X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/fefdce4e24f9e397df5538fe6e7f54b5ece4d841..fdcbc69e0f1f8f4945cd37c1e10be34a910d48c6:/src/librarian/html.py?ds=sidebyside
diff --git a/src/librarian/html.py b/src/librarian/html.py
index f8e4a83..fddeb2f 100644
--- a/src/librarian/html.py
+++ b/src/librarian/html.py
@@ -12,6 +12,7 @@ import copy
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
@@ -32,20 +33,66 @@ def get_stylesheet(name):
def html_has_content(text):
- return etree.ETXPath('//p|//{%(ns)s}p|//h1|//{%(ns)s}h1' % {'ns': str(XHTMLNS)})(text)
+ return etree.ETXPath(
+ '//p|//{%(ns)s}p|//h1|//{%(ns)s}h1' % {'ns': str(XHTMLNS)}
+ )(text)
def transform_abstrakt(abstrakt_element):
style_filename = get_stylesheet('legacy')
style = etree.parse(style_filename)
xml = etree.tostring(abstrakt_element, encoding='unicode')
- document = etree.parse(six.StringIO(xml.replace('abstrakt', 'dlugi_cytat'))) # HACK
+ document = etree.parse(six.StringIO(
+ xml.replace('abstrakt', 'dlugi_cytat')
+ )) # HACK
result = document.xslt(style)
- html = re.sub('', '', etree.tostring(result, encoding='unicode'))
+ html = re.sub('', '',
+ etree.tostring(result, encoding='unicode'))
return re.sub(']*>', '', 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 = six.moves.urllib.parse.urljoin(base_url, rel_path)
+
+ f = six.moves.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,
@@ -67,12 +114,22 @@ def transform(wldoc, stylesheet='legacy', options=None, flags=None, css=None):
document.clean_ed_note()
document.clean_ed_note('abstrakt')
-
+ document.fix_pa_akap()
+
if not options:
options = {}
- options.setdefault('gallery', "''")
- css = css or 'https://static.wolnelektury.pl/css/compressed/book_text.css'
+ try:
+ os.makedirs(gallery_path)
+ except OSError:
+ pass
+
+ add_image_sizes(document.edoc, gallery_path, gallery_url, base_url)
+
+ css = (
+ css
+ or 'https://static.wolnelektury.pl/css/compressed/book_text.css'
+ )
css = "'%s'" % css
result = document.transform(style, css=css, **options)
del document # no longer needed large object :)
@@ -83,7 +140,9 @@ def transform(wldoc, stylesheet='legacy', options=None, flags=None, css=None):
add_table_of_contents(result.getroot())
return OutputFile.from_bytes(etree.tostring(
- result, method='html', xml_declaration=False, pretty_print=True, encoding='utf-8'))
+ result, method='html', xml_declaration=False,
+ pretty_print=True, encoding='utf-8'
+ ))
else:
return None
except KeyError:
@@ -122,7 +181,12 @@ class Fragment(object):
for event, element in self.closed_events():
if event == 'start':
result.append(u'<%s %s>' % (
- element.tag, ' '.join('%s="%s"' % (k, v) for k, v in element.attrib.items())))
+ element.tag,
+ ' '.join(
+ '%s="%s"' % (k, v)
+ for k, v in element.attrib.items()
+ )
+ ))
if element.text:
result.append(element.text)
elif event == 'end':
@@ -146,7 +210,10 @@ def extract_fragments(input_filename):
# iterparse would die on a HTML document
parser = etree.HTMLParser(encoding='utf-8')
buf = six.BytesIO()
- buf.write(etree.tostring(etree.parse(input_filename, parser).getroot()[0][0], encoding='utf-8'))
+ buf.write(etree.tostring(
+ etree.parse(input_filename, parser).getroot()[0][0],
+ encoding='utf-8'
+ ))
buf.seek(0)
for event, element in etree.iterparse(buf, events=('start', 'end')):
@@ -165,6 +232,8 @@ def extract_fragments(input_filename):
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()
@@ -179,12 +248,15 @@ def extract_fragments(input_filename):
try:
fragment = open_fragments[element.get('fid')]
except KeyError:
- print('%s:closed not open fragment #%s' % (input_filename, element.get('fid')))
+ print('%s:closed not open fragment #%s' % (
+ input_filename, element.get('fid')
+ ))
else:
closed_fragments[fragment.id] = fragment
del open_fragments[fragment.id]
- # Append element tail to lost_text (we don't want to lose any text)
+ # Append element tail to lost_text
+ # (we don't want to lose any text)
if element.tail:
for fragment_id in open_fragments:
open_fragments[fragment_id].append('text', element.tail)
@@ -192,19 +264,27 @@ def extract_fragments(input_filename):
# Process all elements except begin and end
else:
# Omit annotation tags
- if (len(element.get('name', '')) or
+ if (len(element.get('name', '')) or
element.get('class', '') in ('annotation', 'anchor')):
if event == 'end' and element.tail:
for fragment_id in open_fragments:
- open_fragments[fragment_id].append('text', element.tail)
+ open_fragments[fragment_id].append(
+ 'text', element.tail
+ )
else:
for fragment_id in open_fragments:
- open_fragments[fragment_id].append(event, copy.copy(element))
+ celem = copy.copy(element)
+ if 'id' in celem.attrib:
+ del celem.attrib['id']
+ open_fragments[fragment_id].append(
+ event, celem
+ )
return closed_fragments, open_fragments
-def add_anchor(element, prefix, with_link=True, with_target=True, link_text=None):
+def add_anchor(element, prefix, with_link=True, with_target=True,
+ link_text=None):
parent = element.getparent()
index = parent.index(element)
@@ -232,20 +312,36 @@ def any_ancestor(element, test):
def add_anchors(root):
counter = 1
+ visible_counter = 1
for element in root.iterdescendants():
def f(e):
- return e.get('class') in ('note', 'motto', 'motto_podpis', 'dedication', 'frame') or \
- e.get('id') == 'nota_red' or e.tag == 'blockquote'
+ return (
+ e.get('class') in (
+ 'note', 'motto', 'motto_podpis', 'dedication', 'frame'
+ )
+ 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):
@@ -261,13 +357,19 @@ def add_table_of_contents(root):
counter = 1
for element in root.iterdescendants():
if element.tag in ('h2', 'h3'):
- if any_ancestor(element,
- lambda e: e.get('id') in ('footnotes', 'nota_red') or e.get('class') in ('person-list',)):
+ if any_ancestor(
+ element,
+ lambda e: e.get('id') in (
+ 'footnotes', 'nota_red'
+ ) or e.get('class') in ('person-list',)):
continue
element_text = raw_printable_text(element)
- if element.tag == 'h3' and len(sections) and sections[-1][1] == 'h2':
- sections[-1][3].append((counter, element.tag, element_text, []))
+ if (element.tag == 'h3' and len(sections)
+ and sections[-1][1] == 'h2'):
+ sections[-1][3].append(
+ (counter, element.tag, element_text, [])
+ )
else:
sections.append((counter, element.tag, element_text, []))
add_anchor(element, "s%d" % counter, with_link=False)
@@ -281,17 +383,19 @@ def add_table_of_contents(root):
for n, section, text, subsections in sections:
section_element = etree.SubElement(toc_list, 'li')
- add_anchor(section_element, "s%d" % n, with_target=False, link_text=text)
+ add_anchor(section_element, "s%d" % n, with_target=False,
+ link_text=text)
if len(subsections):
subsection_list = etree.SubElement(section_element, 'ol')
for n1, subsection, subtext, _ in subsections:
subsection_element = etree.SubElement(subsection_list, 'li')
- add_anchor(subsection_element, "s%d" % n1, with_target=False, link_text=subtext)
+ add_anchor(subsection_element, "s%d" % n1, with_target=False,
+ link_text=subtext)
root.insert(0, toc)
-
+
def add_table_of_themes(root):
try:
from sortify import sortify
@@ -341,8 +445,10 @@ def extract_annotations(html_path):
footnote.text = None
if len(footnote) and footnote[-1].tail == '\n':
footnote[-1].tail = None
- text_str = etree.tostring(footnote, method='text', encoding='unicode').strip()
- html_str = etree.tostring(footnote, method='html', encoding='unicode').strip()
+ text_str = etree.tostring(footnote, method='text',
+ encoding='unicode').strip()
+ html_str = etree.tostring(footnote, method='html',
+ encoding='unicode').strip()
match = re_qualifier.match(text_str)
if match: