Module:Citation/CS1/COinS

--[[ History of changes since last sync: 2015-12-12

]]

local coins = {};

----< F O R W A R D  D E C L A R A T I O N S >--

local is_set, in_array, remove_wiki_link;										-- functions in Module:Citation/CS1/Utilities

local cfg;																		-- table of configuration tables that are defined in Module:Citation/CS1/Configuration

--[[--< S T R I P _ A P O S T R O P H E _ M A R K U P >

Strip wiki italic and bold markup from argument so that it doesn't contaminate COinS metadata. This function strips common patterns of apostrophe markup. We presume that editors who have taken the time to markup a title have, as a result, provided valid markup. When they don't, some single apostrophes are left behind.

]]

local function strip_apostrophe_markup (argument) if not is_set (argument) then return argument; end

if argument:find ( "''", 1, true ) == nil then								-- Is there at least one double apostrophe? If not, exit. return argument; end

while true do		if argument:find ( "'", 1, true ) then								-- bold italic (5) argument=argument:gsub("%'%'%'%'%'", "");							-- remove all instances of it		elseif argument:find ( "", 1, true ) then							-- italic start and end without content (4) argument=argument:gsub("%'%'%'%'", ""); elseif argument:find ( "'''", 1, true ) then							-- bold (3) argument=argument:gsub("%'%'%'", ""); elseif argument:find ( "''", 1, true ) then								-- italic (2) argument=argument:gsub("%'%'", ""); else break; end end return argument;															-- done end

--[[--< M A K E _ C O I N S _ T I T L E >--

Makes a title for COinS from Title and / or ScriptTitle (or any other name-script pairs)

Apostrophe markup (bold, italics) is stripped from each value so that the COinS metadata isn't corrupted with strings of %27%27...

]]

local function make_coins_title (title, script) if is_set (title) then title = strip_apostrophe_markup (title);								-- strip any apostrophe markup else title='';																-- if not set, make sure title is an empty string end if is_set (script) then script = script:gsub ('^%l%l%s*:%s*', '');								-- remove language prefix if present (script value may now be empty string) script = strip_apostrophe_markup (script);								-- strip any apostrophe markup else script='';																-- if not set, make sure script is an empty string end if is_set (title) and is_set (script) then script = ' ' .. script;													-- add a space before we concatenate end return title .. script;														-- return the concatenation end

--[[--< E S C A P E _ L U A _ M A G I C _ C H A R S >--

Returns a string where all of lua's magic characters have been escaped. This is important because functions like string.gsub treat their pattern and replace strings as patterns, not literal strings. ]]

local function escape_lua_magic_chars (argument) argument = argument:gsub("%%", "%%%%");										-- replace % with %% argument = argument:gsub("([%^%$%(%)%.%[%]%*%+%-%?])", "%%%1");				-- replace all other lua magic pattern characters return argument; end

--[[--< G E T _ C O I N S _ P A G E S >

Extract page numbers from external wikilinks in any of the |page=, |pages=, or |at= parameters for use in COinS.

]]

local function get_coins_pages (pages) local pattern; if not is_set (pages) then return pages; end								-- if no page numbers then we're done while true do		pattern = pages:match("%[(%w*:?//[^ ]+%s+)[%w%d].*%]");					-- pattern is the opening bracket, the url and following space(s): "[url " if nil == pattern then break; end										-- no more urls pattern = escape_lua_magic_chars (pattern);								-- pattern is not a literal string; escape lua's magic pattern characters pages = pages:gsub(pattern, "");										-- remove as many instances of pattern as possible end pages = pages:gsub("[%[%]]", "");											-- remove the brackets pages = pages:gsub("–", "-" );							-- replace endashes with hyphens pages = pages:gsub("&%w+;", "-" );						-- and replace html entities (– etc.) with hyphens; do we need to replace numerical entities like &#32; and the like? return pages; end

--[=[-< C O I N S _ R E P L A C E _ M A T H _ S T R I P M A R K E R >--

There are three options for math markup rendering that depend on the editor's math preference settings. These settings are at Special:Preferences and are PNG images TeX source MathML with SVG or PNG fallback

All three are heavy with html and css which doesn't belong in the metadata.

Without this function, the metadata saved in the raw wikitext contained the rendering determined by the settings of the last editor to save the page.

This function gets the rendered form of an equation according to the editor's preference before the page is saved. It then searches the rendering for the text equivalent of the rendered equation and replaces the rendering with that so that the page is saved without extraneous html/css markup and with a reasonably readable text form of the equation.

When a replacement is made, this function returns true and the value with replacement; otherwise false and the intital value. To replace multipe equations it is necesary to call this function from within a loop.

]=]

local function coins_replace_math_stripmarker (value) local stripmarker = '\127UNIQ%-%-math%-[%a%d]+%-QINU\127';					-- math stripmarker pattern local rendering = value:match (stripmarker);								-- is there a math stripmarker

if not rendering then														-- when value doesn't have a math stripmarker, abandon this test return false, value; end rendering = mw.text.unstripNoWiki (rendering);								-- convert stripmarker into rendered value (or nil? ''? when math render error) if rendering:match ('alt="[^"]+"') then										-- if PNG math option		rendering = rendering:match ('alt="([^"]+)"');							-- extract just the math text elseif rendering:match ('$%s+.+%s+%$') then									-- if TeX math option; $ is legit character that is escapes as \$ rendering = rendering:match ('$%s+(.+)%s+%$')							-- extract just the math text elseif rendering:match (']+>.+ ') then			-- if MathML math option rendering = rendering:match (']+>(.+) ')		-- extract just the math text else return false, value;													-- had math stripmarker but not one of the three defined forms end return true, value:gsub (stripmarker, rendering, 1); end

--[[--< C O I N S _ C L E A N U P >

Cleanup parameter values for the metadata by removing or replacing invisible characters and certain html entities.

2015-12-10: there is a bug in mw.text.unstripNoWiki. It replaced math stripmarkers with the appropriate content when it shouldn't. See https://phabricator.wikimedia.org/T121085 and Wikipedia_talk:Lua#stripmarkers_and_mw.text.unstripNoWiki.28.29

TODO: move the replacement patterns and replacement values into a table in /Configuration similar to the invisible characters table?

]]

local function coins_cleanup (value) local replaced = true;														-- default state to get the do loop running while replaced do															-- loop until all math stripmarkers replaced replaced, value = coins_replace_math_stripmarker (value);				-- replace math stripmarker with text representation of the equation end

value = value:gsub ('\127UNIQ%-%-math%-[%a%d]+%-QINU\127', "MATH RENDER ERROR");	-- one or more couldn't be replaced; insert vague error message value = mw.text.unstripNoWiki (value);										-- replace nowiki stripmarkers with their content value = value:gsub (' &#39;s ', "'s");	-- replace template with simple apostrophe-s value = value:gsub ('&zwj;\226\128\138\039\226\128\139', "'");				-- replace ' with simple apostrophe value = value:gsub ('\226\128\138\039\226\128\139', "'");					-- replace ' with simple apostrophe (as of 2015-12-11) value = value:gsub (' ', ' ');											-- replace  entity with plain space value = value:gsub ('\226\128\138', ' ');									-- replace hair space with plain space value = value:gsub ('&zwj;', '');											-- remove &zwj; entities value = value:gsub ('[\226\128\141\226\128\139]', '')						-- remove zero-width joiner, zero-width space value = value:gsub ('[\194\173\009\010\013]', ' ');							-- replace soft hyphen, horizontal tab, line feed, carriage return with plain space return value; end

--[[--< C O I N S >

COinS metadata (see ) allows automated tools to parse the citation information.

]]

local function COinS(data, class) if 'table' ~= type(data) or nil == next(data) then return ''; end

for k, v in pairs (data) do													-- spin through all of the metadata parameter values if 'ID_list' ~= k and 'Authors' ~= k then								-- except the ID_list and Author tables (author nowiki stripmarker done when Author table processed) data[k] = coins_cleanup (v); end end

local ctx_ver = "Z39.88-2004"; -- treat table strictly as an array with only set values. local OCinSoutput = setmetatable( {}, {		__newindex = function(self, key, value)			if is_set(value) then				rawset( self, #self+1, table.concat{ key, '=', mw.uri.encode( remove_wiki_link( value ) ) } );			end		end	}); if in_array (class, {'arxiv', 'journal', 'news'}) or (in_array (class, {'conference', 'interview', 'map', 'press release', 'web'}) and is_set(data.Periodical)) or 		('citation' == class and is_set(data.Periodical) and not is_set (data.Encyclopedia)) then OCinSoutput.rft_val_fmt = "info:ofi/fmt:kev:mtx:journal";			-- journal metadata identifier if 'arxiv' == class then											-- set genre according to the type of citation template we are rendering OCinSoutput["rft.genre"] = "preprint";							-- cite arxiv elseif 'conference' == class then OCinSoutput["rft.genre"] = "conference";						-- cite conference (when Periodical set) elseif 'web' == class then OCinSoutput["rft.genre"] = "unknown";							-- cite web (when Periodical set) else OCinSoutput["rft.genre"] = "article";							-- journal and other 'periodical' articles end OCinSoutput["rft.jtitle"] = data.Periodical;						-- journal only OCinSoutput["rft.atitle"] = data.Title;								-- 'periodical' article titles

-- these used only for periodicals OCinSoutput["rft.ssn"] = data.Season;								-- keywords: winter, spring, summer, fall OCinSoutput["rft.chron"] = data.Chron;								-- free-form date components OCinSoutput["rft.volume"] = data.Volume;							-- does not apply to books OCinSoutput["rft.issue"] = data.Issue; OCinSoutput["rft.pages"] = data.Pages;								-- also used in book metadata

elseif 'thesis' ~= class then												-- all others except cite thesis are treated as 'book' metadata; genre distinguishes OCinSoutput.rft_val_fmt = "info:ofi/fmt:kev:mtx:book";					-- book metadata identifier if 'report' == class or 'techreport' == class then						-- cite report and cite techreport OCinSoutput["rft.genre"] = "report"; elseif 'conference' == class then										-- cite conference when Periodical not set OCinSoutput["rft.genre"] = "conference"; elseif in_array (class, {'book', 'citation', 'encyclopaedia', 'interview', 'map'}) then if is_set (data.Chapter) then OCinSoutput["rft.genre"] = "bookitem"; OCinSoutput["rft.atitle"] = data.Chapter;						-- book chapter, encyclopedia article, interview in a book, or map title else if 'map' == class or 'interview' == class then OCinSoutput["rft.genre"] = 'unknown';						-- standalone map or interview else OCinSoutput["rft.genre"] = 'book';							-- book and encyclopedia end end else	--{'audio-visual', 'AV-media-notes', 'DVD-notes', 'episode', 'interview', 'mailinglist', 'map', 'newsgroup', 'podcast', 'press release', 'serial', 'sign', 'speech', 'web'} OCinSoutput["rft.genre"] = "unknown"; end OCinSoutput["rft.btitle"] = data.Title;									-- book only OCinSoutput["rft.place"] = data.PublicationPlace;						-- book only OCinSoutput["rft.series"] = data.Series;								-- book only OCinSoutput["rft.pages"] = data.Pages;									-- book, journal OCinSoutput["rft.edition"] = data.Edition;								-- book only OCinSoutput["rft.pub"] = data.PublisherName;							-- book and dissertation else																		-- cite thesis OCinSoutput.rft_val_fmt = "info:ofi/fmt:kev:mtx:dissertation";			-- dissertation metadata identifier OCinSoutput["rft.title"] = data.Title;									-- dissertation (also patent but that is not yet supported) OCinSoutput["rft.degree"] = data.Degree;								-- dissertation only OCinSoutput['rft.inst'] = data.PublisherName;							-- book and dissertation end -- and now common parameters (as much as possible) OCinSoutput["rft.date"] = data.Date;										-- book, journal, dissertation for k, v in pairs( data.ID_list ) do										-- what to do about these? For now assume that they are common to all? --		if k == 'ISBN' then v = clean_isbn( v ) end if k == 'ISBN' then v = v:gsub( "[^-0-9X]", "" ); end local id = cfg.id_handlers[k].COinS; if string.sub( id or "", 1, 4 ) == 'info' then							-- for ids that are in the info:registry OCinSoutput["rft_id"] = table.concat{ id, "/", v }; elseif string.sub (id or "", 1, 3 ) == 'rft' then						-- for isbn, issn, eissn, etc that have defined COinS keywords OCinSoutput[ id ] = v;		elseif id then															-- when cfg.id_handlers[k].COinS is not nil OCinSoutput["rft_id"] = table.concat{ cfg.id_handlers[k].prefix, v };	-- others; provide a url end end

--		for k, v in pairs( data.ID_list ) do										-- what to do about these? For now assume that they are common to all?		local id, value = cfg.id_handlers[k].COinS;		if k == 'ISBN' then value = clean_isbn( v ); else value = v; end		if string.sub( id or "", 1, 4 ) == 'info' then			OCinSoutput["rft_id"] = table.concat{ id, "/", v };		else			OCinSoutput[ id ] = value;		end	end local last, first; for k, v in ipairs( data.Authors ) do		last, first = coins_cleanup (v.last), coins_cleanup (v.first or '');	-- replace any nowiki strip markers, non-printing or invisible characers if k == 1 then															-- for the first author name only if is_set(last) and is_set(first) then								-- set these COinS values if |first= and |last= specify the first author name OCinSoutput["rft.aulast"] = last;								-- book, journal, dissertation OCinSoutput["rft.aufirst"] = first;								-- book, journal, dissertation elseif is_set(last) then OCinSoutput["rft.au"] = last;									-- book, journal, dissertation -- otherwise use this form for the first name end else																	-- for all other authors if is_set(last) and is_set(first) then OCinSoutput["rft.au"] = table.concat{ last, ", ", first };		-- book, journal, dissertation elseif is_set(last) then OCinSoutput["rft.au"] = last;									-- book, journal, dissertation end end end

OCinSoutput.rft_id = data.URL; OCinSoutput.rfr_id = table.concat{ "info:sid/", mw.site.server:match( "[^/]*$" ), ":", data.RawPage }; OCinSoutput = setmetatable( OCinSoutput, nil ); -- sort with version string always first, and combine. table.sort( OCinSoutput ); table.insert( OCinSoutput, 1, "ctx_ver=" .. ctx_ver ); -- such as "Z39.88-2004" return table.concat(OCinSoutput, "&"); end

--[[--< S E T _ S E L E C T E D _ M O D U L E S >--

Sets local cfg table and imported functions table to same (live or sandbox) as that used by the other modules.

]]

local function set_selected_modules (cfg_table_ptr, utilities_page_ptr) cfg = cfg_table_ptr;

is_set = utilities_page_ptr.is_set;											-- import functions from select Module:Citation/CS1/Utilities module in_array = utilities_page_ptr.in_array; remove_wiki_link = utilities_page_ptr.remove_wiki_link; end

return { make_coins_title = make_coins_title, get_coins_pages = get_coins_pages, COinS = COinS, set_selected_modules = set_selected_modules, }