模块文档简介
此模块用于把多行wikitext压缩至一行。
用途
此模块最初是为了{{Hid}}编写的。由于MediaWiki的wikitext解析器存在问题，导致把多行wikitext放在列表（*#）缩进（:;）上时会出现错误的结果。
一个著名的例子就是{{Hide}}不能与列表和缩进联用（参见滥用过滤器30）：
| 代码 | 效果 | ||
|---|---|---|---|
* {{Hide}}
* 文本
文本
|
|
由于{{Hide}}展开后是多行wikitext，与列表或缩进连用会导致后续内容全部缩进。而此模块能够预先把多行wikitext压缩至一行，从而避免该问题。
| 代码 | 效果 | ||
|---|---|---|---|
* {{#invoke:Neutralizer|main| {{Hide}} }}
* 文本
文本
|
文本 |
此模块同样适用于面临相同困扰的其他模板，例如{{VersionHistory}}、{{Clade}}等。
技术细节
此模块用Lua部分重写了MediaWiki内置的wikitext解析器，因此能够解析表格、列表以及段落。
传入的wikitext可以用<nowiki>包围，这时此模块将先去掉外面的<nowiki>再进行解析。
然而，此模块尚未经过相对充分的测试，其解析结果可能会与预期存在一定的差别。
local string = string
local table = table
local ipairs = ipairs
local function match_with_offset(pattern, str, match, offset)
offset = offset or 1
match[1] = string.match(str, pattern, offset)
match[2] = string.find(str, pattern, offset)
if match[1] == nil then
return false
else
return true
end
end
local function explode(delim, str, limit)
local result = {}
local m = {0, 0}
local offset = 1
local count = 1
while (limit == nil or count < limit) and match_with_offset(delim, str, m, offset) do
table.insert(result, string.sub(str, offset, m[2]-1))
offset = m[2] + string.len(delim)
count = count + 1
end
table.insert(result, string.sub(str, offset))
return result
end
local function trim(str)
return string.match(str, '^ *(.-) *$')
end
local function rtrim(str)
return string.match(str, '^(.-) *$')
end
local function strspn(str, charlist)
return string.len(string.match(str, '^[' .. charlist .. ']+') or '')
end
-- converted from Parser.php with a few adjustments
local function parseTables(text)
local lines = explode('\n', text)
local out = {}
local td_history = {};
local last_tag_history = {};
local tr_history = {};
local tr_attributes = {};
local has_opened_tr = {};
local indent_level = 0;
for i, outLine in ipairs(lines) do
local line = trim(outLine)
if line == '' then
table.insert(out, outLine .. '\n')
else--CONTINUE
local first_character = string.sub(line, 1, 1)
local first_two = string.sub(line, 1, 2)
local matches = {}
matches[1], matches[3], matches[2] = string.match(line, '^(:*)%s*({|)(.*)$')
if matches[3] ~= nil then
indent_level = string.len(matches[1] or '')
local attributes = matches[2] or ''
-- unstripBoth & fixTagAttributes
outLine = string.rep('<dl><dd>', indent_level) .. '<table ' .. attributes .. '>'
table.insert(td_history, false)
table.insert(last_tag_history, '')
table.insert(tr_history, false)
table.insert(tr_attributes, '')
table.insert(has_opened_tr, false)
elseif #td_history == 0 then
elseif first_two == '|}' then
line = '</table>' .. string.sub(line, 3)
local last_tag = table.remove(last_tag_history)
if not table.remove(has_opened_tr) then
line = '<tr><td></td></tr>' .. line
end
if table.remove(tr_history) then
line = '</tr>' .. line
end
if table.remove(td_history) then
line = '</' .. last_tag .. '>' .. line
end
table.remove(tr_attributes)
if indent_level > 0 then
outLine = rtrim(line) .. string.rep('</dd></dl>', indent_level)
else
outLine = line
end
elseif first_two == '|-' then
line = string.gsub(line, '^|%-+', '')
local attributes = line
-- unstripBoth & fixTagAttributes
table.remove(tr_attributes)
table.insert(tr_attributes, attributes)
line = ''
local last_tag = table.remove(last_tag_history)
table.remove(has_opened_tr)
table.insert(has_opened_tr, true)
if table.remove(tr_history) then
line = '</tr>'
end
if table.remove(td_history) then
line = '</' .. last_tag .. '>' .. line
end
outLine = line
table.insert(tr_history, false)
table.insert(td_history, false)
table.insert(last_tag_history, '')
elseif first_character == '|'
or first_character == '!'
or first_two == '|+' then
if first_two == '|+' then
first_character = '+'
line = string.sub(line, 3)
else
line = string.sub(line, 2)
end
if first_character == '!' then
-- replaceMarkup
line = string.gsub(line, '!!', '||')
end
local cells = explode('||', line)
outLine = ''
for j, cell in ipairs(cells) do
local previous = ''
if first_character ~= '+' then
local tr_after = table.remove(tr_attributes)
if not table.remove(tr_history) then
previous = '<tr ' .. tr_after .. '>'
end
table.insert(tr_history, true)
table.insert(tr_attributes, '')
table.remove(has_opened_tr)
table.insert(has_opened_tr, true)
end
local last_tag = table.remove(last_tag_history)
if table.remove(td_history) then
previous = '</' .. last_tag .. '>\n' .. previous
end
if first_character == '|' then
last_tag = 'td'
elseif first_character == '!' then
last_tag = 'th'
elseif first_character == '+' then
last_tag = 'caption'
else
last_tag = ''
end
table.insert(last_tag_history, last_tag)
local cell_data_iter = explode('|', cell, 2)
local cell_data = {}
for k, item in ipairs(cell_data_iter) do
table.insert(cell_data, item)
end
if string.match('[[', cell_data[1]) or string.match('%-{', cell_data[1]) then
cell = previous .. '<' .. last_tag .. '>' .. trim(cell)
elseif #cell_data == 1 then
cell = previous .. '<' .. last_tag .. '>' .. trim(cell_data[1])
else
local attributes = cell_data[1]
-- unstripBoth & fixTagAttributes
cell = previous .. '<' .. last_tag .. ' ' .. attributes .. '>' .. trim(cell_data[2])
end
outLine = outLine .. cell
table.insert(td_history, true)
end
end
table.insert(out, outLine .. '\n')
end--CONTINUE
end
while #td_history > 0 do
if table.remove(td_history) then
table.insert(out, '</td>')
end
if table.remove(tr_history) then
table.insert(out, '</tr>')
end
if not table.remove(has_opened_tr) then
table.insert(out, '<tr><td></td></tr>')
end
table.insert(out, '</table>')
end
if out[#out] == '\n' then
table.remove(out)
end
--if out == '<table><tr><td></td></tr></table>' then
-- out = ''
--end
return table.concat(out)
end
-- converted from BlockLevelPass.php with a few adjustments
local DTopen = false
local lastParagraph = ''
local COLON_STATE = {
['TEXT']=0,
['TAG']=1,
['TAGSTART']=2,
['CLOSETAG']=3,
['TAGSLASH']=4,
['COMMENT']=5,
['COMMENTDASH']=6,
['COMMENTDASHDASH']=7,
['LC']=8
}
local function hasOpenParagraph()
return lastParagraph ~= ''
end
local function closeParagraph(atTheEnd)
atTheEnd = atTheEnd or false
local result = ''
if hasOpenParagraph() then
result = '</' .. lastParagraph .. '>'
if not atTheEnd then
result = result .. '\n'
end
end
lastParagraph = ''
return result
end
local function getCommon(st1, st2)
local shorter = math.min(string.len(st1), string.len(st2))
local count=0
for i=1, shorter do
if string.sub(st1, i, i) ~= string.sub(st2, i, i) then
break
end
count = count + 1
end
return count
end
local function openList(char)
local result = closeParagraph()
if char == '*' then
result = result .. '<ul><li>'
elseif char == '#' then
result = result .. '<ol><li>'
elseif char == ':' then
result = result .. '<dl><dd>'
elseif char == ';' then
result = result .. '<dl><dt>'
DTopen = true
end
return result
end
local function nextItem(char)
if char == '*' or char == '#' then
return '</li>\n<li>'
elseif char == ':' or char == ';' then
local close = '</dd>\n'
if DTopen then
close = '</dt>\n'
end
if char == ';' then
DTopen = true
return close .. '<dt>'
else
DTopen = false
return close .. '<dd>'
end
end
return ''
end
local function closeList(char)
local text = ''
if char == '*' then
text = '</li></ul>'
elseif char == '#' then
text = '</li></ol>'
elseif char == ':' then
if DTopen then
DTopen = false
text = '</dt></dl>'
else
text = '</dd></dl>'
end
end
return text
end
local function findColonNoLinks(str, before_after)
local m = {0, 0}
if not (match_with_offset(':', str, m) or match_with_offset('<', str, m) or match_with_offset('%-{', str, m)) then
return false
end
if m[1] == ':' then
local colonPos = m[2]
before_after[1] = string.sub(str, 1, colonPos+1)
before_after[2] = string.sub(str, colonPos+2)
return colonPos
end
local state = COLON_STATE.TEXT
local ltLevel = 0
local lcLevel = 0
local len = string.len(str)
local i = m[2]
while i < len do
local c = string.sub(str, i, i)
if state == COLON_STATE.TEXT then
if c == '<' then
state = COLON_STATE.TAGSTART
elseif c == ':' then
if ltLevel == 0 then
before_after[1] = string.sub(str, 1, i+1)
before_after[2] = string.sub(str, i+2)
return i
end
else
if not (match_with_offset(':', str, m) or match_with_offset('<', str, m) or match_with_offset('%-{', str, m)) then
return false
end
if m[1] == '-{' then
state = COLON_STATE.LC
lcLevel = lcLevel + 1
i = m[2] + 1
else
i = m[2] - 1
end
end
elseif state == COLON_STATE.LC then
if not (match_with_offset('%-{', str, m, i+1) or match_with_offset('}%-', str, m, i+1)) then
break
end
if m[1] == '-{' then
i = m[2] + 1
lcLevel = lcLevel + 1
elseif m[1] == '}-' then
i = m[2] + 1
lcLevel = lcLevel - 1
if lcLevel == 0 then
state = COLON_STATE.TEXT
end
end
elseif state == COLON_STATE.TAG then
if c == '>' then
ltLevel = ltLevel + 1
state = COLON_STATE.TEXT
elseif c == '/' then
state = COLON_STATE.TAGSLASH
end
elseif state == COLON_STATE.TAGSTART then
if c == '/' then
state = COLON_STATE.CLOSETAG
elseif c == '!' then
state = COLON_STATE.COMMENT
elseif c == '>' then
state = COLON_STATE.TEXT
else
state = COLON_STATE.TAG
end
elseif state == COLON_STATE.CLOSETAG then
if c == '>' then
if ltLevel > 0 then
ltLevel = ltLevel - 1
end
state = COLON_STATE.TEXT
end
elseif state == COLON_STATE.TAGSLASH then
if c == '-' then
state = COLON_STATE.COMMENTDASH
else
state = COLON_STATE.COMMENT
end
elseif state == COLON_STATE.COMMENTDASH then
if c == '>' then
state = COLON_STATE.TEXT
else
state = COLON_STATE.COMMENT
end
end
end
return false
end
local function parseBlockLevel(text)
local textLines = explode('\n', text)
local lastPrefix = ''
local output = {}
DTopen = false
local inBlockElem = false
local prefixLength = 0
local pendingPTag = false
local inBlockquote = false
local prefix2 = ''
for i, inputLine in ipairs(textLines) do
local lastPrefixLength = string.len(lastPrefix)
prefixLength = strspn(inputLine, '*#:;')
local prefix = string.sub(inputLine, 1, prefixLength)
prefix2 = string.gsub(prefix, ';', ':')
local t = string.sub(inputLine, prefixLength+1)
if prefixLength ~= 0 and lastPrefix == prefix2 then
table.insert(output, nextItem(string.sub(prefix, -1, -1)))
pendingPTag = false
if string.sub(prefix, -1, -1) == ';' then
local term_t2 = {'', ''}
if findColonNoLinks(t, term_t2) ~= false then
t = term_t2[1]
table.insert(output, trim(term_t2[1]) .. nextItem(':'))
end
end
elseif prefixLength ~= 0 or lastPrefixLength ~= 0 then
local commonPrefixLength = getCommon(prefix, lastPrefix)
pendingPTag = false
while commonPrefixLength < lastPrefixLength do
table.insert(output, closeList(string.sub(lastPrefix, lastPrefixLength, lastPrefixLength)))
lastPrefixLength = lastPrefixLength - 1
end
if prefixLength <= commonPrefixLength and commonPrefixLength > 0 then
table.insert(output, nextItem(string.sub(prefix, commonPrefixLength, commonPrefixLength)))
end
if DTopen and commonPrefixLength > 0 and string.sub(prefix, commonPrefixLength, commonPrefixLength) == ':' then
table.insert(output, nextItem(':'))
end
if lastPrefix ~= '' and prefixLength > commonPrefixLength then
table.insert(output, '\n')
end
while prefixLength > commonPrefixLength do
local char = string.sub(prefix, commonPrefixLength + 1, commonPrefixLength + 1)
table.insert(output, openList(char))
if char == ';' then
local term_t2 = {'', ''}
if findColonNoLinks(t, term_t2) ~= false then
t = term_t2[1]
table.insert(output, trim(term_t2[1]) .. nextItem(':'))
end
end
commonPrefixLength = commonPrefixLength + 1
end
if not prefixLength ~= 0 and lastPrefix ~= 0 then
table.insert(output, '\n')
end
lastPrefix = prefix2
end
if prefixLength == 0 then
local blockElems = {'table', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'pre', 'p', 'ul', 'ol', 'dl'}
local antiBlockElems = {'td', 'th'}
local openMatch = false
for j, elem in ipairs(blockElems) do
if string.match(t, '<' .. elem .. '%f[%W]') then
openMatch = true
end
end
if not openMatch then
for j, elem in ipairs(antiBlockElems) do
if string.match(t, '</' .. elem .. '%f[%W]') then
openMatch = true
end
end
end
if not openMatch then
for j, elem in ipairs({'tr', 'caption', 'dt', 'dd', 'li'}) do
if string.match(t, '</?' .. elem .. '%f[%W]') then
openMatch = true
end
end
end
local closeMatch = false
for j, elem in ipairs(blockElems) do
if string.match(t, '</' .. elem .. '%f[%W]') then
closeMatch = true
end
end
if not closeMatch then
for j, elem in ipairs(antiBlockElems) do
if string.match(t, '<' .. elem .. '%f[%W]') then
closeMatch = true
end
end
end
if not closeMatch then
for j, elem in ipairs({'center', 'blockquote', 'div', 'hr', 'mw:', 'aside', 'figure'}) do
if string.match(t, '</?' .. elem .. '%f[%W]') then
closeMatch = true
end
end
end
if openMatch or closeMatch then
pendingPTag = false
closeParagraph()
local bqOffset = 1
local bqMatch = {0, 0}
while match_with_offset('<(/?)blockquote[%s>]', t, bqMatch, bqOffset) do
inBlockquote = not bqMatch[1]
bqOffset = bqMatch[2] + string.len(bqMatch[1])
end
inBlockElem = not closeMatch
elseif not inBlockElem then
if trim(t) ~= ''
and string.sub(t, 1, 2) == ' '
and not inBlockquote then
t = string.sub(t, 2)
elseif string.match(t, '^<style%f[%W][^>]*>.-</style>$')
or string.match(t, '<link%f[%W][^>]*>%s*') then
if pendingPTag ~= '' and pendingPTag ~= false then
table.insert(output, closeParagraph())
pendingPTag = false
end
else
if trim(t) == '' then
if pendingPTag ~= '' and pendingPTag ~= false then
table.insert(output, pendingPTag .. '<br />')
pendingPTag = false
lastParagraph = 'p'
elseif lastParagraph ~= 'p' then
table.insert(output, closeParagraph())
pendingPTag = '<p>'
else
pendingPTag = '</p><p>'
end
elseif pendingPTag ~= '' and pendingPTag ~= false then
table.insert(output, pendingPTag)
pendingPTag = false
lastParagraph = 'p'
elseif lastParagraph ~= 'p' then
table.insert(output, closeParagraph() .. '<p>')
lastParagraph = 'p'
end
end
end
end
if pendingPTag == false then
if prefixLength == 0 then
table.insert(output, t)
if hasOpenParagraph() then
table.insert(output, '\n')
end
else
table.insert(output, trim(t))
end
end
end
while prefixLength > 0 do
table.insert(output, closeList(string.sub(prefix2, prefixLength, prefixLength)))
prefixLength = prefixLength - 1
if prefixLength ~= 0 and hasOpenParagraph() then
table.insert(output, '\n')
end
end
table.insert(output, closeParagraph(true))
return table.concat(output)
end
-- the Module
local p = {}
local getArgs = require('Module:Arguments').getArgs
function p.main(frame)
local args = getArgs(frame, { removeBlanks = false, frameOnly = true })
local rawText = args[1] or ''
rawText = mw.text.unstripNoWiki(rawText)
rawText = mw.text.decode(rawText)
rawText = frame:preprocess(rawText)
local processedText = parseTables(rawText)
processedText = parseBlockLevel(processedText)
processedText = string.gsub(processedText, '\n', '')
return processedText
end
return p