Module:Citation/CS1
Documentation for this module may be created at Module:Citation/CS1/doc
require('Module:No globals');
--[[--------------------------< F O R W A R D D E C L A R A T I O N S >--------------------------------------
each of these counts against the Lua upvalue limit
]]
local validation; -- functions in Module:Citation/CS1/Date_validation
local utilities; -- functions in Module:Citation/CS1/Utilities
local z ={}; -- table of tables in Module:Citation/CS1/Utilities
local identifiers; -- functions and tables in Module:Citation/CS1/Identifiers
local metadata; -- functions in Module:Citation/CS1/COinS
local cfg = {}; -- table of configuration tables that are defined in Module:Citation/CS1/Configuration
local whitelist = {}; -- table of tables listing valid template parameter names; defined in Module:Citation/CS1/Whitelist
--[[------------------< P A G E S C O P E V A R I A B L E S >---------------
declare variables here that have page-wide scope that are not brought in from
other modules; that are created here and used here
]]
local added_deprecated_cat; -- Boolean flag so that the category is added only once
local added_discouraged_cat; -- Boolean flag so that the category is added only once
local added_vanc_errs; -- Boolean flag so we only emit one Vancouver error / category
local Frame; -- holds the module's frame table
--[[--------------------------< F I R S T _ S E T >------------------------------------------------------------
Locates and returns the first set value in a table of values where the order established in the table,
left-to-right (or top-to-bottom), is the order in which the values are evaluated. Returns nil if none are set.
This version replaces the original 'for _, val in pairs do' and a similar version that used ipairs. With the pairs
version the order of evaluation could not be guaranteed. With the ipairs version, a nil value would terminate
the for-loop before it reached the actual end of the list.
]]
local function first_set (list, count)
local i = 1;
while i <= count do -- loop through all items in list
if utilities.is_set( list[i] ) then
return list[i]; -- return the first set list member
end
i = i + 1; -- point to next
end
end
--[[--------------------------< A D D _ V A N C _ E R R O R >----------------------------------------------------
Adds a single Vancouver system error message to the template's output regardless of how many error actually exist.
To prevent duplication, added_vanc_errs is nil until an error message is emitted.
added_vanc_errs is a Boolean declared in page scope variables above
]]
local function add_vanc_error (source, position)
if added_vanc_errs then return end
added_vanc_errs = true; -- note that we've added this category
table.insert( z.message_tail, { utilities.set_message ( 'err_vancouver', {source, position}, true ) } );
end
--[[--------------------------< I S _ S C H E M E >------------------------------------------------------------
does this thing that purports to be a URI scheme seem to be a valid scheme? The scheme is checked to see if it
is in agreement with http://tools.ietf.org/html/std66#section-3.1 which says:
Scheme names consist of a sequence of characters beginning with a
letter and followed by any combination of letters, digits, plus
("+"), period ("."), or hyphen ("-").
returns true if it does, else false
]]
local function is_scheme (scheme)
return scheme and scheme:match ('^%a[%a%d%+%.%-]*:'); -- true if scheme is set and matches the pattern
end
--[=[-------------------------< I S _ D O M A I N _ N A M E >--------------------------------------------------
Does this thing that purports to be a domain name seem to be a valid domain name?
Syntax defined here: http://tools.ietf.org/html/rfc1034#section-3.5
BNF defined here: https://tools.ietf.org/html/rfc4234
Single character names are generally reserved; see https://tools.ietf.org/html/draft-ietf-dnsind-iana-dns-01#page-15;
see also [[Single-letter second-level domain]]
list of TLDs: https://www.iana.org/domains/root/db
RFC 952 (modified by RFC 1123) requires the first and last character of a hostname to be a letter or a digit. Between
the first and last characters the name may use letters, digits, and the hyphen.
Also allowed are IPv4 addresses. IPv6 not supported
domain is expected to be stripped of any path so that the last character in the last character of the TLD. tld
is two or more alpha characters. Any preceding '//' (from splitting a URL with a scheme) will be stripped
here. Perhaps not necessary but retained in case it is necessary for IPv4 dot decimal.
There are several tests:
the first character of the whole domain name including subdomains must be a letter or a digit
internationalized domain name (ASCII characters with .xn-- ASCII Compatible Encoding (ACE) prefix xn-- in the TLD) see https://tools.ietf.org/html/rfc3490
single-letter/digit second-level domains in the .org, .cash, and .today TLDs
q, x, and z SL domains in the .com TLD
i and q SL domains in the .net TLD
single-letter SL domains in the ccTLDs (where the ccTLD is two letters)
two-character SL domains in gTLDs (where the gTLD is two or more letters)
three-plus-character SL domains in gTLDs (where the gTLD is two or more letters)
IPv4 dot-decimal address format; TLD not allowed
returns true if domain appears to be a proper name and TLD or IPv4 address, else false
]=]
local function is_domain_name (domain)
if not domain then
return false; -- if not set, abandon
end
domain = domain:gsub ('^//', ''); -- strip '//' from domain name if present; done here so we only have to do it once
if not domain:match ('^[%w]') then -- first character must be letter or digit
return false;
end
if domain:match ('^%a+:') then -- hack to detect things that look like s:Page:Title where Page: is namespace at Wikisource
return false;
end
local patterns = { -- patterns that look like URLs
'%f[%w][%w][%w%-]+[%w]%.%a%a+$', -- three or more character hostname.hostname or hostname.tld
'%f[%w][%w][%w%-]+[%w]%.xn%-%-[%w]+$', -- internationalized domain name with ACE prefix
'%f[%a][qxz]%.com$', -- assigned one character .com hostname (x.com times out 2015-12-10)
'%f[%a][iq]%.net$', -- assigned one character .net hostname (q.net registered but not active 2015-12-10)
'%f[%w][%w]%.%a%a$', -- one character hostname and ccTLD (2 chars)
'%f[%w][%w][%w]%.%a%a+$', -- two character hostname and TLD
'^%d%d?%d?%.%d%d?%d?%.%d%d?%d?%.%d%d?%d?', -- IPv4 address
}
for _, pattern in ipairs (patterns) do -- loop through the patterns list
if domain:match (pattern) then
return true; -- if a match then we think that this thing that purports to be a URL is a URL
end
end
for _, d in ipairs ({'cash', 'company', 'today', 'org'}) do -- look for single letter second level domain names for these top level domains
if domain:match ('%f[%w][%w]%.' .. d) then
return true
end
end
return false; -- no matches, we don't know what this thing is
end
--[[--------------------------< I S _ U R L >------------------------------------------------------------------
returns true if the scheme and domain parts of a URL appear to be a valid URL; else false.
This function is the last step in the validation process. This function is separate because there are cases that
are not covered by split_url(), for example is_parameter_ext_wikilink() which is looking for bracketted external
wikilinks.
]]
local function is_url (scheme, domain)
if utilities.is_set (scheme) then -- if scheme is set check it and domain
return is_scheme (scheme) and is_domain_name (domain);
else
return is_domain_name (domain); -- scheme not set when URL is protocol-relative
end
end
--[[--------------------------< S P L I T _ U R L >------------------------------------------------------------
Split a URL into a scheme, authority indicator, and domain.
First remove Fully Qualified Domain Name terminator (a dot following TLD) (if any) and any path(/), query(?) or fragment(#).
If protocol-relative URL, return nil scheme and domain else return nil for both scheme and domain.
When not protocol-relative, get scheme, authority indicator, and domain. If there is an authority indicator (one
or more '/' characters immediately following the scheme's colon), make sure that there are only 2.
Any URL that does not have news: scheme must have authority indicator (//). TODO: are there other common schemes
like news: that don't use authority indicator?
Strip off any port and path;
]]
local function split_url (url_str)
local scheme, authority, domain;
url_str = url_str:gsub ('([%a%d])%.?[/%?#].*$', '%1'); -- strip FQDN terminator and path(/), query(?), fragment (#) (the capture prevents false replacement of '//')
if url_str:match ('^//%S*') then -- if there is what appears to be a protocol-relative URL
domain = url_str:match ('^//(%S*)')
elseif url_str:match ('%S-:/*%S+') then -- if there is what appears to be a scheme, optional authority indicator, and domain name
scheme, authority, domain = url_str:match ('(%S-:)(/*)(%S+)'); -- extract the scheme, authority indicator, and domain portions
if utilities.is_set (authority) then
authority = authority:gsub ('//', '', 1); -- replace place 1 pair of '/' with nothing;
if utilities.is_set(authority) then -- if anything left (1 or 3+ '/' where authority should be) then
return scheme; -- return scheme only making domain nil which will cause an error message
end
else
if not scheme:match ('^news:') then -- except for news:..., MediaWiki won't link URLs that do not have authority indicator; TODO: a better way to do this test?
return scheme; -- return scheme only making domain nil which will cause an error message
end
end
domain = domain:gsub ('(%a):%d+', '%1'); -- strip port number if present
end
return scheme, domain;
end
--[[--------------------------< L I N K _ P A R A M _ O K >---------------------------------------------------
checks the content of |title-link=, |series-link=, |author-link=, etc. for properly formatted content: no wikilinks, no URLs
Link parameters are to hold the title of a Wikipedia article, so none of the WP:TITLESPECIALCHARACTERS are allowed:
# < > [ ] | { } _
except the underscore which is used as a space in wiki URLs and # which is used for section links
returns false when the value contains any of these characters.
When there are no illegal characters, this function returns TRUE if value DOES NOT appear to be a valid URL (the
|<param>-link= parameter is ok); else false when value appears to be a valid URL (the |<param>-link= parameter is NOT ok).
]]
local function link_param_ok (value)
local scheme, domain;
if value:find ('[<>%[%]|{}]') then -- if any prohibited characters
return false;
end
scheme, domain = split_url (value); -- get scheme or nil and domain or nil from URL;
return not is_url (scheme, domain); -- return true if value DOES NOT appear to be a valid URL
end
--[[--------------------------< L I N K _ T I T L E _ O K >---------------------------------------------------
Use link_param_ok() to validate |<param>-link= value and its matching |<title>= value.
|<title>= may be wiki-linked but not when |<param>-link= has a value. This function emits an error message when
that condition exists
check <link> for inter-language interwiki-link prefix. prefix must be a MediaWiki-recognized language
code and must begin with a colon.
]]
local function link_title_ok (link, lorig, title, torig)
local orig;
if utilities.is_set (link) then -- don't bother if <param>-link doesn't have a value
if not link_param_ok (link) then -- check |<param>-link= markup
orig = lorig; -- identify the failing link parameter
elseif title:find ('%[%[') then -- check |title= for wikilink markup
orig = torig; -- identify the failing |title= parameter
elseif link:match ('^%a+:') then -- if the link is what looks like an interwiki
local prefix = link:match ('^(%a+):'):lower(); -- get the interwiki prefix
if cfg.inter_wiki_map[prefix] then -- if prefix is in the map, must have preceding colon
orig = lorig; -- flag as error
end
end
end
if utilities.is_set (orig) then
link = ''; -- unset
table.insert( z.message_tail, { utilities.set_message ( 'err_bad_paramlink', orig)}); -- URL or wikilink in |title= with |title-link=;
end
return link; -- link if ok, empty string else
end
--[[--------------------------< C H E C K _ U R L >------------------------------------------------------------
Determines whether a URL string appears to be valid.
First we test for space characters. If any are found, return false. Then split the URL into scheme and domain
portions, or for protocol-relative (//example.com) URLs, just the domain. Use is_url() to validate the two
portions of the URL. If both are valid, or for protocol-relative if domain is valid, return true, else false.
Because it is different from a standard URL, and because this module used external_link() to make external links
that work for standard and news: links, we validate newsgroup names here. The specification for a newsgroup name
is at https://tools.ietf.org/html/rfc5536#section-3.1.4
]]
local function check_url( url_str )
if nil == url_str:match ("^%S+$") then -- if there are any spaces in |url=value it can't be a proper URL
return false;
end
local scheme, domain;
scheme, domain = split_url (url_str); -- get scheme or nil and domain or nil from URL;
if 'news:' == scheme then -- special case for newsgroups
return domain:match('^[%a%d%+%-_]+%.[%a%d%+%-_%.]*[%a%d%+%-_]$');
end
return is_url (scheme, domain); -- return true if value appears to be a valid URL
end
--[=[-------------------------< I S _ P A R A M E T E R _ E X T _ W I K I L I N K >----------------------------
Return true if a parameter value has a string that begins and ends with square brackets [ and ] and the first
non-space characters following the opening bracket appear to be a URL. The test will also find external wikilinks
that use protocol-relative URLs. Also finds bare URLs.
The frontier pattern prevents a match on interwiki-links which are similar to scheme:path URLs. The tests that
find bracketed URLs are required because the parameters that call this test (currently |title=, |chapter=, |work=,
and |publisher=) may have wikilinks and there are articles or redirects like '//Hus' so, while uncommon, |title=[[//Hus]]
is possible as might be [[en://Hus]].
]=]
local function is_parameter_ext_wikilink (value)
local scheme, domain;
if value:match ('%f[%[]%[%a%S*:%S+.*%]') then -- if ext. wikilink with scheme and domain: [xxxx://yyyyy.zzz]
scheme, domain = split_url (value:match ('%f[%[]%[(%a%S*:%S+).*%]'));
elseif value:match ('%f[%[]%[//%S+.*%]') then -- if protocol-relative ext. wikilink: [//yyyyy.zzz]
scheme, domain = split_url (value:match ('%f[%[]%[(//%S+).*%]'));
elseif value:match ('%a%S*:%S+') then -- if bare URL with scheme; may have leading or trailing plain text
scheme, domain = split_url (value:match ('(%a%S*:%S+)'));
elseif value:match ('//%S+') then -- if protocol-relative bare URL: //yyyyy.zzz; may have leading or trailing plain text
scheme, domain = split_url (value:match ('(//%S+)')); -- what is left should be the domain
else
return false; -- didn't find anything that is obviously a URL
end
return is_url (scheme, domain); -- return true if value appears to be a valid URL
end
--[[-------------------------< C H E C K _ F O R _ U R L >-----------------------------------------------------
loop through a list of parameters and their values. Look at the value and if it has an external link, emit an error message.
]]
local function check_for_url (parameter_list)
local error_message = '';
for k, v in pairs (parameter_list) do -- for each parameter in the list
if is_parameter_ext_wikilink (v) then -- look at the value; if there is a URL add an error message
if utilities.is_set(error_message) then -- once we've added the first portion of the error message ...
error_message = error_message .. ", "; -- ... add a comma space separator
end
error_message = error_message .. "|" .. k .. "="; -- add the failed parameter
end
end
if utilities.is_set (error_message) then -- done looping, if there is an error message, display it
table.insert( z.message_tail, { utilities.set_message ( 'err_param_has_ext_link', {error_message}, true ) } );
end
end
--[[--------------------------< S A F E _ F O R _ U R L >------------------------------------------------------
Escape sequences for content that will be used for URL descriptions
]]
local function safe_for_url( str )
if str:match( "%[%[.-%]%]" ) ~= nil then
table.insert( z.message_tail, { utilities.set_message ( 'err_wikilink_in_url', {}, true ) } );
end
return str:gsub( '[%[%]\n]', {
['['] = '[',
[']'] = ']',
['\n'] = ' ' } );
end
--[[--------------------------< E X T E R N A L _ L I N K >----------------------------------------------------
Format an external link with error checking
]]
local function external_link( URL, label, source, access)
local error_str = "";
local domain;
local path;
local base_url;
if not utilities.is_set ( label ) then
label = URL;
if utilities.is_set ( source ) then
error_str = utilities.set_message ( 'err_bare_url_missing_title', { utilities.wrap_style ('parameter', source) }, false, " " );
else
error( cfg.messages["bare_url_no_origin"] );
end
end
if not check_url( URL ) then
error_str = utilities.set_message ( 'err_bad_url', {utilities.wrap_style ('parameter', source)}, false, " " ) .. error_str;
end
domain, path = URL:match ('^([/%.%-%+:%a%d]+)([/%?#].*)$'); -- split the URL into scheme plus domain and path
if path then -- if there is a path portion
path = path:gsub ('[%[%]]', {['['] = '%5b', [']'] = '%5d'}); -- replace '[' and ']' with their percent-encoded values
URL = table.concat ({domain, path}); -- and reassemble
end
base_url = table.concat({ "[", URL, " ", safe_for_url (label), "]" }); -- assemble a wiki-markup URL
if utilities.is_set (access) then -- access level (subscription, registration, limited)
base_url = utilities.substitute (cfg.presentation['ext-link-access-signal'], {cfg.presentation[access].class, cfg.presentation[access].title, base_url}); -- add the appropriate icon
end
return table.concat ({base_url, error_str});
end
--[[--------------------------< D E P R E C A T E D _ P A R A M E T E R >--------------------------------------
Categorize and emit an error message when the citation contains one or more deprecated parameters. The function includes the
offending parameter name to the error message. Only one error message is emitted regardless of the number of deprecated
parameters in the citation.
added_deprecated_cat is a Boolean declared in page scope variables above
]]
local function deprecated_parameter(name)
if not added_deprecated_cat then
added_deprecated_cat = true; -- note that we've added this category
table.insert( z.message_tail, { utilities.set_message ( 'err_deprecated_params', {name}, true ) } ); -- add error message
end
end
--[[--------------------------< D I S C O U R A G E D _ P A R A M E T E R >------------------------------------
Categorize and emit an maintenance message when the citation contains one or more discouraged parameters. Only
one error message is emitted regardless of the number of discouraged parameters in the citation.
added_discouraged_cat is a Boolean declared in page scope variables above
]]
local function discouraged_parameter(name)
if not added_discouraged_cat then
added_discouraged_cat = true; -- note that we've added this category
table.insert( z.message_tail, { utilities.set_message ( 'maint_discouraged', {name}, true ) } ); -- add maint message
end
end
--[=[-------------------------< K E R N _ Q U O T E S >--------------------------------------------------------
Apply kerning to open the space between the quote mark provided by the module and a leading or trailing quote
mark contained in a |title= or |chapter= parameter's value.
This function will positive kern either single or double quotes:
"'Unkerned title with leading and trailing single quote marks'"
" 'Kerned title with leading and trailing single quote marks' " (in real life the kerning isn't as wide as this example)
Double single quotes (italic or bold wiki-markup) are not kerned.
Replaces Unicode quote marks in plain text or in the label portion of a [[L|D]] style wikilink with typewriter
quote marks regardless of the need for kerning. Unicode quote marks are not replaced in simple [[D]] wikilinks.
Call this function for chapter titles, for website titles, etc.; not for book titles.
]=]
local function kern_quotes (str)
local cap = '';
local cap2 = '';
local wl_type, label, link;
wl_type, label, link =