def _word_repl(self, word, groups):
if self.in_dd:
name = groups.get('word_name')
current_page = self.formatter.page.page_name
abs_name = AbsPageName(current_page, name)
if abs_name == current_page:
return self.__real_val(abs_name)
else:
# handle anchors
try:
abs_name, anchor = rsplit(abs_name, "#", 1)
except ValueError:
anchor = ""
if self.cat_re.match(abs_name):
return self.__real_val(abs_name)
else:
if not anchor:
wholename = abs_name
else:
wholename = "%s#%s" % (abs_name, anchor)
return self.__real_val(wholename)
return apply(wikiParser._word_repl, (self, word, groups))
def metas_to_abs_links(request, page, values):
new_values = list()
stripped = False
for value in values:
if is_meta_link(value) != 'link':
new_values.append(value)
continue
if ((value.startswith('[[') and value.endswith(']]')) or
(value.startswith('{{') and value.endswith('}}'))):
stripped = True
value = value.lstrip('[')
value = value.lstrip('{')
attachment = ''
for scheme in ('attachment:', 'inline:', 'drawing:'):
if value.startswith(scheme):
if len(value.split('/')) == 1:
value = ':'.join(value.split(':')[1:])
if not '|' in value:
# If page does not have descriptive text, try
# to shorten the link to the attachment name.
value = "%s|%s" % (value.rstrip(']').rstrip('}'), value)
value = "%s%s/%s" % (scheme, page, value)
else:
att_page = value.split(':')[1]
if (att_page.startswith('./') or
att_page.startswith('/') or
att_page.startswith('../')):
attachment = scheme
value = ':'.join(value.split(':')[1:])
if (value.startswith('./') or
value.startswith('/') or
value.startswith('../')):
value = AbsPageName(page, value)
if value.startswith('#'):
value = page + value
value = attachment + value
if stripped:
if value.endswith(']'):
value = '[[' + value
elif value.endswith('}'):
value = '{{' + value
new_values.append(value)
return new_values
def metas_to_abs_links(request, page, values):
new_values = list()
stripped = False
for value in values:
if is_meta_link(value) != 'link':
new_values.append(value)
continue
if ((value.startswith('[[') and value.endswith(']]'))
or (value.startswith('{{') and value.endswith('}}'))):
stripped = True
value = value.lstrip('[')
value = value.lstrip('{')
attachment = ''
for scheme in ('attachment:', 'inline:', 'drawing:'):
if value.startswith(scheme):
if len(value.split('/')) == 1:
value = ':'.join(value.split(':')[1:])
if not '|' in value:
# If page does not have descriptive text, try
# to shorten the link to the attachment name.
value = "%s|%s" % (value.rstrip(']').rstrip('}'),
value)
value = "%s%s/%s" % (scheme, page, value)
else:
att_page = value.split(':')[1]
if (att_page.startswith('./') or att_page.startswith('/')
or att_page.startswith('../')):
attachment = scheme
value = ':'.join(value.split(':')[1:])
if (value.startswith('./') or value.startswith('/')
or value.startswith('../')):
value = AbsPageName(page, value)
if value.startswith('#'):
value = page + value
value = attachment + value
if stripped:
if value.endswith(']'):
value = '[[' + value
elif value.endswith('}'):
value = '{{' + value
new_values.append(value)
return new_values
def _metatable_parseargs(request, args, cat_re, temp_re):
# Arg placeholders
argset = set([])
keyspec = list()
excluded_keys = list()
orderspec = list()
limitregexps = dict()
limitops = dict()
# Capacity for storing indirection keys in metadata comparisons
# and regexps, eg. k->c=/.+/
indirection_keys = list()
# list styles
styles = dict()
# Flag: were there page arguments?
pageargs = False
# Regex preprocessing
for arg in (x.strip() for x in args.split(',') if x.strip()):
# metadata key spec, move on
if arg.startswith('||') and arg.endswith('||'):
# take order, strip empty ones, look at styles
for key in arg.split('||'):
if not key:
continue
# Grab styles
if key.startswith('<') and '>' in key:
style = parseAttributes(request, key[1:], '>')
key = key[key.index('>') + 1:].strip()
if style:
styles[key] = style[0]
# Grab key exclusions
if key.startswith('!'):
excluded_keys.append(key.lstrip('!'))
continue
keyspec.append(key.strip())
continue
op_match = False
# Check for Python operator comparisons
for op in OPERATORS:
if op in arg:
data = arg.rsplit(op)
# If this is not a comparison but indirection,
# continue. Good: k->s>3, bad: k->s=/.+/
if op == '>' and data[0].endswith('-'):
continue
# Must have real comparison
if not len(data) == 2:
if op == '==':
data.append('')
else:
continue
key, comp = map(string.strip, data)
# Add indirection key
if '->' in key:
indirection_keys.append(key)
limitops.setdefault(key, list()).append((comp, op))
op_match = True
# One of the operators matched, no need to go forward
if op_match:
break
# One of the operators matched, process next arg
if op_match:
continue
# Metadata regexp, move on
if '=' in arg:
data = arg.split("=")
key = data[0]
# Add indirection key
if '->' in key:
indirection_keys.append(key)
val = '='.join(data[1:])
# Assume that value limits are regexps, if
# not, escape them into exact regexp matches
if not REGEX_RE.match(val):
from MoinMoin.parser.text_moin_wiki import Parser
# If the value is a page, make it a non-matching
# regexp so that all link variations will generate a
# match. An alternative would be to match from links
# also, but in this case old-style metalinks, which
# cannot be edited, would appear in metatables, which
# is not wanted (old-style eg. [[Page| key: Page]])
# Only allow non-matching regexp for values if they
# are WikiWords. Eg. 'WikiWord some text' would match
# 'WikiWord', emulating ye olde matching behaviour,
# but 'nonwikiword some text' would not match
# 'nonwikiword'
if re.match(Parser.word_rule_js, val):
re_val = "(%s|" % (re.escape(val))
else:
re_val = "(^%s$|" % (re.escape(val))
# or as bracketed link
re_val += "(?P<sta>\[\[)%s(?(sta)\]\])|" % (re.escape(val))
# or as commented bracketed link
re_val += "(?P<stb>\[\[)%s(?(stb)\|[^\]]*\]\]))" % \
(re.escape(val))
limitregexps.setdefault(key, set()).add(
re.compile(re_val, re.UNICODE))
# else strip the //:s
else:
if len(val) > 1:
val = val[1:-1]
limitregexps.setdefault(key, set()).add(
re.compile(val, re.IGNORECASE | re.UNICODE))
continue
# order spec
if arg.startswith('>>') or arg.startswith('<<'):
# eg. [('<<', 'koo'), ('>>', 'kk')]
orderspec = re.findall('(?:(<<|>>)([^<>]+))', arg)
continue
# Ok, we have a page arg, i.e. a page or page regexp in args
pageargs = True
# Normal pages, check perms, encode and move on
if not REGEX_RE.match(arg):
# Fix relative links
if (arg.startswith('/') or arg.startswith('./')
or arg.startswith('../')):
arg = AbsPageName(request.page.page_name, arg)
argset.add(arg)
continue
# Ok, it's a page regexp
# if there's something wrong with the regexp, ignore it and move on
try:
arg = arg[1:-1]
# Fix relative links
if (arg.startswith('/') or arg.startswith('./')
or arg.startswith('../')):
arg = AbsPageName(request.page.page_name, arg)
page_re = re.compile("%s" % arg)
except:
continue
# Get all pages, check which of them match to the supplied regexp
for page in request.graphdata:
if page_re.match(page):
argset.add(page)
return (argset, pageargs, keyspec, excluded_keys, orderspec, limitregexps,
limitops, indirection_keys, styles)
def add_matching_redirs(request,
loadedPage,
loadedOuts,
loadedMeta,
metakeys,
key,
curpage,
curkey,
prev='',
formatLinks=False,
linkdata=None):
if not linkdata:
linkdata = dict()
args = curkey.split('->')
inlink = False
if args[0] == 'gwikiinlinks':
inlink = True
args = args[1:]
newkey = '->'.join(args[2:])
last = False
if not args:
return
if len(args) in [1, 2]:
last = True
if len(args) == 1:
linked, target_key = prev, args[0]
else:
linked, target_key = args[:2]
if inlink:
pages = request.graphdata.get_in(curpage).get(linked, set())
else:
pages = request.graphdata.get_out(curpage).get(linked, set())
for indir_page in set(pages):
# Relative pages etc
indir_page = AbsPageName(request.page.page_name, indir_page)
if request.user.may.read(indir_page):
pagedata = request.graphdata.getpage(indir_page)
outs = pagedata.get('out', dict())
metas = pagedata.get('meta', dict())
# Add matches at first round
if last:
if target_key in metas:
loadedMeta.setdefault(key, list())
linkdata.setdefault(key, dict())
if formatLinks:
values = metas_to_abs_links(request, indir_page,
metas[target_key])
else:
values = metas[target_key]
loadedMeta[key].extend(values)
linkdata[key].setdefault(indir_page, list()).extend(values)
else:
linkdata.setdefault(key, dict())
linkdata[key].setdefault(indir_page, list())
continue
elif not target_key in outs:
continue
# Handle inlinks separately
if 'gwikiinlinks' in metakeys:
inLinks = inlinks_key(request, loadedPage)
loadedOuts[key] = inLinks
continue
linkdata = add_matching_redirs(request, loadedPage, loadedOuts,
loadedMeta, metakeys, key,
indir_page, newkey, target_key,
formatLinks, linkdata)
return linkdata
def _metatable_parseargs(request, args, cat_re, temp_re):
# Arg placeholders
argset = set([])
keyspec = list()
excluded_keys = list()
orderspec = list()
limitregexps = dict()
limitops = dict()
# Capacity for storing indirection keys in metadata comparisons
# and regexps, eg. k->c=/.+/
indirection_keys = list()
# list styles
styles = dict()
# Flag: were there page arguments?
pageargs = False
# Regex preprocessing
for arg in (x.strip() for x in args.split(',') if x.strip()):
# metadata key spec, move on
if arg.startswith('||') and arg.endswith('||'):
# take order, strip empty ones, look at styles
for key in arg.split('||'):
if not key:
continue
# Grab styles
if key.startswith('<') and '>' in key:
style = parseAttributes(request,
key[1:], '>')
key = key[key.index('>') + 1:].strip()
if style:
styles[key] = style[0]
# Grab key exclusions
if key.startswith('!'):
excluded_keys.append(key.lstrip('!'))
continue
keyspec.append(key.strip())
continue
op_match = False
# Check for Python operator comparisons
for op in OPERATORS:
if op in arg:
data = arg.rsplit(op)
# If this is not a comparison but indirection,
# continue. Good: k->s>3, bad: k->s=/.+/
if op == '>' and data[0].endswith('-'):
continue
# Must have real comparison
if not len(data) == 2:
if op == '==':
data.append('')
else:
continue
key, comp = map(string.strip, data)
# Add indirection key
if '->' in key:
indirection_keys.append(key)
limitops.setdefault(key, list()).append((comp, op))
op_match = True
# One of the operators matched, no need to go forward
if op_match:
break
# One of the operators matched, process next arg
if op_match:
continue
# Metadata regexp, move on
if '=' in arg:
data = arg.split("=")
key = data[0]
# Add indirection key
if '->' in key:
indirection_keys.append(key)
val = '='.join(data[1:])
# Assume that value limits are regexps, if
# not, escape them into exact regexp matches
if not REGEX_RE.match(val):
from MoinMoin.parser.text_moin_wiki import Parser
# If the value is a page, make it a non-matching
# regexp so that all link variations will generate a
# match. An alternative would be to match from links
# also, but in this case old-style metalinks, which
# cannot be edited, would appear in metatables, which
# is not wanted (old-style eg. [[Page| key: Page]])
# Only allow non-matching regexp for values if they
# are WikiWords. Eg. 'WikiWord some text' would match
# 'WikiWord', emulating ye olde matching behaviour,
# but 'nonwikiword some text' would not match
# 'nonwikiword'
if re.match(Parser.word_rule_js, val):
re_val = "(%s|" % (re.escape(val))
else:
re_val = "(^%s$|" % (re.escape(val))
# or as bracketed link
re_val += "(?P<sta>\[\[)%s(?(sta)\]\])|" % (re.escape(val))
# or as commented bracketed link
re_val += "(?P<stb>\[\[)%s(?(stb)\|[^\]]*\]\]))" % \
(re.escape(val))
limitregexps.setdefault(
key, set()).add(re.compile(re_val, re.UNICODE))
# else strip the //:s
else:
if len(val) > 1:
val = val[1:-1]
limitregexps.setdefault(
key, set()).add(re.compile(val,
re.IGNORECASE | re.UNICODE))
continue
# order spec
if arg.startswith('>>') or arg.startswith('<<'):
# eg. [('<<', 'koo'), ('>>', 'kk')]
orderspec = re.findall('(?:(<<|>>)([^<>]+))', arg)
continue
# Ok, we have a page arg, i.e. a page or page regexp in args
pageargs = True
# Normal pages, check perms, encode and move on
if not REGEX_RE.match(arg):
# Fix relative links
if (arg.startswith('/') or arg.startswith('./') or
arg.startswith('../')):
arg = AbsPageName(request.page.page_name, arg)
argset.add(arg)
continue
# Ok, it's a page regexp
# if there's something wrong with the regexp, ignore it and move on
try:
arg = arg[1:-1]
# Fix relative links
if (arg.startswith('/') or arg.startswith('./') or
arg.startswith('../')):
arg = AbsPageName(request.page.page_name, arg)
page_re = re.compile("%s" % arg)
except:
continue
# Get all pages, check which of them match to the supplied regexp
for page in request.graphdata:
if page_re.match(page):
argset.add(page)
return (argset, pageargs, keyspec, excluded_keys, orderspec,
limitregexps, limitops, indirection_keys, styles)
def parse_text(request, page, text):
pagename = page.page_name
newreq = request
newreq.page = lcpage = LinkCollectingPage(newreq, pagename, text)
parserclass = importPlugin(request.cfg, "parser",
'link_collect', "Parser")
myformatter = importPlugin(request.cfg, "formatter",
'nullformatter', "Formatter")
lcpage.formatter = myformatter(newreq)
lcpage.formatter.page = lcpage
p = parserclass(lcpage.get_raw_body(), newreq, formatter=lcpage.formatter)
lcpage.parser = p
lcpage.format(p)
# These are the match types that really should be noted
linktypes = ["wikiname_bracket", "word",
"interwiki", "url", "url_bracket"]
new_data = dict_with_getpage()
# Add the page categories as links too
categories, _, _ = parse_categories(request, text)
# Process ACL:s
pi, _ = get_processing_instructions(text)
for verb, args in pi:
if verb == u'acl':
# Add all ACL:s on multiple lines to an one-lines
acls = new_data.get(pagename, dict()).get('acl', '')
acls = acls.strip() + args
new_data.setdefault(pagename, dict())['acl'] = acls
for metakey, value in p.definitions.iteritems():
for ltype, item in value:
dnode = None
if ltype in ['url', 'wikilink', 'interwiki', 'email']:
dnode = item[1]
if '#' in dnode:
# Fix anchor links to point to the anchor page
url = False
for schema in config.url_schemas:
if dnode.startswith(schema):
url = True
if not url:
# Do not fix URLs
if dnode.startswith('#'):
dnode = pagename
else:
dnode = dnode.split('#')[0]
if (dnode.startswith('/') or
dnode.startswith('./') or
dnode.startswith('../')):
# Fix relative links
dnode = AbsPageName(pagename, dnode)
hit = item[0]
elif ltype == 'category':
# print "adding cat", item, repr(categories)
dnode = item
hit = item
if item in categories:
add_link(new_data, pagename, dnode,
u"gwikicategory")
elif ltype == 'meta':
add_meta(new_data, pagename, (metakey, item))
elif ltype == 'include':
# No support for regexp includes, for now!
if not item[0].startswith("^"):
included = AbsPageName(pagename, item[0])
add_link(new_data, pagename, included, u"gwikiinclude")
if dnode:
add_link(new_data, pagename, dnode, metakey)
return new_data
