def commit_toc(container, toc, lang=None, uid=None):
tocname = find_existing_toc(container)
if tocname is None:
item = container.generate_item("toc.ncx", id_prefix="toc")
tocname = container.href_to_name(item.get("href"), base=container.opf_name)
if not lang:
lang = get_lang()
for l in container.opf_xpath("//dc:language"):
l = canonicalize_lang(xml2text(l).strip())
if l:
lang = l
lang = lang_as_iso639_1(l) or l
break
lang = lang_as_iso639_1(lang) or lang
if not uid:
uid = uuid_id()
eid = container.opf.get("unique-identifier", None)
if eid:
m = container.opf_xpath('//*[@id="%s"]' % eid)
if m:
uid = xml2text(m[0])
title = _("Table of Contents")
m = container.opf_xpath("//dc:title")
if m:
x = xml2text(m[0]).strip()
title = x or title
to_href = partial(container.name_to_href, base=tocname)
root = create_ncx(toc, to_href, title, lang, uid)
container.replace(tocname, root)
container.pretty_print.add(tocname)
def commit_ncx_toc(container, toc, lang=None, uid=None):
tocname = find_existing_ncx_toc(container)
if tocname is None:
item = container.generate_item('toc.ncx', id_prefix='toc')
tocname = container.href_to_name(item.get('href'), base=container.opf_name)
ncx_id = item.get('id')
[s.set('toc', ncx_id) for s in container.opf_xpath('//opf:spine')]
if not lang:
lang = get_lang()
for l in container.opf_xpath('//dc:language'):
l = canonicalize_lang(xml2text(l).strip())
if l:
lang = l
lang = lang_as_iso639_1(l) or l
break
lang = lang_as_iso639_1(lang) or lang
if not uid:
uid = uuid_id()
eid = container.opf.get('unique-identifier', None)
if eid:
m = container.opf_xpath('//*[@id="%s"]'%eid)
if m:
uid = xml2text(m[0])
title = _('Table of Contents')
m = container.opf_xpath('//dc:title')
if m:
x = xml2text(m[0]).strip()
title = x or title
to_href = partial(container.name_to_href, base=tocname)
root = create_ncx(toc, to_href, title, lang, uid)
container.replace(tocname, root)
container.pretty_print.add(tocname)
def detect_chapters(self):
self.detected_chapters = []
def find_matches(expr, doc):
try:
return XPath(expr)(doc)
except:
self.log.warn('Invalid chapter expression, ignoring: %s'%expr)
return []
if self.opts.chapter:
for item in self.oeb.spine:
for x in find_matches(self.opts.chapter, item.data):
self.detected_chapters.append((item, x))
chapter_mark = self.opts.chapter_mark
page_break_before = 'display: block; page-break-before: always'
page_break_after = 'display: block; page-break-after: always'
for item, elem in self.detected_chapters:
text = xml2text(elem).strip()
text = re.sub(r'\s+', ' ', text.strip())
self.log('\tDetected chapter:', text[:50])
if chapter_mark == 'none':
continue
elif chapter_mark == 'rule':
mark = etree.Element(XHTML('hr'))
elif chapter_mark == 'pagebreak':
mark = etree.Element(XHTML('div'), style=page_break_after)
else: # chapter_mark == 'both':
mark = etree.Element(XHTML('hr'), style=page_break_before)
try:
elem.addprevious(mark)
except TypeError:
self.log.exception('Failed to mark chapter')
def remove_first_image(self):
deleted_item = None
for item in self.oeb.spine:
if XPath(JACKET_XPATH)(item.data):
continue
removed = self.remove_images(item)
if removed > 0:
self.log('Removed first image')
body = XPath('//h:body')(item.data)
if body:
raw = xml2text(body[0]).strip()
imgs = XPath('//h:img|//svg:svg')(item.data)
if not raw and not imgs:
self.log('Removing %s as it has no content'%item.href)
self.oeb.manifest.remove(item)
deleted_item = item
break
else:
self.log.warn('Could not find first image to remove')
if deleted_item is not None:
for item in list(self.oeb.toc):
href = urldefrag(item.href)[0]
if href == deleted_item.href:
self.oeb.toc.remove(item)
self.oeb.guide.remove_by_href(deleted_item.href)
def create_toc_from_links(self):
num = 0
for item in self.oeb.spine:
for a in XPath('//h:a[@href]')(item.data):
href = a.get('href')
try:
purl = urlparse(href)
except ValueError:
self.log.warning('Ignoring malformed URL:', href)
continue
if not purl[0] or purl[0] == 'file':
href, frag = purl.path, purl.fragment
href = item.abshref(href)
if frag:
href = '#'.join((href, frag))
if not self.oeb.toc.has_href(href):
text = xml2text(a)
text = text[:100].strip()
if (not self.opts.duplicate_links_in_toc and
self.oeb.toc.has_text(text)):
continue
try:
self.oeb.toc.add(text, href,
play_order=self.oeb.toc.next_play_order())
num += 1
except ValueError:
self.oeb.log.exception('Failed to process link: %r' % href)
continue # Most likely an incorrectly URL encoded link
if self.opts.max_toc_links > 0 and \
num >= self.opts.max_toc_links:
self.log('Maximum TOC links reached, stopping.')
return
def _toc_from_html(self, opf):
if 'toc' not in self.oeb.guide:
return False
self.log.debug('Reading TOC from HTML...')
itempath, frag = urldefrag(self.oeb.guide['toc'].href)
item = self.oeb.manifest.hrefs[itempath]
html = item.data
if frag:
elems = xpath(html, './/*[@id="%s"]' % frag)
if not elems:
elems = xpath(html, './/*[@name="%s"]' % frag)
elem = elems[0] if elems else html
while elem != html and not xpath(elem, './/h:a[@href]'):
elem = elem.getparent()
html = elem
titles = defaultdict(list)
order = []
for anchor in xpath(html, './/h:a[@href]'):
href = anchor.attrib['href']
href = item.abshref(urlnormalize(href))
path, frag = urldefrag(href)
if path not in self.oeb.manifest.hrefs:
continue
title = xml2text(anchor)
title = COLLAPSE_RE.sub(' ', title.strip())
if href not in titles:
order.append(href)
titles[href].append(title)
toc = self.oeb.toc
for href in order:
toc.add(' '.join(titles[href]), href)
return True
def detect_chapters(self):
self.detected_chapters = []
if self.opts.chapter:
chapter_xpath = XPath(self.opts.chapter)
for item in self.oeb.spine:
for x in chapter_xpath(item.data):
self.detected_chapters.append((item, x))
chapter_mark = self.opts.chapter_mark
page_break_before = 'display: block; page-break-before: always'
page_break_after = 'display: block; page-break-after: always'
for item, elem in self.detected_chapters:
text = xml2text(elem).strip()
text = re.sub(r'\s+', ' ', text.strip())
self.log('\tDetected chapter:', text[:50])
if chapter_mark == 'none':
continue
elif chapter_mark == 'rule':
mark = etree.Element(XHTML('hr'))
elif chapter_mark == 'pagebreak':
mark = etree.Element(XHTML('div'), style=page_break_after)
else: # chapter_mark == 'both':
mark = etree.Element(XHTML('hr'), style=page_break_before)
try:
elem.addprevious(mark)
except TypeError:
self.log.exception('Failed to mark chapter')
def description_for_anchor(elem):
def check(x, min_len=4):
if x:
x = x.strip()
if len(x) >= min_len:
return x[:30]
desc = check(elem.get('title'))
if desc is not None:
return desc
desc = check(elem.text)
if desc is not None:
return desc
if len(elem) > 0:
desc = check(elem[0].text)
if desc is not None:
return desc
# Get full text for tags that have only a few descendants
for i, x in enumerate(elem.iterdescendants('*')):
if i > 5:
break
else:
desc = check(xml2text(elem), min_len=1)
if desc is not None:
return desc
def elem_to_toc_text(elem):
text = xml2text(elem).strip()
if not text:
text = elem.get('title', '')
if not text:
text = elem.get('alt', '')
text = re.sub(r'\s+', ' ', text.strip())
text = text[:1000].strip()
if not text:
text = _('(Untitled)')
return text
def elem_to_toc_text(elem):
text = xml2text(elem).strip()
if not text:
text = elem.get("title", "")
if not text:
text = elem.get("alt", "")
text = re.sub(r"\s+", " ", text.strip())
text = text[:1000].strip()
if not text:
text = _("(Untitled)")
return text
def elem_to_link(self, item, elem, counter):
text = xml2text(elem).strip()
if not text:
text = elem.get('title', '')
if not text:
text = elem.get('alt', '')
text = re.sub(r'\s+', ' ', text.strip())
text = text[:1000].strip()
id = elem.get('id', 'calibre_toc_%d'%counter)
elem.set('id', id)
href = '#'.join((item.href, id))
return text, href
def elem_to_link(self, item, elem, counter):
text = xml2text(elem).strip()
if not text:
text = elem.get("title", "")
if not text:
text = elem.get("alt", "")
text = re.sub(r"\s+", " ", text.strip())
text = text[:1000].strip()
id = elem.get("id", "calibre_toc_%d" % counter)
elem.set("id", id)
href = "#".join((item.href, id))
return text, href
def detect_chapters(self):
self.detected_chapters = []
self.chapter_title_attribute = None
def find_matches(expr, doc):
try:
ans = XPath(expr)(doc)
len(ans)
return ans
except:
self.log.warn('Invalid chapter expression, ignoring: %s'%expr)
return []
if self.opts.chapter:
chapter_path, title_attribute = self.get_toc_parts_for_xpath(self.opts.chapter)
self.chapter_title_attribute = title_attribute
for item in self.oeb.spine:
for x in find_matches(chapter_path, item.data):
self.detected_chapters.append((item, x))
chapter_mark = self.opts.chapter_mark
page_break_before = 'display: block; page-break-before: always'
page_break_after = 'display: block; page-break-after: always'
c = Counter()
for item, elem in self.detected_chapters:
c[item] += 1
text = xml2text(elem).strip()
text = re.sub(r'\s+', ' ', text.strip())
self.log('\tDetected chapter:', text[:50])
if chapter_mark == 'none':
continue
if chapter_mark == 'rule':
mark = elem.makeelement(XHTML('hr'))
elif chapter_mark == 'pagebreak':
if c[item] < 3 and at_start(elem):
# For the first two elements in this item, check if they
# are at the start of the file, in which case inserting a
# page break in unnecessary and can lead to extra blank
# pages in the PDF Output plugin. We need to use two as
# feedbooks epubs match both a heading tag and its
# containing div with the default chapter expression.
continue
mark = elem.makeelement(XHTML('div'), style=page_break_after)
else: # chapter_mark == 'both':
mark = elem.makeelement(XHTML('hr'), style=page_break_before)
try:
elem.addprevious(mark)
except TypeError:
self.log.exception('Failed to mark chapter')
def detect_chapters(self):
self.detected_chapters = []
self.chapter_title_attribute = None
def find_matches(expr, doc):
try:
ans = XPath(expr)(doc)
len(ans)
return ans
except:
self.log.warn('Invalid chapter expression, ignoring: %s'%expr)
return []
if self.opts.chapter:
chapter_path, title_attribute = self.get_toc_parts_for_xpath(self.opts.chapter)
self.chapter_title_attribute = title_attribute
for item in self.oeb.spine:
for x in find_matches(chapter_path, item.data):
self.detected_chapters.append((item, x))
chapter_mark = self.opts.chapter_mark
page_break_before = 'display: block; page-break-before: always'
page_break_after = 'display: block; page-break-after: always'
c = Counter()
for item, elem in self.detected_chapters:
c[item] += 1
text = xml2text(elem).strip()
text = re.sub(r'\s+', ' ', text.strip())
self.log('\tDetected chapter:', text[:50])
if chapter_mark == 'none':
continue
if chapter_mark == 'rule':
mark = elem.makeelement(XHTML('hr'))
elif chapter_mark == 'pagebreak':
if c[item] < 3 and at_start(elem):
# For the first two elements in this item, check if they
# are at the start of the file, in which case inserting a
# page break in unnecessary and can lead to extra blank
# pages in the PDF Output plugin. We need to use two as
# feedbooks epubs match both a heading tag and its
# containing div with the default chapter expression.
continue
mark = elem.makeelement(XHTML('div'), style=page_break_after)
else: # chapter_mark == 'both':
mark = elem.makeelement(XHTML('hr'), style=page_break_before)
try:
elem.addprevious(mark)
except TypeError:
self.log.exception('Failed to mark chapter')
def find_text(node):
LIMIT = 200
pat = re.compile(r'\s+')
for child in node:
if isinstance(child, etree._Element):
text = xml2text(child).strip()
text = pat.sub(' ', text)
if len(text) < 1:
continue
if len(text) > LIMIT:
# Look for less text in a child of this node, recursively
ntext = find_text(child)
return ntext or (text[:LIMIT] + '...')
else:
return text
def find_text(node):
LIMIT = 200
pat = re.compile(r'\s+')
for child in node:
if isinstance(child, etree._Element):
text = xml2text(child).strip()
text = pat.sub(' ', text)
if len(text) < 1:
continue
if len(text) > LIMIT:
# Look for less text in a child of this node, recursively
ntext = find_text(child)
return ntext or (text[:LIMIT] + '...')
else:
return text
def elem_to_link(self, item, elem, title_attribute, counter):
text = ''
if title_attribute is not None:
text = elem.get(title_attribute, '')
if not text:
text = xml2text(elem).strip()
if not text:
text = elem.get('title', '')
if not text:
text = elem.get('alt', '')
text = re.sub(r'\s+', ' ', text.strip())
text = text[:1000].strip()
id = elem.get('id', 'calibre_toc_%d' % counter)
elem.set('id', id)
href = '#'.join((item.href, id))
return text, href
def detect_chapters(self):
self.detected_chapters = []
def find_matches(expr, doc):
try:
ans = XPath(expr)(doc)
len(ans)
return ans
except:
self.log.warn("Invalid chapter expression, ignoring: %s" % expr)
return []
if self.opts.chapter:
for item in self.oeb.spine:
for x in find_matches(self.opts.chapter, item.data):
self.detected_chapters.append((item, x))
chapter_mark = self.opts.chapter_mark
page_break_before = "display: block; page-break-before: always"
page_break_after = "display: block; page-break-after: always"
c = Counter()
for item, elem in self.detected_chapters:
c[item] += 1
text = xml2text(elem).strip()
text = re.sub(r"\s+", " ", text.strip())
self.log("\tDetected chapter:", text[:50])
if chapter_mark == "none":
continue
if chapter_mark == "rule":
mark = etree.Element(XHTML("hr"))
elif chapter_mark == "pagebreak":
if c[item] < 3 and at_start(elem):
# For the first two elements in this item, check if they
# are at the start of the file, in which case inserting a
# page break in unnecessary and can lead to extra blank
# pages in the PDF Output plugin. We need to use two as
# feedbooks epubs match both a heading tag and its
# containing div with the default chapter expression.
continue
mark = etree.Element(XHTML("div"), style=page_break_after)
else: # chapter_mark == 'both':
mark = etree.Element(XHTML("hr"), style=page_break_before)
try:
elem.addprevious(mark)
except TypeError:
self.log.exception("Failed to mark chapter")
def find_cover_image_in_page(container, cover_page):
root = container.parsed(cover_page)
body = XPath('//h:body')(root)
if len(body) != 1:
return
body = body[0]
images = []
for img in XPath('descendant::h:img[@src]|descendant::svg:svg/descendant::svg:image')(body):
href = img.get('src') or img.get(XLINK('href'))
if href:
name = container.href_to_name(href, base=cover_page)
images.append(name)
text = re.sub(r'\s+', '', xml2text(body))
if text or len(images) > 1:
# Document has more content than a single image
return
if images:
return images[0]
def find_cover_image_in_page(container, cover_page):
root = container.parsed(cover_page)
body = XPath('//h:body')(root)
if len(body) != 1:
return
body = body[0]
images = []
for img in XPath('descendant::h:img[@src]|descendant::svg:svg/descendant::svg:image')(body):
href = img.get('src') or img.get(XLINK('href'))
if href:
name = container.href_to_name(href, base=cover_page)
images.append(name)
text = re.sub(r'\s+', '', xml2text(body))
if text or len(images) > 1:
# Document has more content than a single image
return
if images:
return images[0]
def remove_first_image(self):
deleted_item = None
for item in self.oeb.spine:
removed = self.remove_images(item)
if removed > 0:
self.log('Removed first image')
body = XPath('//h:body')(item.data)
if body:
raw = xml2text(body[0]).strip()
imgs = XPath('//h:img|//svg:svg')(item.data)
if not raw and not imgs:
self.log('Removing %s as it has no content'%item.href)
self.oeb.manifest.remove(item)
deleted_item = item
break
if deleted_item is not None:
for item in list(self.oeb.toc):
href = urldefrag(item.href)[0]
if href == deleted_item.href:
self.oeb.toc.remove(item)
def remove_first_image(self):
deleted_item = None
for item in self.oeb.spine:
removed = self.remove_images(item)
if removed > 0:
self.log('Removed first image')
body = XPath('//h:body')(item.data)
if body:
raw = xml2text(body[0]).strip()
imgs = XPath('//h:img|//svg:svg')(item.data)
if not raw and not imgs:
self.log('Removing %s as it has no content'%item.href)
self.oeb.manifest.remove(item)
deleted_item = item
break
if deleted_item is not None:
for item in list(self.oeb.toc):
href = urldefrag(item.href)[0]
if href == deleted_item.href:
self.oeb.toc.remove(item)
def create_toc_from_links(self):
num = 0
for item in self.oeb.spine:
for a in XPath("//h:a[@href]")(item.data):
href = a.get("href")
purl = urlparse(href)
if not purl[0] or purl[0] == "file":
href, frag = purl.path, purl.fragment
href = item.abshref(href)
if frag:
href = "#".join((href, frag))
if not self.oeb.toc.has_href(href):
text = xml2text(a)
text = text[:100].strip()
if not self.opts.duplicate_links_in_toc and self.oeb.toc.has_text(text):
continue
num += 1
self.oeb.toc.add(text, href, play_order=self.oeb.toc.next_play_order())
if self.opts.max_toc_links > 0 and num >= self.opts.max_toc_links:
self.log("Maximum TOC links reached, stopping.")
return
def detect_chapters(self):
self.detected_chapters = []
def find_matches(expr, doc):
try:
ans = XPath(expr)(doc)
len(ans)
return ans
except:
self.log.warn('Invalid chapter expression, ignoring: %s'%expr)
return []
if self.opts.chapter:
for item in self.oeb.spine:
for x in find_matches(self.opts.chapter, item.data):
self.detected_chapters.append((item, x))
chapter_mark = self.opts.chapter_mark
page_break_before = 'display: block; page-break-before: always'
page_break_after = 'display: block; page-break-after: always'
for item, elem in self.detected_chapters:
text = xml2text(elem).strip()
text = re.sub(r'\s+', ' ', text.strip())
self.log('\tDetected chapter:', text[:50])
if chapter_mark == 'none':
continue
elif chapter_mark == 'rule':
mark = etree.Element(XHTML('hr'))
elif chapter_mark == 'pagebreak':
mark = etree.Element(XHTML('div'), style=page_break_after)
else: # chapter_mark == 'both':
mark = etree.Element(XHTML('hr'), style=page_break_before)
try:
elem.addprevious(mark)
except TypeError:
self.log.exception('Failed to mark chapter')
def create_toc_from_links(self):
num = 0
for item in self.oeb.spine:
for a in XPath('//h:a[@href]')(item.data):
href = a.get('href')
purl = urlparse(href)
if not purl[0] or purl[0] == 'file':
href, frag = purl.path, purl.fragment
href = item.abshref(href)
if frag:
href = '#'.join((href, frag))
if not self.oeb.toc.has_href(href):
text = xml2text(a)
text = text[:100].strip()
if (not self.opts.duplicate_links_in_toc and
self.oeb.toc.has_text(text)):
continue
num += 1
self.oeb.toc.add(text, href,
play_order=self.oeb.toc.next_play_order())
if self.opts.max_toc_links > 0 and \
num >= self.opts.max_toc_links:
self.log('Maximum TOC links reached, stopping.')
return
def read_inline_toc(self, href, frag):
ans = TOC()
base_href = '/'.join(href.split('/')[:-1])
with open(href.replace('/', os.sep), 'rb') as f:
raw = f.read().decode(self.header.codec)
root = parse_html(raw, log=self.log)
body = XPath('//h:body')(root)
reached = False
if body:
start = body[0]
else:
start = None
reached = True
if frag:
elems = XPath('//*[@id="%s"]' % frag)(root)
if elems:
start = elems[0]
def node_depth(elem):
ans = 0
parent = elem.getparent()
while parent is not None:
parent = parent.getparent()
ans += 1
return ans
# Layer the ToC based on nesting order in the source HTML
current_depth = None
parent = ans
seen = set()
links = []
for elem in root.iterdescendants(etree.Element):
if reached and elem.tag == XHTML('a') and elem.get('href', False):
href = elem.get('href')
href, frag = urldefrag(href)
href = base_href + '/' + href
text = xml2text(elem).strip()
if (text, href, frag) in seen:
continue
seen.add((text, href, frag))
links.append((text, href, frag, node_depth(elem)))
elif elem is start:
reached = True
depths = sorted(set(x[-1] for x in links))
depth_map = {x: i for i, x in enumerate(depths)}
for text, href, frag, depth in links:
depth = depth_map[depth]
if current_depth is None:
current_depth = 0
parent.add_item(href, frag, text)
elif current_depth == depth:
parent.add_item(href, frag, text)
elif current_depth < depth:
parent = parent[-1] if len(parent) > 0 else parent
parent.add_item(href, frag, text)
current_depth += 1
else:
delta = current_depth - depth
while delta > 0 and parent.parent is not None:
parent = parent.parent
delta -= 1
parent.add_item(href, frag, text)
current_depth = depth
return ans
def read_inline_toc(self, href, frag):
ans = TOC()
base_href = '/'.join(href.split('/')[:-1])
with open(href.replace('/', os.sep), 'rb') as f:
raw = f.read().decode(self.header.codec)
root = parse_html(raw, log=self.log)
body = XPath('//h:body')(root)
reached = False
if body:
start = body[0]
else:
start = None
reached = True
if frag:
elems = XPath('//*[@id="%s"]'%frag)(root)
if elems:
start = elems[0]
def node_depth(elem):
ans = 0
parent = elem.getparent()
while parent is not None:
parent = parent.getparent()
ans += 1
return ans
# Layer the ToC based on nesting order in the source HTML
current_depth = None
parent = ans
seen = set()
links = []
for elem in root.iterdescendants(etree.Element):
if reached and elem.tag == XHTML('a') and elem.get('href',
False):
href = elem.get('href')
href, frag = urldefrag(href)
href = base_href + '/' + href
text = xml2text(elem).strip()
if (text, href, frag) in seen:
continue
seen.add((text, href, frag))
links.append((text, href, frag, node_depth(elem)))
elif elem is start:
reached = True
depths = sorted(set(x[-1] for x in links))
depth_map = {x:i for i, x in enumerate(depths)}
for text, href, frag, depth in links:
depth = depth_map[depth]
if current_depth is None:
current_depth = 0
parent.add_item(href, frag, text)
elif current_depth == depth:
parent.add_item(href, frag, text)
elif current_depth < depth:
parent = parent[-1] if len(parent) > 0 else parent
parent.add_item(href, frag, text)
current_depth += 1
else:
delta = current_depth - depth
while delta > 0 and parent.parent is not None:
parent = parent.parent
delta -= 1
parent.add_item(href, frag, text)
current_depth = depth
return ans
