def traverse_tree(self, elem, page=None, __tag_page_href=moin_page.page_href,
__tag_link=_tag_xlink_href, __tag_include=_tag_xinclude_href):
"""
Traverses the tree and handles each element appropriately
"""
new_page_href = elem.get(__tag_page_href)
if new_page_href:
page = Iri(new_page_href)
xlink_href = elem.get(__tag_link)
xinclude_href = elem.get(__tag_include)
if xlink_href:
xlink_href = Iri(xlink_href)
if xlink_href.scheme == 'wiki.local':
self.handle_wikilocal_links(elem, xlink_href, page)
elif xlink_href.scheme == 'wiki':
self.handle_wiki_links(elem, xlink_href)
elif xlink_href.scheme:
self.handle_external_links(elem, xlink_href)
elif xinclude_href:
xinclude_href = Iri(xinclude_href)
if xinclude_href.scheme == 'wiki.local':
self.handle_wikilocal_transclusions(elem, xinclude_href, page)
elif xinclude_href.scheme == 'wiki':
self.handle_wiki_transclusions(elem, xinclude_href)
elif xlink_href == u'':
# ReST link to page fragment
elem.set(self._tag_xlink_href, u'#' + elem.text.replace(' ', '_'))
for child in elem.iter_elements():
self.traverse_tree(child, page)
return elem
def inline_object_repl(self, stack, object, object_url=None, object_item=None,
object_text=None, object_args=None):
"""Handles objects transcluded within the page."""
if object_args:
args = parse_arguments(object_args).keyword # XXX needs different parsing
else:
args = {}
query_keys = {}
attrib = {}
whitelist = ['width', 'height', 'class']
for attr, value in args.iteritems():
if attr.startswith('&'):
query_keys[attr[1:]] = value
elif attr in whitelist:
attrib[html(attr)] = value
if object_text:
attrib[html.alt] = object_text
if object_item is not None:
# img tag
query = url_encode(query_keys, charset=CHARSET, encode_keys=True)
# TODO: moin 1.9 needed this for an attached file; move functionality to script/migration/moin/import19.py
att = 'attachment:'
if object_item.startswith(att):
object_item = '/' + object_item[len(att):] # now we have a subitem
target = Iri(scheme='wiki.local', path=object_item, query=query, fragment=None)
attrib[xinclude.href] = target
element = xinclude.include(attrib=attrib)
stack.top_append(element)
else:
# object tag
target = Iri(object_url)
attrib[xlink.href] = target
element = moin_page.object(attrib)
stack.top_append(element)
def visit_reference(self, node):
refuri = node.get('refuri', u'')
if refuri.startswith(u'<<') and refuri.endswith(u'>>'): # moin macro
macro_name = refuri[2:-2].split(u'(')[0]
if macro_name == u"TableOfContents":
arguments = refuri[2:-2].split(u'(')[1][:-1].split(u',')
node = moin_page.table_of_content()
self.open_moin_page_node(node)
if arguments and arguments[0]:
node.set(moin_page.outline_level, arguments[0])
return
if macro_name == u"Include":
# include macros are expanded by include.py similar to transclusions
# rst include handles only wiki pages and does not support additional arguments like moinwiki
arguments = refuri[2:-2].split(u'(')[1][:-1].split(u',')
link = Iri(scheme=u'wiki.local', path=arguments)
node = xinclude.include(
attrib={
xinclude.href: link,
moin_page.alt: refuri,
moin_page.content_type: 'x-moin/macro;name=' +
macro_name,
})
self.open_moin_page_node(node)
return
try:
arguments = refuri[2:-2].split(u'(')[1][:-1]
except IndexError:
arguments = u'' # <<DateTime>>
self.open_moin_page_node(
moin_page.inline_part(
attrib={
moin_page.content_type:
"x-moin/macro;name={0}".format(macro_name)
}))
if arguments:
self.open_moin_page_node(moin_page.arguments())
self.open_moin_page_node(arguments)
self.close_moin_page_node()
self.close_moin_page_node()
return
if not allowed_uri_scheme(refuri):
self.visit_error(node)
return
if refuri == u'':
# build a link to a heading or an explicitly defined anchor
refuri = Iri(scheme='wiki.local',
fragment=node.attributes['name'].replace(' ', '_'))
self.open_moin_page_node(moin_page.a(attrib={xlink.href: refuri}))
def visit_image(self, node):
"""
Processes images and other transcluded objects.
"""
whitelist = [
'width',
'height',
'alt',
]
attrib = {}
for key in whitelist:
if node.get(key):
attrib[html(key)] = node.get(key)
# there is no 'scale' attribute, hence absent from whitelist, handled separately
if node.get('scale'):
scaling_factor = int(node.get('scale')) / 100.0
for key in ('width', 'height'):
if html(key) in attrib:
attrib[html(key)] = int(
int(attrib[html(key)]) * scaling_factor)
# "align" parameter is invalid in HTML5. Convert it to a class defined in userstyles.css.
userstyles = {
'left': 'left',
'center': 'center',
'right': 'right',
'top':
'top', # rst parser creates error messages for top, bottom, and middle
'bottom': 'bottom',
'middle': 'middle',
}
alignment = userstyles.get(node.get('align'))
if alignment:
attrib[html.class_] = alignment
url = Iri(node['uri'])
if url.scheme is None:
# img
target = Iri(scheme='wiki.local', path=node['uri'], fragment=None)
attrib[xinclude.href] = target
new_node = xinclude.include(attrib=attrib)
else:
# obj
new_node = moin_page.object(attrib)
new_node.set(xlink.href, url)
self.open_moin_page_node(new_node)
def test_Iri_quote_1():
u = Iri(scheme=u'wiki', authority=u'authority_ä%?#',
path=u'/path_ä%?#', query=u'query_ä%?#', fragment=u'fragment_ä%?#')
assert u.scheme == u'wiki'
assert u.authority == u'authority_ä%?#'
authority = u'authority_ä%25%3F%23'
assert u.authority.fullquoted == authority
assert u.authority.quoted == u'authority_ä%25?#'
assert u.authority.urlquoted == u'authority_%C3%A4%25%3F%23'
assert u.path == u'/path_ä%?#'
path = u'/path_ä%25%3F%23'
assert u.path.fullquoted == path
assert u.path.quoted == u'/path_ä%25?#'
assert u.path.urlquoted == u'/path_%C3%A4%25%3F%23'
assert u.query == u'query_ä%?#'
query = u'query_ä%25?%23'
assert u.query.fullquoted == query
assert u.query.quoted == u'query_ä%25?#'
assert u.query.urlquoted == u'query_%C3%A4%25?%23'
assert u.fragment == u'fragment_ä%?#'
fragment = u'fragment_ä%25?%23'
assert u.fragment.fullquoted == fragment
assert u.fragment.quoted == u'fragment_ä%25?#'
assert u.fragment.urlquoted == u'fragment_%C3%A4%25?%23'
assert unicode(u) == u'wiki://{0}{1}?{2}#{3}'.format(authority, path, query, fragment)
def test_Iri_init_1():
u = Iri(scheme='wiki', path='/StartSeite', query='action=raw')
assert u.scheme == 'wiki'
assert u.authority is None
assert u.path == '/StartSeite'
assert u.query == 'action=raw'
assert u.fragment is None
def __call__(self, rev, contenttype=None, arguments=None):
item_name = rev.item.name
query_keys = {'do': 'get', 'rev': rev.revid}
attrib = {}
if arguments:
query = arguments.keyword.get(xinclude.href)
if query and query.query:
# query.query value is similar to "w=75" given a transclusion "{{jpeg||&w=75 class="top"}}"
query_keys.update(url_decode(query.query))
attrib = arguments.keyword
query = url_encode(query_keys, charset=CHARSET, encode_keys=True)
attrib.update({
moin_page.type_:
unicode(self.input_type),
xlink.href:
Iri(scheme='wiki', authority='', path='/' + item_name,
query=query),
})
obj = moin_page.object_(attrib=attrib, children=[
item_name,
])
body = moin_page.body(children=(obj, ))
return moin_page.page(children=(body, ))
def inline_link_repl(self,
stack,
link,
link_url=None,
link_item=None,
link_text=None,
link_interwiki_site=None,
link_interwiki_item=None):
"""Handle all kinds of links."""
if link_interwiki_site:
if is_known_wiki(link_interwiki_site):
link = Iri(scheme='wiki',
authority=link_interwiki_site,
path='/' + link_interwiki_item)
element = moin_page.a(attrib={xlink.href: link})
stack.push(element)
if link_text:
self.parse_inline(link_text, stack, self.inlinedesc_re)
else:
stack.top_append(link_interwiki_item)
stack.pop()
return
else:
# assume local language uses ":" inside of words, set link_item and continue
link_item = '{0}:{1}'.format(link_interwiki_site,
link_interwiki_item)
if link_item is not None:
att = 'attachment:' # moin 1.9 needed this for an attached file
if link_item.startswith(att):
link_item = '/' + link_item[len(att):] # now we have a subitem
# we have Anchor macro, so we support anchor links despite lack of docs in Creole spec
if '#' in link_item:
path, fragment = link_item.rsplit('#', 1)
else:
path, fragment = link_item, None
target = Iri(scheme='wiki.local', path=path, fragment=fragment)
text = link_item
else:
target = Iri(link_url)
text = link_url
element = moin_page.a(attrib={xlink.href: target})
stack.push(element)
if link_text:
self.parse_inline(link_text, stack, self.inlinedesc_re)
else:
stack.top_append(text)
stack.pop()
def test_Iri_init_2():
i = 'wiki://MoinMoin/StartSeite?action=raw#body'
u = Iri(i, scheme='newwiki', path='/newStartSeite', query='action=false')
assert u.scheme == 'newwiki'
assert u.authority == 'MoinMoin'
assert u.path == '/newStartSeite'
assert u.query == 'action=false'
assert u.fragment == 'body'
def __call__(self, rev, contenttype=None, arguments=None):
item_name = rev.item.fqname.value
attrib = {
xlink.href: Iri(scheme='wiki', authority='', path='/' + item_name, query='do=modify'),
}
a = moin_page.a(attrib=attrib, children=[_("%(item_name)s does not exist. Create it?", item_name=item_name)])
body = moin_page.body(children=(a, ))
return moin_page.page(children=(body, ))
def test_Iri_quote_2():
authority = u'authority_ä%25%3F%23'
path = u'/path_ä%25%3F%23'
query = u'query_ä%25?%23'
fragment = u'fragment_ä%25?%23'
i = u'wiki://{0}{1}?{2}#{3}'.format(authority, path, query, fragment)
u = Iri(i)
assert unicode(u) == i
def __call__(self, rev, contenttype=None, arguments=None):
item_name = rev.item.name or rev.meta['name'][0]
attrib = {
xlink.href: Iri(scheme='wiki', authority='', path='/' + item_name,
query='do=get&rev={0}'.format(rev.revid)),
}
a = moin_page.a(attrib=attrib, children=[u"Download {0}.".format(item_name)])
body = moin_page.body(children=(a, ))
return moin_page.page(children=(body, ))
def test_Iri_2():
i = 'wiki://MoinMoin/StartSeite?action=raw#body'
u = Iri(i)
assert u.scheme == 'wiki'
assert u.authority == 'MoinMoin'
assert u.path == '/StartSeite'
assert u.query == 'action=raw'
assert u.fragment == 'body'
assert unicode(u) == i
i = 'wiki:///StartSeite?action=raw#body'
u = Iri(i)
assert u.scheme == 'wiki'
assert u.authority == ''
assert u.path == '/StartSeite'
assert u.query == 'action=raw'
assert u.fragment == 'body'
assert unicode(u) == i
def test_Iri_3():
i = 'wiki.local:StartSeite?action=raw#body'
u = Iri(i)
assert u.scheme == 'wiki.local'
assert u.authority is None
assert u.path == 'StartSeite'
assert u.query == 'action=raw'
assert u.fragment == 'body'
assert unicode(u) == i
def visit_img(self, element):
"""
<img src="URI" /> --> <object xlink:href="URI />
"""
attrib = {}
url = Iri(element.attrib.get('src'))
if element.attrib.get('alt'):
attrib[html.alt] = element.attrib.get('alt')
if element.attrib.get('title'):
attrib[html.title_] = element.attrib.get('title')
if url.scheme is None:
# img tag
target = Iri(scheme='wiki.local', path=element.attrib.get("src"), fragment=None)
attrib[xinclude.href] = target
new_node = xinclude.include(attrib=attrib)
else:
# object tag
attrib[xlink.href] = url
new_node = moin_page.object(attrib)
return new_node
def inline_freelink_repl(self,
stack,
freelink,
freelink_bang=None,
freelink_interwiki_page=None,
freelink_interwiki_ref=None,
freelink_page=None,
freelink_email=None):
if freelink_bang:
stack.top_append(freelink)
return
attrib = {}
if freelink_page:
page = freelink_page.encode('utf-8')
if '#' in page:
path, fragment = page.rsplit('#', 1)
else:
path, fragment = page, None
link = Iri(scheme='wiki.local', path=path, fragment=fragment)
text = freelink_page
elif freelink_email:
link = 'mailto:' + freelink_email
text = freelink_email
else:
if not is_known_wiki(freelink_interwiki_ref):
stack.top_append(freelink)
return
link = Iri(scheme='wiki',
authority=freelink_interwiki_ref,
path='/' + freelink_interwiki_page)
text = freelink_interwiki_page
attrib[xlink.href] = link
element = moin_page.a(attrib, children=[text])
stack.top_append(element)
def process_name(self, member_name):
name = unicode(member_name, 'utf-8')
attrib = {
xlink.href:
Iri(scheme='wiki',
authority='',
path='/' + self.item_name,
query=u'do=get&member={0}'.format(name)),
}
return moin_page.a(attrib=attrib, children=[
name,
])
def handle_wiki_links(self, elem, input):
wiki_name = 'Self'
if input.authority and input.authority.host:
wn = unicode(input.authority.host)
if is_known_wiki(wn):
# interwiki link
elem.set(moin_page.class_, 'moin-interwiki')
wiki_name = wn
item_name = unicode(input.path[1:])
endpoint, rev, query = self._get_do_rev(input.query)
url = url_for_item(item_name, wiki_name=wiki_name, rev=rev, endpoint=endpoint)
link = Iri(url, query=query, fragment=input.fragment)
elem.set(self._tag_xlink_href, link)
def inline_object_repl(self, stack, object, object_page=None, object_url=None, object_text=None):
"""Handles objects included in the page."""
attrib = {}
if object_text:
attrib[html.alt] = object_text
if object_page is not None:
att = 'attachment:' # moin 1.9 needed this for an attached file
if object_page.startswith(att):
object_page = '/' + object_page[len(att):] # now we have a subitem
target = Iri(scheme='wiki.local', path=object_page)
attrib[xinclude.href] = target
element = xinclude.include(attrib=attrib)
else:
attrib[xlink.href] = object_url
element = moin_page.object(attrib)
stack.top_append(element)
def absolute_path(self, path, current_page_path):
"""
Converts a relative iri path into an absolute one
:param path: the relative path to be converted
:type path: Iri.path
:param current_page_path: the path of the page where the link is
:type current_page_path: Iri.path
:returns: the absolute equivalent of the relative path
:rtype: Iri.path
"""
quoted_path = path.quoted
# starts from 1 because 0 is always / for the current page
quoted_current_page_path = current_page_path[1:].quoted
abs_path = AbsItemName(quoted_current_page_path, quoted_path)
abs_path = Iri(abs_path).path
return abs_path
def __call__(self, rev, contenttype=None, arguments=None):
item_name = rev.item.name
attrib = {
moin_page.type_:
unicode(self.input_type),
xlink.href:
Iri(scheme='wiki',
authority='',
path='/' + item_name,
query='do=get&rev={0}'.format(rev.revid)),
}
obj = moin_page.object_(
attrib=attrib,
children=[
u'Your Browser does not support HTML5 audio/video element.',
])
body = moin_page.body(children=(obj, ))
return moin_page.page(children=(body, ))
def handle_wikilocal_links(self, elem, input, page):
if input.path:
# this can be a relative path, make it absolute:
path = input.path
if page:
path = self.absolute_path(path, page.path)
item_name = unicode(path)
if not flaskg.storage.has_item(item_name):
# XXX these index accesses slow down the link converter quite a bit
elem.set(moin_page.class_, 'moin-nonexistent')
else:
item_name = unicode(page.path[1:])
endpoint, rev, query = self._get_do_rev(input.query)
url = url_for_item(item_name, rev=rev, endpoint=endpoint)
if not page:
url = url[1:]
link = Iri(url, query=query, fragment=input.fragment)
elem.set(self._tag_xlink_href, link)
def macro(self, content, arguments, page_url, alternative):
if arguments:
item_count = int(arguments[0])
else:
item_count = 1
all_item_names = self.get_item_names()
# Now select random item from the full list, and if it exists and
# we can read it, save.
random_item_names = []
found = 0
while found < item_count and all_item_names:
# Take one random item from the list
item_name = random.choice(all_item_names)
all_item_names.remove(item_name)
# Filter out items the user may not read.
try:
item = Item.create(item_name)
random_item_names.append(item_name)
found += 1
except AccessDenied:
pass
if not random_item_names:
return
random_item_names.sort()
result = moin_page.span()
for name in random_item_names:
link = unicode(Iri(scheme=u'wiki', authority=u'',
path=u'/' + name))
result.append(
moin_page.a(attrib={xlink.href: link}, children=[name]))
result.append(", ")
del result[-1] # kill last comma
return result
def internal_representation(self, attributes=None):
"""
Return the internal representation of a document using a DOM Tree
"""
hash_name = HASH_ALGORITHM
hash_hexdigest = self.rev.meta.get(hash_name)
if hash_hexdigest:
cid = cache_key(usage="internal_representation",
hash_name=hash_name,
hash_hexdigest=hash_hexdigest,
attrs=repr(attributes))
doc = app.cache.get(cid)
else:
# likely a non-existing item
doc = cid = None
if doc is None:
# We will see if we can perform the conversion:
# FROM_mimetype --> DOM
# if so we perform the transformation, otherwise we don't
from moin.converter import default_registry as reg
input_conv = reg.get(Type(self.contenttype), type_moin_document)
if not input_conv:
raise TypeError("We cannot handle the conversion from {0} to the DOM tree".format(self.contenttype))
smiley_conv = reg.get(type_moin_document, type_moin_document, icon='smiley')
# We can process the conversion
links = Iri(scheme='wiki', authority='', path='/' + self.name)
doc = input_conv(self.rev, self.contenttype, arguments=attributes)
# XXX is the following assuming that the top element of the doc tree
# is a moin_page.page element? if yes, this is the wrong place to do that
# as not every doc will have that element (e.g. for images, we just get
# moin_page.object, for a tar item, we get a moin_page.table):
doc.set(moin_page.page_href, unicode(links))
if self.contenttype.startswith((u'text/x.moin.wiki', u'text/x-mediawiki', u'text/x.moin.creole', )):
doc = smiley_conv(doc)
if cid:
app.cache.set(cid, doc)
return doc
def recurse(self, elem, page_href):
# on first call, elem.tag.name=='page'.
# Descendants (body, div, p, include, page, etc.) are processed by recursing through DOM
# stack is used to detect transclusion loops
page_href_new = elem.get(moin_page.page_href)
if page_href_new:
page_href_new = Iri(page_href_new)
if page_href_new != page_href:
page_href = page_href_new
self.stack.append(page_href)
else:
self.stack.append(None)
else:
self.stack.append(None)
try:
if elem.tag == xinclude.include:
# we have already recursed several levels and found a transclusion: "{{SomePage}}" or <<Include(...)>>
# process the transclusion and add it to the DOM. Subsequent recursions will traverse through
# the transclusion's elements.
href = elem.get(xinclude.href)
xpointer = elem.get(xinclude.xpointer)
xp_include_pages = None
xp_include_sort = None
xp_include_items = None
xp_include_skipitems = None
xp_include_heading = None
xp_include_level = None
if xpointer:
# we are working on an <<Include(abc)>> macro, not a {{transclusion}}
xp = XPointer(xpointer)
xp_include = None
xp_namespaces = {}
for entry in xp:
uri = None
name = entry.name.split(':', 1)
if len(name) > 1:
prefix, name = name
uri = xp_namespaces.get(prefix, False)
else:
name = name[0]
if uri is None and name == 'xmlns':
d_prefix, d_uri = entry.data.split('=', 1)
xp_namespaces[d_prefix] = d_uri
elif uri == moin_page.namespace and name == 'include':
xp_include = XPointer(entry.data)
if xp_include:
for entry in xp_include:
name, data = entry.name, entry.data_unescape
# TODO: These do not include all parameters in moin 1.9 Include macro docs:
# <<Include(pagename, heading, level, from="regex", to="regex", sort=ascending|descending, items=n, skipitems=n, titlesonly, editlink)>>
# these are currently unsupported in moin 2.0: from, to, titlesonly, editlink
if name == 'pages': # pages == pagename in moin 1.9
xp_include_pages = data
elif name == 'sort':
xp_include_sort = data
elif name == 'items':
xp_include_items = int(data)
elif name == 'skipitems':
xp_include_skipitems = int(data)
elif name == 'heading':
xp_include_heading = data
elif name == 'level':
xp_include_level = data
included_elements = []
if href:
# We have a single page to transclude or include
href = Iri(href)
link = Iri(scheme='wiki', authority='')
if href.scheme == 'wiki':
if href.authority:
raise ValueError(
"can't handle xinclude for non-local authority"
)
else:
path = href.path[1:]
elif href.scheme == 'wiki.local':
page = page_href
path = href.path
if path[0] == '':
# /subitem
tmp = page.path[1:]
tmp.extend(path[1:])
path = tmp
elif path[0] == '..':
# ../sisteritem
path = page.path[1:] + path[1:]
else:
raise ValueError(
"can't handle xinclude for schemes other than wiki or wiki.local"
)
link.path = path
if flaskg.user.may.read(unicode(path)):
page = Item.create(unicode(path))
pages = ((page, link), )
else:
# ACLs prevent user from viewing a transclusion - show message
message = moin_page.p(children=(_(
'Access Denied, transcluded content suppressed.')))
attrib = {html.class_: 'warning'}
div = ET.Element(moin_page.div,
attrib,
children=(message, ))
container = ET.Element(moin_page.body,
children=(div, ))
return [
container, 0
] # replace transclusion with container's child
elif xp_include_pages:
# we have regex of pages to include: <<Include(^qqq)>>
query = And([
Term(WIKINAME, app.cfg.interwikiname),
Regex(NAME_EXACT, xp_include_pages)
])
reverse = xp_include_sort == 'descending'
results = flaskg.storage.search(query,
sortedby=NAME_EXACT,
reverse=reverse,
limit=None)
pagelist = [result.name for result in results]
if xp_include_skipitems is not None:
pagelist = pagelist[xp_include_skipitems:]
if xp_include_items is not None:
pagelist = pagelist[xp_include_items + 1:]
pages = ((Item.create(p),
Iri(scheme='wiki', authority='', path='/' + p))
for p in pagelist)
if not pagelist:
msg = _(
'Error: no items found matching "<<Include({0})>>"'
).format(xp_include_pages)
attrib = {html.class_: 'moin-error'}
strong = ET.Element(moin_page.strong, attrib, (msg, ))
included_elements.append(strong)
for page, p_href in pages:
if p_href.path[0] != '/':
p_href.path = IriPath('/' + '/'.join(p_href.path))
if p_href in self.stack:
# we have a transclusion loop, create an error message showing list of pages forming loop
loop = self.stack[self.stack.index(p_href):]
loop = [
u'{0}'.format(ref.path[1:])
for ref in loop if ref is not None
] + [page.name]
msg = u'Error: Transclusion loop via: ' + u', '.join(
loop)
attrib = {html.class_: 'moin-error'}
strong = ET.Element(moin_page.strong, attrib, (msg, ))
included_elements.append(strong)
continue
if xp_include_heading is not None:
attrib = {xlink.href: p_href}
children = (xp_include_heading or page.name, )
elem_a = ET.Element(moin_page.a,
attrib,
children=children)
attrib = {
moin_page.outline_level: xp_include_level or '1'
}
elem_h = ET.Element(moin_page.h,
attrib,
children=(elem_a, ))
included_elements.append(elem_h)
page_doc = page.content.internal_representation(
attributes=Arguments(keyword=elem.attrib))
if isinstance(page.rev.data, file):
page.rev.data.close()
self.recurse(page_doc, page_href)
# The href needs to be an absolute URI, without the prefix "wiki://"
page_doc = mark_item_as_transclusion(page_doc, p_href.path)
included_elements.append(page_doc)
if len(included_elements) > 1:
# use a div as container
result = ET.Element(moin_page.div)
result.extend(included_elements)
elif included_elements:
result = included_elements[0]
else:
result = None
# end of processing for transclusion; the "result" will get inserted into the DOM below
return result
# Traverse the DOM by calling self.recurse with each child of the current elem.
# Starting elem.tag.name=='page'.
container = []
i = 0
while i < len(elem):
child = elem[i]
if isinstance(child, ET.Node):
ret = self.recurse(child, page_href)
if ret:
# Either child or a descendant of child is a transclusion.
# See top of this script for notes on why these DOM adjustments are required.
if isinstance(ret, ET.Node
) and elem.tag.name in NO_BLOCK_CHILDREN:
body = ret[0]
if len(body) == 0:
# the transcluded item is empty, insert an empty span into DOM
attrib = Attributes(ret).convert()
elem[i] = ET.Element(moin_page.span,
attrib=attrib)
elif (isinstance(body[0], ET.Node)
and (len(body) > 1 or body[0].tag.name
not in ('p', 'object', 'a'))):
# Complex case: "some text {{BlockItem}} more text" or "\n{{BlockItem}}\n" where
# the BlockItem body contains multiple p's, a table, preformatted text, etc.
# These block elements cannot be made a child of the current elem, so we create
# a container to replace elem.
# Create nodes to hold any siblings before and after current child (elem[i])
before = copy.deepcopy(elem)
after = copy.deepcopy(elem)
before[:] = elem[0:i]
after[:] = elem[i + 1:]
if len(before):
# there are siblings before transclude, save them in container
container.append(before)
new_trans_ptr = len(container)
# get attributes from page node;
# we expect {class: "moin-transclusion"; data-href: "http://some.org/somepage"}
attrib = Attributes(ret).convert()
# current elem will likely be replaced by container so we need to copy data-lineno attr
if html.data_lineno in elem.attrib:
attrib[html.data_lineno] = elem.attrib[
html.data_lineno]
# make new div node to hold transclusion, copy children, and save in container
div = ET.Element(moin_page.div,
attrib=attrib,
children=body[:])
container.append(
div) # new_trans_ptr is index to this
if len(after):
container.append(after)
if elem.tag.name == 'a':
# invalid input [[MyPage|{{BlockItem}}]],
# best option is to retain A-tag and fail html validation
# TODO: error may not be obvious to user - add error message
elem[i] = div
else:
# move up 1 level in recursion where elem becomes the child and
# is usually replaced by container
return [container, new_trans_ptr]
else:
# default action for inline transclusions or odd things like circular transclusion error messages
classes = child.attrib.get(html.class_,
'').split()
classes += ret.attrib.get(html.class_,
'').split()
ret.attrib[html.class_] = ' '.join(classes)
elem[i] = ret
elif isinstance(ret, types.ListType):
# a container has been returned.
# Note: there are multiple places where a container may be constructed
ret_container, trans_ptr = ret
# trans_ptr points to the transclusion within ret_container.
# Here the transclusion will always contain a block level element
if elem.tag.name in NO_BLOCK_CHILDREN:
# Complex case, transclusion effects grand-parent, great-grand-parent, e.g.:
# "/* comment {{BlockItem}} */" or "text ''italic {{BlockItem}} italic'' text"
# elem is an inline element, build a bigger container to replace elem's parent,
before = copy.deepcopy(elem)
after = copy.deepcopy(elem)
before[:] = elem[0:i] + ret_container[
0:trans_ptr]
after[:] = ret_container[trans_ptr +
1:] + elem[i + 1:]
if len(before):
container.append(before)
new_trans_ptr = len(container)
# child may have classes like "comment" that must be added to transcluded element
classes = child.attrib.get(
moin_page.class_, '').split()
# must use moin_page.class_ above, but use html.class below per html_out.py code
classes += ret_container[trans_ptr].attrib.get(
html.class_, '').split()
ret_container[trans_ptr].attrib[
html.class_] = ' '.join(classes)
container.append(ret_container[trans_ptr]
) # the transclusion
if len(after):
container.append(after)
return [container, new_trans_ptr]
else:
# elem is a block element
for grandchild in child:
if isinstance(
grandchild, ET.Node
) and grandchild.tag.name == u'include':
# the include may have classes that must be added to transcluded element
classes = grandchild.attrib.get(
html.class_, '').split()
classes += ret_container[
trans_ptr].attrib.get(
html.class_, '').split()
ret_container[trans_ptr].attrib[
html.class_] = ' '.join(classes)
# replace child element with the container generated in lower recursion
elem[i:i +
1] = ret_container # elem[i] is the child
else:
# default action for any ret not fitting special cases above,
# e.g. tranclusion is within a table cell
elem[i] = ret
# we are finished with this child, advance to next sibling
i += 1
finally:
self.stack.pop()
def inline_url_repl(self, stack, url, url_target):
url = Iri(url_target)
attrib = {xlink.href: url}
element = moin_page.a(attrib=attrib, children=[url_target])
stack.top_append(element)
def allowed_uri_scheme(uri):
parsed = Iri(uri)
return not parsed.scheme or parsed.scheme in URI_SCHEMES
def test_wikilocal(conv, input_, page, output):
elem = ET.Element(None)
conv.handle_wikilocal_links(elem, Iri(input_), Iri(page))
assert elem.get(xlink.href) == output
def test_wikiexternal(conv, input_, output):
elem = ET.Element(None)
conv.handle_external_links(elem, Iri(input_))
href = elem.get(xlink.href)
assert href == output
def inline_link_repl(self,
stack,
link,
link_url=None,
link_item=None,
link_args=u'',
external_link_url=None,
alt_text=u''):
"""Handle all kinds of links."""
link_text = ''
link_args_list = []
# Remove the first pipe/space, example of link_args : |arg1|arg2 or " arg1 arg2"
parsed_args = self.parse_args(link_args[1:])
query = None
if parsed_args.keyword:
query = url_encode(parsed_args.keyword,
charset=CHARSET,
encode_keys=True,
sort=True)
# Take the last of positional parameters as link_text(caption)
if parsed_args.positional:
link_text = parsed_args.positional.pop()
if link_item is not None:
if '#' in link_item:
path, fragment = link_item.rsplit('#', 1)
else:
path, fragment = link_item, None
target = Iri(scheme='wiki.local',
path=path,
query=query,
fragment=fragment)
text = link_item
else:
if link_url and len(link_url.split(':')) > 0 and link_url.split(
':')[0] == 'File':
object_item = ':'.join(link_url.split(':')[1:])
args = parsed_args.keyword
if object_item is not None:
if 'do' not in args:
# by default, we want the item's get url for transclusion of raw data:
args['do'] = 'get'
query = url_encode(args,
charset=CHARSET,
encode_keys=True,
sort=True)
target = Iri(scheme='wiki.local',
path=object_item,
query=query,
fragment=None)
text = object_item
else:
target = Iri(scheme='wiki.local', path=link_url)
text = link_url
if not link_text:
link_text = text
attrib = {xlink.href: target}
attrib[moin_page.alt] = link_text
element = moin_page.object(attrib)
stack.push(element)
if link_text:
self.preprocessor.push()
self.parse_inline(link_text, stack, self.inlinedesc_re)
self.preprocessor.pop()
else:
stack.top_append(text)
stack.pop()
return
target = Iri(scheme='wiki.local', path=link_url)
text = link_url
if external_link_url:
target = Iri(external_link_url)
text = alt_text
element = moin_page.a(attrib={xlink.href: target})
stack.push(element)
if link_text:
self.preprocessor.push()
self.parse_inline(link_text, stack, self.inlinedesc_re)
self.preprocessor.pop()
else:
stack.top_append(text)
stack.pop()
