diff --git a/src/calibre/ebooks/fb2/fb2ml.py b/src/calibre/ebooks/fb2/fb2ml.py index 252453d25e..f27729ae8c 100644 --- a/src/calibre/ebooks/fb2/fb2ml.py +++ b/src/calibre/ebooks/fb2/fb2ml.py @@ -26,32 +26,6 @@ from calibre.ebooks.oeb.base import XHTML, XHTML_NS, barename, namespace from calibre.ebooks.oeb.stylizer import Stylizer from calibre.ebooks.oeb.base import OEB_RASTER_IMAGES -TAG_MAP = { - 'b' : 'strong', - 'i' : 'emphasis', - 'p' : 'p', - 'li' : 'p', - 'div': 'p', -} - -TAG_SPACE = [] - -TAG_IMAGES = [ - 'img', -] - -TAG_LINKS = [ -] - -BLOCK = [ - 'p', -] - -STYLES = [ - ('font-weight', {'bold' : 'strong', 'bolder' : 'strong'}), - ('font-style', {'italic' : 'emphasis'}), -] - class FB2MLizer(object): ''' Todo: * Include more FB2 specific tags in the conversion. @@ -63,24 +37,32 @@ class FB2MLizer(object): def __init__(self, log): self.log = log self.image_hrefs = {} + self.reset_state() + + def reset_state(self): # Used to ensure text and tags are always within

and

self.in_p = False + # Mapping of image names. OEB allows for images to have the same name but be stored + # in different directories. FB2 images are all in a flat layout so we rename all images + # into a sequential numbering system to ensure there are no collisions between image names. + self.image_hrefs = {} def extract_content(self, oeb_book, opts): self.log.info('Converting XHTML to FB2 markup...') self.oeb_book = oeb_book self.opts = opts + return self.fb2mlize_spine() def fb2mlize_spine(self): - self.image_hrefs = {} - self.link_hrefs = {} + self.reset_state() + output = [self.fb2_header()] output.append(self.get_text()) - output.append(self.fb2_body_footer()) output.append(self.fb2mlize_images()) output.append(self.fb2_footer()) output = self.clean_text(u''.join(output)) + if self.opts.pretty_print: return u'\n%s' % etree.tostring(etree.fromstring(output), encoding=unicode, pretty_print=True) else: @@ -127,24 +109,21 @@ class FB2MLizer(object): '' \ '%s %s' \ '' \ - '' % tuple(map(prepare_string_for_xml, (author_first, author_middle, author_last, + '' % tuple(map(prepare_string_for_xml, (author_first, author_middle, author_last, self.oeb_book.metadata.title[0].value, __appname__, __version__))) + def fb2_footer(self): + return u'' + def get_text(self): - text = [] + text = [''] for item in self.oeb_book.spine: self.log.debug('Converting %s to FictionBook2 XML' % item.href) stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts, self.opts.output_profile) text.append('
') text += self.dump_text(item.data.find(XHTML('body')), stylizer, item) text.append('
') - return ''.join(text) - - def fb2_body_footer(self): - return u'' - - def fb2_footer(self): - return u'' + return ''.join(text) + '' def fb2mlize_images(self): images = [] @@ -178,24 +157,6 @@ class FB2MLizer(object): else: self.in_p = True return ['

'], ['p'] - - def insert_empty_line(self, tags): - if self.in_p: - text = [''] - closed_tags = [] - tags.reverse() - for t in tags: - text.append('' % t) - closed_tags.append(t) - if t == 'p': - break - text.append('') - closed_tags.reverse() - for t in closed_tags: - text.append('<%s>' % t) - return text - else: - return [''] def close_open_p(self, tags): text = [''] @@ -219,84 +180,125 @@ class FB2MLizer(object): self.in_p = True return text, added_p + + def handle_simple_tag(self, tag, tags): + s_out = [] + s_tags = [] + if tag not in tags: + p_out, p_tags = self.ensure_p() + s_out += p_out + s_tags += p_tags + s_out.append('<%s>' % tag) + s_tags.append(tag) + return s_out, s_tags - def dump_text(self, elem, stylizer, page, tag_stack=[]): - if not isinstance(elem.tag, basestring) \ - or namespace(elem.tag) != XHTML_NS: + def dump_text(self, elem_tree, stylizer, page, tag_stack=[]): + ''' + This function is intended to be used in a recursive manner. dump_text will + run though all elements in the elem_tree and call itself on each element. + + self.image_hrefs will be populated by calling this function. + + @param elem_tree: etree representation of XHTML content to be transformed. + @param stylizer: Used to track the style of elements within the tree. + @param page: OEB page used to determine absolute urls. + @param tag_stack: List of open FB2 tags to take into account. + + @return: List of string representing the XHTML converted to FB2 markup. + ''' + # Ensure what we are converting is not a string and that the fist tag is part of the XHTML namespace. + if not isinstance(elem_tree.tag, basestring) or namespace(elem_tree.tag) != XHTML_NS: return [] - style = stylizer.style(elem) - if style['display'] in ('none', 'oeb-page-head', 'oeb-page-foot') \ - or style['visibility'] == 'hidden': + style = stylizer.style(elem_tree) + if style['display'] in ('none', 'oeb-page-head', 'oeb-page-foot') or style['visibility'] == 'hidden': return [] - fb2_text = [] + # FB2 generated output. + fb2_out = [] + # FB2 tags in the order they are opened. This will be used to close the tags. tags = [] + # First tag in tree + tag = barename(elem_tree.tag) - tag = barename(elem.tag) - - if tag in TAG_IMAGES: - if elem.attrib.get('src', None): - if page.abshref(elem.attrib['src']) not in self.image_hrefs.keys(): - self.image_hrefs[page.abshref(elem.attrib['src'])] = '_%s.jpg' % len(self.image_hrefs.keys()) - p_txt, p_tag = self.ensure_p() - fb2_text += p_txt - tags += p_tag - fb2_text.append('' % self.image_hrefs[page.abshref(elem.attrib['src'])]) - + # Process the XHTML tag if it needs to be converted to an FB2 tag. if tag == 'h1' and self.opts.h1_to_title or tag == 'h2' and self.opts.h2_to_title or tag == 'h3' and self.opts.h3_to_title: - fb2_text.append('') + fb2_out.append('<title>') tags.append('title') - if tag == 'br': - fb2_text += self.insert_empty_line(tag_stack+tags) - - fb2_tag = TAG_MAP.get(tag, None) - if fb2_tag == 'p': + if tag == 'img': + # TODO: Check that the image is in the manifest and only write the tag if it is. + if elem_tree.attrib.get('src', None): + if page.abshref(elem_tree.attrib['src']) not in self.image_hrefs.keys(): + self.image_hrefs[page.abshref(elem_tree.attrib['src'])] = '_%s.jpg' % len(self.image_hrefs.keys()) + p_txt, p_tag = self.ensure_p() + fb2_out += p_txt + tags += p_tag + fb2_out.append('<image xlink:href="#%s" />' % self.image_hrefs[page.abshref(elem_tree.attrib['src'])]) + elif tag == 'br': + if self.in_p: + closed_tags = [] + open_tags = tag_stack+tags + open_tags.reverse() + for t in open_tags: + fb2_out.append('</%s>' % t) + closed_tags.append(t) + if t == 'p': + break + fb2_out.append('<empty-line />') + closed_tags.reverse() + for t in closed_tags: + fb2_out.append('<%s>' % t) + else: + fb2_out.append('<empty-line />') + elif tag in ('div', 'li', 'p'): p_text, added_p = self.close_open_p(tag_stack+tags) - fb2_text += p_text + fb2_out += p_text if added_p: tags.append('p') - elif fb2_tag and fb2_tag not in tag_stack+tags: - p_text, p_tags = self.ensure_p() - fb2_text += p_text - tags += p_tags - fb2_text.append('<%s>' % fb2_tag) - tags.append(fb2_tag) + elif tag == 'b': + s_out, s_tags = self.handle_simple_tag('strong', tag_stack+tags) + fb2_out += s_out + tags += s_tags + elif tag == 'i': + s_out, s_tags = self.handle_simple_tag('emphasis', tag_stack+tags) + fb2_out += s_out + tags += s_tags - # Processes style information - for s in STYLES: - style_tag = s[1].get(style[s[0]], None) - if style_tag and style_tag not in tag_stack+tags: - p_text, p_tags = self.ensure_p() - fb2_text += p_text - tags += p_tags - fb2_text.append('<%s>' % style_tag) - tags.append(style_tag) - - if tag in TAG_SPACE: - fb2_text.append(' ') - - if hasattr(elem, 'text') and elem.text: + # Processes style information. + if style['font-style'] == 'italic': + s_out, s_tags = self.handle_simple_tag('emphasis', tag_stack+tags) + fb2_out += s_out + tags += s_tags + elif style['font-weight'] in ('bold', 'bolder'): + s_out, s_tags = self.handle_simple_tag('strong', tag_stack+tags) + fb2_out += s_out + tags += s_tags + + # Process element text. + if hasattr(elem_tree, 'text') and elem_tree.text: if not self.in_p: - fb2_text.append('<p>') - fb2_text.append(prepare_string_for_xml(elem.text)) + fb2_out.append('<p>') + fb2_out.append(prepare_string_for_xml(elem_tree.text)) if not self.in_p: - fb2_text.append('</p>') + fb2_out.append('</p>') - for item in elem: - fb2_text += self.dump_text(item, stylizer, page, tag_stack+tags) + # Process sub-elements. + for item in elem_tree: + fb2_out += self.dump_text(item, stylizer, page, tag_stack+tags) + # Close open FB2 tags. tags.reverse() - fb2_text += self.close_tags(tags) + fb2_out += self.close_tags(tags) - if hasattr(elem, 'tail') and elem.tail: + # Process element text that comes after the close of the XHTML tag but before the next XHTML tag. + if hasattr(elem_tree, 'tail') and elem_tree.tail: if not self.in_p: - fb2_text.append('<p>') - fb2_text.append(prepare_string_for_xml(elem.tail)) + fb2_out.append('<p>') + fb2_out.append(prepare_string_for_xml(elem_tree.tail)) if not self.in_p: - fb2_text.append('</p>') + fb2_out.append('</p>') - return fb2_text + return fb2_out def close_tags(self, tags): text = []