X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/175c5cf4f727162fa5bddd2460d37595251bbe8e..88f39c4f90a67164b2c3da7c7a41df420f3b17ab:/src/librarian/elements/base.py diff --git a/src/librarian/elements/base.py b/src/librarian/elements/base.py index 3e0f898..b0d16ed 100644 --- a/src/librarian/elements/base.py +++ b/src/librarian/elements/base.py @@ -1,14 +1,28 @@ -# -*- coding: utf-8 - +# This file is part of Librarian, licensed under GNU Affero GPLv3 or later. +# Copyright © Fundacja Wolne Lektury. See NOTICE for more information. +# +import copy import re from lxml import etree from librarian import dcparser, RDFNS -from librarian.html import raw_printable_text from librarian.util import get_translation +def last_words(text, n): + words = [] + for w in reversed(text.split()): + words.append(w) + if len(w) > 2: + n -= 1 + if not n: break + if n: + return n, text + else: + return n, ' '.join(reversed(words)) + class WLElement(etree.ElementBase): SECTION_PRECEDENCE = None + ASIDE = False TXT_TOP_MARGIN = 0 TXT_BOTTOM_MARGIN = 0 @@ -28,11 +42,11 @@ class WLElement(etree.ElementBase): STRIP = False text_substitutions = [ - (u'---', u'—'), - (u'--', u'–'), - #(u'...', u'…'), # Temporary turnoff for epub - (u',,', u'„'), - (u'"', u'”'), + ('---', '—'), + ('--', '–'), + #('...', '…'), # Temporary turnoff for epub + (',,', '„'), + ('"', '”'), ('\ufeff', ''), ("'", "\u2019"), # This was enabled for epub. @@ -76,33 +90,44 @@ class WLElement(etree.ElementBase): if parent is not None: parent.signal(signal) - def raw_printable_text(self): + def raw_printable_text(self, builder): + from librarian.html import raw_printable_text + # TODO: podtagi, wyroznienia, etc t = '' - t += self.normalize_text(self.text) + t += self.normalize_text(self.text, builder) for c in self: if not isinstance(c, WLElement): continue if c.tag not in ('pe', 'pa', 'pt', 'pr', 'motyw'): - t += c.raw_printable_text() - t += self.normalize_text(c.tail) + t += c.raw_printable_text(builder) + t += self.normalize_text(c.tail, builder) return t - def normalize_text(self, text): + def normalize_text(self, text, builder): text = text or '' for e, s in self.text_substitutions: text = text.replace(e, s) # FIXME: TEmporary turnoff # text = re.sub(r'\s+', ' ', text) ### TODO: Added now for epub - text = re.sub(r'(?<=\s\w)\s+', u'\u00A0', text) + + if getattr(builder, 'hyphenator', None) is not None: + newt = '' + wlist = re.compile(r'\w+|[^\w]', re.UNICODE).findall(text) + for w in wlist: + newt += builder.hyphenator.inserted(w, '\u00AD') + text = newt + + if builder.orphans: + text = re.sub(r'(?<=\s\w)\s+', '\u00A0', text) return text def _build_inner(self, builder, build_method): child_count = len(self) if self.CAN_HAVE_TEXT and self.text: - text = self.normalize_text(self.text) + text = self.normalize_text(self.text, builder) if self.STRIP: text = text.lstrip() if not child_count: @@ -111,8 +136,10 @@ class WLElement(etree.ElementBase): for i, child in enumerate(self): if isinstance(child, WLElement): getattr(child, build_method)(builder) + elif getattr(builder, 'debug') and child.tag is etree.Comment: + builder.process_comment(child) if self.CAN_HAVE_TEXT and child.tail: - text = self.normalize_text(child.tail) + text = self.normalize_text(child.tail, builder) if self.STRIP and i == child_count - 1: text = text.rstrip() builder.push_text(text) @@ -143,7 +170,7 @@ class WLElement(etree.ElementBase): # always copy the id attribute (?) if self.attrib.get('id'): attr['id'] = self.attrib['id'] - elif '_compat_section_id' in self.attrib: + elif getattr(self, 'SHOULD_HAVE_ID', False) and '_compat_section_id' in self.attrib: attr['id'] = self.attrib['_compat_section_id'] return attr @@ -173,7 +200,7 @@ class WLElement(etree.ElementBase): # TEMPORARY self.CAN_HAVE_TEXT = True self.STRIP = False - + start_chunk = self.EPUB_START_CHUNK and isinstance(self.getparent(), Master) if start_chunk: @@ -187,7 +214,7 @@ class WLElement(etree.ElementBase): builder.add_toc_entry( fragment, - self.raw_printable_text(), + self.raw_printable_text(builder), self.SECTION_PRECEDENCE ) @@ -195,6 +222,11 @@ class WLElement(etree.ElementBase): attr = self.get_epub_attr(builder) if fragment: attr['id'] = fragment + if builder.debug: + chunkno, sourceline = 0, self.sourceline + if builder.splits: + chunkno, sourceline = len(builder.splits), sourceline - builder.splits[-1] + attr['data-debug'] = f'{chunkno}:{sourceline}' builder.start_element( self.EPUB_TAG, attr @@ -212,7 +244,7 @@ class WLElement(etree.ElementBase): if self.SECTION_PRECEDENCE: assert isinstance(self.getparent(), (Master, DlugiCytat, PoezjaCyt, Footnote)), \ 'Header {} inside a <{}> instead of a master.'.format( - etree.tostring(self), self.getparent().tag) + etree.tostring(self, encoding='unicode'), self.getparent().tag) for c in self: if isinstance(c, WLElement): @@ -224,3 +256,75 @@ class WLElement(etree.ElementBase): for e in self: if isinstance(e, WLElement): e.sanitize() + + def snip(self, words, before=None, sub=False): + if sub and self.ASIDE: + return words, [] + + snippet = [] + if before is not None: + i = self.index(before) + else: + i = len(self) + + while i > 0: + i -= 1 + if self[i].tail: + if words: + words, text = last_words(self[i].tail, words) + snippet = [('text', text)] + snippet + + if words: + words, subsnip = self[i].snip(words, sub=True) + snippet = subsnip + snippet + + if words and self.text: + words, text = last_words(self.text, words) + snippet = [('text', text)] + snippet + + snippet = [('start', self.tag, self.attrib)] + snippet + [('end',)] + + if not sub and words and not self.ASIDE: + # do we dare go up? + parent = self.getparent() + if parent is not None and parent.CAN_HAVE_TEXT: + words, parsnip = parent.snip(words, before=self) + return words, parsnip[:-1] + snippet + parsnip[-1:] + + return words, snippet + + def get_snippet(self, words=15): + from librarian.parser import parser + + words, snippet = self.getparent().snip(words=words, before=self) + + cursor = snipelem = parser.makeelement('snippet') + snipelem._meta_object = self.meta + for s in snippet: + if s[0] == 'start': + elem = parser.makeelement(s[1], **s[2]) + cursor.append(elem) + cursor = elem + elif s[0] == 'end': + cursor = cursor.getparent() + else: + if len(cursor): + cursor[-1].tail = (cursor[-1].tail or '') + s[1] + else: + cursor.text = (cursor.text or '') + s[1] + + return snipelem + + def get_link(self): + sec = getattr(self, 'SHOULD_HAVE_ID', False) and self.attrib.get('_compat_section_id') + if sec: + return sec + parent_index = self.getparent().index(self) + if parent_index: + return self.getparent()[parent_index - 1].get_link() + else: + return self.getparent().get_link() + + +class Snippet(WLElement): + pass