def run(self, interactive):
if interactive:
self.run_prompt()
return
manipulate = NlmManipulate(self.gv)
tree = manipulate.load_dom_tree()
# pre-cleanup: remove all empty ext-links as these break the linker
items_to_clean = tree.xpath('//ext-link')
count = 0
for item in items_to_clean:
if '{http://www.w3.org/1999/xlink}href' in item.attrib and \
item.attrib['{http://www.w3.org/1999/xlink}href'] == '':
count += 1
item.tag = 'REMOVE'
etree.strip_tags(item.getparent(), 'REMOVE')
if count > 0:
manipulate.save_tree(tree)
self.debug.print_debug(self, u'Removed {0} blank ext-link tags'.format(count))
ref_items = tree.xpath('//back/ref-list/ref')
self.clean_ref_items(tree, ref_items, manipulate)
# handle numbered reference items
references_and_numbers = {}
for ref in ref_items:
text = manipulate.get_stripped_text(ref)
ref_match = re.compile('^(?P<number>\d+)\.*')
result = ref_match.match(text)
if result:
references_and_numbers[result.group('number')] = ref
parsed = self.process_ibid_authors(ref_items)
if parsed > 0:
manipulate.save_tree(tree)
self.debug.print_debug(self, u'Replace {0} instances of "---." at start of references'.format(parsed))
to_link = []
to_stub = []
square_bracket_count = {}
for p in tree.xpath('//sec//p[not(mml:math)] | //td',
namespaces={'mml': 'http://www.w3.org/1998/Math/MathML'}):
text = manipulate.get_stripped_text(p)
reference_test = re.compile('\((?P<text>[^%]+?)\)')
matches = reference_test.finditer(text)
# exclude any square brackets with numbers inside
sub_match = re.compile('\[(?P<square>\d*[,\-;\d\s]*)\]')
smatch = sub_match.search(text)
if smatch:
smatches = sub_match.finditer(text)
for smatch in smatches:
self.debug.print_debug(self, u'Handling references in square '
u'brackets: [{0}] '.format(smatch.group('square')))
for item in re.split(';|,', smatch.group('square')):
if '-' in item:
parent, tail = manipulate.find_text(p, item)
if parent is not None:
new_string = ''
try:
split_range = item.strip().split('-')
for no in range(int(split_range[0]), int(split_range[1]) + 1):
new_string += str(no) + ','
except:
self.debug.print_debug(self, u'Unable to parse reference '
u'number in range {0}'.format(item))
break
if new_string.endswith(',') and not item.endswith(','):
new_string = new_string[0:len(new_string) - 1]
if tail and new_string != '':
parent.tail = parent.tail.replace(item, new_string)
elif not tail and new_string != '':
parent.text = parent.text.replace(item, new_string)
try:
split_range = item.strip().split('-')
for no in range(int(split_range[0]), int(split_range[1]) + 1):
self.debug.print_debug(self, u'Parsing reference '
u'number in range {0}'.format(str(no)))
to_stub.append(ReplaceStub(self.gv, p, str(no), tree, manipulate,
'TO_LINK_NUMBER', length_ignore=True))
except:
self.debug.print_debug(self, u'Unable to parse reference '
u'number in range {0}'.format(item))
break
else:
# just replace the components
split_range = item.strip().split('-')
for link in split_range:
to_stub.append(ReplaceStub(self.gv, p, link, tree, manipulate,
'TO_LINK_NUMBER', length_ignore=True))
else:
if len(item.strip()) < 60:
to_stub.append(ReplaceStub(self.gv, p, item.strip(), tree, manipulate, 'TO_LINK_NUMBER',
length_ignore=True))
square_bracket_count[item.strip()] = 1
else:
for match in matches:
for item in match.group('text').split(u';'):
if len(item.strip()) < 60:
to_stub.append(ReplaceStub(self.gv, p, item.strip(), tree, manipulate))
for link in to_stub:
link.link(to_stub)
#pass
etree.strip_elements(tree, 'REMOVE')
use_index_method = False
if len(square_bracket_count) != len(references_and_numbers):
# we found more than 3 [1], [2] style references but no reference elements beginning with numbers
# so, we will simply try to use the /index/ of the reference item (-1 for zero-based compensation)
self.debug.print_debug(self, u'Using indexical method for square bracket correlation')
use_index_method = True
if len(ref_items) == 0:
self.debug.print_debug(self, u'Found no references to link')
manipulate.save_tree(tree)
return
for p in tree.xpath('//xref[@rid="TO_LINK_NUMBER"]'):
text = manipulate.get_stripped_text(p)
if not use_index_method:
if text in references_and_numbers:
ReplaceObject(self.gv, p, references_and_numbers[text]).link()
else:
p.attrib['rid'] = 'TO_LINK'
else:
try:
ReplaceObject(self.gv, p, ref_items[int(text) - 1]).link()
except:
self.debug.print_debug(self, u'Failed to link to reference {0} + 1 using '
u'indexical method'.format(text))
p.attrib['rid'] = 'TO_LINK'
for p in tree.xpath('//xref[@rid="TO_LINK"]'):
text = manipulate.get_stripped_text(p)
item = text
bare_items = item.strip().replace(u',', '').split(u' ')
for ref in ref_items:
found = True
bare_ref = manipulate.get_stripped_text(ref)
bare_refs = bare_ref.split(' ')
replace_chars = '[,\.\<\>\(\)\;\:\@\'\#\~\}\{\[\]\"]'
for sub_item in bare_items:
found_ref = False
for sub_ref in bare_refs:
if re.sub(replace_chars, '', sub_item.strip()).strip() == sub_ref.strip(replace_chars):
found_ref = True
break
if not found_ref:
found = False
if len(bare_items) > 0 and found:
to_link.append(ReplaceObject(self.gv, p, ref))
elif len(bare_items) > 0:
replace_chars = '[,\.\<\>\(\)\;\:\@\'\#\~\}\{\[\]\"\d]'
found = True
for sub_item in bare_items:
found_ref = False
subbed_text = re.sub(replace_chars, '', sub_item.strip()).strip()
for sub_ref in bare_refs:
sub_ref = re.sub(replace_chars, '', sub_ref.strip()).strip()
if subbed_text == '' and len(bare_items) > 1:
found_ref = True
break
if subbed_text == sub_ref and subbed_text != '' and sub_ref != '':
found_ref = True
break
if not found_ref:
found = False
# we don't allow linking to the last item here because it is almost universally wrong
if len(bare_items) > 0 and found and ref_items.index(ref) != len(ref_items) - 1:
to_link.append(ReplaceObject(self.gv, p, ref))
if len(to_link) == 0:
self.debug.print_debug(self, u'Found no references to link')
for link in to_link:
link.link()
#pass
manipulate.save_tree(tree)
