Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

DOCX Input: Add support for cross references. Cross references that are used to insert hyperlinks are now converted into hyperlinks in the output document. Features to insert paragraph/page numbers are not supported.

Browse Source
This commit is contained in:
Kovid Goyal 2014-05-08 21:28:49 +05:30
parent d3c67c72fd
commit 24299b12a4
1 changed files with 36 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
src/calibre/ebooks/docx/fields.py
View File

@ -84,6 +84,13 @@ parse_index = parser('index',
' f:entry-type g:page-range-separator h:heading k:crossref-separator' ' f:entry-type g:page-range-separator h:heading k:crossref-separator'
' l:page-number-separator p:letter-range s:sequence-name r:run-together y:yomi z:langcode') ' l:page-number-separator p:letter-range s:sequence-name r:run-together y:yomi z:langcode')
parse_ref = parser('ref',
'd:separator f:footnote h:hyperlink n:number p:position r:relative-number t:suppress w:number-full-context')
parse_noteref = parser('noteref',
'f:footnote h:hyperlink p:position')
class Fields(object): class Fields(object):
def __init__(self): def __init__(self):
@ -117,7 +124,7 @@ class Fields(object):
if stack: if stack:
stack[-1].contents.append(elem) stack[-1].contents.append(elem)
field_types = ('hyperlink', 'xe', 'index') field_types = ('hyperlink', 'xe', 'index', 'ref', 'noteref')
parsers = {x.upper():getattr(self, 'parse_'+x) for x in field_types} parsers = {x.upper():getattr(self, 'parse_'+x) for x in field_types}
field_parsers = {f.upper():globals()['parse_%s' % f] for f in field_types} field_parsers = {f.upper():globals()['parse_%s' % f] for f in field_types}
@ -131,28 +138,42 @@ class Fields(object):
if func is not None: if func is not None:
func(field, field_parsers[field.name], log) func(field, field_parsers[field.name], log)
def get_runs(self, field):
all_runs = []
current_runs = []
# We only handle spans in a single paragraph
# being wrapped in <a>
for x in field.contents:
if x.tag.endswith('}p'):
if current_runs:
all_runs.append(current_runs)
current_runs = []
elif x.tag.endswith('}r'):
current_runs.append(x)
if current_runs:
all_runs.append(current_runs)
return all_runs
def parse_hyperlink(self, field, parse_func, log): def parse_hyperlink(self, field, parse_func, log):
# Parse hyperlink fields # Parse hyperlink fields
hl = parse_func(field.instructions, log) hl = parse_func(field.instructions, log)
if hl: if hl:
if 'target' in hl and hl['target'] is None: if 'target' in hl and hl['target'] is None:
hl['target'] = '_blank' hl['target'] = '_blank'
all_runs = [] for runs in self.get_runs(field):
current_runs = []
# We only handle spans in a single paragraph
# being wrapped in <a>
for x in field.contents:
if x.tag.endswith('}p'):
if current_runs:
all_runs.append(current_runs)
current_runs = []
elif x.tag.endswith('}r'):
current_runs.append(x)
if current_runs:
all_runs.append(current_runs)
for runs in all_runs:
self.hyperlink_fields.append((hl, runs)) self.hyperlink_fields.append((hl, runs))
def parse_ref(self, field, parse_func, log):
ref = parse_func(field.instructions, log)
dest = ref.get(None, None)
if dest is not None and 'hyperlink' in ref:
for runs in self.get_runs(field):
self.hyperlink_fields.append(({'anchor':dest}, runs))
else:
self.log.warn('Unsupported reference field (%s), ignoring: %r' % (field.name, ref))
parse_noteref = parse_ref
def parse_xe(self, field, parse_func, log): def parse_xe(self, field, parse_func, log):
# Parse XE fields # Parse XE fields
if None in (field.start, field.end): if None in (field.start, field.end):