Text converter updated.

[librarian.git] / src / librarian / html.py

diff --git a/src/librarian/html.py b/src/librarian/html.py
index d262198..a96e975 100644 (file)
--- a/src/librarian/html.py
+++ b/src/librarian/html.py
@@ -1,136 +1,64 @@
-# -*- coding: utf-8 -*-
-#

 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
 # 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 os
 import re
 import copy

+import urllib.parse
+import urllib.request

 
 from lxml import etree
 
 from lxml import etree

-from librarian import XHTMLNS, ParseError, OutputFile

 from librarian import functions
 from PIL import Image
 
 from librarian import functions
 from PIL import Image
 

-from lxml.etree import XMLSyntaxError, XSLTApplyError
-import six
-
-
-functions.reg_substitute_entities()
-functions.reg_person_name()
-
-STYLESHEETS = {
-    'legacy': 'xslt/book2html.xslt',
-    'full': 'xslt/wl2html_full.xslt',
-    'partial': 'xslt/wl2html_partial.xslt'
-}
-
-
-def get_stylesheet(name):
-    return os.path.join(os.path.dirname(__file__), STYLESHEETS[name])
-
-
-def html_has_content(text):
-    return etree.ETXPath(
-        '//p|//{%(ns)s}p|//h1|//{%(ns)s}h1' % {'ns': str(XHTMLNS)}
-    )(text)

 
 

+def add_image_sizes(tree, gallery_path, gallery_url, base_url):
+    widths = [360, 600, 1200, 1800, 2400]
+    try:
+        os.makedirs(gallery_path)
+    except:
+        pass

 
 

-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
-    result = document.xslt(style)
-    html = re.sub('<a name="sec[0-9]*"/>', '',
-                  etree.tostring(result, encoding='unicode'))
-    return re.sub('</?blockquote[^>]*>', '', html)
+    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')

 
 

-def add_image_sizes(tree, gallery_path, gallery_url):
-    widths = [360, 600, 1200, 1800]
-    for ilustr in tree.findall('//ilustr'):
-        rel_path = ilustr.attrib['src']
-        img = Image.open(gallery_path + rel_path)

         srcset = []
         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:
         for w in widths:

-            if w < img.size[0]:
+            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))
                 height = round(img.size[1] * w / img.size[0])
                 th = img.resize((w, height))

-
-                fname = ('.W%d.' % w).join(rel_path.rsplit('.', 1))
-                th.save(gallery_path + fname)
-                srcset.append(" ".join((
-                    gallery_url + fname,
-                    '%dw' % w
-                    )))
-        srcset.append(" ".join((
-            gallery_url + rel_path,
-            '%dw' % img.size[0]
-        )))
+                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['srcset'] = ", ".join(srcset)

-        ilustr.attrib['src'] = gallery_url + rel_path
-
+        ilustr.attrib['src'] = largest_url

 
 

-def transform(wldoc, stylesheet='legacy', options=None, flags=None, css=None, gallery_path='img/', gallery_url='img/'):
-    """Transforms the WL document to XHTML.
-
-    If output_filename is None, returns an XML,
-    otherwise returns True if file has been written,False if it hasn't.
-    File won't be written if it has no content.
-    """
-    # Parse XSLT
-    try:
-        style_filename = get_stylesheet(stylesheet)
-        style = etree.parse(style_filename)
-
-        document = copy.deepcopy(wldoc)
-        del wldoc
-        document.swap_endlines()
-
-        if flags:
-            for flag in flags:
-                document.edoc.getroot().set(flag, 'yes')
-
-        document.clean_ed_note()
-        document.clean_ed_note('abstrakt')
-
-        if not options:
-            options = {}
-
-        add_image_sizes(document.edoc, gallery_path, gallery_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 :)
-
-        if html_has_content(result):
-            add_anchors(result.getroot())
-            add_table_of_themes(result.getroot())
-            add_table_of_contents(result.getroot())
-
-            return OutputFile.from_bytes(etree.tostring(
-                result, method='html', xml_declaration=False,
-                pretty_print=True, encoding='utf-8'
-            ))
-        else:
-            return None
-    except KeyError:
-        raise ValueError("'%s' is not a valid stylesheet.")
-    except (XMLSyntaxError, XSLTApplyError) as e:
-        raise ParseError(e)
+        f.close()

 
 
 
 

-@six.python_2_unicode_compatible
-class Fragment(object):
+class Fragment:

     def __init__(self, id, themes):
         super(Fragment, self).__init__()
         self.id = id
     def __init__(self, id, themes):
         super(Fragment, self).__init__()
         self.id = id


@@ -158,7 +86,7 @@ class Fragment(object):
         result = []
         for event, element in self.closed_events():
             if event == 'start':
         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)
                     element.tag,
                     ' '.join(
                         '%s="%s"' % (k, v)


@@ -168,7 +96,7 @@ class Fragment(object):
                 if element.text:
                     result.append(element.text)
             elif event == 'end':
                 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:
                 if element.tail:
                     result.append(element.tail)
             else:


@@ -187,7 +115,7 @@ def extract_fragments(input_filename):
 
     # iterparse would die on a HTML document
     parser = etree.HTMLParser(encoding='utf-8')
 
     # 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'
     buf.write(etree.tostring(
         etree.parse(input_filename, parser).getroot()[0][0],
         encoding='utf-8'


@@ -219,7 +147,8 @@ def extract_fragments(input_filename):
                 for parent in parents:
                     fragment.append('start', parent)
 
                 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:
 
             # Close existing fragment
             else:


@@ -243,7 +172,7 @@ def extract_fragments(input_filename):
         else:
             # Omit annotation tags
             if (len(element.get('name', '')) or
         else:
             # Omit annotation tags
             if (len(element.get('name', '')) or

-                    element.get('class', '') in ('annotation', 'anchor')):
+                    element.get('class', '') in ('annotation-anchor', 'anchor', 'wl-num', 'reference')):

                 if event == 'end' and element.tail:
                     for fragment_id in open_fragments:
                         open_fragments[fragment_id].append(
                 if event == 'end' and element.tail:
                     for fragment_id in open_fragments:
                         open_fragments[fragment_id].append(


@@ -261,24 +190,16 @@ def extract_fragments(input_filename):
     return closed_fragments, open_fragments
 
 
     return closed_fragments, open_fragments
 
 

-def add_anchor(element, prefix, with_link=True, with_target=True,
-               link_text=None):
+def add_anchor(element, prefix, link_text=None):

     parent = element.getparent()
     index = parent.index(element)
 
     parent = element.getparent()
     index = parent.index(element)
 

-    if with_link:
-        if link_text is None:
-            link_text = prefix
-        anchor = etree.Element('a', href='#%s' % prefix)
-        anchor.set('class', 'anchor')
-        anchor.text = six.text_type(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' '
-        parent.insert(index, anchor_target)
+    if link_text is None:
+        link_text = prefix
+    anchor = etree.Element('a', href='#%s' % prefix)
+    anchor.set('class', 'anchor')
+    anchor.text = str(link_text)
+    parent.insert(index, anchor)

 
 
 def any_ancestor(element, test):
 
 
 def any_ancestor(element, test):


@@ -288,32 +209,9 @@ def any_ancestor(element, test):
     return False
 
 
     return False
 
 

-def add_anchors(root):
-    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'
-            )
-        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)
-            counter += 1
-        elif 'paragraph' in element.get('class', ''):
-            add_anchor(element, "f%d" % counter, link_text=counter)
-            counter += 1
-
-

 def raw_printable_text(element):
     working = copy.deepcopy(element)
 def raw_printable_text(element):
     working = copy.deepcopy(element)

-    for e in working.findall('a'):
+    for e in working.findall('.//a'):

         if e.get('class') in ('annotation', 'theme-begin'):
             e.text = ''
     return etree.tostring(working, method='text', encoding='unicode').strip()
         if e.get('class') in ('annotation', 'theme-begin'):
             e.text = ''
     return etree.tostring(working, method='text', encoding='unicode').strip()


@@ -321,7 +219,6 @@ def raw_printable_text(element):
 
 def add_table_of_contents(root):
     sections = []
 
 def add_table_of_contents(root):
     sections = []

-    counter = 1

     for element in root.iterdescendants():
         if element.tag in ('h2', 'h3'):
             if any_ancestor(
     for element in root.iterdescendants():
         if element.tag in ('h2', 'h3'):
             if any_ancestor(


@@ -335,29 +232,30 @@ def add_table_of_contents(root):
             if (element.tag == 'h3' and len(sections)
                     and sections[-1][1] == 'h2'):
                 sections[-1][3].append(
             if (element.tag == 'h3' and len(sections)
                     and sections[-1][1] == 'h2'):
                 sections[-1][3].append(

-                    (counter, element.tag, element_text, [])
+                    (element.attrib['id'], element.tag, element_text, [])

                 )
             else:
                 )
             else:

-                sections.append((counter, element.tag, element_text, []))
-            add_anchor(element, "s%d" % counter, with_link=False)
-            counter += 1
+                sections.append((element.attrib['id'], element.tag, element_text, []))
+
+    if not sections:
+        return

 
     toc = etree.Element('div')
     toc.set('id', 'toc')
     toc_header = etree.SubElement(toc, 'h2')
 
     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:
         section_element = etree.SubElement(toc_list, 'li')
     toc_list = etree.SubElement(toc, 'ol')
 
     for n, section, text, subsections in sections:
         section_element = etree.SubElement(toc_list, 'li')

-        add_anchor(section_element, "s%d" % n, with_target=False,
+        add_anchor(section_element, n,

                    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')
                    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,
+                add_anchor(subsection_element, n1,

                            link_text=subtext)
 
     root.insert(0, toc)
                            link_text=subtext)
 
     root.insert(0, toc)


@@ -388,7 +286,13 @@ def add_table_of_themes(root):
             item = etree.SubElement(themes_li, 'a', href="#%s" % fragment)
             item.text = str(i + 1)
             item.tail = ' '
             item = etree.SubElement(themes_li, 'a', href="#%s" % fragment)
             item.text = str(i + 1)
             item.tail = ' '

+
+    if not len(themes_ol):
+        return
+

     root.insert(0, themes_div)
     root.insert(0, themes_div)

+    themes_div.tail = root.text
+    root.text = None

 
 
 def extract_annotations(html_path):
 
 
 def extract_annotations(html_path):


@@ -425,8 +329,8 @@ def extract_annotations(html_path):
                     candidate = candidate.strip()
                     if candidate in FN_QUALIFIERS:
                         qualifiers.append(candidate)
                     candidate = candidate.strip()
                     if candidate in FN_QUALIFIERS:
                         qualifiers.append(candidate)

-                    elif candidate.startswith('z '):
-                        subcandidate = candidate.split()[1]
+                    elif candidate.startswith('z\u00A0'):
+                        subcandidate = candidate.split('\u00A0')[1].split()[0]

                         if subcandidate in FN_QUALIFIERS:
                             qualifiers.append(subcandidate)
             else:
                         if subcandidate in FN_QUALIFIERS:
                             qualifiers.append(subcandidate)
             else: