Content deleted Content added
switching from :gsub to :match when possible Tag: Reverted |
No edit summary |
||
| (15 intermediate revisions by 6 users not shown) | |||
Line 1:
require ('strict');
local p = {}
Line 11 ⟶ 12:
end
p.ucfirst = function (frame
local s =
if not s or '' == s or s:match ('^%s+$') then -- when <s> is nil, empty, or only whitespace
return s; -- abandon because nothing to do
end
s = mw.text.trim( frame.args[1] or "" )
local s1 = ""
local prefix_patterns_t = { -- sequence of prefix patterns
'^\127[^\127]*UNIQ%-%-%a+%-%x+%-QINU[^\127]*\127', -- stripmarker
'^([%*;:#]+)', -- various list markup
'^(\'\'\'*)', -- bold / italic markup
'^(%b<>)', -- html-like tags because some templates render these
'^(&%a+;)', -- html character entities because some templates render these
'^(&#%d+;)', -- html numeric (decimal) entities because some templates render these
'^(&#x%x+;)', -- html numeric (hexadecimal) entities because some templates render these
'^(%s+)', -- any whitespace characters
'^([%(%)%-%+%?%.%%!~!@%$%^&_={}/`,‘’„“”ʻ|\"\'\\]+)', -- miscellaneous punctuation
}
local prefixes_t = {}; -- list, bold/italic, and html-like markup, & whitespace saved here
local function prefix_strip (s) -- local function to strip prefixes from <s>
for _, pattern in ipairs (prefix_patterns_t) do -- spin through <prefix_patterns_t>
if s:match (pattern) then -- when there is a match
local prefix = s:match (pattern); -- get a copy of the matched prefix
table.insert (prefixes_t, prefix); -- save it
s = s:sub (prefix:len() + 1); -- remove the prefix from <s>
return s, true; -- return <s> without prefix and flag; force restart at top of sequence because misc punct removal can break stripmarker
end
end
return s; -- no prefix found; return <s> with nil flag
end
local prefix_removed; -- flag; boolean true as long as prefix_strip() finds and removes a prefix
repeat -- one by one remove list, bold/italic, html-like markup, whitespace, etc from start of <s>
s, prefix_removed = prefix_strip (s);
until (not prefix_removed); -- until <prefix_removed> is nil
s1 = table.concat (prefixes_t); -- recreate the prefix string for later reattachment
local first_text = mw.ustring.match (s, '^%[%[[^%]]+%]%]'); -- extract wikilink at start of string if present; TODO: this can be string.match()?
local upcased;
if first_text then
if first_text:match ('^%[%[[^|]+|[^%]]+%]%]') then -- if <first_text> is a piped link
upcased = mw.ustring.match (s, '^%[%[[^|]+|%W*(%w)'); -- get first letter character
upcased = mw.ustring.upper (upcased); -- upcase first letter character
s = mw.ustring.gsub (s, '^(%[%[[^|]+|%W*)%w', '%1' .. upcased); -- replace
else -- here when <first_text> is a wikilink but not a piped link
upcased = mw.ustring.match (s, '^%[%[%W*%w'); -- get '[[' and first letter
upcased = mw.ustring.upper (upcased); -- upcase first letter character
s = mw.ustring.gsub (s, '^%[%[%W*%w', upcased); -- replace; no capture needed here
end
elseif s:match ('^%[%S+%s+[^%]]+%]') then -- if <s> is a ext link of some sort; must have label text
upcased = mw.ustring.match (s, '^%[%S+%s+%W*(%w)'); -- get first letter character
upcased = mw.ustring.upper (upcased); -- upcase first letter character
s = mw.ustring.gsub (s, '^(%[%S+%s+%W*)%w', '%1' .. upcased); -- replace
elseif s:match ('^%[%S+%s*%]') then -- if <s> is a ext link without label text; nothing to do
return s1 .. s; -- reattach prefix string (if present) and done
else -- <s> is not a wikilink or ext link; assume plain text
upcased = mw.ustring.match (s, '^%W*%w'); -- get the first letter character
upcased = mw.ustring.upper (upcased); -- upcase first letter character
s = mw.ustring.gsub (s, '^%W*%w', upcased); -- replace; no capture needed here
end
return s1 .. s; -- reattach prefix string (if present) and done
end
p.title = function (frame)
-- http://grammar.yourdictionary.com/capitalization/rules-for-capitalization-in-titles.html
-- recommended by The U.S. Government Printing Office Style Manual:
Line 204 ⟶ 251:
-- Escape Pattern helper function so that all characters are treated as plain text, as per Module:String
function p._escapePattern( pattern_str )
return mw.ustring.gsub( pattern_str, "([%(%)%.%%%+%-%*%?%[%^%$%]])", "%%%1" )
end
Line 317 ⟶ 364:
temp_str, accept = temp_str:gsub ('^%(%((.+)%)%)$', '%1') -- remove accept-this-as-written markup when it wraps all of concatenated out
if accept ~= 0 then
temp_str = str:
end
return temp_str
Line 332 ⟶ 379:
function p.startswith(frame)
return (frame.args[1]:sub(1, frame.args[2]:len()) == frame.args[2]) and 'yes' or ''
end
-- Implements [[Template:Isnumeric]]
function p.isnumeric(frame)
local s = frame.args[1] or frame:getParent().args[1]
local boolean = (frame.args.boolean or frame:getParent().args.boolean) == 'true'
if type(s) == 'string' and mw.getContentLanguage():parseFormattedNumber( s ) then
return boolean and 1 or s
end
return boolean and 0 or ''
end
-- Checks if a value in a group of numbers is not an interger.
-- Allows usage of an |empty= parameter to allow empty values to be skipped.
function p.isInteger(frame)
local values = frame.args or frame:getParent().args
local allow_empty = frame.args.empty or frame:getParent().args.empty
for _, value in ipairs(values) do
-- Trim spaces
value = value and value:gsub("^%s*(.-)%s*$", "%1")
if value == "" or value == nil then
if not allow_empty then
return false -- Empty values are not allowed
end
else
value = tonumber(value)
if not (type(value) == "number" and value == math.floor(value)) then
return false
end
end
end
return true
end
-- Returns an error found in a string.
function p.getError(frame)
local text = frame.args[1] or frame:getParent().args[1]
local error_message = text:match('(<strong class="error">.-</strong>)')
return error_message or nil
end
| |||