X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/0604bdd5f693da9f1c78f9d9fa2276f0c7b6c17b..86becb1ae4d215dd782dfc054471d724f6dcd1d7:/src/librarian/builders/html.py diff --git a/src/librarian/builders/html.py b/src/librarian/builders/html.py index 8015c6a..ed222d3 100644 --- a/src/librarian/builders/html.py +++ b/src/librarian/builders/html.py @@ -1,49 +1,81 @@ +# coding: utf-8 +from __future__ import unicode_literals + +try: + from urllib.request import urlopen +except ImportError: + from urllib2 import urlopen from lxml import etree +from librarian.html import add_anchors, add_table_of_contents, add_table_of_themes from librarian import OutputFile class HtmlBuilder: file_extension = "html" - identifier = "html" + with_anchors = True + with_themes = True + with_toc = True + with_footnotes = True + with_nota_red = True + no_externalities = False + orphans = True - def __init__(self, image_location='https://wolnelektury.pl/media/book/pictures/marcos-historia-kolorow/'): - self.image_location = image_location + def __init__(self, base_url=None): + self._base_url = base_url - #self.tree = etree.Element('html') - #body = etree.SubElement(self.tree, 'body') - #text = etree.SubElement(body, 'div', **{'id': 'book-text'}) self.tree = text = etree.Element('div', **{'id': 'book-text'}) - toc = etree.SubElement(text, 'div', id='toc') - themes = etree.SubElement(text, 'div', id='themes') - h1 = etree.SubElement(text, 'h1') + self.header = etree.SubElement(text, 'h1') + + self.footnotes = etree.Element('div', id='footnotes') + self.footnote_counter = 0 + + self.nota_red = etree.Element('div', id='nota_red') self.cursors = { None: text, - 'toc': toc, - 'themes': themes, - 'header': h1, + 'header': self.header, + 'footnotes': self.footnotes, + 'nota_red': self.nota_red, } - self.current_cursors = [None] + self.current_cursors = [text] + + @property + def base_url(self): + if self._base_url is not None: + return self._base_url + else: + return 'https://wolnelektury.pl/media/book/pictures/{}/'.format(self.document.meta.url.slug) + + @property + def cursor(self): + return self.current_cursors[-1] def enter_fragment(self, fragment): - self.current_cursors.append(fragment) + self.current_cursors.append(self.cursors[fragment]) def exit_fragment(self): self.current_cursors.pop() - - def build(self, document): + + def create_fragment(self, name, element): + assert name not in self.cursors + self.cursors[name] = element + + def forget_fragment(self, name): + del self.cursors[name] + + def preprocess(self, document): + document._compat_assign_ordered_ids() + document._compat_assign_section_ids() + + def build(self, document, **kwargs): + self.document = document + + self.preprocess(document) document.tree.getroot().html_build(self) + self.postprocess(document) + return self.output() - head = etree.Element('head') - self.tree.insert(0, head) - etree.SubElement( - head, - 'link', - href="https://static.wolnelektury.pl/css/compressed/book_text.b15153e56c0a.css", - rel="stylesheet", - type="text/css", - ) - + def output(self): return OutputFile.from_bytes( etree.tostring( self.tree, @@ -53,20 +85,124 @@ class HtmlBuilder: ) ) - def start_element(self, tag, attrib): - self.cursors[self.current_cursors[-1]] = etree.SubElement( - self.cursors[self.current_cursors[-1]], + def postprocess(self, document): + _ = document.tree.getroot().master.gettext + + if document.meta.translators: + self.enter_fragment('header') + self.start_element('span', {'class': 'translator'}) + self.push_text(_("translated by") + " ") + self.push_text( + ", ".join( + translator.readable() + for translator in document.meta.translators + ) + ) + self.exit_fragment() + + if self.with_anchors: + add_anchors(self.tree) + if self.with_nota_red and len(self.nota_red): + self.tree.append(self.nota_red) + if self.with_themes: + add_table_of_themes(self.tree) + if self.with_toc: + add_table_of_contents(self.tree) + + if self.footnote_counter: + fnheader = etree.Element("h3") + fnheader.text = _("Footnotes") + self.footnotes.insert(0, fnheader) + self.tree.append(self.footnotes) + + def start_element(self, tag, attrib=None): + self.current_cursors.append(etree.SubElement( + self.cursor, tag, - **attrib - ) - print(self.cursors) + **(attrib or {}) + )) def end_element(self): - self.cursors[self.current_cursors[-1]] = self.cursors[self.current_cursors[-1]].getparent() + self.current_cursors.pop() def push_text(self, text): - cursor = self.cursors[self.current_cursors[-1]] + cursor = self.cursor if len(cursor): - cursor.tail = (cursor[-1].tail or '') + text + cursor[-1].tail = (cursor[-1].tail or '') + text else: cursor.text = (cursor.text or '') + text + + +class StandaloneHtmlBuilder(HtmlBuilder): + css_url = "https://static.wolnelektury.pl/css/compressed/book_text.css" + + def postprocess(self, document): + super(StandaloneHtmlBuilder, self).postprocess(document) + + tree = etree.Element('html') + body = etree.SubElement(tree, 'body') + body.append(self.tree) + self.tree = tree + + head = etree.Element('head') + tree.insert(0, head) + + + etree.SubElement(head, 'meta', charset='utf-8') + etree.SubElement(head, 'title').text = document.meta.title + + etree.SubElement( + head, + 'meta', + name="viewport", + content="width=device-width, initial-scale=1, maximum-scale=1" + ) + + if self.no_externalities: + etree.SubElement( + head, 'style', + ).text = urlopen(self.css_url).read().decode('utf-8') + else: + etree.SubElement( + head, + 'link', + href=self.css_url, + rel="stylesheet", + type="text/css", + ) + + etree.SubElement( + body, 'script', + src="https://ajax.googleapis.com/ajax/libs/jquery/1/jquery.min.js" + ) + + etree.SubElement( + body, + "script", + src="http://malsup.github.io/min/jquery.cycle2.min.js" + ) + + +class DaisyHtmlBuilder(StandaloneHtmlBuilder): + file_extension = 'xhtml' + with_anchors = False + with_themes = False + with_toc = False + with_footnotes = False + with_nota_red = False + with_deep_identifiers = False + no_externalities = True + + def output(self): + tree = etree.ElementTree(self.tree) + tree.docinfo.public_id = '-//W3C//DTD XHTML 1.0 Transitional//EN' + tree.docinfo.system_url = 'http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd' + return OutputFile.from_bytes( + etree.tostring( + tree, + encoding='utf-8', + pretty_print=True, + xml_declaration=True + ) + ) +