def compare_links(user: str, sitedata: SiteData, sig: str) -> Union[bool, Set[str]]:
"""Compare links in a sig to data in sitedata"""
wikitext = mwph.parse(sig)
user = datasources.normal_name(user)
errors = set()
for link in wikitext.ifilter_wikilinks():
title = str(link.title).partition("#")[0]
# Extract namespace and page.
# Interwiki prefixes are left in the namespace
if ":" in user:
# Colons in usernames break the partitioning
if title.endswith(f":{user}"):
ns = title.replace(f":{user}", "")
sep = ":"
page = title.replace(f"{ns}:", "")
elif title.endswith(f"/{user}"):
raw = title.replace(f"/{user}", "")
ns, sep, page = raw.rpartition(":")
page += f"/{user}"
else:
continue # pragma: no cover
else:
ns, sep, page = title.rpartition(":")
# normalize namespace and strip whitespace from both
ns, page = datasources.normal_name(ns.strip()), page.strip()
# remove leading colon from namespace
if ns.startswith(":"):
ns = ns[1:]
# Check if linking to user or user talk
if not sep:
continue
elif ":" in ns:
errors.add("interwiki-user-link")
elif ns in sitedata.user or ns in sitedata.user_talk:
# Check that it's the right user or user_talk
if datasources.normal_name(page) == user:
return True
else:
errors.add("link-username-mismatch")
continue
elif ns in sitedata.special:
# Could be a contribs page, check
# split page and normalize names
specialpage, slash, target = page.partition("/")
specialpage = datasources.normal_name(specialpage.strip())
target = datasources.normal_name(target.strip())
if specialpage in sitedata.contribs:
# It's contribs
if target == user:
# The right one
return True
else:
errors.add("link-username-mismatch")
continue # pragma: no cover
else:
continue # pragma: no cover
else:
return errors
def check_impersonation(sig: str, user: str, sitedata: SiteData) -> Optional[SigError]:
wikitext = mwph.parse(sig)
problem = False
for link in wikitext.ifilter_wikilinks():
if not link.text:
continue
text = datasources.normal_name(link.text)
if compare_links(user, sitedata, link) is True:
if text == datasources.normal_name(user):
# one link matches, that's good enough
break
elif datasources.check_user_exists(text, sitedata):
problem = True
if problem:
return SigError.LINK_NAME
else:
return None
def check_images(sig: str, sitedata: SiteData) -> Optional[SigError]:
"""Check for displayed images in a signature"""
wikitext = mwph.parse(sig)
for link in wikitext.ifilter_wikilinks():
title = link.title
# if it starts with :, it's not a displayed image
if title.startswith(":"):
continue
# Can't interwiki transclude an image, so the extra safety
# in check_links isn't required
ns, sep, page = title.partition(":")
if not sep:
continue
if datasources.normal_name(ns) in sitedata.file:
return SigError.IMAGES
else:
return None
def get_site_data(hostname: str) -> SiteData:
"""Get metadata about a site from the API"""
url = f"https://{hostname}/w/api.php"
data = dict(
action="query",
meta="siteinfo",
siprop="|".join([
"namespaces",
"namespacealiases",
"specialpagealiases",
"magicwords",
"general",
]),
formatversion="2",
format="json",
)
res_json = backoff_retry("get", url, params=data, output="json")
namespaces: Dict[str, Set[str]] = {}
all_namespaces = res_json["query"]["namespaces"]
namespace_aliases = res_json["query"]["namespacealiases"]
for namespace, nsdata in all_namespaces.items():
namespaces.setdefault(namespace, set()).update([
datasources.normal_name(nsdata.get("canonical", "")),
datasources.normal_name(nsdata.get("name", "")),
])
for nsdata in namespace_aliases:
namespaces.setdefault(str(nsdata["id"]), set()).add(
datasources.normal_name(nsdata.get("alias", "")))
specialpages = {
item["realname"]: item["aliases"]
for item in res_json["query"]["specialpagealiases"]
}
magicwords = {
item["name"]: item["aliases"]
for item in res_json["query"]["magicwords"]
}
general = res_json["query"]["general"]
contribs = {
datasources.normal_name(name)
for name in specialpages["Contributions"]
}
subst = list(
itertools.chain(
magicwords.get("subst", ["SUBST"]),
[item.lower() for item in magicwords.get("subst", ["SUBST"])],
[
item[0] + item[1:].lower()
for item in magicwords.get("subst", ["SUBST"])
],
))
sitedata = SiteData(
user=namespaces["2"] - {""},
user_talk=namespaces["3"] - {""},
file=namespaces["6"] - {""},
special=namespaces["-1"] - {""},
contribs=contribs,
subst=subst,
dbname=general["wikiid"],
hostname=hostname,
)
return sitedata