Module:Sandbox/Hellknowz/Test: Difference between revisions

Revision as of 10:35, 2 September 2013

Latest revision as of 09:24, 16 September 2013

(85 intermediate revisions by the same user not shown)

Line 13:

['october'] = 10,

['november'] = 11,

['december'] = 12,

}

local monthShortIndices = {

['jan'] = 1,

['feb'] = 2,

['mar'] = 3,

['apr'] = 4,

--['may'] = 5, -- long one would have caught this already

['jun'] = 6,

['jul'] = 7,

Line 52 ⟶ 49:

-- February leap year check

if (month == 2) then

-- On Feb 29, if we don't have a year (incomplete date), assume 29 can be valid once every

if (day == 29 and not year) then return true end

-- On Feb 29, check for valid year - every 4 but not 100 but 400

if (day == 29 and not ((year % 4 == 0) and (year % 100 ~= 0) or (year % 400 == 0))) then

return false

Line 84:

local PARSERESULT_UNCRECOGNIZED = 3 -- 14 May 12, 2014 (all elements okay, no pattern)

local PARSERESULT_INCOMPLETE = 4 -- May 3

local PARSERESULT_INVALIDPARSERESULT_INCOMPLETERANGE = 5 -- 3 May 322012 - June 2013

local PARSERESULT_INVALIDRANGEPARSERESULT_INVALID = 6 -- May 3 - May 232

local PARSERESULT_INVALIDRANGE = 7 -- May 3 - May 2

-- This will first verify that we have a valid date and time and then output an ISO datetime string

Line 123 ⟶ 124:

-- Check that the date is actually complete even if valid

if (month and month2 and not year) then return PARSERESULT_INCOMPLETEPARSERESULT_INCOMPLETERANGE end -- Mayany orof May'd-dM', 15'dM-dM', 'Md-d', 'Md-Md'

if (month2month and not year2year) then return PARSERESULT_INCOMPLETE end -- 'May', 'May 15', '15 May'

if (month2 and not year2) then return PARSERESULT_INCOMPLETE end -- same but other end

-- While technically there are more cases, none should have been matched and given to us

-- While technically there are more cases, none should have been matched and been given to us

local date1, time1, date2, time2

Line 179 ⟶ 181:

-- These are the element type "constants" for readability mostly

local ELEMENT_INVALID = 1

local ELEMENT_ONETWODIGITS = 2 -- '1' '12' '01'

local ELEMENT_FOURDIGITS = 3 -- '1234'

local ELEMENT_WHITESPACE = 4 -- ' ' ' '

local ELEMENT_MONTHWORD = 5 -- 'May' 'February' 'Aug'

local ELEMENT_COMMA = 6 -- ',' ', '

local ELEMENT_DASH = 7 -- '-' ' - ' ' — ' '- ' ' -'

local ELEMENT_DATESEPARATOR = 8 -- '-'

local ELEMENT_TIMESEPARATOR = 9 -- ':'

local ELEMENT_TIMEPERIOD = 10 -- 'am' 'p.m.'

local ELEMENT_PERIODWHITESPACE = 11 -- '.' or '. '

local ELEMENT_ONETWODIGITSWITHORDINAL = 12 -- '12th' '3rd'

function seekNextElement()

-- Profiler says mw.ustring.find is the bottleneck, probably because it's unicode; not sure how to improve though besides writing my own pattern matcher

-- Digits with letters

local foundPositionStart, foundPositionEnd, foundMatch, foundMatch2 = mw.ustring.find(seekString, '^([0-9]+)([a-z]+)%.?', currentPosition)

if (foundPositionStart) then

--currentPosition = foundPositionEnd + 1 -- this is our new start ___location -- only if we return

-- Additionally check how many digits we actually have, as arbitrary number isn't valid

if (#foundMatch <= 2) then -- most likely a day number

if (foundMatch2 == 'st' or foundMatch2 == 'nd' or foundMatch2 == 'rd' or foundMatch2 == 'th') then -- won't bother checking against a number, no false positives that I saw in 120k cases

currentPosition = foundPositionEnd + 1 -- this is our new start ___location (forced to do this here, since we don't always return)

return ELEMENT_ONETWODIGITSWITHORDINAL, tonumber(foundMatch), (currentPosition > mw.ustring.len(seekString))

--else -- let it capture digits again, this time '10am' '8p.m.' will be separate

-- return ELEMENT_INVALID -- not a valid ordinal indicator

end

--else -- let it capture digits again, this time '10am' '8p.m.' will be separate

-- return ELEMENT_INVALID -- just the invalid, the number of digits (3+) won't match any patterns

end

end

-- Digits

local foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^([0-9]+)', currentPosition)

Line 197 ⟶ 221:

-- Additionally check how many digits we actually have, as arbitrary number isn't valid

if (#foundMatch <= 2) then -- most likely a day number or time number

return ELEMENT_ONETWODIGITS, tonumber(foundMatch), (currentPosition > mw.ustring.len(seekString))

elseif (#foundMatch == 4) then -- most likely a year

return ELEMENT_FOURDIGITS, tonumber(foundMatch), (currentPosition > mw.ustring.len(seekString))

else

return ELEMENT_INVALID -- just the invalid, the number of digits (3 or 5+) won't match any patternpatterns

end

end

Line 222 ⟶ 246:

-- Find the possible month name index

monthIndex = monthIndices[mw.ustring.lower(foundMatch)]

if (not monthIndex) then monthIndex = monthShortIndices[mw.ustring.lower(foundMatch)] end

if (monthIndex) then

Line 235 ⟶ 258:

end

-- Time separator (colon without whitespace)

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(:)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_TIMESEPARATOR, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Comma and any following whitespace

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(,%s*)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_COMMA, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Period and any following whitespace ('Feb. 2010' or '29. June')

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(%.%s*)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_PERIODWHITESPACE, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Dash with possible whitespace or Date separator (dash without whitespace)

-- Both non-breaking spaces - '&nbsp;-&nbsp;'

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(%s*[%-–—]%s*)', currentPosition)

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(&nbsp;[%-–—]&nbsp;)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_DASH, foundMatch, (currentPosition > mw.ustring.len(seekString))

if (foundMatch == '-') then -- nothing else is date separator, no hyphens no stuff like that

return ELEMENT_DATESEPARATOR, foundMatch, (currentPosition > mw.ustring.len(seekString))

else

return ELEMENT_DASH, foundMatch, (currentPosition > mw.ustring.len(seekString)) -- we will actually need to check for DATESEPARATOR as well, as that one stole the '-' case

end

end

-- SameNon-breaking but entity codespace - '&ndashnbsp;- '

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(%s*&nbsp;[nm%-–—]dash;%s*)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_DASH, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Dash entity code and both non-breaking spaces - '&nbsp;&ndash;&nbsp;'

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(&nbsp;&[nm]dash;&nbsp;)', currentPosition)

-- Time separator (colon without whitespace)

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(:)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_TIMESEPARATORELEMENT_DASH, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Dash entity code and non-breaking space - '&nbsp;&ndash; '

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(&nbsp;&[nm]dash;%s*)', currentPosition)

-- Comma and any following whitespace

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(,%s*)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_COMMAELEMENT_DASH, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Dash entity code - ' &ndash; '

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(%s*&[nm]dash;%s*)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_DASH, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Regular whitespace

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(%s*[%-–—]%s*)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

if (foundMatch == '-') then -- nothing else is date separator, no hyphens no stuff like that

return ELEMENT_DATESEPARATOR, foundMatch, (currentPosition > mw.ustring.len(seekString))

else

return ELEMENT_DASH, foundMatch, (currentPosition > mw.ustring.len(seekString)) -- we will actually need to check for DATESEPARATOR as well, as that one stole the '-' case

end

end

-- Whitespace (after all others that capture whitespace)

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^(%s+)', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

return ELEMENT_WHITESPACE, foundMatch, (currentPosition > mw.ustring.len(seekString))

end

-- Whitespace -- same as above but using &nbsp; (other that for dashes)

foundPositionStart, foundPositionEnd, foundMatch = mw.ustring.find(seekString, '^&nbsp;', currentPosition)

if (foundPositionStart) then

currentPosition = foundPositionEnd + 1 -- this is our new start ___location

Line 318 ⟶ 379:

-- Only immediate big improvement is to only seekNextElement() when actually checking that deep, though this will make a (even bigger) mess

if (elements[1] == ELEMENT_ONETWODIGITS or elements[1] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- '3' or '10' or '12th'

if (elements[2] == ELEMENT_WHITESPACE or elements[2] == ELEMENT_PERIODWHITESPACE) then -- '3 ' or '3. '

if (elements[3] == ELEMENT_MONTHWORD) then -- '3 May'

if (numberOfElements == 3) then return checkAndOutput(nil, values[3], values[1], nil, nil, nil) end

if (elements[4] == ELEMENT_WHITESPACE or elements[4] == ELEMENT_PERIODWHITESPACE or elements[4] == ELEMENT_COMMA) then -- '3 May ' or '3 Feb. ' or '3 May, '

if (elements[5] == ELEMENT_FOURDIGITS) then -- '3 May 2013'

if (numberOfElements == 5) then return checkAndOutput(values[5], values[3], values[1], nil, nil, nil) end

Line 346 ⟶ 407:

end

elseif (elements[6] == ELEMENT_DASH or elements[6] == ELEMENT_DATESEPARATOR) then -- '3 May 2013 - '

if (elements[7] == ELEMENT_ONETWODIGITS or elements[7] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- '3 May 2013 - 12' or '3rd May 2013 - 12th'

if (elements[8] == ELEMENT_WHITESPACE) then -- '3 May 2013 - 12 '

if (elements[9] == ELEMENT_MONTHWORD) then -- '3 May 2013 - 12 February'

Line 360 ⟶ 421:

end

elseif (elements[4] == ELEMENT_DASH or elements[4] == ELEMENT_DATESEPARATOR) then -- '3 May - '

if (elements[5] == ELEMENT_ONETWODIGITS or elements[5] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- '3 May - 12' or '3rd May - 12th'

if (elements[6] == ELEMENT_WHITESPACE) then -- '3 May - 12 '

if (elements[7] == ELEMENT_MONTHWORD) then -- '3 May - 12 October'

if (numberOfElements == 7) then return checkAndOutput(nil, values[3], values[1], nil, nil, nil, nil, values[7], values[5], nil, nil, nil) end

if (elements[8] == ELEMENT_COMMA or elements[8] == ELEMENT_WHITESPACE) then -- '3 May - 12 October '

if (elements[9] == ELEMENT_FOURDIGITS) then -- '3 May - 12 October 2013'

Line 374 ⟶ 436:

end

elseif (elements[2] == ELEMENT_DASH or elements[2] == ELEMENT_DATESEPARATOR) then -- '3 - '

if (elements[3] == ELEMENT_ONETWODIGITS or elements[3] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- '3 - 12' or '3rd - 12th'

if (elements[4] == ELEMENT_WHITESPACE) then -- '3 - 12 '

if (elements[5] == ELEMENT_MONTHWORD) then -- '3 - 12 May'

if (numberOfElements == 5) then return checkAndOutput(nil, values[5], values[1], nil, nil, nil, nil, values[5], values[3], nil, nil, nil) end

if (elements[6] == ELEMENT_COMMA or elements[6] == ELEMENT_WHITESPACE) then -- '3 - 12 May '

if (elements[7] == ELEMENT_FOURDIGITS) then -- '3 - 12 May 2013'

Line 396 ⟶ 459:

local i = 0 -- this is our offset from the closest possible ___location for date seeking

if (elements[2] == ELEMENT_TIMESEPARATOR and elements[1] ~= ELEMENT_ONETWODIGITSWITHORDINAL) then -- '10:' but not '10th:'

possibleHour = values[1] -- only once we see ':' (or 'am' below) it is likely a time

if (elements[3] == ELEMENT_ONETWODIGITS) then -- '10:28'

Line 472 ⟶ 535:

if (elements[2] == ELEMENT_DATESEPARATOR) then -- '2013-'

if (elements[3] == ELEMENT_ONETWODIGITS) then -- '2013-05'

--if (numberOfElements == 3) then return checkAndOutput(values[1], values[3], nil, nil, nil, nil) end

-- This is actually ambiguous -- 2008-12 can be years 2008 to 2012 or it could be Decemeber 2008; few cases, so just ignoring

if (elements[4] == ELEMENT_DATESEPARATOR) then -- '2013-05-'

if (elements[5] == ELEMENT_ONETWODIGITS) then -- '2013-05-03'

Line 505 ⟶ 569:

if (numberOfElements == 3) then return checkAndOutput(values[1], nil, nil, nil, nil, nil, values[3], nil, nil, nil, nil, nil) end

end

elseif (elements[2] == ELEMENT_WHITESPACE or elements[2] == ELEMENT_COMMA) then -- '2013 ' or '2013, '

if (elements[3] == ELEMENT_MONTHWORD) then -- '2013 May'

if (numberOfElements == 3) then return checkAndOutput(values[1], values[3], nil, nil, nil, nil) end

-- 2013 May - 2013 April (let's see first if this is ever used real-world)

if (elements[4] == ELEMENT_WHITESPACE) then -- '2013 May '

if (elements[5] == ELEMENT_ONETWODIGITS) then -- '2013 May 15'

if (numberOfElements == 5) then return checkAndOutput(values[1], values[3], values[5], nil, nil, nil) end

end

end

elseif (elements[3] == ELEMENT_ONETWODIGITS) then -- '2013 15' or '2013, 15'

if (elements[4] == ELEMENT_WHITESPACE) then -- '2013 15 '

if (elements[5] == ELEMENT_MONTHWORD) then -- '2013 15 May'

if (numberOfElements == 5) then return checkAndOutput(values[1], values[5], values[3], nil, nil, nil) end

end

end

end

end

elseif (elements[1] == ELEMENT_MONTHWORD) then -- 'May'

if (numberOfElements == 31) then return checkAndOutput(nil, values[1], nil, nil, nil, nil) end

if (elements[2] == ELEMENT_WHITESPACE or elements[2] == ELEMENT_PERIODWHITESPACE) then -- 'May ' or 'Feb. '

if (elements[3] == ELEMENT_ONETWODIGITS or elements[3] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- 'May 3' or 'May 3rd'

if (numberOfElements == 3) then return checkAndOutput(nil, values[1], values[3], nil, nil, nil) end

if (elements[4] == ELEMENT_COMMA or elements[4] == ELEMENT_WHITESPACE) then -- 'May 3, '

Line 543 ⟶ 618:

if (elements[7] == ELEMENT_MONTHWORD) then -- 'May 3, 2013 - February'

if (elements[8] == ELEMENT_WHITESPACE) then -- 'May 3, 2013 - February '

if (elements[9] == ELEMENT_ONETWODIGITS or elements[3] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- 'May 3, 2013 - February 12' or 'May 3rd, 2013 - February 12th'

if (elements[10] == ELEMENT_COMMA or elements[10] == ELEMENT_WHITESPACE) then -- 'May 3, 2013 - February 12, '

if (elements[11] == ELEMENT_FOURDIGITS) then -- 'May 3, 2013 - February 12, 2014'

Line 557 ⟶ 632:

if (elements[5] == ELEMENT_MONTHWORD) then -- 'May 3 - June'

if (elements[6] == ELEMENT_WHITESPACE) then -- 'May 3 - June '

if (elements[7] == ELEMENT_ONETWODIGITS or elements[3] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- 'May 3 - June 12' or 'May 3rd - June 12th'

if (numberOfElements == 7) then return checkAndOutput(nil, values[1], values[3], nil, nil, nil, nil, values[5], values[7], nil, nil, nil) end

if (elements[8] == ELEMENT_COMMA or elements[8] == ELEMENT_WHITESPACE) then -- 'May 3 - June 12, '

if (elements[9] == ELEMENT_FOURDIGITS) then -- 'May 3 - June 12, 2014'

Line 565 ⟶ 641:

end

end

elseif (elements[5] == ELEMENT_ONETWODIGITS or elements[3] == ELEMENT_ONETWODIGITSWITHORDINAL) then -- 'May 3 - 12' or 'May 3rd - 12th'

if (numberOfElements == 5) then return checkAndOutput(nil, values[1], values[3], nil, nil, nil, nil, values[1], values[5], nil, nil, nil) end

if (elements[6] == ELEMENT_COMMA or elements[6] == ELEMENT_WHITESPACE) then -- 'May 3 - 12, '

if (elements[7] == ELEMENT_FOURDIGITS) then -- 'May 3 - 12, 2013'

Line 592 ⟶ 669:

end

end

end

elseif (elements[2] == ELEMENT_COMMA) then -- 'May, '

if (elements[3] == ELEMENT_FOURDIGITS) then -- 'May, 2012'

if (numberOfElements == 3) then return checkAndOutput(values[3], values[1], nil, nil, nil, nil) end

end

end

Line 601 ⟶ 682:

end

function hasMetadataTemplates(input)

-- This is a basic list of the template names for metadata emiting tempaltes, there are inr eality more templates and more redirects

if (string.match(input, '%{%{[Ss]tart[ %-]?date')) then return true end

if (string.match(input, '%{%{[Ee]nd[ %-]?date')) then return true end

if (string.match(input, '%{%{[Bb]irth[ %-]?date')) then return true end

if (string.match(input, '%{%{[Dd]eath[ %-]?date')) then return true end

if (string.match(input, '%{%{[Bb]irth[ %-]?year')) then return true end

if (string.match(input, '%{%{[Dd]eath[ %-]?year')) then return true end

if (string.match(input, '%{%{[Ff]ilm ?date')) then return true end

if (string.match(input, '%{%{[Rr]elease ?date')) then return true end

if (string.match(input, '%{%{[ISO[ %-]date')) then return true end

return false

end

-- This function will return a raw string for generic checks and unit test, including defined parse errors

function main.parseDateOutputRaw(frame)

local result, startDate, startTime, endDate, endTimeinput = parseDateString(frame.args[1])

-- If desired (default), unstrip and decode to have the raw markup

if (not frame.args.noUnstrip or frame.args.noUnstrip ~= 'yes') then

input = mw.text.decode(mw.text.unstrip(input))

end

-- If desired (not default), strip the field extra stuff

if (frame.args.stripExtras and frame.args.stripExtras == 'yes') then

input = stripFieldExtras(input)

end

-- If there is nothing but whitespace, don't bother

if (result == PARSERESULT_FAIL) then if (frame.args[2] == 'pretty') then return '\'\'{{gray|Failed parse}}\'\'' else return 'Failed parse' end end

if (resultmw.ustring.match(input, == PARSERESULT_UNRECOGNIZED'^%s*$')) then if (frame.args[2] == 'pretty') then return frame:preprocess('\'\'{{gray|UnrecognizedEmpty patterninput}}\'\'') else return 'UnrecognizedEmpty patterninput' end end

if (result == PARSERESULT_INVALID) then if (frame.args[2] == 'pretty') then return '\'\'{{maroon|Invalid date/time}}\'\'' else return 'Invalid date/time' end end

local result, startDate, startTime, endDate, endTime = parseDateString(input)

if (result == PARSERESULT_INVALIDRANGE) then if (frame.args[2] == 'pretty') then return '\'\'{{maroon|Invalid date range}}\'\'' else return 'Invalid date range' end end

if (result == PARSERESULT_INCOMPLETE) then if (frame.args[2] == 'pretty') then return '\'\'{{cyan|Incomplete date}}\'\'' else return 'Incomplete date' end end

if (result == PARSERESULT_FAIL) then if (frame.args[2] == 'pretty') then return frame:preprocess('\'\'{{gray|Failed parse}}\'\'') else return 'Failed parse' end end

if (result == PARSERESULT_UNRECOGNIZED) then

local s

if (hasMetadataTemplates(input)) then s = 'Has metadata template' else s = 'Unrecognized pattern' end

if (frame.args[2] == 'pretty') then return frame:preprocess('\'\'{{gray|'..s..'}}\'\'') else return s end

end

if (result == PARSERESULT_INVALID) then if (frame.args[2] == 'pretty') then return frame:preprocess('\'\'{{maroon|Invalid date/time}}\'\'') else return 'Invalid date/time' end end

if (result == PARSERESULT_INVALIDRANGE) then if (frame.args[2] == 'pretty') then return frame:preprocess('\'\'{{maroon|Invalid date range}}\'\'') else return 'Invalid date range' end end

if (result == PARSERESULT_INCOMPLETE) then if (frame.args[2] == 'pretty') then return frame:preprocess('\'\'{{cyan|Incomplete date}}\'\'') else return 'Incomplete date' end end

if (result == PARSERESULT_INCOMPLETERANGE) then if (frame.args[2] == 'pretty') then return frame:preprocess('\'\'{{cyan|Incomplete date range}}\'\'') else return 'Incomplete date range' end end

local s

Line 622 ⟶ 741:

end

function stripFieldExtras(value)

--[[

function main.parseField(frame)

-- todo: do progressive scan like with that seek string just for ref tags and such

-- note that we can't just replace matches with whitespace and catch them all, because it could be like '3 August<!---->20<ref/>12'

local matchStrip = value:match('^([^<]-)<ref[^>]*>[^<]*</ref><ref[^>]*>[^<]*</ref><ref[^>]*>[^<]*</ref>$') -- basic refs (quite common)

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^([^<]-)<ref[^>]*>[^<]*</ref><ref[^>]*>[^<]*</ref>$') -- basic refs (quite common)

local input = frame.args[1]

if (matchStrip) then return mw.text.trim(matchStrip) end

matchNoRefmatchStrip = inputvalue:match('^([^<]*-)<ref[^>]*>[^<]*</ref>$') -- basic refrefs (quite common)

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^([^<]-)<ref[^>]*/><ref[^>]*/><ref[^>]*/>$') -- basic named ref (quite common)

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^([^<]-)<ref[^>]*/><ref[^>]*/>$') -- basic named ref (quite common)

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^([^<]-)<ref[^>]*/>$') -- basic named ref (quite common)

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^<!--.--->(.+)$') -- comment before (sometimes used for notes to editors [not yet seen metadata-related comment])

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^([^<]-)<!--.--->$') -- comment after (sometimes used for notes to editors)

if (matchNoRef) then

if (matchStrip) then return parseDateStringmw.text.trim(matchNoRefmatchStrip) end

else

return parseDateString(input)

end

matchStrip = value:match('^{{[Ff]lag ?icon[^}]+}}%s*{{[Ff]lag ?icon[^}]+}}(.+)$') -- 2 flag icons (also more common than one would think)

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^{{[Ff]lag ?icon[^}]+}}(.+)$') -- flag icon (quite common, although against MOS:ICON)

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^([^<]-){{[Ff]lag ?icon[^}]+}}%s*{{[Ff]lag ?icon[^}]+}}$') -- after as well

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^([^<]-){{[Ff]lag ?icon[^}]+}}$')

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^{{[Nn]o ?wrap *%|([^}%|]+)}}$') -- no-wrapped value

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^{{[Nn]o ?break *%|([^}%|]+)}}$') -- no-wrapped redirect

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^{{[Nn]obr *%|([^}%|]+)}}$') -- no-wrapped redirect

if (matchStrip) then return mw.text.trim(matchStrip) end

matchStrip = value:match('^{{[Jj] *%|([^}%|]+)}}$') -- no-wrapped redirect, yes '{{j|...}}'

if (matchStrip) then return mw.text.trim(matchStrip) end

return value -- if we didn't match anything, abort now

end

]]

--[[

function main.emitMetadata(frame)

local input = frame.args[1]

-- If desired (default), unstrip and decode to have the raw markup

if (not frame.args.noUnstrip or frame.args.noUnstrip ~= 'yes') then

input = mw.text.decode(mw.text.unstrip(input))

end

input = stripFieldExtras(mw.text.trim(input))

-- First parse the date and see if we get a valid output date

local date = parseDateString(frame.args[1])

if (not date) then return nil end

-- If there is nothing but whitespace, don't bother

local spanClass = frame.args.spanClass or 'bday dtstart published updated'

if (mw.ustring.match(input, '^%s*$')) then return nil end

-- Then parse the date and see if we get a valid output date

return '<span style="display:none">&#160;(<span class="' .. spanClass .. '">' .. date .. '</span>)</span>'

local result, startDate, startTime, endDate, endTime = parseDateString(input)

if (not frame.args.noErrorCats or frame.args.noErrorCats ~= 'yes') then

if (result == PARSERESULT_FAIL and not hasMetadataTemplates(input)) then return frame:preprocess('<includeonly>[[Category:Articles that could not be parsed for automatic date metadata]]</includeonly>') end

if (result == PARSERESULT_UNRECOGNIZED and not hasMetadataTemplates(input)) then return frame:preprocess('<includeonly>[[Category:Articles that could not be parsed for automatic date metadata]]</includeonly>') end

--if (result == PARSERESULT_INVALID) then return frame:preprocess('<includeonly>[[Category:]]</includeonly>') end

--if (result == PARSERESULT_INVALIDRANGE) then return frame:preprocess('<includeonly>[[Category:]]</includeonly>') end

if (result == PARSERESULT_INCOMPLETE) then return frame:preprocess('<includeonly>[[Category:Articles with incomplete dates for automatic metadata]]</includeonly>') end

if (result == PARSERESULT_INCOMPLETERANGE) then return frame:preprocess('<includeonly>[[Category:Articles with incomplete date ranges for automatic metadata]]</includeonly>') end

-- we need to use frame:preprocess() for <includeonly> or it just gets displayed

end

-- We are only doing the rest for a valid date

end

if (result ~= PARSERESULT_OKAY) then return nil end

]]

local dtstartSpan, dtendSpan

--[[

function main.testUnicode(frame)

-- If we have a start value and we're told to output it

if ((startDate or startTime) and frame.args.dtstart and frame.args.dtstart == 'yes') then

if (startDate and startTime) then dtstartSpan = '<span class="dtstart">' .. startDate .. 'T' .. startTime .. '</span>'

elseif (startDate) then dtstartSpan = '<span class="dtstart">' .. startDate .. '</span>'

else dtstartSpan = '<span class="dtstart">' .. startTime .. '</span>' end

end

-- If we have an end value and we're told to output it

local input = frame.args[1]

if ((endDate or endTime) and frame.args.dtend and frame.args.dtend == 'yes') then -- end values only happen when start values happen

if (endDate and endTime) then dtendSpan = '<span class="dtend">' .. endDate .. 'T' .. endTime .. '</span>'

elseif (endDate) then dtendSpan = '<span class="dtend">' .. endDate .. '</span>'

else dtendSpan = '<span class="dtend">' .. endTime .. '</span>' end

end

local s, e, mtrackingCat = mw.ustring.find(input, '^(%s*[%-–—]%s*)')

if (frame.args.trackingCat and frame.args.trackingCat == 'yes') then

if (s) then

trackingCat = '[[Category:Articles with automatically detected infobox date metadata]]'

return 'yes'

else

return 'no'

end

if (dtstartSpan and dtendSpan) then return '<span style="display:none">&#160;(' .. dtstartSpan .. ' - ' .. dtendSpan .. ')</span>' .. trackingCat

elseif (dtstartSpan) then return '<span style="display:none">&#160;(' .. dtstartSpan .. ')</span>' .. trackingCat

elseif (dtendSpan) then return '<span style="display:none">&#160;(' .. dtendSpan .. ')</span>' .. trackingCat

else return nil end

end

function main.outputRawStripped(frame)

return stripFieldExtras(mw.text.trim(mw.text.decode(mw.text.unstrip(frame.args[1]))))

end

function main.reoutputme(frame)

--if (frame.args.preprocess and frame.args.preprocess == 'yes') then

-- return frame:preprocess(mw.text.decode(mw.text.unstrip(frame.args[1])))

--else

-- return mw.text.decode(mw.text.unstrip(frame.args[1]))

--end

----[[

-- frame.args[1]

-- frame:getParent().args[1]

-- frame:getArgument(1)

--do return frame:getArgument(1):expandTo('text/plain') end

local input = mw.text.decode(mw.text.unstrip(frame.args[1]))

s = 'Len=' .. #input .. ' '

for i = 1, string.len(input) do

s = s .. string.sub(input, i, i) .. ' '

end

return s--]]

end

]]

return main