def traverse(iNode):
if (iNode.tag == 'document') or ((iNode.tag == 'content') and (iNode.getparent().tag == 'document')) or ((iNode.tag == 'section') and (iNode.getparent().tag != 'activity')):
children = iNode.getchildren()
for child in children:
if not traverse(child):
iNode.remove(child)
return True
elif (iNode.tag == 'title') and (iNode.getparent().tag == 'section'):
return True
elif iNode.tag == 'exercises':
for correctNode in iNode.xpath('.//correct'):
fullDomPath = get_full_dom_path(correctNode)
if 'solution' in fullDomPath:
correctNode.tail = ''
entryNode = correctNode.getparent()
while entryNode.tag != 'entry':
entryNode = entryNode.getparent()
if 'latex' in fullDomPath:
remove_ampersands(correctNode)
correctNode.tag = 'latex'
entryNode.append(etree.Element('correct'))
entryNode[-1].append(correctNode)
else:
entryNode.append(correctNode)
for tag in ['header', 'footer', 'shortcode']:
for child in iNode.xpath('.//' + tag):
child.getparent().remove(child)
for tag in ['problem', 'solution']:
for child in iNode.xpath('.//' + tag):
child.clear()
for child in iNode.xpath('.//problem-set'):
if child.attrib.get('{http://siyavula.com/cnxml/style/0.1}columns') is not None:
del child.attrib['{http://siyavula.com/cnxml/style/0.1}columns']
return True
else:
if len(iNode.xpath('.//exercises')) > 0:
sys.stderr.write('WARNING: Deleting node that contains exercises.\n')
sys.stderr.write(etree.tostring(iNode) + '\n')
return False
def __get_full_dom_path(self, iNode):
return utils.get_full_dom_path(iNode, iSpec=self.spec)
def find_substitutions(iText, iNode=None, iPreTailTag=None, iPostTailTag=None):
global numberAndCurrencyPattern, lazyMode
oSubstitutions = []
for match in numberAndCurrencyPattern.finditer(iText):
context = [None, None]
contextLength = 20
preContextWarning = 0
postContextWarning = 0
start, stop = match.span()
oldNumber = iText[start:stop]
if oldNumber[0] == 'R':
# Currency
newNode = etree.Element('currency')
newNode.append(etree.Element('number'))
newNode[0].text = oldNumber[1:].replace(' ','').replace(',','.')
# Check whether we need to use a non-default precision
pos = newNode[0].text.find('.')
if (pos != -1) and (pos < len(newNode[0].text)-3):
newNode.attrib['precision'] = str(len(newNode[0].text) - pos - 1)
# Check if the pre-context is funny
while (preContextWarning < contextLength) and (start-preContextWarning > 0) and ('a' <= iText[start-preContextWarning-1].lower() <= 'z'):
preContextWarning += 1
else:
# Number
newNode = etree.Element('number')
newNode.text = oldNumber.replace(' ','').replace(',','.')
# Check if the pre-context is funny
offset = 0
while (start-offset > 0) and (iText[start-offset-1] == ' '):
offset += 1
if (start-offset > 0) and (iText[start-offset-1] in '.,0123456789'):
preContextWarning = min(contextLength, offset+1)
while (preContextWarning < contextLength) and (start-preContextWarning > 0) and (iText[start-preContextWarning-1] in '.,0123456789'):
preContextWarning += 1
# Check if the post-context is funny
if iText[stop:stop+2] in ['st', 'nd', 'rd', 'th']:
postContextWarning = 2
else:
# Check if there are units
match = unitPattern.match(iText[stop:])
if (iNode.tag != 'latex') and (match is not None):
numberNode = newNode
newNode = etree.Element('unit_number')
newNode.append(numberNode)
newNode.append(etree.Element('unit'))
newNode[-1].text = match.group().strip()
oldNumber += iText[stop:stop+len(match.group())]
stop += len(match.group())
else:
# Check if there are units in LaTeX mode
match = latexUnitPattern.match(iText[stop:])
if match is not None:
numberNode = newNode
newNode = etree.Element('unit_number')
newNode.append(numberNode)
newNode.append(etree.Element('unit'))
unitText = match.group().strip()
if unitText == r'\ell':
unitText = u'ℓ'
elif unitText == u'°':
unitText = u'°'
else:
assert unitText[:8] == r'\textrm{'
assert unitText[-1] == '}'
unitText = unitText[8:-1].strip()
newNode[-1].text = unitText
oldNumber += iText[stop:stop+len(match.group())]
stop += len(match.group())
else:
if lazyMode and ((iNode.tag == 'latex') or ((iNode.tag == 'correct') and ('latex' in get_full_dom_path(iNode)))):
# Skip unit-less numbers that are small integers and in LaTeX mode
try:
value = int(oldNumber.replace(' ','').replace(',','.'))
if abs(value) < 1000:
continue
except ValueError:
pass
# Check if the post-context is funny (for both currency and number)
if postContextWarning == 0:
offset = 0
while (stop+offset+1 < len(iText)) and (iText[stop+offset+1] == ' '):
offset += 1
if (stop+offset+1 < len(iText)) and (iText[stop+offset+1] in '.,0123456789'):
postContextWarning = min(contextLength, offset+1)
while (postContextWarning < contextLength) and (stop+postContextWarning+1 < len(iText)) and (iText[stop+postContextWarning+1] in '.,0123456789'):
postContextWarning += 1
if iNode.tag is not None:
context[0] = '<' + iNode.tag + '>'
if iPreTailTag is not None:
context[0] += ' ... </' + iPreTailTag + '>'
else:
context[0] = ''
if start < contextLength:
context[0] += iText[:start]
else:
context[0] += ' ...' + iText[start-contextLength:start]
if stop+contextLength > len(iText):
context[1] = iText[stop:]
if iPostTailTag is not None:
context[1] += '<' + iPostTailTag + '>' + ' ... '
else:
context[1] = iText[stop:stop+contextLength] + '... '
context[1] += '</' + iNode.tag + '>'
if prompt_replace(oldNumber, etree.tostring(newNode, encoding='utf-8').decode('utf-8'), context, (preContextWarning, postContextWarning)):
oSubstitutions.append((start, stop, newNode))
return oSubstitutions
