Module:String2/sandbox: Difference between revisions

Revision as of 00:54, 25 February 2024

Revision as of 20:16, 29 February 2024

Line 1:

require ('strict');

local p = {}

Line 14 ⟶ 15:

local s = mw.text.trim( frame.args[1] or "" )

local s1 = ""

-- if it's a list chop off and (store as s1) everything up to the first <li>

local prefix_patterns_t = { -- sequence of prefix patterns

local lipos = mw.ustring.find(s, "<li>" )

'^([%*;:#]+)', -- various list markup

if lipos then

'^(\'\'\'*)', -- bold / italic markup

s1 = mw.ustring.sub(s, 1, lipos + 3)

'^(%b<>)', -- html-like tags because some templates render these

s = mw.ustring.sub(s, lipos + 4)

'^(&%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 function prefix_strip (s, prefixes_t) -- local function to strip prefixes from <s>

local prefix_removed; -- flag; nil when no prefix removed

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>

prefix_removed = true; -- set the flag

end

end

return s, prefix_removed; -- return <s> without prefixes and flag

end

-- s1 is either "" or the first part of the list markup, so we can continue

-- and prepend s1 to the returned string

local letterpos

local prefixes_t = {}; -- list, bold/italic, and html-like markup, & whitespace saved here

local first_text = mw.ustring.match (s, '^%[%[[^%]]+%]%]'); -- extract

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, prefixes_t);

until (not prefix_removed); -- until <prefix_removed> is nil

s1 = table.concat (prefixes_t); -- recreate the prefix string for later reattachment

local prefix, upcased;

-- if it's a list chop off and (store as s1) everything up to the first <li>

-- why just list markup? Could be any sort of html markup:

-- <span title="Spanish-language text"><i lang="es">casa</i></span> (this from {{lang|es|casa}})

-- what about wikitext italic or bold markup? (''italic text'', '''bold text''') – both may be legitimate

-- local lipos = mw.ustring.find(s, "<li>" ) -- why not constrained to start of string?

-- if lipos then

-- s1 = mw.ustring.sub(s, 1, lipos + 3) -- get <li> tag (assumes that the tag is the first text in the string <s> – it may not be)

-- s = mw.ustring.sub(s, lipos + 4) -- get everything after the <li> tag

-- end

-- -- s1 is either "" or the first part of the list markup, so we can continue

-- -- and prepend s1 to the returned string

-- local letterpos

local first_text = mw.ustring.match (s, '^%[%[[^%]]+%]%]'); -- extract wikilink at start of string if present; TODO: this can be string.match()?

local prefix, upcased;

if first_text then

if mw.ustring.first_text:match (first_text, "'^%[%[[^|]+|[^%]]+%]%]"') then -- if <first_text> is a piped link

prefix, upcased = mw.ustring.match (s, '^(%[%[[^|]+|)(%W*(%w)'); -- get '[[<link text>|' and 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

prefix, 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)', '%1' .. upcased); -- replace; no capture needed here

end

elseif mw.ustring.s:match (s, '^%[%S+%s+[^%]]+%]') then -- if the first text<s> is a ext link of some sort; must have label text

first_textupcased = mw.ustring.match (s, '^%[[^%]]S+%]s+%W*(%w)'); -- get the first textletter character

prefix, upcased = mw.ustring.matchupper (s, '^(%[%S+%s+)(%W*%w)'upcased); -- get '[<url><space>' andupcase first letter character

upcaseds = mw.ustring.uppergsub (s, '^(%[%S+%s+%W*)%w', '%1' .. upcased); -- upcase first letterreplace

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

else

upcasedreturn =s1 mw.ustring.match (s, '(%W*%w)'); -- getreattach theprefix string (if present) firstand letterdone

upcased = mw.ustring.upper (upcased); -- upcase first letter

else -- <s> is not a wikilink or ext link; assume plain text

supcased = mw.ustring.gsubmatch (s, '^(%W*%w)', upcased); -- replaceget the first letter character

upcased = mw.ustring.upper (upcased); -- upcase first letter character

s = mw.ustring.gsub (s, '^(%[%S+%s+)(%W*%w)', '%1' .. upcased); -- replace; no capture needed here

end

return s1 .. s; -- reattach initialprefix list markupstring (if present) and done

-- if mw.ustring.find(s, "^%[%[[^|]+|[^%]]+%]%]") then