def _load_content(self, page_generator):
content = page_generator.content
# Initialize the current entry
current_entry = {}
# Find the <head> and <body> elements.
(headElement,) = page_generator.content.getElementsByTagName('head')
(bodyElement,) = page_generator.content.getElementsByTagName('body')
# Find the <atom:entry> element inside the <head> element.
entries = headElement.getElementsByTagNameNS(ATOM_NAMESPACE, 'entry')
if not entries:
# No Atom feed entry. Do nothing.
self._clear_entry_data(page_generator)
return
elif len(entries) > 1:
# There should only be one atom:entry
raise FGValueError("Too many Atom entries in %s" % (page_generator.path_info.source_filename,))
(entryElement,) = entries
# Store the <atom:entry> element (with all namespace information included)
dummyDocument = minidom.parseString('<dummy/>')
new_entryElement = dummyDocument.importNode(entryElement, True)
dummyDocument.documentElement.appendChild(new_entryElement)
normalize_namespaces(new_entryElement)
current_entry['atom:entry'] = new_entryElement.toxml()
# Find and store the page summary (if any) in the <body> element, and un-set class="feed-summary".
summaryElements = list(find_elements_with_class(content, "feed-summary", remove=True))
if len(summaryElements) > 1:
# There should only be one element with class="feed-summary"
raise FGValueError('Too many elements have class="feed-summary" in %s' % (page_generator.path_info.source_filename,))
elif summaryElements:
# Save the summary
current_entry['summary'] = summaryElements[0].toxml()
else:
# Save an empty summary
current_entry['summary'] = None
# Save the page body
current_entry['body'] = bodyElement.toxml()
# Save the page title
(titleElement,) = getChildElementsNS(headElement, EMPTY_NAMESPACE, 'title')
current_entry['title'] = getChildText(titleElement)
# Save the path_info
current_entry['path_info'] = page_generator.path_info
# Perform some early processing
self._early_process_entry(page_generator, current_entry)
# Write the entry to disk
self._write_entry_data(page_generator, current_entry)
def __invoke_callback(self, page_generator, element, callback):
try:
callback(page_generator, element)
except ReplaceWithNothing as exc:
# Remove the element
element.parentNode.removeChild(element)
except ReplaceWithText as exc:
# Replace the element with the given text
text_node = element.ownerDocument.createTextNode(exc.text)
element.parentNode.replaceChild(text_node, element)
except ReplaceWithNode as exc:
# Replace the element with the given node
new_node = exc.node
if new_node.ownerDocument is not element.ownerDocument:
new_node = element.ownerDocument.importNode(new_node, True)
element.parentNode.replaceChild(new_node, element)
if exc.fix_namespaces:
# page_generator.content uses HTML without specifying a namespace
substitute_namespaces(new_node, {XHTML_NAMESPACE: EMPTY_NAMESPACE})
normalize_namespaces(new_node, strip_dups=True)
except ReplaceWithHTML as exc:
#
# Replace the element with the given HTML code
#
# Parse with TagSoupToXml
p = TagSoupToXml(omit_comments=exc.omit_comments)
p.feed(exc.html)
p.close()
# Get a DOM document
doc = p.todocument()
# Find the <body> element
for node in doc.documentElement.childNodes:
if node.nodeType != node.ELEMENT_NODE:
continue
if node.localName == 'body':
bodyElement = node
break
else:
raise AssertionError("<body> element not found")
# At this stage, HTML code doesn't have a namespace assigned yet.
assert bodyElement.namespaceURI == EMPTY_NAMESPACE
# Replace the placeholder element with the children of the <body> node.
for node in bodyElement.childNodes:
new_node = element.ownerDocument.importNode(node, True)
element.parentNode.insertBefore(new_node, element)
element.parentNode.removeChild(element)
def load_content(self):
self.invoke_filters('load_content:before')
# Convert HTML tag soup to XML
p = TagSoupToXml(omit_comments=True) # Omit commented-out parts in the final output
p.feed(open(self.path_info.source_filename, "rt", encoding="UTF-8").read()) # TODO: support other encodings? (is that safe?)
p.close()
# Return a DOM URL
self.content = p.todocument()
# Drop any "http://www.w3.org/1999/xhtml" namespace declarations
substitute_namespaces(self.content.documentElement, {XHTML_NAMESPACE: EMPTY_NAMESPACE})
normalize_namespaces(self.content.documentElement, strip_dups=True)
self.invoke_filters('load_content:after')
def _handle_news_element(self, page_generator, c_newsElement):
if self._feed_url is None:
raise ValueError("news-here element found before set_news_feed called")
# Find the template element
(c_templateElement,) = getChildElementsNS(c_newsElement, NEWS_NAMESPACE, 'template')
# Create the result document
result_doc = minidom.parseString("<div/>")
# Load the Atom feed
feed = minidom.parseString(open(self._feed_path_info.output_filename, "rb").read())
# Get the content-type of page links
page_content_type = self._framework.plugins['vars'].vars['page_content_type']
# Get the maximum number of entries (if any)
limit = c_newsElement.getAttribute('limit')
if not limit:
limit = None
else:
limit = int(limit)
for i, f_entryElement in enumerate(getChildElementsNS(feed.documentElement, ATOM_NAMESPACE, 'entry')):
# Don't output more than the specified number of articles
if limit is not None and i >= limit:
break
params = {}
# Get entry title
(f_titleElement,) = getChildElementsNS(f_entryElement, ATOM_NAMESPACE, 'title')
assert f_titleElement.getAttribute('type') == 'text'
params['title'] = getChildText(f_titleElement)
# Get entry publication/update dates
params['published'] = getChildText(tuple(getChildElementsNS(f_entryElement, ATOM_NAMESPACE, 'published'))[0])
params['updated'] = getChildText(tuple(getChildElementsNS(f_entryElement, ATOM_NAMESPACE, 'updated'))[0])
# Get entry <link rel="alternate" type="text/html">
ee = getChildElementsNS(f_entryElement, ATOM_NAMESPACE, 'link')
for e in ee:
rel = e.getAttribute('rel')
type = e.getAttribute('type')
hreflang = e.getAttribute('hreflang')
if rel == "alternate" and type == page_content_type and not hreflang:
params['href'] = e.getAttribute('href')
break
else:
raise RuntimeError("link not found")
# Get entry summary
(f_summaryElement,) = getChildElementsNS(f_entryElement, ATOM_NAMESPACE, 'summary')
assert f_summaryElement.getAttribute('type') == 'xhtml'
(f_summaryDiv,) = (n for n in f_summaryElement.childNodes if n.nodeType == n.ELEMENT_NODE)
params['summaryDiv'] = f_summaryDiv
# Create per-entry <div> element
r_divElement = result_doc.createElement('div')
result_doc.documentElement.appendChild(r_divElement)
# Copy the template to the result
self.__copy_template_to_result(c_templateElement, r_divElement, params)
# Namespace normalization
normalize_namespaces(result_doc.documentElement)
# Replace the placeholder
raise ReplaceWithNode(result_doc.documentElement)
def _early_process_entry(self, page_generator, entry):
"""Perform early in-place processing of an entry."""
entryDocument = minidom.parseString(entry['atom:entry'])
entryElement = entryDocument.documentElement
page_content_type = self._framework.plugins['vars'].vars['page_content_type']
# Extract the 'id' of the entry
(idElement,) = getChildElementsNS(entryElement, ATOM_NAMESPACE, 'id')
entry['id'] = getChildText(idElement).strip()
# Extract and normalize the 'published' date of the entry
(publishedElement,) = getChildElementsNS(entryElement, ATOM_NAMESPACE, 'published')
entry['published'] = atom_datetime_to_utc(getChildText(publishedElement).strip())
# Extract and normalize the 'updated' date of the entry; Create it if it doesn't exist.
ee = tuple(getChildElementsNS(entryElement, ATOM_NAMESPACE, 'updated'))
if ee:
(updatedElement,) = ee # there should be only one
else:
# Create an <updated> element using the 'published' date
updatedElement = entryDocument.createElementNS(ATOM_NAMESPACE, 'updated')
replaceChildText(updatedElement, entry['published'])
entryElement.appendChild(updatedElement)
entry['updated'] = atom_datetime_to_utc(getChildText(updatedElement).strip())
# Create a <title> element if one does not already exist.
ee = tuple(getChildElementsNS(entryElement, ATOM_NAMESPACE, 'title'))
if not ee:
titleElement = entryDocument.createElementNS(ATOM_NAMESPACE, 'title')
titleElement.setAttribute('type', 'text')
titleElement.appendChild(entryDocument.createTextNode(entry['title']))
entryElement.appendChild(titleElement)
# Create a <link rel="alternate"> element if one does not already exist.
ee = getChildElementsNS(entryElement, ATOM_NAMESPACE, 'link')
linkElement = None
for e in ee:
rel = e.getAttribute('rel')
type = e.getAttribute('type')
hreflang = e.getAttribute('hreflang')
if rel == "alternate" and type == page_content_type and not hreflang:
if linkElement is not None:
raise FGValueError('Conflicting <link rel="alternate" type=%r hreflang=%r> entries in %s' % (
page_content_type, hreflang, page_generator.path_info.source_filename,))
linkElement = e
if not linkElement:
linkElement = entryDocument.createElementNS(ATOM_NAMESPACE, 'link')
linkElement.setAttribute('rel', 'alternate')
linkElement.setAttribute('href', page_generator.path_info.target_url)
linkElement.setAttribute('type', page_content_type)
entryElement.appendChild(linkElement)
# Rewrite URLs in the atom:entry element
rewrite_links(entryElement, ATOM_CRITERIA, page_generator.path_info.target_url, page_generator.path_info.base_url, always_absolute=True)
# Add a <summary> element, if applicable
if entry['summary']:
summaryDocument = minidom.parseString(entry['summary'])
# Rewrite URLs in the summary
rewrite_links(summaryDocument.documentElement, HTML_CRITERIA,
entry['path_info'].target_url, entry['path_info'].base_url, always_absolute=True)
# Create Atom <summary> element
summaryElement = entryElement.ownerDocument.createElementNS(ATOM_NAMESPACE, 'summary')
summaryElement.setAttribute('type', 'xhtml')
entryElement.appendChild(summaryElement)
# Create XHTML <div> element
divElement = entryElement.ownerDocument.createElementNS(XHTML_NAMESPACE, 'div')
divElement.setAttributeNS(XMLNS_NAMESPACE, 'xmlns', XHTML_NAMESPACE)
summaryElement.appendChild(divElement)
# Add data
for n in summaryDocument.documentElement.childNodes:
divElement.appendChild(divElement.ownerDocument.importNode(n, True))
# Elements with no namespace become XHTML elements
substitute_namespaces(divElement, {EMPTY_NAMESPACE: XHTML_NAMESPACE})
# Clean up
data = None
summaryDocument.unlink()
summaryDocument = None
del entry['summary']
# Add a <content> element
if True:
bodyDocument = minidom.parseString(entry['body'])
# Rewrite URLs in the body
rewrite_links(bodyDocument.documentElement, HTML_CRITERIA,
entry['path_info'].target_url, entry['path_info'].base_url, always_absolute=True)
# Create Atom <content> element
contentElement = entryElement.ownerDocument.createElementNS(ATOM_NAMESPACE, 'content')
contentElement.setAttribute('type', 'xhtml')
entryElement.appendChild(contentElement)
# Create XHTML <div> element
divElement = entryElement.ownerDocument.createElementNS(XHTML_NAMESPACE, 'div')
divElement.setAttributeNS(XMLNS_NAMESPACE, 'xmlns', XHTML_NAMESPACE)
contentElement.appendChild(divElement)
# Add data
for n in bodyDocument.documentElement.childNodes:
divElement.appendChild(divElement.ownerDocument.importNode(n, True))
# Elements with no namespace become XHTML elements
substitute_namespaces(divElement, {EMPTY_NAMESPACE: XHTML_NAMESPACE})
# Clean up
data = None
bodyDocument.unlink()
bodyDocument = None
del entry['body']
# Perform xmlns normalization
normalize_namespaces(entryDocument.documentElement, strip_dups=True)
# Update the new atom:entry document
entry['atom:entry'] = entryDocument.toxml()
