from __future__ import division
import os
import sys
import xml.dom.minidom
import re
import math
import base64
import urllib
from svg_regex import svg_parser
import gzip
import optparse
from yocto_css import parseCssString
try:
from decimal import *
except ImportError:
from fixedpoint import *
Decimal = FixedPoint
try:
import psyco
psyco.full()
except ImportError:
pass
APP = 'scour'
VER = '0.25r171'
COPYRIGHT = 'Copyright Jeff Schiller, 2010'
NS = { 'SVG': 'http://www.w3.org/2000/svg',
'XLINK': 'http://www.w3.org/1999/xlink',
'SODIPODI': 'http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd',
'INKSCAPE': 'http://www.inkscape.org/namespaces/inkscape',
'ADOBE_ILLUSTRATOR': 'http://ns.adobe.com/AdobeIllustrator/10.0/',
'ADOBE_GRAPHS': 'http://ns.adobe.com/Graphs/1.0/',
'ADOBE_SVG_VIEWER': 'http://ns.adobe.com/AdobeSVGViewerExtensions/3.0/',
'ADOBE_VARIABLES': 'http://ns.adobe.com/Variables/1.0/',
'ADOBE_SFW': 'http://ns.adobe.com/SaveForWeb/1.0/',
'ADOBE_EXTENSIBILITY': 'http://ns.adobe.com/Extensibility/1.0/',
'ADOBE_FLOWS': 'http://ns.adobe.com/Flows/1.0/',
'ADOBE_IMAGE_REPLACEMENT': 'http://ns.adobe.com/ImageReplacement/1.0/',
'ADOBE_CUSTOM': 'http://ns.adobe.com/GenericCustomNamespace/1.0/',
'ADOBE_XPATH': 'http://ns.adobe.com/XPath/1.0/'
}
unwanted_ns = [ NS['SODIPODI'], NS['INKSCAPE'], NS['ADOBE_ILLUSTRATOR'],
NS['ADOBE_GRAPHS'], NS['ADOBE_SVG_VIEWER'], NS['ADOBE_VARIABLES'],
NS['ADOBE_SFW'], NS['ADOBE_EXTENSIBILITY'], NS['ADOBE_FLOWS'],
NS['ADOBE_IMAGE_REPLACEMENT'], NS['ADOBE_CUSTOM'], NS['ADOBE_XPATH'] ]
svgAttributes = [
'clip-rule',
'display',
'fill',
'fill-opacity',
'fill-rule',
'filter',
'font-family',
'font-size',
'font-stretch',
'font-style',
'font-variant',
'font-weight',
'line-height',
'marker',
'opacity',
'overflow',
'stop-color',
'stop-opacity',
'stroke',
'stroke-dasharray',
'stroke-dashoffset',
'stroke-linecap',
'stroke-linejoin',
'stroke-miterlimit',
'stroke-opacity',
'stroke-width',
'visibility'
]
colors = {
'aliceblue': 'rgb(240, 248, 255)',
'antiquewhite': 'rgb(250, 235, 215)',
'aqua': 'rgb( 0, 255, 255)',
'aquamarine': 'rgb(127, 255, 212)',
'azure': 'rgb(240, 255, 255)',
'beige': 'rgb(245, 245, 220)',
'bisque': 'rgb(255, 228, 196)',
'black': 'rgb( 0, 0, 0)',
'blanchedalmond': 'rgb(255, 235, 205)',
'blue': 'rgb( 0, 0, 255)',
'blueviolet': 'rgb(138, 43, 226)',
'brown': 'rgb(165, 42, 42)',
'burlywood': 'rgb(222, 184, 135)',
'cadetblue': 'rgb( 95, 158, 160)',
'chartreuse': 'rgb(127, 255, 0)',
'chocolate': 'rgb(210, 105, 30)',
'coral': 'rgb(255, 127, 80)',
'cornflowerblue': 'rgb(100, 149, 237)',
'cornsilk': 'rgb(255, 248, 220)',
'crimson': 'rgb(220, 20, 60)',
'cyan': 'rgb( 0, 255, 255)',
'darkblue': 'rgb( 0, 0, 139)',
'darkcyan': 'rgb( 0, 139, 139)',
'darkgoldenrod': 'rgb(184, 134, 11)',
'darkgray': 'rgb(169, 169, 169)',
'darkgreen': 'rgb( 0, 100, 0)',
'darkgrey': 'rgb(169, 169, 169)',
'darkkhaki': 'rgb(189, 183, 107)',
'darkmagenta': 'rgb(139, 0, 139)',
'darkolivegreen': 'rgb( 85, 107, 47)',
'darkorange': 'rgb(255, 140, 0)',
'darkorchid': 'rgb(153, 50, 204)',
'darkred': 'rgb(139, 0, 0)',
'darksalmon': 'rgb(233, 150, 122)',
'darkseagreen': 'rgb(143, 188, 143)',
'darkslateblue': 'rgb( 72, 61, 139)',
'darkslategray': 'rgb( 47, 79, 79)',
'darkslategrey': 'rgb( 47, 79, 79)',
'darkturquoise': 'rgb( 0, 206, 209)',
'darkviolet': 'rgb(148, 0, 211)',
'deeppink': 'rgb(255, 20, 147)',
'deepskyblue': 'rgb( 0, 191, 255)',
'dimgray': 'rgb(105, 105, 105)',
'dimgrey': 'rgb(105, 105, 105)',
'dodgerblue': 'rgb( 30, 144, 255)',
'firebrick': 'rgb(178, 34, 34)',
'floralwhite': 'rgb(255, 250, 240)',
'forestgreen': 'rgb( 34, 139, 34)',
'fuchsia': 'rgb(255, 0, 255)',
'gainsboro': 'rgb(220, 220, 220)',
'ghostwhite': 'rgb(248, 248, 255)',
'gold': 'rgb(255, 215, 0)',
'goldenrod': 'rgb(218, 165, 32)',
'gray': 'rgb(128, 128, 128)',
'grey': 'rgb(128, 128, 128)',
'green': 'rgb( 0, 128, 0)',
'greenyellow': 'rgb(173, 255, 47)',
'honeydew': 'rgb(240, 255, 240)',
'hotpink': 'rgb(255, 105, 180)',
'indianred': 'rgb(205, 92, 92)',
'indigo': 'rgb( 75, 0, 130)',
'ivory': 'rgb(255, 255, 240)',
'khaki': 'rgb(240, 230, 140)',
'lavender': 'rgb(230, 230, 250)',
'lavenderblush': 'rgb(255, 240, 245)',
'lawngreen': 'rgb(124, 252, 0)',
'lemonchiffon': 'rgb(255, 250, 205)',
'lightblue': 'rgb(173, 216, 230)',
'lightcoral': 'rgb(240, 128, 128)',
'lightcyan': 'rgb(224, 255, 255)',
'lightgoldenrodyellow': 'rgb(250, 250, 210)',
'lightgray': 'rgb(211, 211, 211)',
'lightgreen': 'rgb(144, 238, 144)',
'lightgrey': 'rgb(211, 211, 211)',
'lightpink': 'rgb(255, 182, 193)',
'lightsalmon': 'rgb(255, 160, 122)',
'lightseagreen': 'rgb( 32, 178, 170)',
'lightskyblue': 'rgb(135, 206, 250)',
'lightslategray': 'rgb(119, 136, 153)',
'lightslategrey': 'rgb(119, 136, 153)',
'lightsteelblue': 'rgb(176, 196, 222)',
'lightyellow': 'rgb(255, 255, 224)',
'lime': 'rgb( 0, 255, 0)',
'limegreen': 'rgb( 50, 205, 50)',
'linen': 'rgb(250, 240, 230)',
'magenta': 'rgb(255, 0, 255)',
'maroon': 'rgb(128, 0, 0)',
'mediumaquamarine': 'rgb(102, 205, 170)',
'mediumblue': 'rgb( 0, 0, 205)',
'mediumorchid': 'rgb(186, 85, 211)',
'mediumpurple': 'rgb(147, 112, 219)',
'mediumseagreen': 'rgb( 60, 179, 113)',
'mediumslateblue': 'rgb(123, 104, 238)',
'mediumspringgreen': 'rgb( 0, 250, 154)',
'mediumturquoise': 'rgb( 72, 209, 204)',
'mediumvioletred': 'rgb(199, 21, 133)',
'midnightblue': 'rgb( 25, 25, 112)',
'mintcream': 'rgb(245, 255, 250)',
'mistyrose': 'rgb(255, 228, 225)',
'moccasin': 'rgb(255, 228, 181)',
'navajowhite': 'rgb(255, 222, 173)',
'navy': 'rgb( 0, 0, 128)',
'oldlace': 'rgb(253, 245, 230)',
'olive': 'rgb(128, 128, 0)',
'olivedrab': 'rgb(107, 142, 35)',
'orange': 'rgb(255, 165, 0)',
'orangered': 'rgb(255, 69, 0)',
'orchid': 'rgb(218, 112, 214)',
'palegoldenrod': 'rgb(238, 232, 170)',
'palegreen': 'rgb(152, 251, 152)',
'paleturquoise': 'rgb(175, 238, 238)',
'palevioletred': 'rgb(219, 112, 147)',
'papayawhip': 'rgb(255, 239, 213)',
'peachpuff': 'rgb(255, 218, 185)',
'peru': 'rgb(205, 133, 63)',
'pink': 'rgb(255, 192, 203)',
'plum': 'rgb(221, 160, 221)',
'powderblue': 'rgb(176, 224, 230)',
'purple': 'rgb(128, 0, 128)',
'red': 'rgb(255, 0, 0)',
'rosybrown': 'rgb(188, 143, 143)',
'royalblue': 'rgb( 65, 105, 225)',
'saddlebrown': 'rgb(139, 69, 19)',
'salmon': 'rgb(250, 128, 114)',
'sandybrown': 'rgb(244, 164, 96)',
'seagreen': 'rgb( 46, 139, 87)',
'seashell': 'rgb(255, 245, 238)',
'sienna': 'rgb(160, 82, 45)',
'silver': 'rgb(192, 192, 192)',
'skyblue': 'rgb(135, 206, 235)',
'slateblue': 'rgb(106, 90, 205)',
'slategray': 'rgb(112, 128, 144)',
'slategrey': 'rgb(112, 128, 144)',
'snow': 'rgb(255, 250, 250)',
'springgreen': 'rgb( 0, 255, 127)',
'steelblue': 'rgb( 70, 130, 180)',
'tan': 'rgb(210, 180, 140)',
'teal': 'rgb( 0, 128, 128)',
'thistle': 'rgb(216, 191, 216)',
'tomato': 'rgb(255, 99, 71)',
'turquoise': 'rgb( 64, 224, 208)',
'violet': 'rgb(238, 130, 238)',
'wheat': 'rgb(245, 222, 179)',
'white': 'rgb(255, 255, 255)',
'whitesmoke': 'rgb(245, 245, 245)',
'yellow': 'rgb(255, 255, 0)',
'yellowgreen': 'rgb(154, 205, 50)',
}
def isSameSign(a,b): return (a <= 0 and b <= 0) or (a >= 0 and b >= 0)
coord = re.compile("\\-?\\d+\\.?\\d*")
scinumber = re.compile("[\\-\\+]?(\\d*\\.?)?\\d+[eE][\\-\\+]?\\d+")
number = re.compile("[\\-\\+]?(\\d*\\.?)?\\d+")
sciExponent = re.compile("[eE]([\\-\\+]?\\d+)")
unit = re.compile("(em|ex|px|pt|pc|cm|mm|in|\\%){1,1}$")
class Unit(object):
INVALID = -1
NONE = 0
PCT = 1
PX = 2
PT = 3
PC = 4
EM = 5
EX = 6
CM = 7
MM = 8
IN = 9
def get(str):
if str == None or str == '': return Unit.NONE
elif str == '%': return Unit.PCT
elif str == 'px': return Unit.PX
elif str == 'pt': return Unit.PT
elif str == 'pc': return Unit.PC
elif str == 'em': return Unit.EM
elif str == 'ex': return Unit.EX
elif str == 'cm': return Unit.CM
elif str == 'mm': return Unit.MM
elif str == 'in': return Unit.IN
return Unit.INVALID
def str(u):
if u == Unit.NONE: return ''
elif u == Unit.PCT: return '%'
elif u == Unit.PX: return 'px'
elif u == Unit.PT: return 'pt'
elif u == Unit.PC: return 'pc'
elif u == Unit.EM: return 'em'
elif u == Unit.EX: return 'ex'
elif u == Unit.CM: return 'cm'
elif u == Unit.MM: return 'mm'
elif u == Unit.IN: return 'in'
return 'INVALID'
get = staticmethod(get)
str = staticmethod(str)
class SVGLength(object):
def __init__(self, str):
try:
self.value = float(str)
if int(self.value) == self.value:
self.value = int(self.value)
self.units = Unit.NONE
except ValueError:
self.value = 0
unitBegin = 0
scinum = scinumber.match(str)
if scinum != None:
numMatch = number.match(str)
expMatch = sciExponent.search(str, numMatch.start(0))
self.value = (float(numMatch.group(0)) *
10 ** float(expMatch.group(1)))
unitBegin = expMatch.end(1)
else:
numMatch = number.match(str)
if numMatch != None:
self.value = float(numMatch.group(0))
unitBegin = numMatch.end(0)
if int(self.value) == self.value:
self.value = int(self.value)
if unitBegin != 0 :
unitMatch = unit.search(str, unitBegin)
if unitMatch != None :
self.units = Unit.get(unitMatch.group(0))
else:
self.value = 0
self.units = Unit.INVALID
def getSVGLength(value):
try:
v = float(value)
except ValueError:
coordMatch = coord.match(value)
if coordMatch != None:
unitMatch = unit.search(value, coordMatch.start(0))
v = value
return v
def findElementById(node, id):
if node == None or node.nodeType != 1: return None
if node.getAttribute('id') == id: return node
for child in node.childNodes :
e = findElementById(child,id)
if e != None: return e
return None
def findElementsWithId(node, elems=None):
"""
Returns all elements with id attributes
"""
if elems is None:
elems = {}
id = node.getAttribute('id')
if id != '' :
elems[id] = node
if node.hasChildNodes() :
for child in node.childNodes:
if child.nodeType == 1 :
findElementsWithId(child, elems)
return elems
referencingProps = ['fill', 'stroke', 'filter', 'clip-path', 'mask', 'marker-start',
'marker-end', 'marker-mid']
def findReferencedElements(node, ids=None):
"""
Returns the number of times an ID is referenced as well as all elements
that reference it.
Currently looks at fill, stroke, clip-path, mask, marker, and
xlink:href attributes.
"""
global referencingProps
if ids is None:
ids = {}
if node.nodeName == 'style' and node.namespaceURI == NS['SVG']:
if node.firstChild != None:
cssRules = parseCssString(node.firstChild.nodeValue)
for rule in cssRules:
for propname in rule['properties']:
propval = rule['properties'][propname]
findReferencingProperty(node, propname, propval, ids)
return ids
href = node.getAttributeNS(NS['XLINK'],'href')
if href != '' and len(href) > 1 and href[0] == '#':
id = href[1:]
if id in ids:
ids[id][0] += 1
ids[id][1].append(node)
else:
ids[id] = [1,[node]]
styles = node.getAttribute('style').split(';')
for attr in referencingProps:
styles.append(':'.join([attr, node.getAttribute(attr)]))
for style in styles:
propval = style.split(':')
if len(propval) == 2 :
prop = propval[0].strip()
val = propval[1].strip()
findReferencingProperty(node, prop, val, ids)
if node.hasChildNodes() :
for child in node.childNodes:
if child.nodeType == 1 :
findReferencedElements(child, ids)
return ids
def findReferencingProperty(node, prop, val, ids):
global referencingProps
if prop in referencingProps and val != '' :
if len(val) >= 7 and val[0:5] == 'url(#' :
id = val[5:val.find(')')]
if ids.has_key(id) :
ids[id][0] += 1
ids[id][1].append(node)
else:
ids[id] = [1,[node]]
elif len(val) >= 8 :
id = None
if val[0:6] == 'url("#' :
id = val[6:val.find('")')]
elif val[0:6] == "url('#" :
id = val[6:val.find("')")]
if id != None:
if ids.has_key(id) :
ids[id][0] += 1
ids[id][1].append(node)
else:
ids[id] = [1,[node]]
numIDsRemoved = 0
numElemsRemoved = 0
numAttrsRemoved = 0
numRastersEmbedded = 0
numPathSegmentsReduced = 0
numCurvesStraightened = 0
numBytesSavedInPathData = 0
numBytesSavedInColors = 0
numPointsRemovedFromPolygon = 0
def removeUnusedDefs(doc, defElem, elemsToRemove=None):
if elemsToRemove is None:
elemsToRemove = []
identifiedElements = findElementsWithId(doc.documentElement)
referencedIDs = findReferencedElements(doc.documentElement)
keepTags = ['font', 'style', 'metadata', 'script', 'title', 'desc']
for elem in defElem.childNodes:
if elem.nodeType == 1 and (elem.getAttribute('id') == '' or \
(not elem.getAttribute('id') in referencedIDs)):
if elem.nodeName == 'g' and elem.namespaceURI == NS['SVG']:
elemsToRemove = removeUnusedDefs(doc, elem, elemsToRemove)
elif not elem.nodeName in keepTags:
elemsToRemove.append(elem)
return elemsToRemove
def removeUnreferencedElements(doc):
"""
Removes all unreferenced elements except for <svg>, <font>, <metadata>, <title>, and <desc>.
Also vacuums the defs of any non-referenced renderable elements.
Returns the number of unreferenced elements removed from the document.
"""
global numElemsRemoved
num = 0
removeTags = ['linearGradient', 'radialGradient', 'pattern']
identifiedElements = findElementsWithId(doc.documentElement)
referencedIDs = findReferencedElements(doc.documentElement)
for id in identifiedElements:
if not id in referencedIDs:
goner = findElementById(doc.documentElement, id)
if goner != None and goner.parentNode != None and goner.nodeName in removeTags:
goner.parentNode.removeChild(goner)
num += 1
numElemsRemoved += 1
num = 0
defs = doc.documentElement.getElementsByTagName('defs')
for aDef in defs:
elemsToRemove = removeUnusedDefs(doc, aDef)
for elem in elemsToRemove:
elem.parentNode.removeChild(elem)
numElemsRemoved += 1
num += 1
return num
def removeUnreferencedIDs(referencedIDs, identifiedElements):
"""
Removes the unreferenced ID attributes.
Returns the number of ID attributes removed
"""
global numIDsRemoved
keepTags = ['font']
num = 0;
for id in identifiedElements.keys():
node = identifiedElements[id]
if referencedIDs.has_key(id) == False and not node.nodeName in keepTags:
node.removeAttribute('id')
numIDsRemoved += 1
num += 1
return num
def removeNamespacedAttributes(node, namespaces):
global numAttrsRemoved
num = 0
if node.nodeType == 1 :
attrList = node.attributes
attrsToRemove = []
for attrNum in range(attrList.length):
attr = attrList.item(attrNum)
if attr != None and attr.namespaceURI in namespaces:
attrsToRemove.append(attr.nodeName)
for attrName in attrsToRemove :
num += 1
numAttrsRemoved += 1
node.removeAttribute(attrName)
for child in node.childNodes:
num += removeNamespacedAttributes(child, namespaces)
return num
def removeNamespacedElements(node, namespaces):
global numElemsRemoved
num = 0
if node.nodeType == 1 :
childList = node.childNodes
childrenToRemove = []
for child in childList:
if child != None and child.namespaceURI in namespaces:
childrenToRemove.append(child)
for child in childrenToRemove :
num += 1
numElemsRemoved += 1
node.removeChild(child)
for child in node.childNodes:
num += removeNamespacedElements(child, namespaces)
return num
def removeNestedGroups(node):
"""
This walks further and further down the tree, removing groups
which do not have any attributes or a title/desc child and
promoting their children up one level
"""
global numElemsRemoved
num = 0
groupsToRemove = []
for child in node.childNodes:
if child.nodeName == 'g' and child.namespaceURI == NS['SVG'] and len(child.attributes) == 0:
for grandchild in child.childNodes:
if grandchild.nodeType == 1 and grandchild.namespaceURI == NS['SVG'] and \
grandchild.nodeName in ['title','desc']:
break
else:
groupsToRemove.append(child)
for g in groupsToRemove:
while g.childNodes.length > 0:
g.parentNode.insertBefore(g.firstChild, g)
g.parentNode.removeChild(g)
numElemsRemoved += 1
num += 1
for child in node.childNodes:
if child.nodeType == 1:
num += removeNestedGroups(child)
return num
def moveCommonAttributesToParentGroup(elem):
"""
This recursively calls this function on all children of the passed in element
and then iterates over all child elements and removes common inheritable attributes
from the children and places them in the parent group. But only if the parent contains
nothing but element children and whitespace.
"""
num = 0
childElements = []
for child in elem.childNodes:
if child.nodeType == 1:
childElements.append(child)
num += moveCommonAttributesToParentGroup(child)
elif child.nodeType == 3 and child.nodeValue.strip():
return num
if len(childElements) <= 1: return num
commonAttrs = {}
attrList = childElements[0].attributes
for num in range(attrList.length):
attr = attrList.item(num)
if attr.nodeName in ['clip-rule',
'display-align',
'fill', 'fill-opacity', 'fill-rule',
'font', 'font-family', 'font-size', 'font-size-adjust', 'font-stretch',
'font-style', 'font-variant', 'font-weight',
'letter-spacing',
'pointer-events', 'shape-rendering',
'stroke', 'stroke-dasharray', 'stroke-dashoffset', 'stroke-linecap', 'stroke-linejoin',
'stroke-miterlimit', 'stroke-opacity', 'stroke-width',
'text-anchor', 'text-decoration', 'text-rendering', 'visibility',
'word-spacing', 'writing-mode']:
commonAttrs[attr.nodeName] = attr.nodeValue
for childNum in range(len(childElements)):
if childNum == 0:
continue
child = childElements[childNum]
if child.localName in ['set', 'animate', 'animateColor', 'animateTransform', 'animateMotion']:
continue
distinctAttrs = []
for name in commonAttrs.keys():
if child.getAttribute(name) != commonAttrs[name]:
distinctAttrs.append(name)
for name in distinctAttrs:
del commonAttrs[name]
for name in commonAttrs.keys():
for child in childElements:
child.removeAttribute(name)
elem.setAttribute(name, commonAttrs[name])
num += (len(childElements)-1) * len(commonAttrs)
return num
def removeUnusedAttributesOnParent(elem):
"""
This recursively calls this function on all children of the element passed in,
then removes any unused attributes on this elem if none of the children inherit it
"""
num = 0
childElements = []
for child in elem.childNodes:
if child.nodeType == 1:
childElements.append(child)
num += removeUnusedAttributesOnParent(child)
if len(childElements) <= 1: return num
attrList = elem.attributes
unusedAttrs = {}
for num in range(attrList.length):
attr = attrList.item(num)
if attr.nodeName in ['clip-rule',
'display-align',
'fill', 'fill-opacity', 'fill-rule',
'font', 'font-family', 'font-size', 'font-size-adjust', 'font-stretch',
'font-style', 'font-variant', 'font-weight',
'letter-spacing',
'pointer-events', 'shape-rendering',
'stroke', 'stroke-dasharray', 'stroke-dashoffset', 'stroke-linecap', 'stroke-linejoin',
'stroke-miterlimit', 'stroke-opacity', 'stroke-width',
'text-anchor', 'text-decoration', 'text-rendering', 'visibility',
'word-spacing', 'writing-mode']:
unusedAttrs[attr.nodeName] = attr.nodeValue
for childNum in range(len(childElements)):
child = childElements[childNum]
inheritedAttrs = []
for name in unusedAttrs.keys():
val = child.getAttribute(name)
if val == '' or val == None or val == 'inherit':
inheritedAttrs.append(name)
for a in inheritedAttrs:
del unusedAttrs[a]
for name in unusedAttrs.keys():
elem.removeAttribute(name)
num += 1
return num
def removeDuplicateGradientStops(doc):
global numElemsRemoved
num = 0
for gradType in ['linearGradient', 'radialGradient']:
for grad in doc.getElementsByTagName(gradType):
stops = {}
stopsToRemove = []
for stop in grad.getElementsByTagName('stop'):
offsetU = SVGLength(stop.getAttribute('offset'))
if offsetU.units == Unit.PCT:
offset = offsetU.value / 100.0
elif offsetU.units == Unit.NONE:
offset = offsetU.value
else:
offset = 0
if int(offset) == offset: stop.setAttribute('offset', str(int(offset)))
else: stop.setAttribute('offset', str(offset))
color = stop.getAttribute('stop-color')
opacity = stop.getAttribute('stop-opacity')
if stops.has_key(offset) :
oldStop = stops[offset]
if oldStop[0] == color and oldStop[1] == opacity:
stopsToRemove.append(stop)
stops[offset] = [color, opacity]
for stop in stopsToRemove:
stop.parentNode.removeChild(stop)
num += 1
numElemsRemoved += 1
return num
def collapseSinglyReferencedGradients(doc):
global numElemsRemoved
num = 0
for rid,nodeCount in findReferencedElements(doc.documentElement).iteritems():
count = nodeCount[0]
nodes = nodeCount[1]
if count == 1:
elem = findElementById(doc.documentElement,rid)
if elem != None and elem.nodeType == 1 and elem.nodeName in ['linearGradient', 'radialGradient'] \
and elem.namespaceURI == NS['SVG']:
refElem = nodes[0]
if refElem.nodeType == 1 and refElem.nodeName in ['linearGradient', 'radialGradient'] \
and refElem.namespaceURI == NS['SVG']:
if len(refElem.getElementsByTagName('stop')) == 0:
stopsToAdd = elem.getElementsByTagName('stop')
for stop in stopsToAdd:
refElem.appendChild(stop)
for attr in ['gradientUnits','spreadMethod','gradientTransform']:
if refElem.getAttribute(attr) == '' and not elem.getAttribute(attr) == '':
refElem.setAttributeNS(None, attr, elem.getAttribute(attr))
if elem.nodeName == 'radialGradient' and refElem.nodeName == 'radialGradient':
for attr in ['fx','fy','cx','cy','r']:
if refElem.getAttribute(attr) == '' and not elem.getAttribute(attr) == '':
refElem.setAttributeNS(None, attr, elem.getAttribute(attr))
if elem.nodeName == 'linearGradient' and refElem.nodeName == 'linearGradient':
for attr in ['x1','y1','x2','y2']:
if refElem.getAttribute(attr) == '' and not elem.getAttribute(attr) == '':
refElem.setAttributeNS(None, attr, elem.getAttribute(attr))
refElem.removeAttributeNS(NS['XLINK'], 'href')
elem.parentNode.removeChild(elem)
numElemsRemoved += 1
num += 1
return num
def removeDuplicateGradients(doc):
global numElemsRemoved
num = 0
gradientsToRemove = {}
duplicateToMaster = {}
for gradType in ['linearGradient', 'radialGradient']:
grads = doc.getElementsByTagName(gradType)
for grad in grads:
for ograd in grads:
if grad == ograd: continue
someGradAttrsDoNotMatch = False
for attr in ['gradientUnits','spreadMethod','gradientTransform','x1','y1','x2','y2','cx','cy','fx','fy','r']:
if grad.getAttribute(attr) != ograd.getAttribute(attr):
someGradAttrsDoNotMatch = True
break;
if someGradAttrsDoNotMatch: continue
if grad.getAttributeNS(NS['XLINK'], 'href') != ograd.getAttributeNS(NS['XLINK'], 'href'):
continue
stops = grad.getElementsByTagName('stop')
ostops = ograd.getElementsByTagName('stop')
if stops.length != ostops.length: continue
stopsNotEqual = False
for i in range(stops.length):
if stopsNotEqual: break
stop = stops.item(i)
ostop = ostops.item(i)
for attr in ['offset', 'stop-color', 'stop-opacity']:
if stop.getAttribute(attr) != ostop.getAttribute(attr):
stopsNotEqual = True
break
if stopsNotEqual: continue
if not gradientsToRemove.has_key(ograd):
if not duplicateToMaster.has_key(ograd):
if not gradientsToRemove.has_key(grad):
gradientsToRemove[grad] = []
gradientsToRemove[grad].append( ograd )
duplicateToMaster[ograd] = grad
referencedIDs = findReferencedElements(doc.documentElement)
for masterGrad in gradientsToRemove.keys():
master_id = masterGrad.getAttribute('id')
for dupGrad in gradientsToRemove[masterGrad]:
if not dupGrad.parentNode: continue
dup_id = dupGrad.getAttribute('id')
for elem in referencedIDs[dup_id][1]:
for attr in ['fill', 'stroke']:
v = elem.getAttribute(attr)
if v == 'url(#'+dup_id+')' or v == 'url("#'+dup_id+'")' or v == "url('#"+dup_id+"')":
elem.setAttribute(attr, 'url(#'+master_id+')')
if elem.getAttributeNS(NS['XLINK'], 'href') == '#'+dup_id:
elem.setAttributeNS(NS['XLINK'], 'href', '#'+master_id)
dupGrad.parentNode.removeChild(dupGrad)
numElemsRemoved += 1
num += 1
return num
def repairStyle(node, options):
num = 0
if node.nodeType == 1 and len(node.getAttribute('style')) > 0 :
styleMap = { }
rawStyles = node.getAttribute('style').split(';')
for style in rawStyles:
propval = style.split(':')
if len(propval) == 2 :
styleMap[propval[0].strip()] = propval[1].strip()
for prop in ['fill', 'stroke'] :
if styleMap.has_key(prop) :
chunk = styleMap[prop].split(') ')
if len(chunk) == 2 and (chunk[0][:5] == 'url(#' or chunk[0][:6] == 'url("#' or chunk[0][:6] == "url('#") and chunk[1] == 'rgb(0, 0, 0)' :
styleMap[prop] = chunk[0] + ')'
num += 1
if styleMap.has_key('opacity') :
opacity = float(styleMap['opacity'])
if opacity == 1.0 :
del styleMap['opacity']
num += 1
elif opacity == 0.0 :
for uselessStyle in ['fill', 'fill-opacity', 'fill-rule', 'stroke', 'stroke-linejoin',
'stroke-opacity', 'stroke-miterlimit', 'stroke-linecap', 'stroke-dasharray',
'stroke-dashoffset', 'stroke-opacity'] :
if styleMap.has_key(uselessStyle):
del styleMap[uselessStyle]
num += 1
if styleMap.has_key('stroke') and styleMap['stroke'] == 'none' :
for strokestyle in [ 'stroke-width', 'stroke-linejoin', 'stroke-miterlimit',
'stroke-linecap', 'stroke-dasharray', 'stroke-dashoffset', 'stroke-opacity'] :
if styleMap.has_key(strokestyle) :
del styleMap[strokestyle]
num += 1
del styleMap['stroke']
if styleMap.has_key('fill') and styleMap['fill'] == 'none' :
for fillstyle in [ 'fill-rule', 'fill-opacity' ] :
if styleMap.has_key(fillstyle) :
del styleMap[fillstyle]
num += 1
if styleMap.has_key('stop-opacity') :
if float(styleMap['stop-opacity']) == 1.0 :
del styleMap['stop-opacity']
num += 1
if styleMap.has_key('fill-opacity') :
fillOpacity = float(styleMap['fill-opacity'])
if fillOpacity == 1.0 :
del styleMap['fill-opacity']
num += 1
elif fillOpacity == 0.0 :
for uselessFillStyle in [ 'fill', 'fill-rule' ] :
if styleMap.has_key(uselessFillStyle):
del styleMap[uselessFillStyle]
num += 1
if styleMap.has_key('stroke-opacity') :
strokeOpacity = float(styleMap['stroke-opacity'])
if strokeOpacity == 1.0 :
del styleMap['stroke-opacity']
num += 1
elif strokeOpacity == 0.0 :
for uselessStrokeStyle in [ 'stroke', 'stroke-width', 'stroke-linejoin', 'stroke-linecap',
'stroke-dasharray', 'stroke-dashoffset' ] :
if styleMap.has_key(uselessStrokeStyle):
del styleMap[uselessStrokeStyle]
num += 1
if styleMap.has_key('stroke-width') :
strokeWidth = getSVGLength(styleMap['stroke-width'])
if strokeWidth == 0.0 :
for uselessStrokeStyle in [ 'stroke', 'stroke-linejoin', 'stroke-linecap',
'stroke-dasharray', 'stroke-dashoffset', 'stroke-opacity' ] :
if styleMap.has_key(uselessStrokeStyle):
del styleMap[uselessStrokeStyle]
num += 1
if node.nodeName in ['rect', 'circle', 'ellipse', 'line', 'polyline', 'polygon', 'path']:
for fontstyle in [ 'font-family', 'font-size', 'font-stretch', 'font-size-adjust',
'font-style', 'font-variant', 'font-weight',
'letter-spacing', 'line-height', 'kerning',
'text-anchor', 'text-decoration', 'text-rendering',
'unicode-bidi', 'word-spacing', 'writing-mode'] :
if styleMap.has_key(fontstyle) :
del styleMap[fontstyle]
num += 1
for inkscapeStyle in ['-inkscape-font-specification']:
if styleMap.has_key(inkscapeStyle):
del styleMap[inkscapeStyle]
num += 1
if styleMap.has_key('visibility') :
if styleMap['visibility'] == 'visible':
del styleMap['visibility']
num += 1
if styleMap.has_key('display') :
if styleMap['display'] == 'inline':
del styleMap['display']
num += 1
if styleMap.has_key('overflow') :
if styleMap['overflow'] == 'visible' or node.nodeName in ['svg','marker','pattern']:
del styleMap['overflow']
num += 1
if styleMap.has_key('marker') :
if styleMap['marker'] == 'none':
del styleMap['marker']
num += 1
if options.style_to_xml:
for propName in styleMap.keys() :
if propName in svgAttributes :
node.setAttribute(propName, styleMap[propName])
del styleMap[propName]
fixedStyle = ''
for prop in styleMap.keys() :
fixedStyle += prop + ':' + styleMap[prop] + ';'
if fixedStyle != '' :
node.setAttribute('style', fixedStyle)
else:
node.removeAttribute('style')
for child in node.childNodes :
num += repairStyle(child,options)
return num
def removeDefaultAttributeValues(node, options):
num = 0
if node.nodeType != 1: return 0
if node.getAttribute('gradientUnits') == 'objectBoundingBox':
node.removeAttribute('gradientUnits')
num += 1
if node.getAttribute('spreadMethod') == 'pad':
node.removeAttribute('spreadMethod')
num += 1
if node.getAttribute('x1') != '':
x1 = SVGLength(node.getAttribute('x1'))
if x1.value == 0:
node.removeAttribute('x1')
num += 1
if node.getAttribute('y1') != '':
y1 = SVGLength(node.getAttribute('y1'))
if y1.value == 0:
node.removeAttribute('y1')
num += 1
if node.getAttribute('x2') != '':
x2 = SVGLength(node.getAttribute('x2'))
if (x2.value == 100 and x2.units == Unit.PCT) or (x2.value == 1 and x2.units == Unit.NONE):
node.removeAttribute('x2')
num += 1
if node.getAttribute('y2') != '':
y2 = SVGLength(node.getAttribute('y2'))
if y2.value == 0:
node.removeAttribute('y2')
num += 1
if node.getAttribute('fx') != '':
if node.getAttribute('fx') == node.getAttribute('cx'):
node.removeAttribute('fx')
num += 1
if node.getAttribute('fy') != '':
if node.getAttribute('fy') == node.getAttribute('cy'):
node.removeAttribute('fy')
num += 1
if node.getAttribute('cx') != '':
cx = SVGLength(node.getAttribute('cx'))
if (cx.value == 50 and cx.units == Unit.PCT) or (cx.value == 0.5 and cx.units == Unit.NONE):
node.removeAttribute('cx')
num += 1
if node.getAttribute('cy') != '':
cy = SVGLength(node.getAttribute('cy'))
if (cy.value == 50 and cy.units == Unit.PCT) or (cy.value == 0.5 and cy.units == Unit.NONE):
node.removeAttribute('cy')
num += 1
if node.getAttribute('r') != '':
r = SVGLength(node.getAttribute('r'))
if (r.value == 50 and r.units == Unit.PCT) or (r.value == 0.5 and r.units == Unit.NONE):
node.removeAttribute('r')
num += 1
for child in node.childNodes :
num += removeDefaultAttributeValues(child,options)
return num
rgb = re.compile("\\s*rgb\\(\\s*(\\d+)\\s*\\,\\s*(\\d+)\\s*\\,\\s*(\\d+)\\s*\\)\\s*")
rgbp = re.compile("\\s*rgb\\(\\s*(\\d*\\.?\\d+)\\%\\s*\\,\\s*(\\d*\\.?\\d+)\\%\\s*\\,\\s*(\\d*\\.?\\d+)\\%\\s*\\)\\s*")
def convertColor(value):
"""
Converts the input color string and returns a #RRGGBB (or #RGB if possible) string
"""
s = value
if s in colors.keys():
s = colors[s]
rgbpMatch = rgbp.match(s)
if rgbpMatch != None :
r = int(float(rgbpMatch.group(1)) * 255.0 / 100.0)
g = int(float(rgbpMatch.group(2)) * 255.0 / 100.0)
b = int(float(rgbpMatch.group(3)) * 255.0 / 100.0)
s = 'rgb(%d,%d,%d)' % (r,g,b)
rgbMatch = rgb.match(s)
if rgbMatch != None :
r = hex( int( rgbMatch.group(1) ) )[2:].upper()
g = hex( int( rgbMatch.group(2) ) )[2:].upper()
b = hex( int( rgbMatch.group(3) ) )[2:].upper()
if len(r) == 1: r='0'+r
if len(g) == 1: g='0'+g
if len(b) == 1: b='0'+b
s = '#'+r+g+b
if s[0] == '#' and len(s)==7 and s[1]==s[2] and s[3]==s[4] and s[5]==s[6]:
s = s.upper()
s = '#'+s[1]+s[3]+s[5]
return s
def convertColors(element) :
"""
Recursively converts all color properties into #RRGGBB format if shorter
"""
numBytes = 0
if element.nodeType != 1: return 0
attrsToConvert = []
if element.nodeName in ['rect', 'circle', 'ellipse', 'polygon', \
'line', 'polyline', 'path', 'g', 'a']:
attrsToConvert = ['fill', 'stroke']
elif element.nodeName in ['stop']:
attrsToConvert = ['stop-color']
elif element.nodeName in ['solidColor']:
attrsToConvert = ['solid-color']
for attr in attrsToConvert:
oldColorValue = element.getAttribute(attr)
if oldColorValue != '':
newColorValue = convertColor(oldColorValue)
oldBytes = len(oldColorValue)
newBytes = len(newColorValue)
if oldBytes > newBytes:
element.setAttribute(attr, newColorValue)
numBytes += (oldBytes - len(element.getAttribute(attr)))
for child in element.childNodes :
numBytes += convertColors(child)
return numBytes
def cleanPath(element) :
"""
Cleans the path string (d attribute) of the element
"""
global numBytesSavedInPathData
global numPathSegmentsReduced
global numCurvesStraightened
oldPathStr = element.getAttribute('d')
pathObj = svg_parser.parse(oldPathStr)
path = []
for (cmd,dataset) in pathObj:
if cmd in ['M','m','L','l','T','t']:
nums = []
for t in dataset:
nums.append(Decimal(str(t[0])) * Decimal(1))
nums.append(Decimal(str(t[1])) * Decimal(1))
if nums:
path.append( (cmd, nums) )
elif cmd in ['V','v','H','h']:
nums = []
for n in dataset:
nums.append(Decimal(str(n)))
if nums:
path.append( (cmd, nums) )
elif cmd in ['C','c']:
nums = []
for t in dataset:
for pair in t:
nums.append(Decimal(str(pair[0])) * Decimal(1))
nums.append(Decimal(str(pair[1])) * Decimal(1))
path.append( (cmd, nums) )
elif cmd in ['S','s','Q','q']:
nums = []
for t in dataset:
for pair in t:
nums.append(Decimal(str(pair[0])) * Decimal(1))
nums.append(Decimal(str(pair[1])) * Decimal(1))
path.append( (cmd, nums) )
elif cmd in ['A','a']:
nums = []
for t in dataset:
nums.append( Decimal(str(t[0][0])) * Decimal(1) )
nums.append( Decimal(str(t[0][1])) * Decimal(1) )
nums.append( Decimal(str(t[1])) * Decimal(1))
if t[2]: nums.append( Decimal(1) )
else: nums.append( Decimal(0) )
if t[3]: nums.append( Decimal(1) )
else: nums.append( Decimal(0) )
nums.append( Decimal(str(t[4][0])) * Decimal(1) )
nums.append( Decimal(str(t[4][1])) * Decimal(1) )
path.append( (cmd, nums) )
elif cmd in ['Z','z']:
path.append( (cmd, []) )
if len(path[0][1]) == 2:
(x,y) = path[0][1]
else:
N = len(path[0][1])
if path[0][0] == 'M':
x = path[0][1][N-2]
y = path[0][1][N-1]
else:
(x,y) = path[0][1][0],path[0][1][1]
n = 2
while n < N:
x += path[0][1][n]
y += path[0][1][n+1]
n += 2
(startx,starty) = (x,y)
newPath = [path[0]]
for (cmd,data) in path[1:]:
i = 0
newCmd = cmd
newData = data
if cmd == 'A':
newCmd = 'a'
newData = []
while i < len(data):
newData.append(data[i])
newData.append(data[i+1])
newData.append(data[i+2])
newData.append(data[i+3])
newData.append(data[i+4])
newData.append(data[i+5]-x)
newData.append(data[i+6]-y)
x = data[i+5]
y = data[i+6]
i += 7
elif cmd == 'a':
while i < len(data):
x += data[i+5]
y += data[i+6]
i += 7
elif cmd == 'H':
newCmd = 'h'
newData = []
while i < len(data):
newData.append(data[i]-x)
x = data[i]
i += 1
elif cmd == 'h':
while i < len(data):
x += data[i]
i += 1
elif cmd == 'V':
newCmd = 'v'
newData = []
while i < len(data):
newData.append(data[i] - y)
y = data[i]
i += 1
elif cmd == 'v':
while i < len(data):
y += data[i]
i += 1
elif cmd in ['M']:
newCmd = cmd.lower()
newData = []
startx = data[0]
starty = data[1]
while i < len(data):
newData.append( data[i] - x )
newData.append( data[i+1] - y )
x = data[i]
y = data[i+1]
i += 2
elif cmd in ['L','T']:
newCmd = cmd.lower()
newData = []
while i < len(data):
newData.append( data[i] - x )
newData.append( data[i+1] - y )
x = data[i]
y = data[i+1]
i += 2
elif cmd in ['m']:
startx += data[0]
starty += data[1]
while i < len(data):
x += data[i]
y += data[i+1]
i += 2
elif cmd in ['l','t']:
while i < len(data):
x += data[i]
y += data[i+1]
i += 2
elif cmd in ['S','Q']:
newCmd = cmd.lower()
newData = []
while i < len(data):
newData.append( data[i] - x )
newData.append( data[i+1] - y )
newData.append( data[i+2] - x )
newData.append( data[i+3] - y )
x = data[i+2]
y = data[i+3]
i += 4
elif cmd in ['s','q']:
while i < len(data):
x += data[i+2]
y += data[i+3]
i += 4
elif cmd == 'C':
newCmd = 'c'
newData = []
while i < len(data):
newData.append( data[i] - x )
newData.append( data[i+1] - y )
newData.append( data[i+2] - x )
newData.append( data[i+3] - y )
newData.append( data[i+4] - x )
newData.append( data[i+5] - y )
x = data[i+4]
y = data[i+5]
i += 6
elif cmd == 'c':
while i < len(data):
x += data[i+4]
y += data[i+5]
i += 6
elif cmd in ['z','Z']:
x = startx
y = starty
newCmd = 'z'
newPath.append( (newCmd, newData) )
path = newPath
newPath = [path[0]]
for (cmd,data) in path[1:]:
if cmd in ['m','l','t']:
newData = []
i = 0
while i < len(data):
if data[i] != 0 or data[i+1] != 0:
newData.append(data[i])
newData.append(data[i+1])
else:
numPathSegmentsReduced += 1
i += 2
if newData:
newPath.append( (cmd,newData) )
elif cmd == 'c':
newData = []
i = 0
while i < len(data):
if data[i+4] != 0 or data[i+5] != 0:
newData.append(data[i])
newData.append(data[i+1])
newData.append(data[i+2])
newData.append(data[i+3])
newData.append(data[i+4])
newData.append(data[i+5])
else:
numPathSegmentsReduced += 1
i += 6
if newData:
newPath.append( (cmd,newData) )
elif cmd == 'a':
newData = []
i = 0
while i < len(data):
if data[i+5] != 0 or data[i+6] != 0:
newData.append(data[i])
newData.append(data[i+1])
newData.append(data[i+2])
newData.append(data[i+3])
newData.append(data[i+4])
newData.append(data[i+5])
newData.append(data[i+6])
else:
numPathSegmentsReduced += 1
i += 7
if newData:
newPath.append( (cmd,newData) )
elif cmd == 'q':
newData = []
i = 0
while i < len(data):
if data[i+2] != 0 or data[i+3] != 0:
newData.append(data[i])
newData.append(data[i+1])
newData.append(data[i+2])
newData.append(data[i+3])
else:
numPathSegmentsReduced += 1
i += 4
if newData:
newPath.append( (cmd,newData) )
elif cmd in ['h','v']:
newData = []
i = 0
while i < len(data):
if data[i] != 0:
newData.append(data[i])
else:
numPathSegmentsReduced += 1
i += 1
if newData:
newPath.append( (cmd,newData) )
else:
newPath.append( (cmd,data) )
path = newPath
newPath = [path[0]]
for (cmd,data) in path[1:]:
i = 0
newData = data
if cmd == 'c':
newData = []
while i < len(data):
(p1x,p1y) = (data[i],data[i+1])
(p2x,p2y) = (data[i+2],data[i+3])
dx = data[i+4]
dy = data[i+5]
foundStraightCurve = False
if dx == 0:
if p1x == 0 and p2x == 0:
foundStraightCurve = True
else:
m = dy/dx
if p1y == m*p1x and p2y == m*p2y:
foundStraightCurve = True
if foundStraightCurve:
if newData:
newPath.append( (cmd,newData) )
newData = []
newPath.append( ('l', [dx,dy]) )
numCurvesStraightened += 1
else:
newData.append(data[i])
newData.append(data[i+1])
newData.append(data[i+2])
newData.append(data[i+3])
newData.append(data[i+4])
newData.append(data[i+5])
i += 6
if newData or cmd == 'z' or cmd == 'Z':
newPath.append( (cmd,newData) )
path = newPath
prevCmd = ''
prevData = []
newPath = [path[0]]
for (cmd,data) in path[1:]:
if prevCmd != '':
if cmd != prevCmd or cmd == 'm':
newPath.append( (prevCmd, prevData) )
prevCmd = ''
prevData = []
if cmd == prevCmd and cmd != 'm':
for coord in data:
prevData.append(coord)
else:
prevCmd = cmd
prevData = data
if prevCmd != '':
newPath.append( (prevCmd, prevData) )
path = newPath
newPath = [path[0]]
for (cmd,data) in path[1:]:
if cmd == 'l':
i = 0
lineTuples = []
while i < len(data):
if data[i] == 0:
if lineTuples:
newPath.append( ('l', lineTuples) )
lineTuples = []
newPath.append( ('v', [data[i+1]]) )
numPathSegmentsReduced += 1
elif data[i+1] == 0:
if lineTuples:
newPath.append( ('l', lineTuples) )
lineTuples = []
newPath.append( ('h', [data[i]]) )
numPathSegmentsReduced += 1
else:
lineTuples.append(data[i])
lineTuples.append(data[i+1])
i += 2
if lineTuples:
newPath.append( ('l', lineTuples) )
elif cmd == 'c':
bez_ctl_pt = (0,0)
i = 0
curveTuples = []
while i < len(data):
if bez_ctl_pt[0] == data[i] and bez_ctl_pt[1] == data[i+1]:
if curveTuples:
newPath.append( ('c', curveTuples) )
curveTuples = []
newPath.append( ('s', [data[i+2], data[i+3], data[i+4], data[i+5]]) )
numPathSegmentsReduced += 1
else:
j = 0
while j <= 5:
curveTuples.append(data[i+j])
j += 1
bez_ctl_pt = (data[i+4]-data[i+2], data[i+5]-data[i+3])
i += 6
if curveTuples:
newPath.append( ('c', curveTuples) )
elif cmd == 'q':
quad_ctl_pt = (0,0)
i = 0
curveTuples = []
while i < len(data):
if quad_ctl_pt[0] == data[i] and quad_ctl_pt[1] == data[i+1]:
if curveTuples:
newPath.append( ('q', curveTuples) )
curveTuples = []
newPath.append( ('t', [data[i+2], data[i+3]]) )
numPathSegmentsReduced += 1
else:
j = 0;
while j <= 3:
curveTuples.append(data[i+j])
j += 1
quad_ctl_pt = (data[i+2]-data[i], data[i+3]-data[i+1])
i += 4
if curveTuples:
newPath.append( ('q', curveTuples) )
else:
newPath.append( (cmd, data) )
path = newPath
newPath = [path[0]]
for (cmd,data) in path[1:]:
if cmd in ['h','v'] and len(data) > 1:
newData = []
prevCoord = data[0]
for coord in data[1:]:
if isSameSign(prevCoord, coord):
prevCoord += coord
numPathSegmentsReduced += 1
else:
newData.append(prevCoord)
prevCoord = coord
newData.append(prevCoord)
newPath.append( (cmd, newData) )
else:
newPath.append( (cmd, data) )
path = newPath
prevCmd = ''
prevData = []
newPath = [path[0]]
for (cmd,data) in path[1:]:
if prevCmd != '':
if cmd != prevCmd or cmd == 'm':
newPath.append( (prevCmd, prevData) )
prevCmd = ''
prevData = []
if cmd == prevCmd and cmd != 'm':
for coord in data:
prevData.append(coord)
else:
prevCmd = cmd
prevData = data
if prevCmd != '':
newPath.append( (prevCmd, prevData) )
path = newPath
newPathStr = serializePath(path)
numBytesSavedInPathData += ( len(oldPathStr) - len(newPathStr) )
element.setAttribute('d', newPathStr)
def parseListOfPoints(s):
"""
Parse string into a list of points.
Returns a list of containing an even number of coordinate strings
"""
i = 0
points = []
ws_nums = re.split("\\s*\\,?\\s*", s.strip())
nums = []
for i in range(len(ws_nums)):
negcoords = re.split("\\-", ws_nums[i]);
if len(negcoords) == 1:
nums.append(negcoords[0])
else:
for j in range(len(negcoords)):
if j == 0:
if negcoords[0] != '':
nums.append(negcoords[0])
else:
prev = nums[len(nums)-1]
if prev[len(prev)-1] == 'e' or prev[len(prev)-1] == 'E':
nums[len(nums)-1] = prev + '-' + negcoords[j]
else:
nums.append( '-'+negcoords[j] )
i = 0
while i < len(nums):
x = SVGLength(nums[i])
if i == len(nums)-1: return []
else: y = SVGLength(nums[i+1])
if x.units != Unit.NONE or y.units != Unit.NONE: return []
points.append( str(x.value) )
points.append( str(y.value) )
i += 2
return points
def cleanPolygon(elem):
"""
Remove unnecessary closing point of polygon points attribute
"""
global numPointsRemovedFromPolygon
pts = parseListOfPoints(elem.getAttribute('points'))
N = len(pts)/2
if N >= 2:
(startx,starty) = (pts[0],pts[0])
(endx,endy) = (pts[len(pts)-2],pts[len(pts)-1])
if startx == endx and starty == endy:
pts = pts[:-2]
numPointsRemovedFromPolygon += 1
elem.setAttribute('points', scourCoordinates(pts,True))
def cleanPolyline(elem):
"""
Scour the polyline points attribute
"""
pts = parseListOfPoints(elem.getAttribute('points'))
elem.setAttribute('points', scourCoordinates(pts,True))
def serializePath(pathObj):
"""
Reserializes the path data with some cleanups.
"""
pathStr = ""
for (cmd,data) in pathObj:
pathStr += cmd
pathStr += scourCoordinates(data, (cmd == 'a'))
return pathStr
def scourCoordinates(data, forceCommaWsp = False):
"""
Serializes coordinate data with some cleanups:
- removes all trailing zeros after the decimal
- integerize coordinates if possible
- removes extraneous whitespace
- adds commas between values in a subcommand if required (or if forceCommaWsp is True)
"""
coordsStr = ""
if data != None:
c = 0
for coord in data:
coordsStr += scourLength(coord)
if c < len(data)-1 and (forceCommaWsp or Decimal(data[c+1]) >= 0):
coordsStr += ','
c += 1
return coordsStr
def scourLength(str):
length = SVGLength(str)
coord = length.value
coord = Decimal(unicode(coord)) * Decimal(1)
if int(coord) == coord: coord = Decimal(unicode(int(coord)))
try:
coord = coord.trim()
except AttributeError:
s = unicode(coord)
dec = s.find('.')
if dec != -1:
while s[-1] == '0':
s = s[:-1]
coord = Decimal(s)
normd = coord.normalize()
if len(unicode(normd)) < len(unicode(coord)):
coord = normd
return unicode(coord)+Unit.str(length.units)
def embedRasters(element, options) :
"""
Converts raster references to inline images.
NOTE: there are size limits to base64-encoding handling in browsers
"""
global numRastersEmbedded
href = element.getAttributeNS(NS['XLINK'],'href')
if href != '' and len(href) > 1:
ext = os.path.splitext(os.path.basename(href))[1].lower()[1:]
if ext == 'png' or ext == 'jpg' or ext == 'gif':
if os.path.isfile(href) == False :
if href[:7] != 'http://' and os.path.isfile(href) == False :
infilename = '.'
if options.infilename: infilename = options.infilename
href = os.path.join(os.path.dirname(infilename), href)
rasterdata = ''
if os.path.isfile(href) == True :
raster = open( href, "rb")
rasterdata = raster.read()
elif href[:7] == 'http://':
webFile = urllib.urlopen( href )
rasterdata = webFile.read()
webFile.close()
if rasterdata != '' :
b64eRaster = base64.b64encode( rasterdata )
if b64eRaster != '':
if ext == 'jpg':
ext = 'jpeg'
element.setAttributeNS(NS['XLINK'], 'href', 'data:image/' + ext + ';base64,' + b64eRaster)
numRastersEmbedded += 1
del b64eRaster
def properlySizeDoc(docElement):
w = SVGLength(docElement.getAttribute('width'))
h = SVGLength(docElement.getAttribute('height'))
if ((w.units != Unit.NONE and w.units != Unit.PX) or
(w.units != Unit.NONE and w.units != Unit.PX)):
return
vbSep = re.split("\\s*\\,?\\s*", docElement.getAttribute('viewBox'), 3)
vbWidth,vbHeight = 0,0
if len(vbSep) == 4:
try:
vbX = float(vbSep[0])
vbY = float(vbSep[1])
if vbX != 0 or vbY != 0:
return
vbWidth = float(vbSep[2])
vbHeight = float(vbSep[3])
if vbWidth != w.value or vbHeight != h.value:
return
except ValueError:
pass
docElement.setAttribute('viewBox', '0 0 %s %s' % (w.value, h.value))
docElement.removeAttribute('width')
docElement.removeAttribute('height')
def remapNamespacePrefix(node, oldprefix, newprefix):
if node == None or node.nodeType != 1: return
if node.prefix == oldprefix:
localName = node.localName
namespace = node.namespaceURI
doc = node.ownerDocument
parent = node.parentNode
newNode = None
if newprefix != '':
newNode = doc.createElementNS(namespace, newprefix+":"+localName)
else:
newNode = doc.createElement(localName);
attrList = node.attributes
for i in range(attrList.length):
attr = attrList.item(i)
newNode.setAttributeNS( attr.namespaceURI, attr.localName, attr.nodeValue)
for child in node.childNodes:
newNode.appendChild(child.cloneNode(True))
parent.replaceChild( newNode, node )
node = newNode
for child in node.childNodes :
remapNamespacePrefix(child, oldprefix, newprefix)
def makeWellFormed(str):
xml_ents = { '<':'<', '>':'>', '&':'&', "'":''', '"':'"'}
return ''.join([xml_ents[c] if c in xml_ents else c for c in str])
def serializeXML(element, options, ind = 0, preserveWhitespace = False):
indent = ind
I=''
if options.indent_type == 'tab': I='\t'
elif options.indent_type == 'space': I=' '
outString = (I * ind) + '<' + element.nodeName
if element.getAttribute('id') != '':
id = element.getAttribute('id')
quot = '"'
if id.find('"') != -1:
quot = "'"
outString += ' ' + 'id=' + quot + id + quot
if element.getAttribute('xml:id') != '':
id = element.getAttribute('xml:id')
quot = '"'
if id.find('"') != -1:
quot = "'"
outString += ' ' + 'xml:id=' + quot + id + quot
attrList = element.attributes
for num in range(attrList.length) :
attr = attrList.item(num)
if attr.nodeName == 'id' or attr.nodeName == 'xml:id': continue
quot = '"'
if attr.nodeValue.find('"') != -1:
quot = "'"
attrValue = makeWellFormed( attr.nodeValue )
outString += ' '
if attr.prefix != None:
outString += attr.prefix + ':'
elif attr.namespaceURI != None:
if attr.namespaceURI == 'http://www.w3.org/2000/xmlns/' and attr.nodeName.find('xmlns') == -1:
outString += 'xmlns:'
elif attr.namespaceURI == 'http://www.w3.org/1999/xlink':
outString += 'xlink:'
outString += attr.localName + '=' + quot + attrValue + quot
if attr.nodeName == 'xml:space':
if attrValue == 'preserve':
preserveWhitespace = True
elif attrValue == 'default':
preserveWhitespace = False
children = element.childNodes
if children.length > 0:
outString += '>'
onNewLine = False
for child in element.childNodes:
if child.nodeType == 1:
if preserveWhitespace:
outString += serializeXML(child, options, 0, preserveWhitespace)
else:
outString += os.linesep + serializeXML(child, options, indent + 1, preserveWhitespace)
onNewLine = True
elif child.nodeType == 3:
if preserveWhitespace:
outString += makeWellFormed(child.nodeValue)
else:
outString += makeWellFormed(child.nodeValue.strip())
elif child.nodeType == 4:
outString += '<![CDATA[' + child.nodeValue + ']]>'
elif child.nodeType == 8:
outString += '<!--' + child.nodeValue + '-->'
else:
pass
if onNewLine: outString += (I * ind)
outString += '</' + element.nodeName + '>'
if indent > 0: outString += os.linesep
else:
outString += '/>'
if indent > 0: outString += os.linesep
return outString
def scourString(in_string, options=None):
if options is None:
options = _options_parser.get_default_values()
getcontext().prec = options.digits
global numAttrsRemoved
global numStylePropsFixed
global numElemsRemoved
global numBytesSavedInColors
doc = xml.dom.minidom.parseString(in_string)
if options.keep_editor_data == False:
while removeNamespacedElements( doc.documentElement, unwanted_ns ) > 0 :
pass
while removeNamespacedAttributes( doc.documentElement, unwanted_ns ) > 0 :
pass
xmlnsDeclsToRemove = []
attrList = doc.documentElement.attributes
for num in range(attrList.length) :
if attrList.item(num).nodeValue in unwanted_ns :
xmlnsDeclsToRemove.append(attrList.item(num).nodeName)
for attr in xmlnsDeclsToRemove :
doc.documentElement.removeAttribute(attr)
numAttrsRemoved += 1
if doc.documentElement.getAttribute('xmlns') != 'http://www.w3.org/2000/svg':
doc.documentElement.setAttribute('xmlns', 'http://www.w3.org/2000/svg')
attrList = doc.documentElement.attributes
xmlnsDeclsToRemove = []
redundantPrefixes = []
for i in range(attrList.length):
attr = attrList.item(i)
name = attr.nodeName
val = attr.nodeValue
if name[0:6] == 'xmlns:' and val == 'http://www.w3.org/2000/svg':
redundantPrefixes.append(name[6:])
xmlnsDeclsToRemove.append(name)
for attrName in xmlnsDeclsToRemove:
doc.documentElement.removeAttribute(attrName)
for prefix in redundantPrefixes:
remapNamespacePrefix(doc.documentElement, prefix, '')
numStylePropsFixed = repairStyle(doc.documentElement, options)
if options.simple_colors:
numBytesSavedInColors = convertColors(doc.documentElement)
elemsToRemove = []
for tag in ['defs', 'metadata', 'g'] :
for elem in doc.documentElement.getElementsByTagName(tag) :
removeElem = not elem.hasChildNodes()
if removeElem == False :
for child in elem.childNodes :
if child.nodeType in [1, 3, 4, 8] :
break
else:
removeElem = True
if removeElem :
elem.parentNode.removeChild(elem)
numElemsRemoved += 1
while removeUnreferencedElements(doc) > 0:
pass
if options.strip_ids:
bContinueLooping = True
while bContinueLooping:
identifiedElements = findElementsWithId(doc.documentElement)
referencedIDs = findReferencedElements(doc.documentElement)
bContinueLooping = (removeUnreferencedIDs(referencedIDs, identifiedElements) > 0)
if options.group_collapse:
while removeNestedGroups(doc.documentElement) > 0:
pass
while removeDuplicateGradientStops(doc) > 0:
pass
while collapseSinglyReferencedGradients(doc) > 0:
pass
while removeDuplicateGradients(doc) > 0:
pass
numAttrsRemoved += moveCommonAttributesToParentGroup(doc.documentElement)
numAttrsRemoved += removeUnusedAttributesOnParent(doc.documentElement)
for elem in doc.documentElement.getElementsByTagName('path') :
if elem.getAttribute('d') == '':
elem.parentNode.removeChild(elem)
else:
cleanPath(elem)
for polygon in doc.documentElement.getElementsByTagName('polygon') :
cleanPolygon(polygon)
for polyline in doc.documentElement.getElementsByTagName('polyline') :
cleanPolygon(polyline)
for type in ['svg', 'image', 'rect', 'circle', 'ellipse', 'line', 'linearGradient', 'radialGradient', 'stop']:
for elem in doc.getElementsByTagName(type):
for attr in ['x', 'y', 'width', 'height', 'cx', 'cy', 'r', 'rx', 'ry',
'x1', 'y1', 'x2', 'y2', 'fx', 'fy', 'offset', 'opacity',
'fill-opacity', 'stroke-opacity', 'stroke-width', 'stroke-miterlimit']:
if elem.getAttribute(attr) != '':
elem.setAttribute(attr, scourLength(elem.getAttribute(attr)))
numAttrsRemoved += removeDefaultAttributeValues(doc.documentElement, options)
if options.embed_rasters:
for elem in doc.documentElement.getElementsByTagName('image') :
embedRasters(elem, options)
if options.enable_viewboxing:
properlySizeDoc(doc.documentElement)
out_string = serializeXML(doc.documentElement, options) + os.linesep
lines = []
for line in out_string.splitlines(True):
if line.strip():
lines.append(line)
if options.strip_xml_prolog == False:
total_output = '<?xml version="1.0" encoding="UTF-8" standalone="no"?>' + os.linesep
else:
total_output = ""
for child in doc.childNodes:
if child.nodeType == 1:
total_output += "".join(lines)
else:
total_output += child.toxml() + os.linesep
return total_output
def scourXmlFile(filename, options=None):
in_string = open(filename).read()
out_string = scourString(in_string, options)
return xml.dom.minidom.parseString(out_string.encode('utf-8'))
02349 class HeaderedFormatter(optparse.IndentedHelpFormatter):
"""
Show application name, version number, and copyright statement
above usage information.
"""
def format_usage(self, usage):
return "%s %s\n%s\n%s" % (APP, VER, COPYRIGHT,
optparse.IndentedHelpFormatter.format_usage(self, usage))
_options_parser = optparse.OptionParser(
usage="%prog [-i input.svg] [-o output.svg] [OPTIONS]",
description=("If the input/output files are specified with a svgz"
" extension, then compressed SVG is assumed. If the input file is not"
" specified, stdin is used. If the output file is not specified, "
" stdout is used."),
formatter=HeaderedFormatter(max_help_position=30),
version=VER)
_options_parser.add_option("--disable-simplify-colors",
action="store_false", dest="simple_colors", default=True,
help="won't convert all colors to #RRGGBB format")
_options_parser.add_option("--disable-style-to-xml",
action="store_false", dest="style_to_xml", default=True,
help="won't convert styles into XML attributes")
_options_parser.add_option("--disable-group-collapsing",
action="store_false", dest="group_collapse", default=True,
help="won't collapse <g> elements")
_options_parser.add_option("--enable-id-stripping",
action="store_true", dest="strip_ids", default=False,
help="remove all un-referenced ID attributes")
_options_parser.add_option("--disable-embed-rasters",
action="store_false", dest="embed_rasters", default=True,
help="won't embed rasters as base64-encoded data")
_options_parser.add_option("--keep-editor-data",
action="store_true", dest="keep_editor_data", default=False,
help="won't remove Inkscape, Sodipodi or Adobe Illustrator elements and attributes")
_options_parser.add_option("--strip-xml-prolog",
action="store_true", dest="strip_xml_prolog", default=False,
help="won't output the <?xml ?> prolog")
_options_parser.add_option("--enable-viewboxing",
action="store_true", dest="enable_viewboxing", default=False,
help="changes document width/height to 100%/100% and creates viewbox coordinates")
_options_parser.add_option("-p", "--set-precision",
action="store", type=int, dest="digits", default=5,
help="set number of significant digits (default: %default)")
_options_parser.add_option("-i",
action="store", dest="infilename", help=optparse.SUPPRESS_HELP)
_options_parser.add_option("-o",
action="store", dest="outfilename", help=optparse.SUPPRESS_HELP)
_options_parser.add_option("--indent",
action="store", type="string", dest="indent_type", default="space",
help="indentation of the output: none, space, tab (default: %default)")
def maybe_gziped_file(filename, mode="r"):
if os.path.splitext(filename)[1].lower() in (".svgz", ".gz"):
return gzip.GzipFile(filename, mode)
return file(filename, mode)
def parse_args(args=None):
options, rargs = _options_parser.parse_args(args)
if rargs:
_options_parser.error("Additional arguments not handled: %r, see --help" % rargs)
if options.digits < 0:
_options_parser.error("Can't have negative significant digits, see --help")
if not options.indent_type in ["tab", "space", "none"]:
_options_parser.error("Invalid value for --indent, see --help")
if options.infilename and options.outfilename and options.infilename == options.outfilename:
_options_parser.error("Input filename is the same as output filename")
if options.infilename:
infile = maybe_gziped_file(options.infilename)
else:
infile = sys.stdin
if options.outfilename:
outfile = maybe_gziped_file(options.outfilename, "w")
else:
outfile = sys.stdout
return options, [infile, outfile]
def getReport():
return ' Number of elements removed: ' + str(numElemsRemoved) + os.linesep + \
' Number of attributes removed: ' + str(numAttrsRemoved) + os.linesep + \
' Number of unreferenced id attributes removed: ' + str(numIDsRemoved) + os.linesep + \
' Number of style properties fixed: ' + str(numStylePropsFixed) + os.linesep + \
' Number of raster images embedded inline: ' + str(numRastersEmbedded) + os.linesep + \
' Number of path segments reduced/removed: ' + str(numPathSegmentsReduced) + os.linesep + \
' Number of bytes saved in path data: ' + str(numBytesSavedInPathData) + os.linesep + \
' Number of bytes saved in colors: ' + str(numBytesSavedInColors) + os.linesep + \
' Number of points removed from polygons: ' + str(numPointsRemovedFromPolygon)
if __name__ == '__main__':
if sys.platform == "win32":
from time import clock as get_tick
else:
def get_tick():
return os.times()[0]
start = get_tick()
options, (input, output) = parse_args()
print >>sys.stderr, "%s %s\n%s" % (APP, VER, COPYRIGHT)
in_string = input.read()
out_string = scourString(in_string, options).encode("UTF-8")
output.write(out_string)
input.close()
output.close()
end = get_tick()
print >>sys.stderr, ' File:', input.name, \
os.linesep + ' Time taken:', str(end-start) + 's' + os.linesep, \
getReport()
oldsize = len(in_string)
newsize = len(out_string)
sizediff = (newsize / oldsize) * 100
print >>sys.stderr, ' Original file size:', oldsize, 'bytes;', \
'new file size:', newsize, 'bytes (' + str(sizediff)[:5] + '%)'