingStyles = dict()
shadingStyles['FLAT'] = None
shadingStyles['TOON'] = None

# A dictionary to collect different edge style methods
edgeStyles = dict()
edgeStyles['MESH'] = MeshUtils.isMeshEdge
edgeStyles['SILHOUETTE'] = MeshUtils.isSilhouetteEdge

# A dictionary to collect the supported output formats
outputWriters = dict()
outputWriters['SVG'] = SVGVectorWriter
if SWFSupported:
    outputWriters['SWF'] = SWFVectorWriter
if PDFSupported:
    outputWriters['PDF'] = PDFVectorWriter


class Renderer:
    """Render a scene viewed from the active camera.

    This class is responsible of the rendering process, transformation and
    projection of the objects in the scene are invoked by the renderer.

    The rendering is done using the active camera for the current scene.
    """

    def __init__(self):
        """Make the rendering process only for the current scene by default.

        We will work on a copy of the scene, to be sure that the current scene do
        not get modified in any way.
        """

        # Render the current Scene, this should be a READ-ONLY property
        self._SCENE = Scene.GetCurrent()

        # Use the aspect ratio of the scene rendering context
        context = self._SCENE.getRenderingContext()

        aspect_ratio = float(context.imageSizeX()) / float(context.imageSizeY())
        self.canvasRatio = (float(context.aspectRatioX()) * aspect_ratio,
                            float(context.aspectRatioY())
                            )

        # Render from the currently active camera
        #self.cameraObj = self._SCENE.objects.camera

        self.lights = []

    ##
    # Public Methods
    #

    def doRendering(self, outputWriter, animation=False):
        """Render picture or animation and write it out.

        The parameters are:
            - a Vector writer object that will be used to output the result.
            - a flag to tell if we want to render an animation or only the
              current frame.
        """

        context = self._SCENE.getRenderingContext()
        origCurrentFrame = context.currentFrame()

        # Handle the animation case
        if not animation:
            startFrame = origCurrentFrame
            endFrame = startFrame
            outputWriter.open()
        else:
            startFrame = context.startFrame()
            endFrame = context.endFrame()
            outputWriter.open(startFrame, endFrame)

        # Do the rendering process frame by frame
        print("Start Rendering of %d frames" % (endFrame - startFrame + 1))
        for f in xrange(startFrame, endFrame + 1):
            print("\n\nFrame: %d" % f)

            # FIXME To get the correct camera position we have to use +1 here.
            # Is there a bug somewhere in the Scene module?
            context.currentFrame(f + 1)
            self.cameraObj = self._SCENE.objects.camera

            # Use some temporary workspace, a full copy of the scene
            inputScene = self._SCENE.copy(2)

            # To get the objects at this frame remove the +1 ...
            ctx = inputScene.getRenderingContext()
            ctx.currentFrame(f)

            # Get a projector for this camera.
            # NOTE: the projector wants object in world coordinates,
            # so we should remember to apply modelview transformations
            # _before_ we do projection transformations.
            self.proj = Projector(self.cameraObj, self.canvasRatio)

            try:
                renderedScene = self.doRenderScene(inputScene)
            except:
                print("There was an error! Aborting.")
                import traceback
                print(traceback.print_exc())

                self._SCENE.makeCurrent()
                Scene.Unlink(inputScene)
                del inputScene
                return

            outputWriter.printCanvas(renderedScene,
                    doPrintPolygons=config.polygons['SHOW'],
                    doPrintEdges=config.edges['SHOW'],
                    showHiddenEdges=config.edges['SHOW_HIDDEN'])

            # delete the rendered scene
            self._SCENE.makeCurrent()
            Scene.Unlink(renderedScene)
            del renderedScene

        outputWriter.close()
        print("Done!")
        context.currentFrame(origCurrentFrame)

    def doRenderScene(self, workScene):
        """Control the rendering process.

        Here we control the entire rendering process invoking the operation
        needed to transform and project the 3D scene in two dimensions.
        """

        # global processing of the scene

        self._filterHiddenObjects(workScene)

        self._buildLightSetup(workScene)

        self._doSceneClipping(workScene)

        self._doConvertGeometricObjsToMesh(workScene)

        if config.output['JOIN_OBJECTS']:
            self._joinMeshObjectsInScene(workScene)

        self._doSceneDepthSorting(workScene)

        # Per object activities

        Objects = workScene.objects

        print("Total Objects: %d" % len(Objects))
        for i, obj in enumerate(Objects):
            print("\n\n-------")
            print("Rendering Object: %d" % i)

            if obj.getType() != 'Mesh':
                print("Only Mesh supported! - Skipping type:", obj.getType())
                continue

            print("Rendering: ", obj.getName())

            mesh = obj.getData(mesh=1)

            self._doModelingTransformation(mesh, obj.matrix)

            self._doBackFaceCulling(mesh)

            # When doing HSR with NEWELL we may want to flip all normals
            # toward the viewer
            if config.polygons['HSR'] == "NEWELL":
                for f in mesh.faces:
                    f.sel = 1 - f.sel
                mesh.flipNormals()
                for f in mesh.faces:
                    f.sel = 1

            self._doLighting(mesh)

            # Do "projection" now so we perform further processing
            # in Normalized View Coordinates
            self._doProjection(mesh, self.proj)

            self._doViewFrustumClipping(mesh)

            self._doHiddenSurfaceRemoval(mesh)

            self._doEdgesStyle(mesh, edgeStyles[config.edges['STYLE']])

            # Update the object data, important! :)
            mesh.update()

        return workScene

    ##
    # Private Methods
    #

    # Utility methods

    def _getObjPosition(self, obj):
        """Return the obj position in World coordinates.
        """
        return obj.matrix.translationPart()

    def _cameraViewVector(self):
        """Get the View Direction form the camera matrix.
        """
        return Vector(self.cameraObj.matrix[2]).resize3D()

    # Faces methods

    def _isFaceVisible(self, face):
        """Determine if a face of an object is visible from the current camera.

        The view vector is calculated from the camera location and one of the
        vertices of the face (expressed in World coordinates, after applying
        modelview transformations).

        After those transformations we determine if a face is visible by
        computing the angle between the face normal and the view vector, this
        angle has to be between -90 and 90 degrees for the face to be visible.
        This corresponds somehow to the dot product between the two, if it
        results > 0 then the face is visible.

        There is no need to normalize those vectors since we are only interested in
        the sign of the cross product and not in the product value.

        NOTE: here we assume the face vertices are in WorldCoordinates, so
        please transform the object _before_ doing the test.
        """

        normal = Vector(face.no)
        camPos = self._getObjPosition(self.cameraObj)
        view_vect = None

        # View Vector in orthographics projections is the view Direction of
        # the camera
        if self.cameraObj.data.getType() == 1:
            view_vect = self._cameraViewVector()

        # View vector in perspective projections can be considered as
        # the difference between the camera position and one point of
        # the face, we choose the farthest point from the camera.
        if self.cameraObj.data.getType() == 0:
            vv = max([((camPos - Vector(v.co)).length, (camPos - Vector(v.co))) for v in face])
            view_vect = vv[1]

        # if d > 0 the face is visible from the camera
        d = view_vect * normal

        if d > 0:
            return True
        else:
            return False

    # Scene methods

    def _filterHiddenObjects(self, scene):
        """Discard object that are on hidden layers in the scene.
        """

        Objects = scene.objects

        visible_obj_list = [obj for obj in Objects if
                set(obj.layers).intersection(set(scene.getLayers()))]

        for o in Objects:
            if o not in visible_obj_list:
                scene.objects.unlink(o)

        scene.update()

    def _buildLightSetup(self, scene):
        # Get the list of lighting sources
        obj_lst = scene.objects
        self.lights = [o for o in obj_lst if o.getType() == 'Lamp']

        # When there are no lights we use a default lighting source
        # that have the same position of the camera
        if len(self.lights) == 0:
            l = Lamp.New('Lamp')
            lobj = Object.New('Lamp')
            lobj.loc = self.cameraObj.loc
            lobj.link(l)
            self.lights.append(lobj)

    def _doSceneClipping(self, scene):
        """Clip whole objects against the View Frustum.

        For now clip away only objects according to their center position.
        """

        cam_pos = self._getObjPosition(self.cameraObj)
        view_vect = self._cameraViewVector()

        near = self.cameraObj.data.clipStart
        far = self.cameraObj.data.clipEnd

        aspect = float(self.canvasRatio[0]) / float(self.canvasRatio[1])
        fovy = atan(0.5 / aspect / (self.cameraObj.data.lens / 32))
        fovy = fovy * 360.0 / pi

        Objects = scene.objects

        for o in Objects:
            if o.getType() != 'Mesh':
                continue

            """
            obj_vect = Vector(cam_pos) - self._getObjPosition(o)

            d = obj_vect*view_vect
            theta = AngleBetweenVecs(obj_vect, view_vect)

            # if the object is outside the view frustum, clip it away
            if (d < near) or (d > far) or (theta > fovy):
                scene.objects.unlink(o)
            """

            # Use the object bounding box
            # (whose points are already in WorldSpace Coordinate)

            bb = o.getBoundBox()

            points_outside = 0
            for p in bb:
                p_vect = Vector(cam_pos) - Vector(p)

                d = p_vect * view_vect
                theta = AngleBetweenVecs(p_vect, view_vect)

                # Is this point outside the view frustum?
                if (d < near) or (d > far) or (theta > fovy):
                    points_outside += 1

            # If the bb is all outside the view frustum we clip the whole
            # object away
            if points_outside == len(bb):
                scene.objects.unlink(o)

    def _doConvertGeometricObjsToMesh(self, scene):
        """Convert all "geometric" objects to mesh ones.
        """
        geometricObjTypes = ['Mesh', 'Surf', 'Curve', 'Text']
        #geometricObjTypes = ['Mesh', 'Surf', 'Curve']

        Objects = scene.objects

        objList = [o for o in Objects if o.getType() in geometricObjTypes]
        for obj in objList:
            old_obj = obj
            obj = self._convertToRawMeshObj(obj)
            scene.objects.link(obj)
            scene.objects.unlink(old_obj)

            # XXX Workaround for Text and Curve which have some normals
            # inverted when they are converted to Mesh, REMOVE that when
            # blender will fix that!!
            if old_obj.getType() in ['Curve', 'Text']:
                me = obj.getData(mesh=1)

                for f in me.faces:
                    f.sel = 1
                for v in me.verts:
                    v.sel = 1

                me.remDoubles(0)
                me.triangleToQuad()
                me.recalcNormals()
                me.update()

    def _doSceneDepthSorting(self, scene):
        """Sort objects in the scene.

        The object sorting is done accordingly to the object centers.
        """

        c = self._getObjPosition(self.cameraObj)

        by_obj_center_pos = (lambda o1, o2:
                (o1.getType() == 'Mesh' and o2.getType() == 'Mesh') and
                cmp((self._getObjPosition(o1) - Vector(c)).length,
                    (self._getObjPosition(o2) - Vector(c)).length)
            )

        # Implement sorting by bounding box, the object with the bb
        # nearest to the camera should be drawn as last.
        by_nearest_bbox_point = (lambda o1, o2:
                (o1.getType() == 'Mesh' and o2.getType() == 'Mesh') and
                cmp(min([(Vector(p) - Vector(c)).length for p in o1.getBoundBox()]),
                    min([(Vector(p) - Vector(c)).length for p in o2.getBoundBox()])
                )
            )

        Objects = list(scene.objects)

        #Objects.sort(by_obj_center_pos)
        Objects.sort(by_nearest_bbox_point)

        # update the scene
        for o in Objects:
            scene.objects.unlink(o)
            scene.objects.link(o)

    def _joinMeshObjectsInScene(self, scene):
        """Merge all the Mesh Objects in a scene into a single Mesh Object.
        """

        oList = [o for o in scene.objects if o.getType() == 'Mesh']

        # FIXME: Object.join() do not work if the list contains 1 object
        if len(oList) == 1:
            return

        mesh = Mesh.New('BigOne')
        bigObj = Object.New('Mesh', 'BigOne')
        bigObj.link(mesh)

        scene.objects.link(bigObj)

        try:
            bigObj.join(oList)
        except RuntimeError:
            print("\nWarning! - Can't Join Objects\n")
            scene.objects.unlink(bigObj)
            return
        except TypeError:
            print("Objects Type error?")

        for o in oList:
            scene.objects.unlink(o)

        scene.update()

    # Per object/mesh methods

    def _convertToRawMeshObj(self, object):
        """Convert geometry based object to a mesh object.
        """
        me = Mesh.New('RawMesh_' + object.name)
        me.getFromObject(object.name)

        newObject = Object.New('Mesh', 'RawMesh_' + object.name)
        newObject.link(me)

        # If the object has no materials set a default material
        if not me.materials:
            me.materials = [Material.New()]
            #for f in me.faces: f.mat = 0

        newObject.setMatrix(object.getMatrix())

        return newObject

    def _doModelingTransformation(self, mesh, matrix):
        """Transform object coordinates to world coordinates.

        This step is done simply applying to the object its tranformation
        matrix and recalculating its normals.
        """
        # XXX FIXME: blender do not transform normals in the right way when
        # there are negative scale values
        if matrix[0][0] < 0 or matrix[1][1] < 0 or matrix[2][2] < 0:
            print("WARNING: Negative scales, expect incorrect results!")

        mesh.transform(matrix, True)

    def _doBackFaceCulling(self, mesh):
        """Simple Backface Culling routine.

        At this level we simply do a visibility test face by face and then
        select the vertices belonging to visible faces.
        """

        # Select all vertices, so edges can be displayed even if there are no
        # faces
        for v in mesh.verts:
            v.sel = 1

        Mesh.Mode(Mesh.SelectModes['FACE'])
        # Loop on faces
        for f in mesh.faces:
            f.sel = 0
            if self._isFaceVisible(f):
                f.sel = 1

    def _doLighting(self, mesh):
        """Apply an Illumination and shading model to the object.

        The model used is the Phong one, it may be inefficient,
        but I'm just learning about rendering and starting from Phong seemed
        the most natural way.
        """

        # If the mesh has vertex colors already, use them,
        # otherwise turn them on and do some calculations
        if mesh.vertexColors:
            return
        mesh.vertexColors = 1

        materials = mesh.materials

        camPos = self._getObjPosition(self.cameraObj)

        # We do per-face color calculation (FLAT Shading), we can easily turn
        # to a per-vertex calculation if we want to implement some shading
        # technique. For an example see:
        # http://www.miralab.unige.ch/papers/368.pdf
        for f in mesh.faces:
            if not f.sel:
                continue

            mat = None
            if materials:
                mat = materials[f.mat]

            # A new default material
            if mat == None:
                mat = Material.New('defMat')

            # Check if it is a shadeless material
            elif mat.getMode() & Material.Modes['SHADELESS']:
                I = mat.getRGBCol()
                # Convert to a value between 0 and 255
                tmp_col = [int(c * 255.0) for c in I]

                for c in f.col:
                    c.r = tmp_col[0]
                    c.g = tmp_col[1]
                    c.b = tmp_col[2]
                    #c.a = tmp_col[3]

                continue

            # do vertex color calculation

            TotDiffSpec = Vector([0.0, 0.0, 0.0])

            for l in self.lights:
                light_obj = l
                light_pos = self._getObjPosition(l)
                light = light_obj.getData()

                L = Vector(light_pos).normalize()

                V = (Vector(camPos) - Vector(f.cent)).normalize()

                N = Vector(f.no).normalize()

                if config.polygons['SHADING'] == 'TOON':
                    NL = ShadingUtils.toonShading(N * L)
                else:
                    NL = (N * L)

                # Should we use NL instead of (N*L) here?
                R = 2 * (N * L) * N - L

                Ip = light.getEnergy()

                # Diffuse co-efficient
                kd = mat.getRef() * Vector(mat.getRGBCol())
                for i in [0, 1, 2]:
                    kd[i] *= light.col[i]

                Idiff = Ip * kd * max(0, NL)

                # Specular component
                ks = mat.getSpec() * Vector(mat.getSpecCol())
                ns = mat.getHardness()
                Ispec = Ip * ks * pow(max(0, (V * R)), ns)

                TotDiffSpec += (Idiff + Ispec)

            # Ambient component
            Iamb = Vector(Blender.World.Get()[0].getAmb())
            ka = mat.getAmb()

            # Emissive component (convert to a triplet)
            ki = Vector([mat.getEmit()] * 3)

            #I = ki + Iamb + (Idiff + Ispec)
            I = ki + (ka * Iamb) + TotDiffSpec

            # Set Alpha component
            I = list(I)
            I.append(mat.getAlpha())

            # Clamp I values between 0 and 1
            I = [min(c, 1) for c in I]
            I = [max(0, c) for c in I]

            # Convert to a value between 0 and 255
            tmp_col = [int(c * 255.0) for c in I]

            for c in f.col:
                c.r = tmp_col[0]
                c.g = tmp_col[1]
                c.b = tmp_col[2]
                c.a = tmp_col[3]

    def _doProjection(self, mesh, projector):
        """Apply Viewing and Projection tranformations.
        """

        for v in mesh.verts:
            p = projector.doProjection(v.co[:])
            v.co[0] = p[0]
            v.co[1] = p[1]
            v.co[2] = p[2]

        #mesh.recalcNormals()
        #mesh.update()

        # We could reeset Camera matrix, since now
        # we are in Normalized Viewing Coordinates,
        # but doung that would affect World Coordinate
        # processing for other objects

        #self.cameraObj.data.type = 1
        #self.cameraObj.data.scale = 2.0
        #m = Matrix().identity()
        #self.cameraObj.setMatrix(m)

    def _doViewFrustumClipping(self, mesh):
        """Clip faces against the View Frustum.
        """

        # The Canonical View Volume, 8 vertices, and 6 faces,
        # We consider its face normals pointing outside

        v1 = NMesh.Vert(1, 1, -1)
        v2 = NMesh.Vert(1, -1, -1)
        v3 = NMesh.Vert(-1, -1, -1)
        v4 = NMesh.Vert(-1, 1, -1)
        v5 = NMesh.Vert(1, 1, 1)
        v6 = NMesh.Vert(1, -1, 1)
        v7 = NMesh.Vert(-1, -1, 1)
        v8 = NMesh.Vert(-1, 1, 1)

        cvv = []
        f1 = NMesh.Face([v1, v4, v3, v2])
        cvv.append(f1)
        f2 = NMesh.Face([v5, v6, v7, v8])
        cvv.append(f2)
        f3 = NMesh.Face([v1, v2, v6, v5])
        cvv.append(f3)
        f4 = NMesh.Face([v2, v3, v7, v6])
        cvv.append(f4)
        f5 = NMesh.Face([v3, v4, v8, v7])
        cvv.append(f5)
        f6 = NMesh.Face([v4, v1, v5, v8])
        cvv.append(f6)

        nmesh = NMesh.GetRaw(mesh.name)
        clippedfaces = nmesh.faces[:]
        facelist = clippedfaces[:]

        for clipface in cvv:

            clippedfaces = []

            for f in facelist:

                #newfaces = HSR.splitOn(clipface, f, return_positive_faces=False)
                newfaces = None

                if not newfaces:
                    # Check if the face is all outside the view frustum
                    # TODO: Do this test before, it is more efficient
                    points_outside = 0
                    for v in f:
                        if abs(v[0]) > (1 - EPS) or abs(v[1]) > (1 - EPS) or abs(v[2]) > (1 - EPS):
                            points_outside += 1

                    if points_outside != len(f):
                        clippedfaces.append(f)
                else:
                    for nf in newfaces:
                        for v in nf:
                            nmesh.verts.append(v)

                        nf.mat = f.mat
                        nf.sel = f.sel
                        nf.col = [f.col[0]] * len(nf.v)

                        clippedfaces.append(nf)
            facelist = clippedfaces[:]

        nmesh.faces = facelist
        nmesh.update()

    # HSR routines
    def __simpleDepthSort(self, mesh):
        """Sort faces by the furthest vertex.

        This simple mesthod is known also as the painter algorithm, and it
        solves HSR correctly only for convex meshes.
        """

        #global progress

        # The sorting requires circa n*log(n) steps
        n = len(mesh.faces)
        progress.setActivity("HSR: Painter", n * log(n))

        by_furthest_z = (lambda f1, f2: progress.update() and
                cmp(max([v.co[2] for v in f1]), max([v.co[2] for v in f2]) + EPS)
                )

        # FIXME: using NMesh to sort faces. We should avoid that!
        nmesh = NMesh.GetRaw(mesh.name)

        # remember that _higher_ z values mean further points
        nmesh.faces.sort(by_furthest_z)
        nmesh.faces.reverse()

        nmesh.update()

    def __newellDepthSort(self, mesh):
        """Newell's depth sorting.

        """

        #global progress

        # Find non planar quads and convert them to triangle
        #for f in mesh.faces:
        #    f.sel = 0
        #    if is_nonplanar_quad(f.v):
        #        print "NON QUAD??"
        #        f.sel = 1

        # Now reselect all faces
        for f in mesh.faces:
            f.sel = 1
        mesh.quadToTriangle()

        # FIXME: using NMesh to sort faces. We should avoid that!
        nmesh = NMesh.GetRaw(mesh.name)

        # remember that _higher_ z values mean further points
        nmesh.faces.sort(by_furthest_z)
        nmesh.faces.reverse()

        # Begin depth sort tests

        # use the smooth flag to set marked faces
        for f in nmesh.faces:
            f.smooth = 0

        facelist = nmesh.faces[:]
        maplist = []

        # The steps are _at_least_ equal to len(facelist), we do not count the
        # feces coming out from splitting!!
        progress.setActivity("HSR: Newell", len(facelist))
        #progress.setQuiet(True)

        while len(facelist):
            debug("\n----------------------\n")
            debug("len(facelits): %d\n" % len(facelist))
            P = facelist[0]

            pSign = sign(P.normal[2])

            # We can discard faces parallel to the view vector
            #if P.normal[2] == 0:
            #    facelist.remove(P)
            #    continue

            split_done = 0
            face_marked = 0

            for Q in facelist[1:]:

                debug("P.smooth: " + str(P.smooth) + "\n")
                debug("Q.smooth: " + str(Q.smooth) + "\n")
                debug("\n")

                qSign = sign(Q.normal[2])
                # TODO: check also if Q is parallel??

                # Test 0: We need to test only those Qs whose furthest vertex
                # is closer to the observer than the closest vertex of P.

                zP = [v.co[2] for v in P.v]
                zQ = [v.co[2] for v in Q.v]
                notZOverlap = min(zP) > max(zQ) + EPS

                if notZOverlap:
                    debug("\nTest 0\n")
                    debug("NOT Z OVERLAP!\n")
                    if Q.smooth == 0:
                        # If Q is not marked then we can safely print P
                        break
                    else:
                        debug("met a marked face\n")
                        continue

                # Test 1: X extent overlapping
                xP = [v.co[0] for v in P.v]
                xQ = [v.co[0] for v in Q.v]
                #notXOverlap = (max(xP) <= min(xQ)) or (max(xQ) <= min(xP))
                notXOverlap = min(xQ) >= (max(xP) - EPS) or min(xP) >= (max(xQ) - EPS)

                if notXOverlap:
                    debug("\nTest 1\n")
                    debug("NOT X OVERLAP!\n")
                    continue

                # Test 2: Y extent Overlapping
                yP = [v.co[1] for v in P.v]
                yQ = [v.co[1] for v in Q.v]
                #notYOverlap = max(yP) <= min(yQ) or max(yQ) <= min(yP)
                notYOverlap = min(yQ) >= (max(yP) - EPS) or min(yP) >= (max(yQ) - EPS)

                if notYOverlap:
                    debug("\nTest 2\n")
                    debug("NOT Y OVERLAP!\n")
                    continue

                # Test 3: P vertices are all behind the plane of Q
                n = 0
                for Pi in P:
                    d = qSign * HSR.Distance(Vector(Pi), Q)
                    if d <= EPS:
                        n += 1
                pVerticesBehindPlaneQ = (n == len(P))

                if pVerticesBehindPlaneQ:
                    debug("\nTest 3\n")
                    debug("P BEHIND Q!\n")
                    continue

                # Test 4: Q vertices in front of the plane of P
                n = 0
                for Qi in Q:
                    d = pSign * HSR.Distance(Vector(Qi), P)
                    if d >= -EPS:
                        n += 1
                qVerticesInFrontPlaneP = (n == len(Q))

                if qVerticesInFrontPlaneP:
                    debug("\nTest 4\n")
                    debug("Q IN FRONT OF P!\n")
                    continue

                # Test 5: Check if projections of polygons effectively overlap,
                # in previous tests we checked only bounding boxes.

                #if not projectionsOverlap(P, Q):
                if not (HSR.projectionsOverlap(P, Q) or HSR.projectionsOverlap(Q, P)):
                    debug("\nTest 5\n")
                    debug("Projections do not overlap!\n")
                    continue

                # We still can't say if P obscures Q.

                # But if Q is marked we do a face-split trying to resolve a
                # difficulty (maybe a visibility cycle).
                if Q.smooth == 1:
                    # Split P or Q
                    debug("Possibly a cycle detected!\n")
                    debug("Split here!!\n")

                    facelist = HSR.facesplit(P, Q, facelist, nmesh)
                    split_done = 1
                    break

                # The question now is: Does Q obscure P?

                # Test 3bis: Q vertices are all behind the plane of P
                n = 0
                for Qi in Q:
                    d = pSign * HSR.Distance(Vector(Qi), P)
                    if d <= EPS:
                        n += 1
                qVerticesBehindPlaneP = (n == len(Q))

                if qVerticesBehindPlaneP:
                    debug("\nTest 3bis\n")
                    debug("Q BEHIND P!\n")

                # Test 4bis: P vertices in front of the plane of Q
                n = 0
                for Pi in P:
                    d = qSign * HSR.Distance(Vector(Pi), Q)
                    if d >= -EPS:
                        n += 1
                pVerticesInFrontPlaneQ = (n == len(P))

                if pVerticesInFrontPlaneQ:
                    debug("\nTest 4bis\n")
                    debug("P IN FRONT OF Q!\n")

                # We don't even know if Q does obscure P, so they should
                # intersect each other, split one of them in two parts.
                if not qVerticesBehindPlaneP and not pVerticesInFrontPlaneQ:
                    debug("\nSimple Intersection?\n")
                    debug("Test 3bis or 4bis failed\n")
                    debug("Split here!!2\n")

                    facelist = HSR.facesplit(P, Q, facelist, nmesh)
                    split_done = 1
                    break

                facelist.remove(Q)
                facelist.insert(0, Q)
                Q.smooth = 1
                face_marked = 1
                debug("Q marked!\n")
                break

            # Write P!
            if split_done == 0 and face_marked == 0:
                facelist.remove(P)
                maplist.append(P)
                dumpfaces(maplist, "dump" + str(len(maplist)).zfill(4) + ".svg")

                progress.update()

            if len(facelist) == 870:
                dumpfaces([P, Q], "loopdebug.svg")

            #if facelist == None:
            #    maplist = [P, Q]
            #    print [v.co for v in P]
            #    print [v.co for v in Q]
            #    break

            # end of while len(facelist)

        nmesh.faces = maplist
        #for f in nmesh.faces:
        #    f.sel = 1

        nmesh.update()

    def _doHiddenSurfaceRemoval(self, mesh):
        """Do HSR for the given mesh.
        """
        if len(mesh.faces) == 0:
            return

        if config.polygons['HSR'] == 'PAINTER':
            print("\nUsing the Painter algorithm for HSR.")
            self.__simpleDepthSort(mesh)

        elif config.polygons['HSR'] == 'NEWELL':
            print("\nUsing the Newell's algorithm for HSR.")
            self.__newellDepthSort(mesh)

    def _doEdgesStyle(self, mesh, edgestyleSelect):
        """Process Mesh Edges accroding to a given selection style.

        Examples of algorithms:

        Contours:
            given an edge if its adjacent faces have the same normal (that is
            they are complanar), than deselect it.

        Silhouettes:
            given an edge if one its adjacent faces is frontfacing and the
            other is backfacing, than select it, else deselect.
        """

        Mesh.Mode(Mesh.SelectModes['EDGE'])

        edge_cache = MeshUtils.buildEdgeFaceUsersCache(mesh)

        for i, edge_faces in enumerate(edge_cache):
            mesh.edges[i].sel = 0
            if edgestyleSelect(edge_faces):
                mesh.edges[i].sel = 1

        """
        for e in mesh.edges:

            e.sel = 0
            if edgestyleSelect(e, mesh):
                e.sel = 1
        """
        #


# ---------------------------------------------------------------------
#
## GUI Class and Main Program
#
# ---------------------------------------------------------------------


class GUI:

    def _init():

        # Output Format menu
        output_format = config.output['FORMAT']
        default_value = outputWriters.keys().index(output_format) + 1
        GUI.outFormatMenu = Draw.Create(default_value)
        GUI.evtOutFormatMenu = 0

        # Animation toggle button
        GUI.animToggle = Draw.Create(config.output['ANIMATION'])
        GUI.evtAnimToggle = 1

        # Join Objects toggle button
        GUI.joinObjsToggle = Draw.Create(config.output['JOIN_OBJECTS'])
        GUI.evtJoinObjsToggle = 2

        # Render filled polygons
        GUI.polygonsToggle = Draw.Create(config.polygons['SHOW'])

        # Shading Style menu
        shading_style = config.polygons['SHADING']
        default_value = shadingStyles.keys().index(shading_style) + 1
        GUI.shadingStyleMenu = Draw.Create(default_value)
        GUI.evtShadingStyleMenu = 21

        GUI.evtPolygonsToggle = 3
        # We hide the config.polygons['EXPANSION_TRICK'], for now

        # Render polygon edges
        GUI.showEdgesToggle = Draw.Create(config.edges['SHOW'])
        GUI.evtShowEdgesToggle = 4

        # Render hidden edges
        GUI.showHiddenEdgesToggle = Draw.Create(config.edges['SHOW_HIDDEN'])
        GUI.evtShowHiddenEdgesToggle = 5

        # Edge Style menu
        edge_style = config.edges['STYLE']
        default_value = edgeStyles.keys().index(edge_style) + 1
        GUI.edgeStyleMenu = Draw.Create(default_value)
        GUI.evtEdgeStyleMenu = 6

        # Edge Width slider
        GUI.edgeWidthSlider = Draw.Create(config.edges['WIDTH'])
        GUI.evtEdgeWidthSlider = 7

        # Edge Color Picker
        c = config.edges['COLOR']
        GUI.edgeColorPicker = Draw.Create(c[0] / 255.0, c[1] / 255.0, c[2] / 255.0)
        GUI.evtEdgeColorPicker = 71

        # Render Button
        GUI.evtRenderButton = 8

        # Exit Button
        GUI.evtExitButton = 9

        # Save default button
        GUI.evtSaveDefaultButton = 99

    def draw():

        # initialize static members
        GUI._init()

        glClear(GL_COLOR_BUFFER_BIT)
        glColor3f(0.0, 0.0, 0.0)
        glRasterPos2i(10, 380)
        Draw.Text("VRM: Vector Rendering Method script. Version %s." %
                __version__)
        glRasterPos2i(10, 365)
        Draw.Text("%s (c) 2012" % __author__)

        glRasterPos2i(10, 335)
        Draw.Text("Press Q or ESC to quit.")

        # Build the output format menu
        glRasterPos2i(10, 310)
        Draw.Text("Select the output Format:")
        outMenuStruct = "Output Format %t"
        for t in outputWriters.keys():
            outMenuStruct = outMenuStruct + "|%s" % t
        GUI.outFormatMenu = Draw.Menu(outMenuStruct, GUI.evtOutFormatMenu,
            10, 285, 160, 18, GUI.outFormatMenu.val, "Choose the Output Format")

        # Animation toggle
        GUI.animToggle = Draw.Toggle("Animation", GUI.evtAnimToggle,
            10, 260, 160, 18, GUI.animToggle.val,
            "Toggle rendering of animations")

        # Join Objects toggle
        GUI.joinObjsToggle = Draw.Toggle("Join objects", GUI.evtJoinObjsToggle,
            10, 235, 160, 18, GUI.joinObjsToggle.val,
            "Join objects in the rendered file")

        # Render Button
        Draw.Button("Render", GUI.evtRenderButton, 10, 210 - 25, 75, 25 + 18,
            "Start Rendering")
        Draw.Button("Exit", GUI.evtExitButton, 95, 210 - 25, 75, 25 + 18, "Exit!")

        Draw.Button("Save settings as default", GUI.evtSaveDefaultButton, 10, 210 - 50, 160, 18,
            "Save settings as default")

        # Rendering Styles
        glRasterPos2i(200, 310)
        Draw.Text("Rendering Style:")

        # Render Polygons
        GUI.polygonsToggle = Draw.Toggle("Filled Polygons", GUI.evtPolygonsToggle,
            200, 285, 160, 18, GUI.polygonsToggle.val,
            "Render filled polygons")

        if GUI.polygonsToggle.val == 1:

            # Polygon Shading Style
            shadingStyleMenuStruct = "Shading Style %t"
            for t in shadingStyles.keys():
                shadingStyleMenuStruct = shadingStyleMenuStruct + "|%s" % t.lower()
            GUI.shadingStyleMenu = Draw.Menu(shadingStyleMenuStruct, GUI.evtShadingStyleMenu,
                200, 260, 160, 18, GUI.shadingStyleMenu.val,
                "Choose the shading style")

        # Render Edges
        GUI.showEdgesToggle = Draw.Toggle("Show Edges", GUI.evtShowEdgesToggle,
            200, 235, 160, 18, GUI.showEdgesToggle.val,
            "Render polygon edges")

        if GUI.showEdgesToggle.val == 1:

            # Edge Style
            edgeStyleMenuStruct = "Edge Style %t"
            for t in edgeStyles.keys():
                edgeStyleMenuStruct = edgeStyleMenuStruct + "|%s" % t.lower()
            GUI.edgeStyleMenu = Draw.Menu(edgeStyleMenuStruct, GUI.evtEdgeStyleMenu,
                200, 210, 160, 18, GUI.edgeStyleMenu.val,
                "Choose the edge style")

            # Edge size
            GUI.edgeWidthSlider = Draw.Slider("Width: ", GUI.evtEdgeWidthSlider,
                200, 185, 140, 18, GUI.edgeWidthSlider.val,
                0.0, 10.0, 0, "Change Edge Width")

            # Edge Color
            GUI.edgeColorPicker = Draw.ColorPicker(GUI.evtEdgeColorPicker,
                342, 185, 18, 18, GUI.edgeColorPicker.val, "Choose Edge Color")

            # Show Hidden Edges
            GUI.showHiddenEdgesToggle = Draw.Toggle("Show Hidden Edges",
                GUI.evtShowHiddenEdgesToggle,
                200, 160, 160, 18, GUI.showHiddenEdgesToggle.val,
                "Render hidden edges as dashed lines")

    def event(evt, val):

        if evt == Draw.ESCKEY or evt == Draw.QKEY:
            Draw.Exit()
        else:
            return

        Draw.Redraw(1)

    def button_event(evt):

        if evt == GUI.evtExitButton:
            Draw.Exit()

        elif evt == GUI.evtOutFormatMenu:
            i = GUI.outFormatMenu.val - 1
            config.output['FORMAT'] = outputWriters.keys()[i]
            # Set the new output file
            global outputfile
            outputfile = Blender.sys.splitext(basename)[0] + "." + str(config.output['FORMAT']).lower()

        elif evt == GUI.evtAnimToggle:
            config.output['ANIMATION'] = bool(GUI.animToggle.val)

        elif evt == GUI.evtJoinObjsToggle:
            config.output['JOIN_OBJECTS'] = bool(GUI.joinObjsToggle.val)

        elif evt == GUI.evtPolygonsToggle:
            config.polygons['SHOW'] = bool(GUI.polygonsToggle.val)

        elif evt == GUI.evtShadingStyleMenu:
            i = GUI.shadingStyleMenu.val - 1
            config.polygons['SHADING'] = shadingStyles.keys()[i]

        elif evt == GUI.evtShowEdgesToggle:
            config.edges['SHOW'] = bool(GUI.showEdgesToggle.val)

        elif evt == GUI.evtShowHiddenEdgesToggle:
            config.edges['SHOW_HIDDEN'] = bool(GUI.showHiddenEdgesToggle.val)

        elif evt == GUI.evtEdgeStyleMenu:
            i = GUI.edgeStyleMenu.val - 1
            config.edges['STYLE'] = edgeStyles.keys()[i]

        elif evt == GUI.evtEdgeWidthSlider:
            config.edges['WIDTH'] = float(GUI.edgeWidthSlider.val)

        elif evt == GUI.evtEdgeColorPicker:
            config.edges['COLOR'] = [int(c * 255.0) for c in GUI.edgeColorPicker.val]

        elif evt == GUI.evtRenderButton:
            label = "Save %s" % config.output['FORMAT']
            # Show the File Selector
            global outputfile
            Blender.Window.FileSelector(vectorize, label, outputfile)

        elif evt == GUI.evtSaveDefaultButton:
            config.saveToRegistry()

        else:
            print("Event: %d not handled!" % evt)

        if evt:
            Draw.Redraw(1)
            #GUI.conf_debug()

    def conf_debug():
        from pprint import pprint
        print("\nConfig")
        pprint(config.output)
        pprint(config.polygons)
        pprint(config.edges)

    _init = staticmethod(_init)
    draw = staticmethod(draw)
    event = staticmethod(event)
    button_event = staticmethod(button_event)
    conf_debug = staticmethod(conf_debug)


# A wrapper function for the vectorizing process
def vectorize(filename):
    """The vectorizing process is as follows:

     - Instanciate the writer and the renderer
     - Render!
     """

    if filename == "":
        print("\nERROR: invalid file name!")
        return

    from Blender import Window
    editmode = Window.EditMode()
    if editmode:
        Window.EditMode(0)

    actualWriter = outputWriters[config.output['FORMAT']]
    writer = actualWriter(filename)

    renderer = Renderer()
    renderer.doRendering(writer, config.output['ANIMATION'])

    if editmode:
        Window.EditMode(1)


# Here the main
if __name__ == "__main__":

    global progress

    # initialize writer setting also here to configure writer specific
    # settings on startup
    actualWriter = outputWriters[config.output['FORMAT']]
    writer = actualWriter("")

    outputfile = ""
    basename = bpy.path.basename(bpy.data.filepath)
    if basename != "":
        outputfile = Blender.sys.splitext(basename)[0] + "." + str(config.output['FORMAT']).lower()

    progress = ConsoleProgressIndicator()
    vectorize(outputfile)
[/code]

Post Reply

Who is online

Users browsing this forum: No registered users and 0 guests