X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/2c73c162844627d26991915fecc7e39f02bc34b8..665c88719c108626edee864affe07b085d28b7b4:/librarian/html.py diff --git a/librarian/html.py b/librarian/html.py index e084ed2..ed7b4d6 100644 --- a/librarian/html.py +++ b/librarian/html.py @@ -3,9 +3,10 @@ # This file is part of Librarian, licensed under GNU Affero GPLv3 or later. # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information. # +from __future__ import print_function, unicode_literals + import os import re -import cStringIO import copy from lxml import etree @@ -13,6 +14,8 @@ from librarian import XHTMLNS, ParseError, OutputFile from librarian import functions from lxml.etree import XMLSyntaxError, XSLTApplyError +import six + functions.reg_substitute_entities() functions.reg_person_name() @@ -23,12 +26,25 @@ STYLESHEETS = { '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 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('', '', etree.tostring(result, encoding='unicode')) + return re.sub(']*>', '', html) + + def transform(wldoc, stylesheet='legacy', options=None, flags=None): """Transforms the WL document to XHTML. @@ -50,26 +66,30 @@ def transform(wldoc, stylesheet='legacy', options=None, flags=None): document.edoc.getroot().set(flag, 'yes') document.clean_ed_note() + document.clean_ed_note('abstrakt') if not options: options = {} + options.setdefault('gallery', "''") result = document.transform(style, **options) - del document # no longer needed large object :) + 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_string(etree.tostring(result, method='html', - xml_declaration=False, pretty_print=True, encoding='utf-8')) + 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), e: + except (XMLSyntaxError, XSLTApplyError) as e: raise ParseError(e) + +@six.python_2_unicode_compatible class Fragment(object): def __init__(self, id, themes): super(Fragment, self).__init__() @@ -89,7 +109,7 @@ class Fragment(object): try: stack.pop() except IndexError: - print 'CLOSED NON-OPEN TAG:', element + print('CLOSED NON-OPEN TAG:', element) stack.reverse() return self.events + stack @@ -98,7 +118,8 @@ class Fragment(object): result = [] 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()))) + result.append(u'<%s %s>' % ( + element.tag, ' '.join('%s="%s"' % (k, v) for k, v in element.attrib.items()))) if element.text: result.append(element.text) elif event == 'end': @@ -110,7 +131,7 @@ class Fragment(object): return ''.join(result) - def __unicode__(self): + def __str__(self): return self.to_string() @@ -121,14 +142,15 @@ def extract_fragments(input_filename): # iterparse would die on a HTML document parser = etree.HTMLParser(encoding='utf-8') - buf = cStringIO.StringIO() + buf = six.BytesIO() 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')): # Process begin and end elements if element.get('class', '') in ('theme-begin', 'theme-end'): - if not event == 'end': continue # Process elements only once, on end event + if not event == 'end': + continue # Process elements only once, on end event # Open new fragment if element.get('class', '') == 'theme-begin': @@ -154,7 +176,7 @@ 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] @@ -164,7 +186,6 @@ def extract_fragments(input_filename): for fragment_id in open_fragments: open_fragments[fragment_id].append('text', element.tail) - # Process all elements except begin and end else: # Omit annotation tags @@ -189,7 +210,7 @@ def add_anchor(element, prefix, with_link=True, with_target=True, link_text=None link_text = prefix anchor = etree.Element('a', href='#%s' % prefix) anchor.set('class', 'anchor') - anchor.text = unicode(link_text) + anchor.text = six.text_type(link_text) parent.insert(index, anchor) if with_target: @@ -209,9 +230,10 @@ def any_ancestor(element, test): def add_anchors(root): counter = 1 for element in root.iterdescendants(): - if any_ancestor(element, lambda e: e.get('class') in ('note', 'motto', 'motto_podpis', 'dedication') - or e.get('id') == 'nota_red' - or e.tag == 'blockquote'): + 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 == 'p' and 'verse' in element.get('class', ''): @@ -228,7 +250,7 @@ def raw_printable_text(element): 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() + return etree.tostring(working, method='text', encoding='unicode').strip() def add_table_of_contents(root): @@ -236,7 +258,8 @@ 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) @@ -259,9 +282,9 @@ def add_table_of_contents(root): if len(subsections): subsection_list = etree.SubElement(section_element, 'ol') - for n, subsection, text, _ in subsections: + for n1, subsection, subtext, _ in subsections: subsection_element = etree.SubElement(subsection_list, 'li') - add_anchor(subsection_element, "s%d" % n, with_target=False, link_text=text) + add_anchor(subsection_element, "s%d" % n1, with_target=False, link_text=subtext) root.insert(0, toc) @@ -270,7 +293,8 @@ def add_table_of_themes(root): try: from sortify import sortify except ImportError: - sortify = lambda x: x + def sortify(x): + return x book_themes = {} for fragment in root.findall('.//a[@class="theme-begin"]'): @@ -279,7 +303,7 @@ def add_table_of_themes(root): theme_names = [s.strip() for s in fragment.text.split(',')] for theme_name in theme_names: book_themes.setdefault(theme_name, []).append(fragment.get('name')) - book_themes = book_themes.items() + book_themes = list(book_themes.items()) book_themes.sort(key=lambda s: sortify(s[0])) themes_div = etree.Element('div', id="themes") themes_ol = etree.SubElement(themes_div, 'ol') @@ -294,11 +318,18 @@ def add_table_of_themes(root): def extract_annotations(html_path): - """For each annotation, yields a tuple: anchor, text, html.""" + """Extracts annotations from HTML for annotations dictionary. + + For each annotation, yields a tuple of: + anchor, footnote type, valid qualifiers, text, html. + + """ + from .fn_qualifiers import FN_QUALIFIERS + parser = etree.HTMLParser(encoding='utf-8') tree = etree.parse(html_path, parser) footnotes = tree.find('//*[@id="footnotes"]') - re_qualifier = re.compile(ur'[^\u2014]+\s+\((.+)\)\s+\u2014') + re_qualifier = re.compile(r'[^\u2014]+\s+\(([^\)]+)\)\s+\u2014') if footnotes is not None: for footnote in footnotes.findall('div'): fn_type = footnote.get('class').split('-')[1] @@ -307,12 +338,22 @@ 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() - qualifier = None + 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: - qualifier = match.group(1) - - yield anchor, fn_type, qualifier, text_str, html_str + qualifier_str = match.group(1) + qualifiers = [] + for candidate in re.split('[;,]', qualifier_str): + candidate = candidate.strip() + if candidate in FN_QUALIFIERS: + qualifiers.append(candidate) + elif candidate.startswith('z '): + subcandidate = candidate.split()[1] + if subcandidate in FN_QUALIFIERS: + qualifiers.append(subcandidate) + else: + qualifiers = [] + yield anchor, fn_type, qualifiers, text_str, html_str