Module:Citation/CS1/Configuration: Difference between revisions

Line 1:

local lang_obj = mw.language.getContentLanguage(); -- make a language object for the local language; used here for languages and dates

--[[--------------------------< S E T T I N G S >--------------------------------------------------------------

boolean settings used to control various things. these setting located here to make them easy to find

]]

-- these settings local to this module only

local local_digits_from_mediawiki = false; -- for i18n; when true, module fills date_names['local_digits'] from MediaWiki; manual fill required else; always false at en.wiki

local local_date_names_from_mediawiki = false; -- for i18n; when true, module fills date_names['local']['long'] and date_names['local']['short'] from MediaWiki;

-- manual translation required else; ; always false at en.wiki

-- these settings exported to other modules

local use_identifier_redirects = true; -- when true use redirect name for identifier label links; always true at en.wiki

local local_lang_cat_enable = false; -- when true categorizes pages where |language=<local wiki's language>; always false at en.wiki

local date_name_auto_xlate_enable = false; -- when true translates English month-names to the local-wiki's language month names; always false at en.wiki

local date_digit_auto_xlate_enable = false; -- when true translates Western date digit to the local-wiki's language digits (date_names['local_digits']); always false at en.wiki

--[[--------------------------< U N C A T E G O R I Z E D _ N A M E S P A C E S >------------------------------

Line 34:

Line 52:

['archived-dead'] = 'Archived from $1 on $2',

['archived-live'] = '$1 from the original on $2',

~~['archived-missing'] = 'Archived from the original $1 on $2',~~

['archived-unfit'] = 'Archived from the original on ',

['archived'] = 'Archived',

Line 48:

Line 65:

['inset'] = '$1 inset',

['interview'] = 'Interviewed by $1',

~~['lay summary'] = 'Lay summary',~~

['mismatch'] = '<code class="cs1-code">|$1=</code> / <code class="cs1-code">|$2=</code> mismatch', -- $1 is year param name; $2 is date param name

['newsgroup'] = '[[Usenet newsgroup|Newsgroup]]: $1',

Line 125:

Line 141:

local citation_class_map_t = { -- TODO: if kept, these and all other config.CitationClass 'names' require some sort of i18n

['arxiv'] = 'arXiv',

['audio-visual'] = 'AV media',

['AV-media-notes'] = 'AV media notes',

['biorxiv'] = 'bioRxiv',

['citeseerx'] = 'CiteSeerX',

['encyclopaedia'] = 'encyclopedia',

['mailinglist'] = 'mailing list',

['pressrelease'] = 'press release'

['medrxiv'] = 'medRxiv',

['pressrelease'] = 'press release',

['ssrn'] = 'SSRN',

['techreport'] = 'tech report',

}

Line 275:

Line 297:

['Issue'] = {'issue', 'number'},

['Language'] = {'language', 'lang'},

~~['LayDate'] = 'lay-date',~~

~~['LayFormat'] = 'lay-format',~~

~~['LaySource'] = 'lay-source',~~

~~['LayURL'] = 'lay-url',~~

['MailingList'] = {'mailing-list', 'mailinglist'}, -- cite mailing list only

['Map'] = 'map', -- cite map only

Line 326:

Line 344:

['Title'] = 'title', -- Used by InternetArchiveBot

['TitleLink'] = {'title-link', 'episode-link', 'episodelink'}, -- Used by InternetArchiveBot

['TitleNote'] = 'department',

['TitleNote'] = {'title-note', 'department'},

['TitleType'] = {'type', 'medium'},

['TransChapter'] = {'trans-article', 'trans-chapter', 'trans-contribution',

Line 332:

Line 350:

['Transcript'] = 'transcript',

['TranscriptFormat'] = 'transcript-format',

['TranscriptURL'] = {'transcript-url', ~~'transcripturl'},~~ -- Used by InternetArchiveBot

['TranscriptURL'] = 'transcript-url', -- Used by InternetArchiveBot

['TransMap'] = 'trans-map', -- cite map only

['TransPeriodical'] = {'trans-journal', 'trans-magazine', 'trans-newspaper',

Line 347:

Line 365:

['Year'] = 'year',

['AuthorList-First'] = {"first#", "author-first#", "author#-first", "given#",

['AuthorList-First'] = {"first#", "author-first#", "author#-first", "author-given#", "author#-given",

"~~author~~-given#", "~~author~~#-given"},

"subject-first#", "subject#-first", "subject-given#", "subject#-given",

['AuthorList-Last'] = {"last#", "author-last#", "author#-last", "surname#",

"given#"},

"~~author~~-surname#", "~~author~~#-surname", "author#", "subject#", ~~'host~~#'},

['AuthorList-Last'] = {"last#", "author-last#", "author#-last", "author-surname#", "author#-surname",

"subject-last#", "subject#-last", "subject-surname#", "subject#-surname",

"author#", 'host#', "subject#", "surname#"},

['AuthorList-Link'] = {"author-link#", "author#-link", "subject-link#",

"subject#-link", "authorlink#", "author#link"},

Line 394:

Line 414:

'AuthorList-Mask', 'ContributorList-Mask', 'EditorList-Mask', 'InterviewerList-Mask', 'TranslatorList-Mask', -- name-list mask may have name separators

'PostScript', 'Quote', 'ScriptQuote', 'TransQuote', 'Ref', -- miscellaneous

'ArchiveURL', 'ChapterURL', 'ConferenceURL~~', 'LayURL~~', 'MapURL', 'TranscriptURL', 'URL', -- URL-holding parameters

'ArchiveURL', 'ChapterURL', 'ConferenceURL', 'MapURL', 'TranscriptURL', 'URL', -- URL-holding parameters

}

local url_meta_params = { -- table of aliases[] keys (meta parameters); each key has a table of parameter names for a value

'ArchiveURL', 'ChapterURL', 'ConferenceURL', 'ID~~', 'LayURL~~', 'MapURL', 'TranscriptURL', 'URL', -- parameters allowed to hold urls

'ArchiveURL', 'ChapterURL', 'ConferenceURL', 'ID', 'MapURL', 'TranscriptURL', 'URL', -- parameters allowed to hold urls

'Page', 'Pages', 'At', 'QuotePage', 'QuotePages', -- insource locators allowed to hold urls

}

Line 617:

Line 637:

date_names[invert_t[2]][i] = name; -- invert to get [i] = 'name' for conversions from ymd

end

if local_digits_from_mediawiki then -- if fetching local digits from MediaWiki is enabled

local digits_t = {};

for i=0, 9 do -- loop 10x and

digits_t [lang_obj:formatNum (i)] = tostring (i); -- format the loop indexer as local lang table index and assign loop indexer (a string) as the value

end

date_names['local_digits'] = digits_t;

end

Line 638:

Line 666:

'{{ *([Mm]DY) *[|}]', -- 0

}

local title_object = mw.title.getCurrentTitle();

local content; -- done this way so that unused templates appear in unused-template-reports; self-transcluded makes them look like they are used

if 10 ~= title_object.namespace then -- all namespaces except Template

content = title_object:getContent() or ''; -- get the content of the article or ''; new pages edited w/ve do not have 'content' until saved; ve does not preview; phab:T221625

end

local function get_date_format ()

~~local title_object = mw.title.getCurrentTitle();~~

-- if title_object.namespace == 10 then -- not in template space so that unused templates appear in unused-template-reports;

if title_object.namespace == 10 then -- not in template space so that unused templates appear in unused-template-reports;

if not content then -- nil content when we're in template

return nil; -- auto-formatting does not work in Template space so don't set global_df

end

~~local content = title_object:getContent() or ''; -- get the content of the article or ''; new pages edited w/ve do not have 'content' until saved; ve does not preview; phab:T221625~~

for _, pattern in ipairs (df_template_patterns) do -- loop through the patterns looking for {{Use dmy dates}} or {{Use mdy dates}} or any of their redirects

local start, _, match = content:find(pattern); -- match is the three letters indicating desired date format

if match then

~~content~~ = content:match ('%b{}', start); -- get the whole template

local use_dates_template = content:match ('%b{}', start); -- get the whole template

if ~~content~~:match ('| *cs1%-dates *= *[lsy][sy]?') then -- look for |cs1-dates=publication date length access-/archive-date length

if use_dates_template:match ('| *cs1%-dates *= *[lsy][sy]?') then -- look for |cs1-dates=publication date length access-/archive-date length

return match:lower() .. '-' .. ~~content~~:match ('| *cs1%-dates *= *([lsy][sy]?)');

return match:lower() .. '-' .. use_dates_template:match ('| *cs1%-dates *= *([lsy][sy]?)');

else

return match:lower() .. '-all'; -- no |cs1-dates= k/v pair; return value appropriate for use in |df=

Line 658:

Line 691:

end

local global_df;

local global_df; -- TODO: add this to <global_cs1_config_t>?

Line 850:

Line 883:

--[[---------------------< S ~~T R~~ I ~~P M A R K~~ E ~~R S~~ >----------------------------

--[[--------------------------< C S 1 _ C O N F I G _ G E T >--------------------------------------------------

fetch and validate values from {{cs1 config}} template to fill <global_cs1_config_t>

no error messages; when errors are detected, the parameter value from {{cs1 config}} is blanked.

~~Common pattern definition location for stripmarkers so that we don't have to go~~

Supports all parameters and aliases associated with the metaparameters: DisplayAuthors, DisplayContributors,

~~hunting for them if~~ (~~when~~) ~~MediaWiki changes their form~~.

DisplayEditors, DisplayInterviewers, DisplayTranslators, NameListStyle, and Mode. The DisplayWhatever metaparameters

accept numeric values only (|display-authors=etal and the like is not supported).

]]

local ~~stripmarkers~~ = {

local global_cs1_config_t = {}; -- TODO: add value returned from get_date_format() to this table?

~~['any'] = '\127[^\127]*UNIQ%-%-(%a+)%-[%a%d]+%-QINU[^\127]*\127', -- capture returns name of stripmarker~~

~~['math'] = '\127[^\127]*UNIQ%~~-%-~~math%-[%a%d]+%-QINU[^\127]*\127' -- math stripmarkers used in coins_cleanup() and coins_replace_math_stripmarker~~()

}

local function get_cs1_config ()

-- if title_object.namespace == 10 then -- not in template space so that unused templates appear in unused-template-reports;

if not content then -- nil content when we're in template

return nil; -- auto-formatting does not work in Template space so don't set global_df

end

--[[--~~----------~~< ~~I N V I S I B L E _ C H A R A C T E R S~~ >--~~-------------------~~

local start = content:find('{{ *[Cc][Ss]1 config *[|}]'); -- <start> is offset into <content> when {{cs1 config}} found; nil else

if start then

local cs1_config_template = content:match ('%b{}', start); -- get the whole template

~~This table holds non-printing or invisible characters indexed either by name or~~

if not cs1_config_template then

~~by Unicode group. Values are decimal representations of UTF-8 codes. The table~~

return nil;

~~is organized as a table of tables because the Lua pairs keyword returns table~~

end

~~data in an arbitrary order. Here, we want to process the table from top to bottom~~

~~because the entries at the top of the table are also found in the ranges specified~~

~~by the entries at the bottom of the table.~~

Also here is a pattern that recognizes stripmarkers that begin and end with the

local params_t = mw.text.split (cs1_config_template:gsub ('^{{%s*', ''):gsub ('%s*}}$', ''), '%s*|%s*'); -- remove '{{' and '}}'; make a sequence of parameter/value pairs (split on the pipe)

delete characters. The nowiki stripmarker is not an error but some others are

table.remove (params_t, 1); -- remove the template name because it isn't a parameter/value pair

because the parameter values that include them become part of the template's

metadata before stripmarker replacement.

local config_meta_params_t = {'DisplayAuthors', 'DisplayContributors', 'DisplayEditors', 'DisplayInterviewers', 'DisplayTranslators', 'NameListStyle', 'Mode'};

local meta_param_map_t = {}; -- list of accepted parameter names usable in {{cs1 config}} goes here

]]

for _, meta_param in ipairs (config_meta_params_t) do -- for i18n using <config_meta_params_t>, map template parameter names to their metaparameter equivalents

local invisible_defs = {

if 'table' == type (aliases[meta_param]) then -- if <meta_param> is a sequence,

del = '\127', -- used to distinguish between stripmarker and del char

for _, param in ipairs (aliases[meta_param]) do -- extract its contents

zwj = '\226\128\141', -- used with capture because zwj may be allowed

meta_param_map_t[param] = meta_param; -- and add to <meta_param_map_t>

}

end

else

local invisible_chars = {

meta_param_map_t[aliases[meta_param]] = meta_param; -- not a sequence so just add the parameter to <meta_param_map_t>

{'replacement', '\239\191\189'}, -- U+FFFD, EF BF BD

end

{'zero width joiner', '('.. invisible_defs.zwj .. ')'}, -- U+200D, E2 80 8D; capture because zwj may be allowed

end

{'zero width space', '\226\128\139'}, -- U+200B, E2 80 8B

{'hair space', '\226\128\138'}, -- U+200A, E2 80 8A

local keywords_t = {}; -- map valid keywords to their associate metaparameter; reverse form of <keyword_lists[key] for these metaparameters

{'soft hyphen', '\194\173'}, -- U+00AD, C2 AD

for _, metaparam_t in ipairs ({{'NameListStyle', 'name-list-style'}, {'Mode', 'mode'}}) do -- only these metaparameter / keywords_lists key pairs

{'horizontal tab', '\009'}, -- U+0009 (HT), 09

for _, keyword in ipairs (keywords_lists[metaparam_t[2]]) do -- spin through the list of keywords

{'line feed', '\010'}, -- U+000A (LF), 0A

keywords_t[keyword] = metaparam_t[1]; -- add [keyword] = metaparameter to the map

{'no-break space', '\194\160'}, -- U+00A0 (NBSP), C2 A0

end

{'carriage return', '\013'}, -- U+000D (CR), 0D

end

{'stripmarker', stripmarkers.any}, -- stripmarker; may or may not be an error; capture returns the stripmaker type

{'delete', '('.. invisible_defs.del .. ')'}, -- U+007F (DEL), 7F; must be done after stripmarker test; capture to distinguish isolated del chars not part of stripmarker

for _, param in ipairs (params_t) do -- spin through the {{cs1 config}} parameters and fill <global_cs1_config_t>

{'C0 control', '[\000-\008\011\012\014-\031]'}, -- U+0000–U+001F (NULL–US), 00–1F (except HT, LF, CR (09, 0A, 0D))

local k, v = param:match ('([^=]-)%s*=%s*(.+)'); -- <k> is the parameter name; <v> is parameter's assigned value

{'C1 control', '[\194\128-\194\159]'}, -- U+0080–U+009F (XXX–APC), C2 80 – C2 9F

if k then

-- {'Specials', '[\239\191\185-\239\191\191]'}, -- U+FFF9-U+FFFF, EF BF B9 – EF BF BF

if k:find ('^display') then -- if <k> is one of the |display-<namelist>= parameters

-- {'Private use area', '[\238\128\128-\239\163\191]'}, -- U+E000–U+F8FF, EE 80 80 – EF A3 BF

if v:match ('%d+') then -- the assigned value must be digits; doesn't accept 'etal'

-- {'Supplementary Private Use Area-A', '[\243\176\128\128-\243\191\191\189]'}, -- U+F0000–U+FFFFD, F3 B0 80 80 – F3 BF BF BD

global_cs1_config_t[meta_param_map_t[k]]=v; -- add the display param and its value to globals table

-- {'Supplementary Private Use Area-B', '[\244\128\128\128-\244\143\191\189]'}, -- U+100000–U+10FFFD, F4 80 80 80 – F4 8F BF BD

end

else

if keywords_t[v] == meta_param_map_t[k] then -- keywords_t[v] returns nil or the metaparam name; these must be the same

global_cs1_config_t[meta_param_map_t[k]]=v; -- add the parameter and its value to globals table

end

get_cs1_config (); -- fill <global_cs1_config_t>

--[[---------------------< S T R I P M A R K E R S >----------------------------

Common pattern definition location for stripmarkers so that we don't have to go

hunting for them if (when) MediaWiki changes their form.

]]

local stripmarkers = {

['any'] = '\127[^\127]*UNIQ%-%-(%a+)%-[%a%d]+%-QINU[^\127]*\127', -- capture returns name of stripmarker

['math'] = '\127[^\127]*UNIQ%-%-math%-[%a%d]+%-QINU[^\127]*\127' -- math stripmarkers used in coins_cleanup() and coins_replace_math_stripmarker()

}

--[[------------< I N V I S I B L E _ C H A R A C T E R S >---------------------

This table holds non-printing or invisible characters indexed either by name or

by Unicode group. Values are decimal representations of UTF-8 codes. The table

is organized as a table of tables because the Lua pairs keyword returns table

data in an arbitrary order. Here, we want to process the table from top to bottom

because the entries at the top of the table are also found in the ranges specified

by the entries at the bottom of the table.