def __setitem__(self, item, value):
if isinstance(value,strTypes):
# very simple case, no defaults, no empty case
value = MapNode(asUnicode(value))
if isinstance(item,strTypes):
item = asUnicode(item)
self.nodemappers[item] = value
def test17(self):
self.assertEqual(asUnicode(u'abc'),u'abc')
self.assertEqual(asUnicode(b'abc'),u'abc')
self.assertRaises(AttributeError,asUnicode,['abc'])
self.myAssertRaisesRegex(AttributeError,r"asUnicode\(.*'list' object has no attribute 'decode'", asUnicode,['abc'])
self.assertEqual(asUnicodeEx(u'abc'),u'abc')
self.assertEqual(asUnicodeEx(b'abc'),u'abc')
self.assertEqual(asUnicodeEx(123),u'123')
self.assertEqual(asBytes(u'abc'),b'abc')
self.assertEqual(asBytes(b'abc'),b'abc')
self.assertRaises(AttributeError,asBytes,['abc'])
self.myAssertRaisesRegex(AttributeError,"asBytes\(.*'list' object has no attribute 'encode'", asBytes,['abc'])
def draw(self):
# general widget bits
s = float(self.size) # abbreviate as we will use this a lot
g = shapes.Group()
ig = self.innerGap
x = self.x + self.dx
y = self.y + self.dy
hsize = 0.5 * self.size
if not ig:
L = [(x - hsize, y, x + hsize, y), (x, y - hsize, x, y + hsize)]
else:
if isStr(ig):
ig = asUnicode(ig)
if ig.endswith(u'%'):
gs = hsize * float(ig[:-1]) / 100.0
else:
gs = float(ig) * 0.5
else:
gs = ig * 0.5
L = [(x - hsize, y, x - gs, y), (x + gs, y, x + hsize, y),
(x, y - hsize, x, y - gs), (x, y + gs, x, y + hsize)]
P = shapes.Path(strokeWidth=self.strokeWidth,
strokeColor=self.strokeColor)
for x0, y0, x1, y1 in L:
P.moveTo(x0, y0)
P.lineTo(x1, y1)
g.add(P)
return g
def addParas(words):
words = [asUnicode(w) for w in words]
txt = u' '.join([(len(w) > 5 and u'<index item=%s/>%s' %
(quoteattr(commajoin([w[:2], w[:3], w])), w)
or w) for w in words])
para = Paragraph(txt, makeBodyStyle())
story.append(para)
def xhtmlDocFromXhtmlFragment(xhtmlFrag,enc='utf8'):
r = []
#r.append('<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">')
r.append('<html><head><title></title></head><body>')
xhtmlFrag=xhtmlFrag.replace('<p><table>','<table>').replace('</table></p>','</table>')
r.append(asUnicode(xhtmlFrag,enc=enc))
r.append('</body></html>')
r = ''.join(r)
return r
def findImages(xml):
"Returns lists of all images referred to in markup"
xml = asUnicode(xml)
validDoc = xhtmlDocFromXhtmlFragment(xml) #adds 'html','head etc. to our fragment
parser = pyRXPU.Parser(ReturnUTF8=1, ExpandCharacterEntities=0, ExpandGeneralEntities=0)
tree = parser.parse(validDoc)
walker = ImageFinder(tree)
walker.go()
return walker.images
def _build(self,availWidth,availHeight):
_tempEntries = [(tuple(asUnicode(t) for t in texts),pageNumbers)
for texts, pageNumbers in self._getlastEntries()]
def getkey(seq):
return [''.join((c for c in unicodedata.normalize('NFD', x.upper()) if unicodedata.category(c) != 'Mn')) for x in seq[0]]
_tempEntries.sort(key=getkey)
leveloffset = self.headers and 1 or 0
def drawIndexEntryEnd(canvas, kind, label):
'''Callback to draw dots and page numbers after each entry.'''
style = self.getLevelStyle(leveloffset)
pages = [(p[1],k) for p,k in sorted(decode_label(label))]
drawPageNumbers(canvas, style, pages, availWidth, availHeight, self.dot)
self.canv.drawIndexEntryEnd = drawIndexEntryEnd
alpha = ''
tableData = []
lastTexts = []
alphaStyle = self.getLevelStyle(0)
for texts, pageNumbers in _tempEntries:
texts = list(texts)
#track when the first character changes; either output some extra
#space, or the first letter on a row of its own. We cannot do
#widow/orphan control, sadly.
nalpha = ''.join((c for c in unicodedata.normalize('NFD', texts[0][0].upper()) if unicodedata.category(c) != 'Mn'))
if alpha != nalpha:
alpha = nalpha
if self.headers:
header = alpha
else:
header = ' '
tableData.append([Spacer(1, alphaStyle.spaceBefore),])
tableData.append([Paragraph(header, alphaStyle),])
tableData.append([Spacer(1, alphaStyle.spaceAfter),])
i, diff = listdiff(lastTexts, texts)
if diff:
lastTexts = texts
texts = texts[i:]
label = encode_label(list(pageNumbers))
texts[-1] = '%s<onDraw name="drawIndexEntryEnd" label="%s"/>' % (texts[-1], label)
for text in texts:
#Platypus and RML differ on how parsed XML attributes are escaped.
#e.g. <index item="M&S"/>. The only place this seems to bite us is in
#the index entries so work around it here.
text = escapeOnce(text)
style = self.getLevelStyle(i+leveloffset)
para = Paragraph(text, style)
if style.spaceBefore:
tableData.append([Spacer(1, style.spaceBefore),])
tableData.append([para,])
i += 1
self._flowable = Table(tableData, colWidths=[availWidth], style=self.tableStyle)
def _build(self,availWidth,availHeight):
_tempEntries = [(tuple(asUnicode(t) for t in texts),pageNumbers)
for texts, pageNumbers in self._getlastEntries()]
def getkey(seq):
return [''.join((c for c in unicodedata.normalize('NFD', x.upper()) if unicodedata.category(c) != 'Mn')) for x in seq[0]]
_tempEntries.sort(key=getkey)
leveloffset = self.headers and 1 or 0
def drawIndexEntryEnd(canvas, kind, label):
'''Callback to draw dots and page numbers after each entry.'''
style = self.getLevelStyle(leveloffset)
pages = [(p[1],k) for p,k in sorted(decode_label(label))]
drawPageNumbers(canvas, style, pages, availWidth, availHeight, self.dot)
self.canv.drawIndexEntryEnd = drawIndexEntryEnd
alpha = ''
tableData = []
lastTexts = []
alphaStyle = self.getLevelStyle(0)
for texts, pageNumbers in _tempEntries:
texts = list(texts)
#track when the first character changes; either output some extra
#space, or the first letter on a row of its own. We cannot do
#widow/orphan control, sadly.
nalpha = ''.join((c for c in unicodedata.normalize('NFD', texts[0][0].upper()) if unicodedata.category(c) != 'Mn'))
if alpha != nalpha:
alpha = nalpha
if self.headers:
header = alpha
else:
header = ' '
tableData.append([Spacer(1, alphaStyle.spaceBefore),])
tableData.append([Paragraph(header, alphaStyle),])
tableData.append([Spacer(1, alphaStyle.spaceAfter),])
i, diff = listdiff(lastTexts, texts)
if diff:
lastTexts = texts
texts = texts[i:]
label = encode_label(list(pageNumbers))
texts[-1] = '%s<onDraw name="drawIndexEntryEnd" label="%s"/>' % (texts[-1], label)
for text in texts:
#Platypus and RML differ on how parsed XML attributes are escaped.
#e.g. <index item="M&S"/>. The only place this seems to bite us is in
#the index entries so work around it here.
text = escapeOnce(text)
style = self.getLevelStyle(i+leveloffset)
para = Paragraph(text, style)
if style.spaceBefore:
tableData.append([Spacer(1, style.spaceBefore),])
tableData.append([para,])
i += 1
self._flowable = Table(tableData, colWidths=[availWidth], style=self.tableStyle)
def parse(self, text, style):
"attempt replacement for parse"
self._setup_for_parse(style)
text = asUnicode(text)
if not(len(text)>=6 and text[0]=='<' and _re_para.match(text)):
text = u"<para>"+text+u"</para>"
try:
self.feed(text)
except:
annotateException('paragraph text %s caused exception' % ascii(text))
return self._complete_parse()
def processContent(self, elts, overrides=None, containerdict=None, top_level=1):
if containerdict is None:
containerdict = self.topDict
if overrides is None:
overrides = {} # no overrides yet
if isinstance(elts,strTypes):
return self.processString(asUnicode(elts))
if isinstance(elts,tuple):
return self.processTuple(elts, overrides, containerdict)
else:
L = []
for e in elts:
if isinstance(e,strTypes):
e2 = self.processString(asUnicode(e))
else:
if not isinstance(e,tuple):
raise ValueError("bad content type %s" % type(e))
e2 = self.processTuple(e, overrides, containerdict)
L.append(e2)
return self.joinTranslatedContent(L)
def xhtmlDocFromXhtmlFragment(xhtmlFrag, enc='utf8'):
r = []
#r.append('<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">')
r.append('<html><head><title></title></head><body>')
xhtmlFrag = xhtmlFrag.replace('<p><table>',
'<table>').replace('</table></p>',
'</table>')
r.append(asUnicode(xhtmlFrag, enc=enc))
r.append('</body></html>')
r = ''.join(r)
return r
def text_decode(source):
"""
try decoding ``source`` with various known codepoints to unicode
for more information on the various codepoints, see
https://docs.python.org/2/library/codecs.html#standard-encodings
:see: https://github.com/prjemian/assign_gup/issues/55
:see: http://stackoverflow.com/questions/9942594/unicodeencodeerror-ascii-codec-cant-encode-character-u-xa0-in-position-20
"""
return asUnicode(source)
def parse(self, text, style):
"attempt replacement for parse"
self._setup_for_parse(style)
text = asUnicode(text)
if not(len(text)>=6 and text[0]=='<' and _re_para.match(text)):
text = u"<para>"+text+u"</para>"
try:
self.feed(text)
except:
annotateException('paragraph text %s caused exception' % ascii(text))
return self._complete_parse()
def run(pagesize=None, verbose=0, outDir=None):
import sys, os
from reportlab.lib.utils import open_and_read, asUnicode
cwd = os.getcwd()
docsDir = os.path.dirname(os.path.dirname(sys.argv[0]) or cwd)
topDir = os.path.dirname(docsDir)
if not outDir: outDir = docsDir
G = {}
sys.path.insert(0, topDir)
from reportlab.pdfbase.pdfmetrics import registerFontFamily
from reportlab.pdfbase import pdfmetrics
from reportlab.pdfbase.ttfonts import TTFont
pdfmetrics.registerFont(TTFont('Vera', 'Vera.ttf'))
pdfmetrics.registerFont(TTFont('VeraBd', 'VeraBd.ttf'))
pdfmetrics.registerFont(TTFont('VeraIt', 'VeraIt.ttf'))
pdfmetrics.registerFont(TTFont('VeraBI', 'VeraBI.ttf'))
registerFontFamily('Vera',
normal='Vera',
bold='VeraBd',
italic='VeraIt',
boldItalic='VeraBI')
from tools.docco.rl_doc_utils import setStory, getStory, RLDocTemplate, defaultPageSize, H1, H2, H3, H4
from tools.docco import rl_doc_utils
exec('from tools.docco.rl_doc_utils import *', G, G)
destfn = os.path.join(outDir, 'reportlab-userguide.pdf')
doc = RLDocTemplate(destfn, pagesize=pagesize or defaultPageSize)
#this builds the story
setStory()
for f in (
'ch1_intro',
'ch2_graphics',
'ch2a_fonts',
'ch3_pdffeatures',
'ch4_platypus_concepts',
'ch5_paragraphs',
'ch6_tables',
'ch7_custom',
'graph_intro',
'graph_concepts',
'graph_charts',
'graph_shapes',
'graph_widgets',
'app_demos',
):
#python source is supposed to be utf8 these days
exec(asUnicode(open_and_read(f + '.py')), G, G)
del G
story = getStory()
if verbose: print('Built story contains %d flowables...' % len(story))
doc.multiBuild(story)
if verbose: print('Saved "%s"' % destfn)
def findImages(xml):
"Returns lists of all images referred to in markup"
xml = asUnicode(xml)
validDoc = xhtmlDocFromXhtmlFragment(
xml) #adds 'html','head etc. to our fragment
parser = pyRXPU.Parser(ReturnUTF8=1,
ExpandCharacterEntities=0,
ExpandGeneralEntities=0)
tree = parser.parse(validDoc)
walker = ImageFinder(tree)
walker.go()
return walker.images
def run(pagesize=None, verbose=0, outDir=None):
import sys,os
from reportlab.lib.utils import open_and_read, asUnicode
cwd = os.getcwd()
docsDir=os.path.dirname(os.path.dirname(sys.argv[0]) or cwd)
topDir=os.path.dirname(docsDir)
if not outDir: outDir=docsDir
G = {}
sys.path.insert(0,topDir)
from reportlab.pdfbase.pdfmetrics import registerFontFamily
from reportlab.pdfbase import pdfmetrics
from reportlab.pdfbase.ttfonts import TTFont
pdfmetrics.registerFont(TTFont('Vera', 'Vera.ttf'))
pdfmetrics.registerFont(TTFont('VeraBd', 'VeraBd.ttf'))
pdfmetrics.registerFont(TTFont('VeraIt', 'VeraIt.ttf'))
pdfmetrics.registerFont(TTFont('VeraBI', 'VeraBI.ttf'))
registerFontFamily('Vera',normal='Vera',bold='VeraBd',italic='VeraIt',boldItalic='VeraBI')
from tools.docco.rl_doc_utils import setStory, getStory, RLDocTemplate, defaultPageSize, H1, H2, H3, H4
from tools.docco import rl_doc_utils
exec('from tools.docco.rl_doc_utils import *', G, G)
destfn = os.path.join(outDir,'reportlab-userguide.pdf')
doc = RLDocTemplate(destfn,pagesize = pagesize or defaultPageSize)
#this builds the story
setStory()
for f in (
'ch1_intro',
'ch2_graphics',
'ch2a_fonts',
'ch3_pdffeatures',
'ch4_platypus_concepts',
'ch5_paragraphs',
'ch6_tables',
'ch7_custom',
'graph_intro',
'graph_concepts',
'graph_charts',
'graph_shapes',
'graph_widgets',
'app_demos',
):
#python source is supposed to be utf8 these days
exec(asUnicode(open_and_read(f+'.py')), G, G)
del G
story = getStory()
if verbose: print('Built story contains %d flowables...' % len(story))
doc.multiBuild(story)
if verbose: print('Saved "%s"' % destfn)
def normalizeName(self, name):
if not isUnicode(name):
for enc in ('utf8', 'latin1'):
try:
name = asUnicode(name, enc)
break
except:
pass
else:
raise ValueError('Cannot normalize name %r' % name)
r = name.strip().lower()
nns = getattr(self, 'normalizeNameSpaces', None)
if isStr(nns):
r = nns.join(filter(None, r.split()))
return r
def drawString(self, x, y, s, angle=0):
if self._fillColor != None:
fontSize = self._fontSize
fontObj = getFont(self._font)
dynamicFont = fontObj._dynamicFont
embedding = self._ttf_embed and dynamicFont
if not embedding and not self.code[self._fontCodeLoc]:
psName = fontObj.face.name
self.code[
self.
_fontCodeLoc] = '(%s) findfont %s scalefont setfont' % (
psName, fp_str(fontSize))
if psName not in self._fontsUsed:
self._fontsUsed.append(psName)
self.setColor(self._fillColor)
if angle != 0:
self.code_append('gsave %s translate %s rotate' %
(fp_str(x, y), fp_str(angle)))
x = y = 0
if embedding:
i = 0
s = asUnicode(s)
for subset, t in fontObj.splitString(s, self):
if subset != self._curSubset:
psName = fontObj.getSubsetInternalName(subset,
self)[1:]
sf = '(%s) findfont %s scalefont setfont' % (
psName, fp_str(fontSize))
if not self.code[self._fontCodeLoc]:
self.code[self._fontCodeLoc] = sf
else:
self.code_append(sf)
self._curSubset = subset
self.code_append('%s m (%s) show ' %
(fp_str(x, y), self._escape(t)))
j = i + len(t)
x += fontObj.stringWidth(s[i:j], fontSize)
i = j
elif dynamicFont:
s = self._escape(s)
self.code_append('%s m (%s) show ' % (fp_str(x, y), s))
else:
self._issueT1String(fontObj, x, y, s)
if angle != 0:
self.code_append('grestore')
def textLines(self, stuff, trim=1):
"""prints multi-line or newlined strings, moving down. One
comon use is to quote a multi-line block in your Python code;
since this may be indented, by default it trims whitespace
off each line and from the beginning; set trim=0 to preserve
whitespace."""
if isStr(stuff):
lines = asUnicode(stuff).strip().split(u'\n')
if trim == 1:
lines = [s.strip() for s in lines]
elif isinstance(stuff, (tuple, list)):
lines = stuff
else:
assert 1 == 0, "argument to textlines must be string,, list or tuple"
# Output each line one at a time. This used to be a long-hand
# copy of the textLine code, now called as a method.
for line in lines:
self.textLine(line)
def textLines(self, stuff, trim=1):
"""prints multi-line or newlined strings, moving down. One
comon use is to quote a multi-line block in your Python code;
since this may be indented, by default it trims whitespace
off each line and from the beginning; set trim=0 to preserve
whitespace."""
if isStr(stuff):
lines = asUnicode(stuff).strip().split(u'\n')
if trim==1:
lines = [s.strip() for s in lines]
elif isinstance(stuff,(tuple,list)):
lines = stuff
else:
assert 1==0, "argument to textlines must be string,, list or tuple"
# Output each line one at a time. This used to be a long-hand
# copy of the textLine code, now called as a method.
for line in lines:
self.textLine(line)
def drawString(self, x, y, s, angle=0):
if self._fillColor != None:
fontSize = self._fontSize
fontObj = getFont(self._font)
dynamicFont = fontObj._dynamicFont
embedding = self._ttf_embed and dynamicFont
if not embedding and not self.code[self._fontCodeLoc]:
psName = fontObj.face.name
self.code[self._fontCodeLoc]='(%s) findfont %s scalefont setfont' % (psName,fp_str(fontSize))
if psName not in self._fontsUsed:
self._fontsUsed.append(psName)
self.setColor(self._fillColor)
if angle!=0:
self.code_append('gsave %s translate %s rotate' % (fp_str(x,y),fp_str(angle)))
x = y = 0
if embedding:
i = 0
s = asUnicode(s)
for subset, t in fontObj.splitString(s, self):
if subset!=self._curSubset:
psName = fontObj.getSubsetInternalName(subset, self)[1:]
sf = '(%s) findfont %s scalefont setfont' % (psName,fp_str(fontSize))
if not self.code[self._fontCodeLoc]:
self.code[self._fontCodeLoc] = sf
else:
self.code_append(sf)
self._curSubset = subset
self.code_append('%s m (%s) show ' % (fp_str(x,y),self._escape(t)))
j = i + len(t)
x += fontObj.stringWidth(s[i:j],fontSize)
i = j
elif dynamicFont:
s = self._escape(s)
self.code_append('%s m (%s) show ' % (fp_str(x,y),s))
else:
self._issueT1String(fontObj,x,y,s)
if angle!=0:
self.code_append('grestore')
def __init__(self,
target="block",
breaksAllowed=True,
stripComments=False,
stripUnknownEntities=True,
allowImages=True,
allowTables=True,
allowAtags=True,
allowStyleAttr=True, #passed through to para and headings
allowAlignAttr=True, #passed through to para and headings
aHrefTr=None,
imgSrcTr=None,
substitutions=[],
maxLines=None,
lineWidth=40,
entities = known_entities.keys(),
encoding = None,
):
"""Initialising defines your language options.
You can re-use the same parser many times.
if breaksAllowed, they will be written to output.
if not, in inline mode they vanish, and in block
mode they end the block.
substitutions is a a singleton or list containing
pat pat --> ''
(pat,str) pat --> str
callable c(src) --> src
pat may be a str or a compiled pattern. These substitutions
are done before parsing.
"""
target = self.asUnicode(target)
self.stripUnknownEntities = stripUnknownEntities
self.allowImages = allowImages
self.allowTables = allowTables
self.allowAtags = allowAtags
self.aHrefTr = aHrefTr
self.imgSrcTr = imgSrcTr
self.encoding = encoding
self.allowAlignAttr = allowAlignAttr
self.allowStyleAttr = allowStyleAttr
self._setupGrammar()
assert target in (u"block", u"inline"), "unexpected block '%s', must be 'block' or 'inline'" % target
self.target = target
HTMLParser.__init__(self)
self.breaksAllowed = breaksAllowed
self.stripComments = stripComments
self.entities = set(entities).union(('lt', 'gt', 'amp'))
#prefix up the substitutions list
if not isinstance(substitutions,(list,tuple)):
substitutions = (substitutions,)
S=[].append
for s in substitutions:
if isinstance(s,strTypes):
s = lambda x,pat=re.compile(asUnicode(s)): pat.sub('',x)
elif hasattr(s,'sub'):
s = lambda x,pat=s: pat.sub('',x)
elif isinstance(s,(tuple,list)) and len(s)==2:
p=s[0]
if isinstance(p,str):
s = lambda x,pat=re.compile(p),s=s[1]: pat.sub(s,x)
elif hasattr(p,'sub'):
s = lambda x,pat=p,s=s[1]: pat.sub(s,x)
else:
raise ValueError('Invalid value %r in substitions list' % s)
elif not isinstance(s, collections.Callable):
raise ValueError('Invalid value %r in substitions list' % s)
S(s)
self.substitutions = S.__self__
self.remainingLines = maxLines
self.lineWidth = lineWidth
self.textLength = 0
def equalStrings(a,b,enc='utf8'):
return a==b if type(a)==type(b) else asUnicode(a,enc)==asUnicode(b,enc)
def addParas(words):
words = [asUnicode(w) for w in words]
txt = u' '.join([(len(w) > 5 and u'<index item=%s/>%s' % (quoteattr(commajoin([w[:2], w[:3], w])), w) or w) for w in words])
para = Paragraph(txt, makeBodyStyle())
story.append(para)
def bullet(text):
text = u'<bullet><font name="Symbol">\u2022</font></bullet>' + asUnicode(quickfix(text))
P = Paragraph(text, BU)
getStory().append(P)
def uniDict(d):
r = {}
for k in d:
r[asUnicode(k)] = d[k]
return r
def _processLine(line, sep=',', conv=0):
if isUnicode(line):
space = u' '
dquot = u'"'
empty = u''
speol = u' \r\n'
sep = asUnicode(sep)
else:
space = b' '
dquot = b'"'
empty = b''
speol = b' \r\n'
sep = asBytes(sep)
fields = []
p = 0
ll = len(line)
ls = len(sep)
line += space
while (ll > 0 and (line[ll-1] in speol)): ll -= 1
while p < ll:
#Skip unquoted space at the start of a field
while p<ll and line[p]==space: p += 1
field = empty
ql = 0
while p < ll:
#Skip unquoted space at the end of a field
if ql == 0 and line[p] == space:
q = p
while q < ll and line[q] == space:
q += 1
if q >= ll:
break
elif line[q:q+ls] == sep:
p = q
if ql == 0 and line[p:p+ls] == sep:
break
elif line[p:p+1] == dquot:
if ql == 0:
ql = 1
elif line[p+1:p+2]==dquot:
field += dquot
p += 1
else:
ql = 0
else:
field += line[p:p+1]
p += 1
p += ls
if conv:
try:
fields.append(int(field))
except ValueError:
try:
fields.append(float(field))
except ValueError:
fields.append(field)
else:
fields.append(field)
if line[ll-ls:ll]==sep:
fields.append(empty) #extra field when there's a separator at the end
return fields
def __dict_replace(s, d):
# This is grabbed from xml.sax.saxutils
"""Replace substrings of a string using a dictionary."""
for key, value in list(d.items()):
s = s.replace(asUnicode(key), asUnicode(value))
return s
def bullet(text):
text = u'<bullet><font name="Symbol">\u2022</font></bullet>' + asUnicode(
quickfix(text))
P = Paragraph(text, BU)
getStory().append(P)
def equalStrings(a, b, enc='utf8'):
return a == b if type(a) == type(b) else asUnicode(a, enc) == asUnicode(
b, enc)
def __dict_replace(s, d):
# This is grabbed from xml.sax.saxutils
"""Replace substrings of a string using a dictionary."""
for key, value in list(d.items()):
s = s.replace(asUnicode(key), asUnicode(value))
return s