Module:Lyrics2

Documentation for this module may be created at Module:Lyrics2/doc

-- A port of [[Template:Lyrics]] to Lua
-- written by K
-- rewritten by DennouNeko

-- == HELPER FUNCTIONS ==
local common = require("Module:Common")
local utf8   = require("Module:Lib UTF8")

-- minor bugfixes, mw.text seem to be undefined at the time of this update
mw.text2 = require("Module:MW.text")

-- "Module:Common" candidate?
--[[
Builds a simple HTML table from a 'tbl' elements with 'options' as a list of options for table.
Useful for quick tests or if a complex table is not needed.
Currently available options are:
* cols - max column count
* class - text containing classes that should be added to the table
* style - text that should be added to the table's style (like "empty-cells: hide;")
* spacing - separation of table cells
--]]
function table.build_table(tbl, options)
  local cols = nil
  local ret = {}
  -- to avoid script crashing when skipping options in param list
  if options == nil then options = {} end
  if options['cols'] ~= nil then cols = tonumber(options['cols']) end

  ret[#ret+1] = '<table'
  if common.isset(options['class']) then ret[#ret+1] = ' class="' .. options['class'] .. '"' end
  if common.isset(options['style']) then ret[#ret+1] = ' style="' .. options['style'] .. '"' end
  if common.isset(options['spacing']) then ret[#ret+1] = ' cellspacing="' .. options['spacing'] .. '"' end
  ret[#ret+1] = '>'

  local colnum = 0
  if cols ~= nil then
    -- build a table with max 'cols' columns in row, content of cells are the values of 'tbl' elements
    ret[#ret+1] = '<tr>'
    for k,v in pairs(tbl) do
      if colnum > 0 and math.fmod(colnum, cols) == 0 then
        ret[#ret+1] = '</tr><tr>'
      end
      ret[#ret+1] = '<td><!--' .. k .. '-->\n' .. v .. '</td>'
      colnum = colnum + 1
    end

    if cols ~= nil then
      while math.fmod(colnum, cols) ~= 0 do
        ret[#ret+1] = '<td></td>'
        colnum = colnum + 1
      end
    end
    ret[#ret+1] = '</tr>'
  else
    -- build a simple table with 'key | value' rows
    for k,v in pairs(tbl) do
      ret[#ret+1] = '<tr><td>\n' .. k .. '</td><td>\n' .. v .. '</td></tr>'
    end
  end
  ret[#ret+1] = '</table>'
  return table.concat(ret)
end

--[[ Search the list of template arguments for arguments matching the pattern 
     "album#", where # is a number, and then return them as an indexed, sorted table.]]
local function scanForAlbums(frame)
  local idx = {}
  local args = frame.args
  local albumList = {}

  for k in pairs(args) do
    s,e,t = string.find(k, '^album(%d+)$')
    if s ~= nil then
      table.insert(idx, t)
    end
  end

  table.sort(idx)

  for k in pairs(idx) do
    table.insert (albumList, args["album"..k])
  end

  return albumList
end

--[[ Extracts list of names from a albumList.
     Indexes of found names match indexes in album table. ]]
local function getAlbumNames(albumList)
  local cats = {}
  local found
  local link

  for k,v in pairs(albumList) do
    found = nil
    link = nil
    s,e,t = string.find(v, '%[%[(.+)%]%]')
    if s ~= nil then
      tl = mw.text2.split(t, '|')
      for k1,v1 in pairs(tl) do
        s,e,t = string.find(v1, '^%s-link%s-=%s-(.+)%s-$')
        if s ~= nil then
          -- found a image tag
          found = t
          break
        end
      end
      if found == nil then
        -- not an image link then assume it's usual link, so get the last part as album name, first part as link
        if #tl > 0 then
          found = tl[#tl]
          if #tl > 1 then link = tl[1] end
        end
      end
    else
      -- TODO: not a link - assume it's a name?
      -- found = t
    end

    if found ~= nil then
      cats[#cats+1] = {name = string.gsub(found, '_', ' '), link = link, idx = k}
    end
  end

  return cats
end

--[[ Partial name normalization. ]]
local function normalize(text)
  if common.isset(text) then
    -- full-width characters to ASCII equivalents
    text = utf8.replace_char(text, common.normalization_table)
    -- remove any HTML tags
    text = common.stripTags(text)
    -- any whitespace chain into single space
    text = string.gsub(text, '%s+', ' ')
    -- remove preceding and trailing spaces
    text = mw.text.trim(text)
  end
  return text
end

--[[ Loads and parses a JavaScript table "song_info" from a script with given name. ]]
local function load_info(name)
  local tstr = common.getPage(name)
  if tstr == nil then return ret end

  local ret = {}
  local s,e,t,t1,t2

  s,e,t = string.find(tstr, '{(.+)}')
  if s ~= nil then
    for n,lst in string.gmatch(t, '"(.-[^\\])"%s-:%s-{(.-)}') do
      n = normalize(n)
      ret[n] = {}
      for t1,t2 in string.gmatch(lst, '"(.-[^\\])"%s-:%s-"(.-[^\\])"') do
        t1 = normalize(t1)
        -- unescape the string
        ret[n][t1] = string.gsub(t2, '\\(.)', {['\\'] = '\\', ['n'] = '\n', ['"'] = '"'})
      end
    end
  else
    error('song_info not found!')
  end

  return ret
end

--[[ Search for "source: " and "original title: ", add a hover text to title and a "notes: " below. ]]
local function wrap_titles(frame, elems, info_list)
  local source,sourcek = nil,nil
  local title,titlek,titlen = nil,nil,nil

  local s,e,t,t1,t2,tt

  for k,v in pairs(elems) do
    s,e,t1,t = string.find(v, '^(%*%s*original title:%s*)(.+)$')
    if s ~= nil then
      title = t
      titlen = normalize(t)
      titlek = k
      tt = t1
    else
      s,e,t = string.find(v, '^%*%s*source:%s*(.*)$')
      if s ~= nil then
        s,e,t1 = string.find(t, '%[%[.+|(.+)%]%]')
        if s == nil then s,e,t1 = string.find(t, '%[%[(.+)%]%]') end
        if s ~= nil then t = t1 end
        source = normalize(t)
        sourcek = k
      end
    end
    if common.isset(source) and common.isset(titlen) then
      if common.isset(info_list) and info_list[source] ~= nil then
        if common.isset(info_list[source][titlen]) then
          local extra = info_list[source][titlen]
          s,e,t,t1 = string.find(extra, '^(.*)%s-%[(.+)%]')

          if s == nil then t = extra end

          if s ~= nil and common.isset(t1) then
            local lst = mw.text2.split(t1, ';')
            if common.exists(lst[#lst]) then
              lst[#lst] = '[[' .. lst[#lst] .. ']]'
            end
            extra = table.concat(lst,'; ')
            --if common.isset(extra) then table.insert(elems, sourcek+1, '* notes: ' .. extra) end
            if common.isset(extra) then elems[sourcek] = elems[sourcek] .. '<br/>' .. extra end
          end
        
          if common.isset(t) then
            elems[titlek] = tt .. frame:expandTemplate{title = 'h:title', args = {title, t} }
          end
        end
      end
      source = nil
      titlen = nil
    end
  end
end

--[[ Add links for staff where linking is possible. ]]
local function link_staff(frame, elems)
  -- nothing to do for now
end

--[[ Search the list of template arguments for arguments matching the patterns 
     "eng#", "kan#", or "rom#", where # is a number. Entries are sorted and
     grouped together into appropriate tables within an indexed table. ]]
local function scanForStanzas(frame)
  local idx = {}
  local eng = {}
  local kan = {}
  local rom = {}
  local cols = {['colnum'] = 0, ['single'] = false, ['full'] = false, ['eng'] = false, ['rom'] = false, ['kan'] = false}
  local stanzaList = {}

  for k, v in frame:argumentPairs() do
    if common.isset(v) then
      s,e,t = string.find(k, '^eng(%d+)$')
      if s ~= nil then
        local i = tonumber(t)
        idx[#idx+1] = i
        eng[i] = v
        cols['eng'] = true
      end
      s,e,t = string.find(k, '^kan(%d+)$')
      if s ~= nil then
        local i = tonumber(t)
        idx[#idx+1] = i
        kan[i] = v
        cols['kan'] = true
      end
      s,e,t = string.find(k, '^rom(%d+)$')
      if s ~= nil then
        local i = tonumber(t)
        idx[#idx+1] = i
        rom[i] = v
        cols['rom'] = true
      end
    end
  end
  
  -- sort indexes and remove repeating ones
  common.trunkTable(idx)

  for k,v in pairs(idx) do
    local tmp = {}
    tmp['eng'] = common.cv(common.isset(eng[v]), eng[v], '&nbsp;')
    tmp['rom'] = common.cv(common.isset(rom[v]), rom[v], '&nbsp;')
    tmp['kan'] = common.cv(common.isset(kan[v]), kan[v], '&nbsp;')
    stanzaList[#stanzaList+1] = tmp
  end

  if cols['eng'] then cols['colnum'] = cols['colnum'] + 1 end
  if cols['rom'] then cols['colnum'] = cols['colnum'] + 1 end
  if cols['kan'] then cols['colnum'] = cols['colnum'] + 1 end

  cols['single'] = (cols['colnum'] < 2)
  cols['full'] = (cols['colnum'] > 2)

  return stanzaList, cols
end

--[[ Creates the sort key used for DEFAULTSORT. A sort key is created by
     taking the English title, romanized title, or Japanese title (takes the 
     first it can find), setting it to lowercase, and capitalizing the first
     letter in the title. ]]
local function calculateDefaultSort(frame)
  local sortkey = ""

  if(common.isset(frame.args.titleen)) then
    sortkey = frame.args.titleen
  elseif(common.isset(frame.args.titlerom)) then
    sortkey = frame.args.titlerom
  elseif(common.isset(frame.args.titlejp)) then
    sortkey = frame.args.titlejp
  end

  return common.cv(common.isset(sortkey), "{{DEFAULTSORT:" .. string.gsub(string.lower(sortkey), '^%l', string.upper) .. "}}", "")
end


-- == PAGE SECTIONS ==
--[[ Generates the header, which contains the title and group name. ]]
local function lyricsHeaderRow(frame, cols)
  local hr = {}

  hr[#hr+1] = '<tr>\n<th class="incell_top" style="font-weight: normal;" colspan="' .. cols['colnum'] .. '">'

  -- display title
  if common.isset(frame.args.titleen) or common.isset(frame.args.titlejp) then
    hr[#hr+1] = "'''" .. common.cv(frame.args.titleen, frame.args.titleen, frame.args.titlejp) .. "'''"
  end
  if common.isset(frame.args.group) then hr[#hr+1] = " by " .. frame.args.group end

  hr[#hr+1] = '</th>\n</tr>'

  return table.concat(hr)
end

--[[ Generates the main info section, which contains song info as well as a
     gallery of albums the song is featured on. ]]
local function lyricsInfoRow(frame, cols, albumList, albumNames, info_list)
  local ir = {}
  local cc = {}
  
  local s,e,t

  ir[#ir+1] = '<tr>\n<td colspan="' .. cols['colnum'] .. '">'

  -- albums first, since it's a floating frame
  if #albumList > 0 then
    local alist = mw.clone(albumList)
    local lnk
    for k,v in pairs(albumNames) do
      if common.isset(alist[v.idx]) then
        local link = v.name
        if common.isset(v.link) then link = v.link .. '|' .. v.name end
        alist[v.idx] = alist[v.idx] .. '<br/>[[' .. link .. ']]'
      end
    end

    ir[#ir+1] = '\n<div class="lyrics_albums" style="float: right; clear: right;">'
    ir[#ir+1] = "Featured in: \n:"
    ir[#ir+1] = table.build_table(alist, {cols = 3, spacing = 4, style="text-align: center;"})
    ir[#ir+1] = '</div>'
  end

  -- yes, these newlines are required for the MW parser to correctly interpret start-of-line entities like * or # for lists
  if common.isset(frame.args.titleen) and common.isset(frame.args.titlejp) then
    s,e,t = string.find(frame.args.titlejp, '<span lang=') -- lang template output is already expanded
    if s ~= nil then -- lang template already included
        cc[#cc+1] = frame:preprocess("\n*'''" .. frame.args.titlejp .. "'''")
    else
        cc[#cc+1] = frame:preprocess("\n*'''{{lang|ja|" .. frame.args.titlejp .. "}}'''")
    end
  end
  if common.isset(frame.args.titlerom) then
    cc[#cc+1] = frame:preprocess("\n*''" .. frame.args.titlerom .. "''")
  end
  if common.isset(frame.args.length) then
    cc[#cc+1] = frame:preprocess("\n*length: " .. frame.args.length)
  end
  if common.isset(frame.args.arranger) then
    cc[#cc+1] = frame:preprocess("\n*arrangement: " .. frame.args.arranger)
  end
  if common.isset(frame.args.lyricist) then
    cc[#cc+1] = frame:preprocess("\n*lyrics: " .. frame.args.lyricist)
  end
  if common.isset(frame.args.vocalist) then
    cc[#cc+1] = frame:preprocess("\n*vocals: " .. frame.args.vocalist)
  end
  if common.isset(frame.args.other_staff) then
    cc[#cc+1] = frame:preprocess("\n" .. frame.args.other_staff)
  end
  if common.isset(frame.args.source) then
    cc[#cc+1] = frame:preprocess("\n" .. frame.args.source)
  end

  local elems = mw.text2.split(table.concat(cc), '\n') -- to make sure that it's a "one element per line" table

  -- process the elements
  link_staff(frame, elems)
  wrap_titles(frame, elems, info_list)

  ir[#ir+1] = table.concat(elems, '\n') -- assemble it back into a string

  -- just to make sure that everything stays in the cell
  ir[#ir+1] = '<div style="float: none; clear: both;></div>'

  ir[#ir+1] = "</td>\n</tr>"

  return table.concat(ir)
end

--[[ Generates the extra info section, which contains extra information as
     specified in the extra_info argument. ]]
local function lyricsExtraInfoRow(frame, cols)
  local eir = {}

  if common.isset(frame.args.extra_info) then
    eir[#eir+1] = '\n<tr>\n<th class="incell" colspan="' .. cols['colnum'] .. '">Additional Info</th>\n</tr>'
    eir[#eir+1] = '\n<tr>\n<td style="text-align: justify" colspan="' .. cols['colnum'] .. '">'

    eir[#eir+1] = frame:preprocess('\n' .. frame.args.extra_info)

    eir[#eir+1] = '</td>\n</tr>'
  end

  return table.concat(eir)
end

--[[ Generates stanzas upon stanzas of lyrics. ]]
local function lyricsStanzaRows(frame, cols, slist)
  local srow = {}
  local lang = 'ja'
  if common.isset(frame.args['lang']) then
    lang = common.cv(frame.args['lang'] == 'none', '', frame.args['lang'])
  end

  -- header row for stanzas
  srow[#srow+1] = '\n<tr>'

  if cols['kan'] then
    srow[#srow+1] = '\n<th class="incell"'
    -- if not cols['single'] then
      srow[#srow+1] = ' style="width: 30%"'
    -- end
    srow[#srow+1] = '>Original</th>'
  end

  if cols['rom'] then
    srow[#srow+1] = '\n<th class="incell"'
    -- if not cols['single'] then
      srow[#srow+1] = ' style="width: 35%"'
    -- end
    srow[#srow+1] = '>Romanized</th>'
  end

  if cols['eng'] then
    srow[#srow+1] = '\n<th class="incell"'
    -- if not cols['single'] then
      srow[#srow+1] = ' style="width: 35%"'
    -- end
    srow[#srow+1] = common.cv(cols['single'], '>English</th>', '>Translation</th>')
  end

  srow[#srow+1] = '\n</tr>'

  for i, stanza in pairs(slist) do
    srow[#srow+1] = '\n<!-- row:' .. i .. '-->\n<tr class="lyrics_row">'

    if cols['kan'] then
      srow[#srow+1] = '<td'
      if(common.isset(lang)) then srow[#srow+1] = ' lang="' .. lang .. '" xml:lang="' .. lang .. '"' end
      srow[#srow+1] = '>\n' .. stanza['kan'] .. '\n</td>'
    end

    if cols['rom'] then
      srow[#srow+1] = '<td>\n' .. stanza['rom'] .. '\n</td>'
    end

    if cols['eng'] then
      srow[#srow+1] = '<td>\n' .. stanza['eng'] .. '\n</td>'
    end

    srow[#srow+1] = '\n</tr>'
  end

  if common.isset(frame.args['lyrics_source']) then
    srow[#srow+1] = '<tr><td colspan="' .. cols['colnum'] .. '" style="text-align: center; font-size: 83%;">'
    srow[#srow+1] = "\n:''Lyrics source: " .. frame.args['lyrics_source'] .. "''"
    srow[#srow+1] = '</td></tr>'
  end

  return table.concat(srow)
end

--[[ Generates the notes sections, based on the notes argument. ]]
local function lyricsNotesRow(frame, cols)
  local nrow = {}

  if common.isset(frame.args['notes']) then
    nrow[#nrow+1] = '\n<tr><th class="incell" colspan="' .. cols['colnum'] .. '"></th></tr>'
    nrow[#nrow+1] = '\n<tr><td style="text-align: justify" colspan="' .. cols['colnum'] .. '">'
    nrow[#nrow+1] = frame:preprocess('\n' .. frame.args['notes'])
    nrow[#nrow+1] = '\n</td></tr>'
    cols['close_single'] = true
  end

  return table.concat(nrow)
end

--[[ Generates the shaded row at the bottom of the lyric sheet. ]]
local function lyricsClosingRow(frame, cols)
  local crow = {}

  crow[#crow+1] = '\n<tr>\n'

  if cols['close_single'] then
    crow[#crow+1] = '<td class="incell_bottom" colspan="' .. cols['colnum'] .. '"></td>'
  else
    if cols['single'] then
      crow[#crow+1] = '<td class="incell_bottom"></td>'
    else
      crow[#crow+1] = '<td class="incell_bottomleft"></td>'
      if cols['full'] then crow[#crow+1] = '<td class="incell"></td>' end
      crow[#crow+1] = '<td class="incell_bottomright"></td>'
    end
  end

  crow[#crow+1] = '\n</tr>'

  return table.concat(crow)
end

--[[ Categorizes the article based on its transcription/translation statuses. ]]
local function lyricsCategories(frame, cols, albumNames)
  local cats = {''} -- so the first category won't disappear
  local group

  local romanizable = not common.isset(frame.args['lang']) or (common.isInTable(common.romanizable, frame.args['lang']) ~= nil)
  local skip_rom = false
  if common.isset(frame.args['eng_only']) then skip_rom = true end
  if common.isset(frame.args['lang']) and not romanizable then skip_rom = true end

  cats[#cats+1] = frame:preprocess(calculateDefaultSort(frame))
  -- escaping the second '[', so parser won't assume it's a category
  cats[#cats+1] = '[[Category:Lyrics]]'

  if common.isset(frame.args['group_en']) then
      group = frame.args['group_en']
  elseif common.isset(frame.args['group']) then
      group = frame.args['group']
  end
  
  if common.isset(group) then
    local groups = {}
    local s1, s,e,t

    -- extract the group names from links
    s1 = 0
    while true do
      s,e,t = string.find(group, '%[%[.-|(.-)%]%]', s1 + 1)
      
      -- in case of simple links
      if s == nil then
        s,e,t = string.find(group, '%[%[(.-)%]%]', s1 + 1)
      end
      
      if s == nil then break end
      
      s1 = e
      groups[#groups+1] = t
    end

    -- if group is not a link, assume it's plain name
    if #groups < 1 then groups[1] = group end

    for k,v in pairs(groups) do
      cats[#cats+1] = '[[Category:' .. v .. ']]'
    end
  end

  -- for k,v in pairs(albumNames) do
    -- TODO: scan album pages for categories?
    -- cats[#cats+1] = '[[Category:' .. v['name'] .. ' (Album)]]'
  -- end

  if cols['eng'] then
    if common.isset(frame.args['eng_only']) then cats[#cats+1] = '[[Category:Lyrics in English]]' end
  else
    cats[#cats+1] = '[[Category:Untranslated/Lyrics]]'
  end

  if cols['kan'] then
    if romanizable then cats[#cats+1] = '[[Category:Lyrics in Kanji]]' end
  elseif not common.isset(frame.args['eng_only']) then
    cats[#cats+1] = '[[Category:Untranscribed/Lyrics]]'
  end

  if not cols['rom'] and not skip_rom then
    cats[#cats+1] = '[[Category:Unromanized/Lyrics]]'
  end

  -- MediaWiki would ignore the repeating categories anyway, but just to be safe...
  common.trunkTable(cats)

  return table.concat(cats, '\n')
end

local function replace_double_quote(info_list)
    for k, v in pairs(info_list) do
        k = string.gsub(k, "\\", "")
        info_list[k] = v
    end
end


-- == PUBLIC FUNCTIONS (see export table at bottom) ==
--[[ Assembles the whole template for a Lyrics page. ]]
local function outputLyrics(frame)
  local tpl = {}
  tpl[#tpl+1] = '<table class="template_lyrics outcell" style="width:98%;">'

  -- first prepare all the necessary data
  local albumList = scanForAlbums(frame)
  local albumNames = getAlbumNames(albumList)
  local slist,cols = scanForStanzas(frame)
  local info_list = load_info('Touhou_Wiki:SongSource.json')
  replace_double_quote(info_list)
  -- allows to force a single-cell closing row
  cols['close_single'] = false

  tpl[#tpl+1] = lyricsHeaderRow(frame, cols)
  tpl[#tpl+1] = lyricsInfoRow(frame, cols, albumList, albumNames, info_list)
  tpl[#tpl+1] = lyricsExtraInfoRow(frame, cols)
  tpl[#tpl+1] = lyricsStanzaRows(frame, cols, slist)
  tpl[#tpl+1] = lyricsNotesRow(frame, cols)
  tpl[#tpl+1] = lyricsClosingRow(frame, cols)

  tpl[#tpl+1] = '</table>'

  tpl[#tpl+1] = lyricsCategories(frame, cols, albumNames)

  return table.concat(tpl, '\n')
end

local function dump_info(frame)
  local ret = {''}
  lst = load_info(frame.args['name'])
  ret[#ret+1] = '<table class="wikitable" style="width: 100%;">'
  for k,v in pairs(lst) do
    ret[#ret+1] = '<tr><th colspan="2" style="text-align: center;">' .. k .. '</th></tr>'
    for k1,v1 in pairs(v) do
      ret[#ret+1] = '<tr><td>' .. k1 .. '</td><td>' .. v1 .. '</td></tr>'
    end
  end
  ret[#ret+1] = '</table>'
  return table.concat(ret, '\n')
end

local function uni_test(frame)
  local text = frame.args[1]
  local search = frame.args['search']
  local replace = frame.args['replace']
  if common.isset(text) then
    if common.isset(search) and common.isset(replace) then
      return utf8.replace_char(text, {[search] = replace})
    else
      local pts = {''}
      for s,e,v in utf8.iter(text) do
        pts[#pts+1] = '* (' .. s .. ',' .. e .. ") = '" .. v .. "'"
      end
      return table.concat(pts, "\n")
    end
  end
  return ''
end

local function uni_test2(frame)
  local text = frame.args[1]
  local search = frame.args['search']
  local replace = frame.args['replace']
  if common.isset(text) and common.isset(search) and common.isset(replace) then
    return utf8.replace(text, { {search, replace} })
  else
    return uni_test(frame)
  end
end

function lookup_song_info(frame)
  local lst = load_info(frame.args['src'])
  local k1 = normalize(frame.args['game'])
  local k2 = normalize(frame.args['name'])

  if common.isset(lst[k1]) and common.isset(lst[k1][k2]) then
    local trimmed_info = lst[k1][k2]
    local s,e,title,info = string.find(trimmed_info, '^(.*)%s*%[(.+)%]')
    
    if s ~= nil and common.isset(info) then
      local expinfo = mw.text2.split(info, ';')
      local charname = expinfo[#expinfo] -- last element of table
      if common.exists(charname) then
        expinfo[#expinfo] = '[[' .. charname .. ']]'
      end
      trimmed_info = title .. ' [' .. table.concat(expinfo, "; ") .. ']'
    end

    return trimmed_info
  end
  
  return '' -- so we won't get a 'nil' as result if can't find a match
end

return {
  ['outputLyrics'] = outputLyrics,
  ['outputLyricsFromTemplate'] = function(frame) return outputLyrics(frame:getParent()) end,
  ['dump_info'] = dump_info,
  ['unitest'] = uni_test,
  ['unitest2'] = uni_test2,
  ['lookup_song_info'] = lookup_song_info,
}