[vrm.git] / vrm.py

#!BPY
"""
Name: 'VRM'
Blender: 242
Group: 'Render'
Tooltip: 'Vector Rendering Method script'
"""

__author__ = "Antonio Ospite"
__url__ = ["http://projects.blender.org/projects/vrm"]
__version__ = "0.3.beta"

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

# ---------------------------------------------------------------------
#    Copyright (c) 2006 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:
#   - FIX the issue with negative scales in object tranformations!
#   - Use a better depth sorting algorithm
#   - Implement clipping of primitives and do handle object intersections.
#     (for now only clipping away whole objects is supported).
#   - 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.
#   - Check memory use!!
#   - Support Indexed palettes!! (Useful for ILDA FILES, for example,
#     see http://www.linux-laser.org/download/autotrace/ilda-output.patch)
#
# ---------------------------------------------------------------------
#
# Changelog:
#
#   vrm-0.3.py  - ...
#     * First release after code restucturing.
#       Now the script offers a useful set of functionalities
#       and it can render animations, too.
#     * Optimization in Renderer.doEdgeStyle(), build a topology cache
#       so to speed up the lookup of adjacent faces of an edge.
#       Thanks ideasman42.
#     * The SVG output is now SVG 1.0 valid.
#       Checked with: http://jiggles.w3.org/svgvalidator/ValidatorURI.html
#     * Progress indicator during HSR.
#     * Initial SWF output support
#     * Fixed a bug in the animation code, now the projection matrix is
#       recalculated at each frame!
#
# ---------------------------------------------------------------------

import Blender
from Blender import Scene, Object, Mesh, NMesh, Material, Lamp, Camera, Window
from Blender.Mathutils import *
from math import *
import sys, time

# Constants
EPS = 10e-5

# We use a global progress Indicator Object
progress = None


# Some global settings

class config:
    polygons = dict()
    polygons['SHOW'] = True
    polygons['SHADING'] = 'FLAT'
    #polygons['HSR'] = 'PAINTER' # 'PAINTER' or 'NEWELL'
    polygons['HSR'] = 'PAINTER'
    # 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' # or SILHOUETTE
    edges['STYLE'] = 'SILHOUETTE'
    edges['WIDTH'] = 2
    edges['COLOR'] = [0, 0, 0]

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



# Utility functions
def sign(x):

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


# ---------------------------------------------------------------------
#
## 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
        
        # What projection do we want?
        if camera.type == 0:
            mP = self._calcPerspectiveMatrix(fovy, aspect, near, far) 
        elif camera.type == 1:
            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"""

        self.outputFileName = fileName
        self.file = None
        
        context = Scene.GetCurrent().getRenderingContext()
        self.canvasSize = ( context.imageSizeX(), context.imageSizeY() )

        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

        self.file = open(self.outputFileName, "w")
        print "Outputting to: ", self.outputFileName

        return

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


    ##
    # Public Methods
    #

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

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

        # 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.
        """

        Objects = scene.getChildren()

        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:

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

            /* FIXME: Use 1000 as interval as lower values gives problems */
            timerID = setInterval("NextFrame()", 1000);
            globalFrameCounter=%d;

            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""" % (self.startFrame, self.endFrame, self.startFrame) )
                
    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])
            self.file.write("M %g,%g L " % (p[0], p[1]))

            for v in face.verts[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

from ming import *

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])
        # set fps
        self.movie.setRate(25)
        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.getChildren()

        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)


            """
            # 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")

            """

    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)

            # FIXME: this is just a qorkaround, remove that after the
            # implementation of propoer Viewport clipping
            if abs(p1[0]) < 3000 and abs(p2[0]) < 3000 and abs(p1[1]) < 3000 and abs(p1[2]) < 3000:
                s.movePenTo(p1[0], p1[1])
                s.drawLineTo(p2[0], p2[1])
            

        s.end()
        sprite.add(s)
            


# ---------------------------------------------------------------------
#
## 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
outputWriters['SWF'] = SWFVectorWriter


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.getCurrentCamera()

        # Get the list of lighting sources
        obj_lst = self._SCENE.getChildren()
        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)


    ##
    # 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)
        for f in xrange(startFrame, endFrame+1):
            print "\n\nFrame: %d" % f
            context.currentFrame(f)

            # Use some temporary workspace, a full copy of the scene
            inputScene = self._SCENE.copy(2)
            # And Set our camera accordingly
            self.cameraObj = inputScene.getCurrentCamera()

            # 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._doSceneClipping(workScene)

        self._doConvertGeometricObjsToMesh(workScene)

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

        self._doSceneDepthSorting(workScene)
        
        # Per object activities

        Objects = workScene.getChildren()
        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 _doSceneClipping(self, scene):
        """Clip whole objects against the View Frustum.

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

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

            obj_vect = Vector(cpos) - 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.unlink(o)

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

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

        # TODO: implement sorting by bounding box, if obj1.bb is inside obj2.bb,
        # then ob1 goes farther than obj2, useful when obj2 has holes
        by_bbox = None
        
        Objects = scene.getChildren()
        Objects.sort(by_center_pos)
        
        # update the scene
        for o in Objects:
            scene.unlink(o)
            scene.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.getChildren() 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.link(bigObj)

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

    # 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.

        """
        from hsrtk import *

        #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 * 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 * 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):
                    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 = 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 * 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 * 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 = 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)

                progress.update()

            #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
#
# ---------------------------------------------------------------------


from Blender import BGL, Draw
from Blender.BGL import *

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

    def draw():

        # initialize static members
        GUI._init()

        glClear(GL_COLOR_BUFFER_BIT)
        glColor3f(0.0, 0.0, 0.0)
        glRasterPos2i(10, 350)
        Draw.Text("VRM: Vector Rendering Method script. Version %s." %
                __version__)
        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!")

        # 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")

        glRasterPos2i(10, 160)
        Draw.Text("%s (c) 2006" % __author__)

    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]

        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)

        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

    outputfile = ""
    basename = Blender.sys.basename(Blender.Get('filename'))
    if basename != "":
        outputfile = Blender.sys.splitext(basename)[0] + "." + str(config.output['FORMAT']).lower()

    if Blender.mode == 'background':
        progress = ConsoleProgressIndicator()
        vectorize(outputfile)
    else:
        progress = GraphicalProgressIndicator()
        Draw.Register(GUI.draw, GUI.event, GUI.button_event)