Documentation for this module may be created at Module:Citation/CS1/doc
--[[--------------------------< F O R W A R D D E C L A R A T I O N S >--------------------------------------
]]
local dates, year_date_check, reformat_dates, date_hyphen_to_dash, -- functions in Module:Citation/CS1/Date_validation
date_name_xlate
local is_set, in_array, substitute, error_comment, set_error, select_one, -- functions in Module:Citation/CS1/Utilities
add_maint_cat, wrap_style, safe_for_italics, is_wikilink, make_wikilink,
strip_apostrophe_markup;
local z ={}; -- tables in Module:Citation/CS1/Utilities
local extract_ids, extract_id_access_levels, build_id_list, is_embargoed; -- functions in Module:Citation/CS1/Identifiers
local make_coins_title, get_coins_pages, COinS; -- 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 >--------------------------------------
delare variables here that have page-wide scope that are not brought in from other modules; thatare created here
and used here
]]
local added_deprecated_cat; -- boolean flag so that the category is added only once
local added_prop_cats = {}; -- list of property categories that have been added to z.properties_cats
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 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 _ P R O P _ C A T >--------------------------------------------------------
Adds a category to z.properties_cats using names from the configuration file with additional text if any.
foreign_lang_source and foreign_lang_source_2 keys have a language code appended to them so that multiple languages
may be categorized but multiples of the same language are not categorized.
added_prop_cats is a table declared in page scope variables above
]]
local function add_prop_cat (key, arguments)
if not added_prop_cats [key] then
added_prop_cats [key] = true; -- note that we've added this category
key = key:gsub ('(foreign_lang_source_?2?)%a%a%a?[%a%-]*', '%1'); -- strip lang code from keyname
table.insert( z.properties_cats, substitute (cfg.prop_cats [key], arguments)); -- make name then add to table
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)
if not added_vanc_errs then
added_vanc_errs = true; -- note that we've added this category
table.insert( z.message_tail, { set_error( 'vancouver', {source}, true ) } );
end
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
rfc952 (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 incase 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 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.
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
authority = authority:gsub ('//', '', 1); -- replace place 1 pair of '/' with nothing;
if 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
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 wikilinked but not when |<param>-link= has a value. This function emits an error message when
that condition exists
]]
local function link_title_ok (link, lorig, title, torig)
local orig;
if 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
end
end
if is_set (orig) then
table.insert( z.message_tail, { set_error( 'bad_paramlink', orig)}); -- url or wikilink in |title= with |title-link=;
end
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 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 is_set (error_message) then -- done looping, if there is an error message, display it
table.insert( z.message_tail, { set_error( '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, { set_error( '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 is_set( label ) then
label = URL;
if is_set( source ) then
error_str = set_error( 'bare_url_missing_title', { wrap_style ('parameter', source) }, false, " " );
else
error( cfg.messages["bare_url_no_origin"] );
end
end
if not check_url( URL ) then
error_str = set_error( 'bad_url', {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 wikimarkup url
if is_set (access) then -- access level (subscription, registration, limited)
base_url = 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, { set_error( 'deprecated_params', {name}, true ) } ); -- add error 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 wikimarkup) are not kerned.
Replaces unicode quotemarks 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 = is_wikilink (str); -- wl_type is: 0, no wl (text in label variable); 1, [[D]]; 2, [[L|D]]
if 1 == wl_type then -- [[D]] simple wikilink with or without quote marks
if mw.ustring.match (str, '%[%[[\"“”\'‘’].+[\"“”\'‘’]%]%]') then -- leading and trailing quote marks
str = substitute (cfg.presentation['kern-wl-both'], str);
elseif mw.ustring.match (str, '%[%[[\"“”\'‘’].+%]%]') then -- leading quote marks
str = substitute (cfg.presentation['kern-wl-left'], str);
elseif mw.ustring.match (str, '%[%[.+[\"“”\'‘’]%]%]') then -- trailing quote marks
str = substitute (cfg.presentation['kern-wl-right'], str);
end
else -- plain text or [[L|D]]; text in label variable
label= mw.ustring.gsub (label, '[“”]', '\"'); -- replace “” (U+201C & U+201D) with " (typewriter double quote mark)
label= mw.ustring.gsub (label, '[‘’]', '\''); -- replace ‘’ (U+2018 & U+2019) with ' (typewriter single quote mark)
cap, cap2 = mw.ustring.match (label, "^([\"\'])([^\'].+)"); -- match leading double or single quote but not doubled single quotes (italic markup)
if is_set (cap) then
label = substitute (cfg.presentation['kern-left'], {cap, cap2});
end
cap, cap2 = mw.ustring.match (label, "^(.+[^\'])([\"\'])$") -- match trailing double or single quote but not doubled single quotes (italic markup)
if is_set (cap) then
label = substitute (cfg.presentation['kern-right'], {cap, cap2});
end
if 2 == wl_type then
str = make_wikilink (link, label); -- reassemble the wikilink
else
str = label;
end
end
return str;
end
--[[--------------------------< F O R M A T _ S C R I P T _ V A L U E >----------------------------------------
|script-title= holds title parameters that are not written in Latin based scripts: Chinese, Japanese, Arabic, Hebrew, etc. These scripts should
not be italicized and may be written right-to-left. The value supplied by |script-title= is concatenated onto Title after Title has been wrapped
in italic markup.
Regardless of language, all values provided by |script-title= are wrapped in <bdi>...</bdi> tags to isolate rtl languages from the English left to right.
|script-title= provides a unique feature. The value in |script-title= may be prefixed with a two-character ISO639-1 language code and a colon:
|script-title=ja:*** *** (where * represents a Japanese character)
Spaces between the two-character code and the colon and the colon and the first script character are allowed:
|script-title=ja : *** ***
|script-title=ja: *** ***
|script-title=ja :*** ***
Spaces preceding the prefix are allowed: |script-title = ja:*** ***
The prefix is checked for validity. If it is a valid ISO639-1 language code, the lang attribute (lang="ja") is added to the <bdi> tag so that browsers can
know the language the tag contains. This may help the browser render the script more correctly. If the prefix is invalid, the lang attribute
is not added. At this time there is no error message for this condition.
Supports |script-title=, |script-chapter=, |script-<periodical>=
]]
local function format_script_value (script_value, script_param)
local lang=''; -- initialize to empty string
local name;
if script_value:match('^%l%l%l?%s*:') then -- if first 3 or 4 non-space characters are script language prefix
lang = script_value:match('^(%l%l%l?)%s*:%s*%S.*'); -- get the language prefix or nil if there is no script
if not is_set (lang) then
table.insert( z.message_tail, { set_error( 'script_parameter', {script_param, 'missing title part'}, true ) } ); -- prefix without 'title'; add error message
return ''; -- script_value was just the prefix so return empty string
end
-- if we get this far we have prefix and script
name = cfg.lang_code_remap[lang] or mw.language.fetchLanguageName( lang, cfg.this_wiki_code ); -- get language name so that we can use it to categorize
if is_set (name) then -- is prefix a proper ISO 639-1 language code?
script_value = script_value:gsub ('^%l+%s*:%s*', ''); -- strip prefix from script
-- is prefix one of these language codes?
if in_array (lang, cfg.script_lang_codes) then
add_prop_cat ('script_with_name', {name, lang})
else
table.insert( z.message_tail, { set_error( 'script_parameter', {script_param, 'unknown language code'}, true ) } ); -- unknown script-language; add error message
end
lang = ' lang="' .. lang .. '" '; -- convert prefix into a lang attribute
else
table.insert( z.message_tail, { set_error( 'script_parameter', {script_param, 'invalid language code'}, true ) } ); -- invalid language code; add error message
lang = ''; -- invalid so set lang to empty string
end
else
table.insert( z.message_tail, { set_error( 'script_parameter', {script_param, 'missing prefix'}, true ) } ); -- no language code prefix; add error message
end
script_value = substitute (cfg.presentation['bdi'], {lang, script_value}); -- isolate in case script is rtl
return script_value;
end
--[[--------------------------< S C R I P T _ C O N C A T E N A T E >------------------------------------------
Initially for |title= and |script-title=, this function concatenates those two parameter values after the script value has been
wrapped in <bdi> tags.
]]
local function script_concatenate (title, script, script_param)
if is_set (script) then
script = format_script_value (script, script_param); -- <bdi> tags, lang atribute, categorization, etc; returns empty string on error
if is_set (script) then
title = title .. ' ' .. script; -- concatenate title and script title
end
end
return title;
end
--[[--------------------------< W R A P _ M S G >--------------------------------------------------------------
Applies additional message text to various parameter values. Supplied string is wrapped using a message_list
configuration taking one argument. Supports lower case text for {{citation}} templates. Additional text taken
from citation_config.messages - the reason this function is similar to but separate from wrap_style().
]]
local function wrap_msg (key, str, lower)
if not is_set( str ) then
return "";
end
if true == lower then
local msg;
msg = cfg.messages[key]:lower(); -- set the message to lower case before
return substitute( msg, str ); -- including template text
else
return substitute( cfg.messages[key], str );
end
end
--[[--------------------------< W I K I S O U R C E _ U R L _ M A K E >----------------------------------------
makes a wikisource url from wikisource interwiki link. returns the url and appropriate label; nil else.
str is the value assigned to |chapter= (or aliases) or |title= or |title-link=
]]
local function wikisource_url_make (str)
local wl_type, D, L;
local ws_url, ws_label;
local wikisource_prefix = table.concat ({'https://', cfg.this_wiki_code, '.wikisource.org/wiki/'});
wl_type, D, L = is_wikilink (str); -- wl_type is 0 (not a wikilink), 1 (simple wikilink), 2 (complex wikilink)
if 0 == wl_type then -- not a wikilink; might be from |title-link=
str = D:match ('^[Ww]ikisource:(.+)') or D:match ('^[Ss]:(.+)'); -- article title from interwiki link with long-form or short-form namespace
if is_set (str) then
ws_url = table.concat ({ -- build a wikisource url
wikisource_prefix, -- prefix
str, -- article title
});
ws_label = str; -- label for the url
end
elseif 1 == wl_type then -- simple wikilink: [[Wikisource:ws article]]
str = D:match ('^[Ww]ikisource:(.+)') or D:match ('^[Ss]:(.+)'); -- article title from interwiki link with long-form or short-form namespace
if is_set (str) then
ws_url = table.concat ({ -- build a wikisource url
wikisource_prefix, -- prefix
str, -- article title
});
ws_label = str; -- label for the url
end
elseif 2 == wl_type then -- non-so-simple wikilink: [[Wikisource:ws article|displayed text]] ([[L|D]])
str = L:match ('^[Ww]ikisource:(.+)') or L:match ('^[Ss]:(.+)'); -- article title from interwiki link with long-form or short-form namespace
if is_set (str) then
ws_label = D; -- get ws article name from display portion of interwiki link
ws_url = table.concat ({ -- build a wikisource url
wikisource_prefix, -- prefix
str, -- article title without namespace from link portion of wikilink
});
end
end
if ws_url then
ws_url = mw.uri.encode (ws_url, 'WIKI'); -- make a usable url
ws_url = ws_url:gsub ('%%23', '#'); -- undo percent encoding of fragment marker
end
return ws_url, ws_label, L or D; -- return proper url or nil and a label or nil
end
--[[--------------------------< F O R M A T _ P E R I O D I C A L >--------------------------------------------
Format the three periodical parameters: |script-<periodical>=, |<periodical>=, and |trans-<periodical>= into a single Periodical meta-
parameter.
]]
local function format_periodical (script_periodical, script_periodical_source, periodical, trans_periodical)
local periodical_error = '';
if not is_set (periodical) then
periodical = ''; -- to be safe for concatenation
else
periodical = wrap_style ('italic-title', periodical); -- style
end
periodical = script_concatenate (periodical, script_periodical, script_periodical_source); -- <bdi> tags, lang atribute, categorization, etc; must be done after title is wrapped
if is_set (trans_periodical) then
trans_periodical = wrap_style ('trans-italic-title', trans_periodical);
if is_set (periodical) then
periodical = periodical .. ' ' .. trans_periodical;
else -- here when trans-periodical without periodical or script-periodical
periodical = trans_periodical;
periodical_error = ' ' .. set_error ('trans_missing_title', {'periodical'});
end
end
return periodical .. periodical_error;
end
--[[--------------------------< F O R M A T _ C H A P T E R _ T I T L E >--------------------------------------
Format the four chapter parameters: |script-chapter=, |chapter=, |trans-chapter=, and |chapter-url= into a single Chapter meta-
parameter (chapter_url_source used for error messages).
]]
local function format_chapter_title (scriptchapter, script_chapter_source, chapter, transchapter, chapterurl, chapter_url_source, no_quotes, access)
local chapter_error = '';
local ws_url, ws_label, L = wikisource_url_make (chapter); -- make a wikisource url and label from a wikisource interwiki link
if ws_url then
ws_label = ws_label:gsub ('_', ''); -- replace underscore separaters with space characters
chapter = ws_label;
end
if not is_set (chapter) then
chapter = ''; -- to be safe for concatenation
else
if false == no_quotes then
chapter = kern_quotes (chapter); -- if necessary, separate chapter title's leading and trailing quote marks from Module provided quote marks
chapter = wrap_style ('quoted-title', chapter);
end
end
chapter = script_concatenate (chapter, scriptchapter, script_chapter_source); -- <bdi> tags, lang atribute, categorization, etc; must be done after title is wrapped
if is_set (chapterurl) then
chapter = external_link (chapterurl, chapter, chapter_url_source, access); -- adds bare_url_missing_title error if appropriate
elseif ws_url then
chapter = external_link (ws_url, chapter .. ' ', 'ws link in chapter'); -- adds bare_url_missing_title error if appropriate; space char to move icon away from chap text; TODO: better way to do this?
chapter = substitute (cfg.presentation['interwiki-icon'], {cfg.presentation['class-wikisource'], L, chapter});
end
if is_set (transchapter) then
transchapter = wrap_style ('trans-quoted-title', transchapter);
if is_set (chapter) then
chapter = chapter .. ' ' .. transchapter;
else -- here when transchapter without chapter or script-chapter
chapter = transchapter;
chapter_error = ' ' .. set_error ('trans_missing_title', {'chapter'});
end
end
return chapter .. chapter_error;
end
--[[--------------------------< H A S _ I N V I S I B L E _ C H A R S >----------------------------------------
This function searches a parameter's value for nonprintable or invisible characters. The search stops at the
first match.
This function will detect the visible replacement character when it is part of the wikisource.
Detects but ignores nowiki and math stripmarkers. Also detects other named stripmarkers (gallery, math, pre, ref)
and identifies them with a slightly different error message. See also coins_cleanup().
Output of this function is an error message that identifies the character or the Unicode group, or the stripmarker
that was detected along with its position (or, for multi-byte characters, the position of its first byte) in the
parameter value.
]]
local function has_invisible_chars (param, v)
local position = ''; -- position of invisible char or starting position of stripmarker
local dummy; -- end of matching string; not used but required to hold end position when a capture is returned
local capture; -- used by stripmarker detection to hold name of the stripmarker
local i=1;
local stripmarker, apostrophe;
capture = string.match (v, '[%w%p ]*'); -- Test for values that are simple ASCII text and bypass other tests if true
if capture == v then -- if same there are no unicode characters
return;
end
while cfg.invisible_chars[i] do
local char=cfg.invisible_chars[i][1] -- the character or group name
local pattern=cfg.invisible_chars[i][2] -- the pattern used to find it
position, dummy, capture = mw.ustring.find (v, pattern) -- see if the parameter value contains characters that match the pattern
if position and (char == 'zero width joiner') then -- if we found a zero width joiner character
if mw.ustring.find (v, cfg.indic_script) then -- its ok if one of the indic scripts
position = nil; -- unset position
end
end
if position then
if 'nowiki' == capture or 'math' == capture or -- nowiki and math stripmarkers (not an error condition)
('templatestyles' == capture and in_array (param, {'id', 'quote'})) then -- templatestyles stripmarker allowed in these parameters
stripmarker = true; -- set a flag
elseif true == stripmarker and 'delete' == char then -- because stripmakers begin and end with the delete char, assume that we've found one end of a stripmarker
position = nil; -- unset
else
local err_msg;
if capture then
err_msg = capture .. ' ' .. char;
else
err_msg = char .. ' ' .. 'character';
end
table.insert( z.message_tail, { set_error( 'invisible_char', {err_msg, wrap_style ('parameter', param), position}, true ) } ); -- add error message
return; -- and done with this parameter
end
end
i=i+1; -- bump our index
end
end
--[[--------------------------< A R G U M E N T _ W R A P P E R >----------------------------------------------
Argument wrapper. This function provides support for argument mapping defined in the configuration file so that
multiple names can be transparently aliased to single internal variable.
]]
local function argument_wrapper( args )
local origin = {};
return setmetatable({
ORIGIN = function( self, k )
local dummy = self[k]; --force the variable to be loaded.
return origin[k];
end
},
{
__index = function ( tbl, k )
if origin[k] ~= nil then
return nil;
end
local args, list, v = args, cfg.aliases[k];
if type( list ) == 'table' then
v, origin[k] = select_one( args, list, 'redundant_parameters' );
if origin[k] == nil then
origin[k] = ''; -- Empty string, not nil
end
elseif list ~= nil then
v, origin[k] = args[list], list;
else
-- maybe let through instead of raising an error?
-- v, origin[k] = args[k], k;
error( cfg.messages['unknown_argument_map'] .. ': ' .. k);
end
-- Empty strings, not nil;
if v == nil then
-- v = cfg.defaults[k] or '';
v = '';
origin[k] = '';
end
tbl = rawset( tbl, k, v );
return v;
end,
});
end
--[[--------------------------< N O W R A P _ D A T E >--------------------------------------------------------
When date is YYYY-MM-DD format wrap in nowrap span: <span ...>YYYY-MM-DD</span>. When date is DD MMMM YYYY or is
MMMM DD, YYYY then wrap in nowrap span: <span ...>DD MMMM</span> YYYY or <span ...>MMMM DD,</span> YYYY
DOES NOT yet support MMMM YYYY or any of the date ranges.
]]
local function nowrap_date (date)
local cap='';
local cap2='';
if date:match("^%d%d%d%d%-%d%d%-%d%d$") then
date = substitute (cfg.presentation['nowrap1'], date);
elseif date:match("^%a+%s*%d%d?,%s+%d%d%d%d$") or date:match ("^%d%d?%s*%a+%s+%d%d%d%d$") then
cap, cap2 = string.match (date, "^(.*)%s+(%d%d%d%d)$");
date = substitute (cfg.presentation['nowrap2'], {cap, cap2});
end
return date;
end
--[[--------------------------< S E T _ T I T L E T Y P E >----------------------------------------------------
This function sets default title types (equivalent to the citation including |type=<default value>) for those templates that have defaults.
Also handles the special case where it is desirable to omit the title type from the rendered citation (|type=none).
]]
local function set_titletype (cite_class, title_type)
if is_set(title_type) then
if cfg.keywords_xlate[title_type] == 'none' then
title_type = ""; -- if |type=none then type parameter not displayed
end
return title_type; -- if |type= has been set to any other value use that value
end
return cfg.title_types [cite_class] or ''; -- set template's default title type; else empty string for concatenation
end
--[[--------------------------< H Y P H E N _ T O _ D A S H >--------------------------------------------------
Converts a hyphen to a dash under certain conditions. The hyphen must separate like items; unlike items are
returned unmodified. These forms are modified:
letter - letter (A - B)
digit - digit (4-5)
digit separator digit - digit separator digit (4.1-4.5 or 4-1-4-5)
letterdigit - letterdigit (A1-A5) (an optional separator between letter and digit is supported – a.1-a.5 or a-1-a-5)
digitletter - digitletter (5a - 5d) (an optional separator between letter and digit is supported – 5.a-5.d or 5-a-5-d)
any other forms are returned unmodified.
str may be a comma- or semicolon-separated list
]]
local function hyphen_to_dash( str )
if not is_set (str) then
return str;
end
str, count = str:gsub ('^%(%((.+)%)%)$', '%1'); -- remove accept-this-as-written markup when it wraps all of str
if 0 ~= count then -- non-zero when markup removed; zero else
return str; -- nothing to do, we're done
end
str = str:gsub ('&[nm]dash;', {['–'] = '–', ['—'] = '—'}); -- replace — and – entities with their characters; semicolon mucks up the text.split
str = str:gsub ('-', '-'); -- replace html numeric entity with hyphen character
str = str:gsub (' ', ' '); -- replace entity with generic keyboard space character
local out = {};
local list = mw.text.split (str, '%s*[,;]%s*'); -- split str at comma or semicolon separators if there are any
for _, item in ipairs (list) do -- for each item in the list
if mw.ustring.match (item, '^%w*[%.%-]?%w+%s*[%-–—]%s*%w*[%.%-]?%w+$') then -- if a hyphenated range or has endash or emdash separators
if item:match ('^%a+[%.%-]?%d+%s*%-%s*%a+[%.%-]?%d+$') or -- letterdigit hyphen letterdigit (optional separator between letter and digit)
item:match ('^%d+[%.%-]?%a+%s*%-%s*%d+[%.%-]?%a+$') or -- digitletter hyphen digitletter (optional separator between digit and letter)
item:match ('^%d+[%.%-]%d+%s*%-%s*%d+[%.%-]%d+$') or -- digit separator digit hyphen digit separator digit
item:match ('^%d+%s*%-%s*%d+$') or -- digit hyphen digit
item:match ('^%a+%s*%-%s*%a+$') then -- letter hyphen letter
item = item:gsub ('(%w*[%.%-]?%w+)%s*%-%s*(%w*[%.%-]?%w+)', '%1–%2'); -- replace hyphen, remove extraneous space characters
else
item = mw.ustring.gsub (item, '%s*[–—]%s*', '–'); -- for endash or emdash separated ranges, replace em with en, remove extraneous white space
end
end
item = item:gsub ('^%(%((.+)%)%)$', '%1'); -- remove the accept-this-as-written markup
table.insert (out, item); -- add the (possibly modified) item to the output table
end
return table.concat (out, ', '); -- concatenate the output table into a comma separated string
end
--[[--------------------------< S A F E _ J O I N >------------------------------------------------------------
Joins a sequence of strings together while checking for duplicate separation characters.
]]
local function safe_join( tbl, duplicate_char )
local f = {}; -- create a function table appropriate to type of 'dupicate character'
if 1 == #duplicate_char then -- for single byte ascii characters use the string library functions
f.gsub=string.gsub
f.match=string.match
f.sub=string.sub
else -- for multi-byte characters use the ustring library functions
f.gsub=mw.ustring.gsub
f.match=mw.ustring.match
f.sub=mw.ustring.sub
end
local str = ''; -- the output string
local comp = ''; -- what does 'comp' mean?
local end_chr = '';
local trim;
for _, value in ipairs( tbl ) do
if value == nil then value = ''; end
if str == '' then -- if output string is empty
str = value; -- assign value to it (first time through the loop)
elseif value ~= '' then
if value:sub(1,1) == '<' then -- Special case of values enclosed in spans and other markup.
comp = value:gsub( "%b<>", "" ); -- remove html markup (<span>string</span> -> string)
else
comp = value;
end
-- typically duplicate_char is sepc
if f.sub(comp, 1,1) == duplicate_char then -- is first character same as duplicate_char? why test first character?
-- Because individual string segments often (always?) begin with terminal punct for the
-- preceding segment: 'First element' .. 'sepc next element' .. etc?
trim = false;
end_chr = f.sub(str, -1,-1); -- get the last character of the output string
-- str = str .. "<HERE(enchr=" .. end_chr.. ")" -- debug stuff?
if end_chr == duplicate_char then -- if same as separator
str = f.sub(str, 1,-2); -- remove it
elseif end_chr == "'" then -- if it might be wikimarkup
if f.sub(str, -3,-1) == duplicate_char .. "''" then -- if last three chars of str are sepc''
str = f.sub(str, 1, -4) .. "''"; -- remove them and add back ''
elseif f.sub(str, -5,-1) == duplicate_char .. "]]''" then -- if last five chars of str are sepc]]''
trim = true; -- why? why do this and next differently from previous?
elseif f.sub(str, -4,-1) == duplicate_char .. "]''" then -- if last four chars of str are sepc]''
trim = true; -- same question
end
elseif end_chr == "]" then -- if it might be wikimarkup
if f.sub(str, -3,-1) == duplicate_char .. "]]" then -- if last three chars of str are sepc]] wikilink
trim = true;
elseif f.sub(str, -3,-1) == duplicate_char .. '"]' then -- if last three chars of str are sepc"] quoted external link
trim = true;
elseif f.sub(str, -2,-1) == duplicate_char .. "]" then -- if last two chars of str are sepc] external link
trim = true;
elseif f.sub(str, -4,-1) == duplicate_char .. "'']" then -- normal case when |url=something & |title=Title.
trim = true;
end
elseif end_chr == " " then -- if last char of output string is a space
if f.sub(str, -2,-1) == duplicate_char .. " " then -- if last two chars of str are <sepc><space>
str = f.sub(str, 1,-3); -- remove them both
end
end
if trim then
if value ~= comp then -- value does not equal comp when value contains html markup
local dup2 = duplicate_char;
if f.match(dup2, "%A" ) then dup2 = "%" .. dup2; end -- if duplicate_char not a letter then escape it
value = f.gsub(value, "(%b<>)" .. dup2, "%1", 1 ) -- remove duplicate_char if it follows html markup
else
value = f.sub(value, 2, -1 ); -- remove duplicate_char when it is first character
end
end
end
str = str .. value; --add it to the output string
end
end
return str;
end
--[[--------------------------< I S _ S U F F I X >------------------------------------------------------------
returns true is suffix is properly formed Jr, Sr, or ordinal in the range 2–9. Puncutation not allowed.
]]
local function is_suffix (suffix)
if in_array (suffix, {'Jr', 'Sr', '2nd', '3rd'}) or suffix:match ('^%dth$') then
return true;
end
return false;
end
--[[--------------------------< I S _ G O O D _ V A N C _ N A M E >--------------------------------------------
For Vancouver Style, author/editor names are supposed to be rendered in Latin (read ASCII) characters. When a name
uses characters that contain diacritical marks, those characters are to converted to the corresponding Latin character.
When a name is written using a non-Latin alphabet or logogram, that name is to be transliterated into Latin characters.
These things are not currently possible in this module so are left to the editor to do.
This test allows |first= and |last= names to contain any of the letters defined in the four Unicode Latin character sets
[http://www.unicode.org/charts/PDF/U0000.pdf C0 Controls and Basic Latin] 0041–005A, 0061–007A
[http://www.unicode.org/charts/PDF/U0080.pdf C1 Controls and Latin-1 Supplement] 00C0–00D6, 00D8–00F6, 00F8–00FF
[http://www.unicode.org/charts/PDF/U0100.pdf Latin Extended-A] 0100–017F
[http://www.unicode.org/charts/PDF/U0180.pdf Latin Extended-B] 0180–01BF, 01C4–024F
|lastn= also allowed to contain hyphens, spaces, and apostrophes. (http://www.ncbi.nlm.nih.gov/books/NBK7271/box/A35029/)
|firstn= also allowed to contain hyphens, spaces, apostrophes, and periods
This original test:
if nil == mw.ustring.find (last, "^[A-Za-zÀ-ÖØ-öø-ƿDŽ-ɏ%-%s%']*$") or nil == mw.ustring.find (first, "^[A-Za-zÀ-ÖØ-öø-ƿDŽ-ɏ%-%s%'%.]+[2-6%a]*$") then
was written ouside of the code editor and pasted here because the code editor gets confused between character insertion point and cursor position.
The test has been rewritten to use decimal character escape sequence for the individual bytes of the unicode characters so that it is not necessary
to use an external editor to maintain this code.
\195\128-\195\150 – À-Ö (U+00C0–U+00D6 – C0 controls)
\195\152-\195\182 – Ø-ö (U+00D8-U+00F6 – C0 controls)
\195\184-\198\191 – ø-ƿ (U+00F8-U+01BF – C0 controls, Latin extended A & B)
\199\132-\201\143 – DŽ-ɏ (U+01C4-U+024F – Latin extended B)
]]
local function is_good_vanc_name (last, first)
local first, suffix = first:match ('(.-),?%s*([%dJS][%drndth]+)%.?$') or first; -- if first has something that looks like a generational suffix, get it
if is_set (suffix) then
if not is_suffix (suffix) then
add_vanc_error (cfg.err_msg_supl.suffix);
return false; -- not a name with an appropriate suffix
end
end
if nil == mw.ustring.find (last, "^[A-Za-z\195\128-\195\150\195\152-\195\182\195\184-\198\191\199\132-\201\143%-%s%']*$") or
nil == mw.ustring.find (first, "^[A-Za-z\195\128-\195\150\195\152-\195\182\195\184-\198\191\199\132-\201\143%-%s%'%.]*$") then
add_vanc_error (cfg.err_msg_supl['non-Latin character']);
return false; -- not a string of latin characters; Vancouver requires Romanization
end;
return true;
end
--[[--------------------------< R E D U C E _ T O _ I N I T I A L S >------------------------------------------
Attempts to convert names to initials in support of |name-list-format=vanc.
Names in |firstn= may be separated by spaces or hyphens, or for initials, a period. See http://www.ncbi.nlm.nih.gov/books/NBK7271/box/A35062/.
Vancouver style requires family rank designations (Jr, II, III, etc) to be rendered as Jr, 2nd, 3rd, etc. See http://www.ncbi.nlm.nih.gov/books/NBK7271/box/A35085/.
This code only accepts and understands generational suffix in the Vancouver format because Roman numerals look like, and can be mistaken for, initials.
This function uses ustring functions because firstname initials may be any of the unicode Latin characters accepted by is_good_vanc_name ().
]]
local function reduce_to_initials(first)
local name, suffix = mw.ustring.match(first, "^(%u+) ([%dJS][%drndth]+)$");
if not name then -- if not initials and a suffix
name = mw.ustring.match(first, "^(%u+)$"); -- is it just intials?
end
if name then -- if first is initials with or without suffix
if 3 > mw.ustring.len (name) then -- if one or two initials
if suffix then -- if there is a suffix
if is_suffix (suffix) then -- is it legitimate?
return first; -- one or two initials and a valid suffix so nothing to do
else
add_vanc_error (cfg.err_msg_supl.suffix); -- one or two initials with invalid suffix so error message
return first; -- and return first unmolested
end
else
return first; -- one or two initials without suffix; nothing to do
end
end
end -- if here then name has 3 or more uppercase letters so treat them as a word
local initials, names = {}, {}; -- tables to hold name parts and initials
local i = 1; -- counter for number of initials
names = mw.text.split (first, '[%s,]+'); -- split into a table of names and possible suffix
while names[i] do -- loop through the table
if 1 < i and names[i]:match ('[%dJS][%drndth]+%.?$') then -- if not the first name, and looks like a suffix (may have trailing dot)
names[i] = names[i]:gsub ('%.', ''); -- remove terminal dot if present
if is_suffix (names[i]) then -- if a legitimate suffix
table.insert (initials, ' ' .. names[i]); -- add a separator space, insert at end of initials table
break; -- and done because suffix must fall at the end of a name
end -- no error message if not a suffix; possibly because of Romanization
end
if 3 > i then
table.insert (initials, mw.ustring.sub(names[i],1,1)); -- insert the intial at end of initials table
end
i = i+1; -- bump the counter
end
return table.concat(initials) -- Vancouver format does not include spaces.
end
Community content is available under CC-BY-SA unless otherwise noted.