-# -*- coding: utf-8 -*-
-#
# This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
-# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
+# Copyright © Fundacja Wolne Lektury. See NOTICE for more information.
#
-from __future__ import print_function, unicode_literals
-
+import io
import os
import re
import copy
+import urllib.parse
+import urllib.request
from lxml import etree
from librarian import XHTMLNS, ParseError, OutputFile
from librarian import functions
+from PIL import Image
from lxml.etree import XMLSyntaxError, XSLTApplyError
-import six
functions.reg_substitute_entities()
STYLESHEETS = {
'legacy': 'xslt/book2html.xslt',
- 'full': 'xslt/wl2html_full.xslt',
- 'partial': 'xslt/wl2html_partial.xslt'
}
style_filename = get_stylesheet('legacy')
style = etree.parse(style_filename)
xml = etree.tostring(abstrakt_element, encoding='unicode')
- document = etree.parse(six.StringIO(
+ document = etree.parse(io.StringIO(
xml.replace('abstrakt', 'dlugi_cytat')
)) # HACK
result = document.xslt(style)
return re.sub('</?blockquote[^>]*>', '', html)
-def transform(wldoc, stylesheet='legacy', options=None, flags=None, css=None):
+def add_image_sizes(tree, gallery_path, gallery_url, base_url):
+ widths = [360, 600, 1200, 1800, 2400]
+
+ for i, ilustr in enumerate(tree.findall('//ilustr')):
+ rel_path = ilustr.attrib['src']
+ img_url = urllib.parse.urljoin(base_url, rel_path)
+
+ f = urllib.request.urlopen(img_url)
+ img = Image.open(f)
+ ext = {'GIF': 'gif', 'PNG': 'png'}.get(img.format, 'jpg')
+
+ srcset = []
+ # Needed widths: predefined and original, limited by
+ # whichever is smaller.
+ img_widths = [
+ w for w in
+ sorted(
+ set(widths + [img.size[0]])
+ )
+ if w <= min(widths[-1], img.size[0])
+ ]
+ largest = None
+ for w in widths:
+ fname = '%d.W%d.%s' % (i, w, ext)
+ fpath = gallery_path + fname
+ if not os.path.exists(fpath):
+ height = round(img.size[1] * w / img.size[0])
+ th = img.resize((w, height))
+ th.save(fpath)
+ th_url = gallery_url + fname
+ srcset.append(" ".join((
+ th_url,
+ '%dw' % w
+ )))
+ largest_url = th_url
+ ilustr.attrib['srcset'] = ", ".join(srcset)
+ ilustr.attrib['src'] = largest_url
+
+ f.close()
+
+
+def transform(wldoc, stylesheet='legacy', options=None, flags=None, css=None, gallery_path='img/', gallery_url='img/', base_url='file://./'):
"""Transforms the WL document to XHTML.
If output_filename is None, returns an XML,
document.clean_ed_note()
document.clean_ed_note('abstrakt')
-
+ document.fix_pa_akap()
+
if not options:
options = {}
- options.setdefault('gallery', "''")
+
+ try:
+ os.makedirs(gallery_path)
+ except OSError:
+ pass
+
+ add_image_sizes(document.edoc, gallery_path, gallery_url, base_url)
css = (
css
raise ParseError(e)
-@six.python_2_unicode_compatible
-class Fragment(object):
+class Fragment:
def __init__(self, id, themes):
super(Fragment, self).__init__()
self.id = id
result = []
for event, element in self.closed_events():
if event == 'start':
- result.append(u'<%s %s>' % (
+ result.append('<%s %s>' % (
element.tag,
' '.join(
'%s="%s"' % (k, v)
if element.text:
result.append(element.text)
elif event == 'end':
- result.append(u'</%s>' % element.tag)
+ result.append('</%s>' % element.tag)
if element.tail:
result.append(element.tail)
else:
# iterparse would die on a HTML document
parser = etree.HTMLParser(encoding='utf-8')
- buf = six.BytesIO()
+ buf = io.BytesIO()
buf.write(etree.tostring(
etree.parse(input_filename, parser).getroot()[0][0],
encoding='utf-8'
while parent.get('id', None) != 'book-text':
cparent = copy.deepcopy(parent)
cparent.text = None
+ if 'id' in cparent.attrib:
+ del cparent.attrib['id']
parents.append(cparent)
parent = parent.getparent()
for parent in parents:
fragment.append('start', parent)
- open_fragments[fragment.id] = fragment
+ if fragment.id not in open_fragments:
+ open_fragments[fragment.id] = fragment
# Close existing fragment
else:
)
else:
for fragment_id in open_fragments:
+ celem = copy.copy(element)
+ if 'id' in celem.attrib:
+ del celem.attrib['id']
open_fragments[fragment_id].append(
- event, copy.copy(element)
+ event, celem
)
return closed_fragments, open_fragments
link_text = prefix
anchor = etree.Element('a', href='#%s' % prefix)
anchor.set('class', 'anchor')
- anchor.text = six.text_type(link_text)
+ anchor.text = str(link_text)
parent.insert(index, anchor)
if with_target:
anchor_target = etree.Element('a', name='%s' % prefix)
anchor_target.set('class', 'target')
- anchor_target.text = u' '
+ anchor_target.text = ' '
parent.insert(index, anchor_target)
def add_anchors(root):
counter = 1
+ visible_counter = 1
for element in root.iterdescendants():
def f(e):
return (
)
or e.get('id') == 'nota_red'
or e.tag == 'blockquote'
+ or e.get('id') == 'footnotes'
)
+
+ if element.get('class') == 'numeracja':
+ try:
+ visible_counter = int(element.get('data-start'))
+ except ValueError:
+ visible_counter = 1
+
if any_ancestor(element, f):
continue
if element.tag == 'div' and 'verse' in element.get('class', ''):
- if counter == 1 or counter % 5 == 0:
- add_anchor(element, "f%d" % counter, link_text=counter)
+ if visible_counter == 1 or visible_counter % 5 == 0:
+ add_anchor(element, "f%d" % counter, link_text=visible_counter)
counter += 1
+ visible_counter += 1
elif 'paragraph' in element.get('class', ''):
- add_anchor(element, "f%d" % counter, link_text=counter)
+ add_anchor(element, "f%d" % counter, link_text=visible_counter)
counter += 1
+ visible_counter += 1
def raw_printable_text(element):
toc = etree.Element('div')
toc.set('id', 'toc')
toc_header = etree.SubElement(toc, 'h2')
- toc_header.text = u'Spis treści'
+ toc_header.text = 'Spis treści'
toc_list = etree.SubElement(toc, 'ol')
for n, section, text, subsections in sections:
fnp / librarian.git / blobdiff