local p = {}
local data = mw.loadData( 'Module:Interlinear/data' )
local gloss_override = {} -- for custom gloss abbreviations
local getArgs = require('Module:Arguments').getArgs
local yesno = require('Module:Yesno')
local lang_data = mw.loadData( 'Module:Lang/data' )
local gcl = require('Module:Interlinear/sandbox2/gcl').gcl
--------------------------
-- Almost-global variables
--------------------------
local frame, glossing_type, displaying_messages, free_translation, msg, buffer
-------------------
-- General settings
-------------------
local conf = { --settings
WordSeparator = " \n\r\t", -- Don't replace with %s as this would include non-breaking spaces
GlossAbbrPattern = "^([Ø0-9A-Z]+)$", -- this isn't a full regex, but a Lua pattern
-- NOTE: The following characters must be formatted for use in a pattern set.
GlossAbbrBoundary = "-.,;:<>‹›/\\~+=%?%s%[%]()%_\127'",
GlossExcludeTable = {I = true,}, --strings not be treated as glossing abbreviations
GlossExcludePattern = '^[0-9][0-9]+$', -- excludes strings consisting entirely of digits
ErrorCategory = "[[Category:Pages with errors in interlinear text]]",
AmbiguousGlossCategory = "[[Category:Articles with ambiguous glossing abbreviations]]",
MessageGlossingError = "Error(s) in interlinear glossing",
LowerCaseGlosses = {
["1sg"] = true, ["2sg"] = true, ["3sg"] = true,
["1du"] = true, ["2du"] = true, ["3du"] = true,
["1pl"] = true, ["2pl"] = true, ["3pl"] = true,
["Fsg"] = true, ["Fpl"] = true,
["Msg"] = true, ["Mpl"] = true,
},
ErrorHelpLocation = "Template:Interlinear",
}
---------------------
-- CSS styles and classes
---------------------
conf.class = { --CSS classes
GlossAbbr = "gloss-abbr",
ErrorMessage = "error",
}
---------------------
-- Sundry small functions
---------------------
local function normalise(str)
return mw.ustring.gsub(str,"[" .. conf.WordSeparator .. "]+"," ")
end
local function tidyCss(str)
str = mw.ustring.gsub(str, '^[\"\']*(.-)[\"\']*$', "%1") -- trims quotation marks
if mw.ustring.sub(str, -1) ~= ";" then str = str .. ";" end -- appends ";" if missing
return str
end
local function highlight(text)
if text then
return '<b>' .. text .. '</b>'
else return "" end
end
local function tone_sup(str)
return mw.ustring.gsub(str, "([^%p%s0-9])([0-9]+)", "%1<sup>%2</sup>")
end
local function is_empty(str) -- returns "false" if its argument is a string containing chars other than spaces &c.
if not str then return true end
if mw.ustring.find(str, "[^" .. conf.WordSeparator .. "]")
then return false
else return true end
end
local function help_link (anchor)
if anchor then
return " ([[" .. conf.ErrorHelpLocation .. "#" .. anchor .. "|help]])"
else return "" end
end
---------------------
-- Apply other template
---------------------
local function expand_template(str, template, langarg)
return frame:expandTemplate{
title = template,
args = { langarg, str }
}
end
local function template_to_line(arr, n, template, args)
if template and arr then
for i = 0, n do
arr[i] = expand_template(arr[i], template, args)
end
end
return arr;
end
---------------------
-- The UserMessages object contains and processes error messages and warnings
---------------------
local UserMessages = {errors = {}, warnings = {}, gloss_messages = {}}
function UserMessages:add(msgtype, text, gloss)
if msgtype == "gloss_message" then
self.gloss_messages[gloss] = text
elseif msgtype == "warning" then
table.insert(self.warnings, text)
elseif msgtype == "non-repeating error" then
self.errors.nre = text
elseif msgtype == "ambiguous gloss" then
self.if_ambiguous_glosses = true
elseif msgtype == "error" then
table.insert(self.errors, text)
else return error("UserMessages:add(): unknown message type", 2)
end
end
function UserMessages:print_errors()
local out = ""
local namespace = mw.title.getCurrentTitle().namespace
if next(self.errors) or self.warnings[1] then
local err_span = mw.html.create("span")
err_span:addClass(conf.class.ErrorMessage)
for _,v in pairs(self.errors) do
err_span:wikitext(" " .. v .. ";") end
if namespace % 2 == 0 and namespace ~= 2 -- non-talk namespaces, excluding user pages; if modifying please update the description on the category page
then err_span:wikitext(conf.ErrorCategory)
end
out = tostring(err_span)
mw.addWarning(conf.MessageGlossingError)
end
if self.if_ambiguous_glosses then
if namespace == 0 -- article namespace
then out = out .. conf.AmbiguousGlossCategory -- this category will only track articles
end
end
return out
end
function UserMessages:print_warnings()
local out = ""
-- Messages and warnings get displayed only if the page is being viewed in "preview" mode:
if displaying_messages and (next(self.gloss_messages) or next(self.warnings)) then
local div = mw.html.create("div")
div:addClass("interlinear-preview-warning")
:wikitext("<i>This message box is shown only in preview:</i>")
:newline()
for _,v in ipairs(self.warnings) do
local p = div:tag("p")
p:addClass(conf.class.ErrorMessage)
p:wikitext(v)
end
if self.gloss_messages then
div:wikitext("<p>To change any of the following default expansions, see [[Template:Interlinear/doc#Custom abbreviations|the template's documentation]]:</p>")
end
for _,v in pairs(self.gloss_messages) do
div:wikitext("<p>" .. v .. "</p>")
end
out = out .. "\n\n" .. tostring(div)
end
return out
end
---------------------
-- find_gloss() parses a word into morphemes, and it calls format_gloss()
-- for anything that looks like a glossing abbreviation.
---------------------
local function set_glossing_type(glossing)
if glossing then
local GlossingType
glossing = mw.ustring.lower(mw.text.trim(glossing))
if mw.ustring.find(glossing, 'link') then
GlossingType = "wikilink"
elseif mw.ustring.find(glossing, 'label')
or mw.ustring.find(glossing, 'no link') then
GlossingType = 'label'
elseif mw.ustring.find(glossing, 'no abbr') then
GlossingType = "no abbr"
elseif yesno(glossing) == false then
GlossingType = nil
elseif yesno(glossing) then
GlossingType = 'label' --conf.GlossingType
else
msg:add('error', 'Glossing type "' .. glossing .. '" not recognised') end
return GlossingType
else
error("set_glossing_type: 'glossing' is nil or false", 2)
end
end
local function find_gloss(word)
local function scan_gloss(boundary, gloss_abbr) -- checks a morpheme if it is a gloss abbreviation
if (mw.ustring.match(gloss_abbr, conf.GlossAbbrPattern)
or conf.LowerCaseGlosses[gloss_abbr])
and not (conf.GlossExcludeTable[gloss_abbr]
or mw.ustring.match(gloss_abbr, conf.GlossExcludePattern))
then --gloss_abbr = '{{gcl|' .. gloss_abbr .. '}}'
local gcl_args = { gloss_abbr }
if glossing_type then
gcl_args = { gloss_abbr, glossing = glossing_type }
end
-- gloss_abbr = frame.gcl(gcl_args)
gloss_abbr = frame:expandTemplate{
title = 'gcl/sandbox', args = gcl_args
}
end
return boundary .. gloss_abbr
end
local word = mw.text.decode(word, true)
-- for the case of the English word "I", the 1SG pronoun
if word == "I" then return word end
local pattern = "([" .. conf.GlossAbbrBoundary .. "]?)([^" .. conf.GlossAbbrBoundary .. "]+)"
word = mw.ustring.gsub(word, pattern, scan_gloss) -- splits into morphemes
return word
end
---------------------
-- The main purpose of the bletcherous parse() is to split a line into words and and then for each eligible word
-- to call find_gloss(). The parser outputs the individual words (with any gloss abbreviation formatting applied).
-- The simple job of splitting at whitespaces has been made complicated by a) the fact that the input can contain
-- whitespaces inside the various html elements that are the result of the application of various formatting templates;
-- and b) the need to be able to recognise the output of the template that formats custom gloss abbreviations
-- (and hence skip passing it on to find_gloss). See talk for a suggestion about its future.
---------------------
local function parse(cline, i, tags_found,ifglossing)
local function issue_error(message, culprit)
UserMessages:add("error", message .. ": ''" .. mw.ustring.sub(cline.whole, 1, i-1) .. "'''" .. culprit .. "'''''")
end
if i > cline.length then return i end --this will only be triggered if the current line has less words than line 1
local next_step, j, _, chunk
local probe = mw.ustring.sub(cline.whole,i,i)
if mw.ustring.match(probe,"[" .. conf.WordSeparator .. "]") and tags_found == 0
then next_step = i-1
elseif probe == "[" then --Wikilink?
if mw.ustring.sub(cline.whole,i+1,i+1) == "[" then
_,j,chunk = mw.ustring.find(cline.whole,"(%[%[.-%]%])", i)
else chunk = "["; j = i end --not a wikilink then
buffer = buffer .. chunk
next_step = parse(cline, j+1,tags_found,ifglossing)
elseif probe == "{" and tags_found == 0 then --curly brackets enclose a sequence of words to be treated as a single unit
_,j,chunk = mw.ustring.find(cline.whole,"(.-)(})", i+1)
if not chunk then
issue_error("Unclosed curly bracket", "{")
chunk = highlight("{"); j = i
elseif ifglossing==true then
chunk = find_gloss(chunk)
elseif
cline.tone_sup then chunk = tone_sup(chunk)
end
buffer = buffer .. chunk
next_step = parse(cline, j+1,tags_found,ifglossing)
elseif probe == "<" then -- We've encountered an HTML tag. What do we do now?
local _,j,chunk = mw.ustring.find(cline.whole,"(<.->)",i)
if not chunk then
issue_error("Unclosed angle bracket", "<")
chunk = highlight("<"); j = i
elseif mw.ustring.sub(cline.whole,i,i+1) == "</" then -- It's a CLOSING tag
if cline.glossing
and ifglossing==false
and mw.ustring.match(chunk,"</abbr>")
then ifglossing=true end
tags_found = tags_found - 1
elseif not mw.ustring.match(chunk, "/>$") -- It's an OPENING tag, unless it opens a self-closing element (in which case the element is ignored)
then if ifglossing == true -- the following checks for the output of {{ggl}}:
and mw.ustring.find(chunk, conf.class.GlossAbbr, 1, true) -- it's important that the "find" function uses literal strings and not patterns
then ifglossing = false end
tags_found = tags_found + 1
end
buffer = buffer .. chunk
next_step = parse(cline, j+1,tags_found,ifglossing)
else -- No HTML tags, so we only need to find where the word ends
local _,k,chunk = mw.ustring.find(cline.whole,"(..-)([ <[])",i)
if k then --ordinary text
if ifglossing==true then
buffer = buffer .. find_gloss(chunk)
else
if cline.tone_sup then chunk = tone_sup(chunk) end
buffer = buffer .. chunk
end
next_step = parse(cline, k, tags_found, ifglossing)
else -- reached end of string
if ifglossing == true then
chunk = find_gloss(mw.ustring.sub(cline.whole,i))
else
chunk = mw.ustring.sub(cline.whole,i)
if cline.tone_sup then chunk = tone_sup(chunk) end
end
buffer = buffer .. chunk
next_step = cline.length
end
end
return next_step
end
---------------------
-- HTML stuff
---------------------
local function build_interlinear_html(args, number_of_words, line)
local interlinear_wrapper = mw.html.create("div")
interlinear_wrapper:addClass("interlinear")
-- right-to-left script
if yesno(args.rtl) == true then
interlinear_wrapper:addClass("right_to_left")
end
-- box
if yesno(args.box) == true then
interlinear_wrapper:addClass("box")
end
-- numbering and/or indent in the left margin
local number, indent = nil, nil
if args.number and args.number ~= ""
then number = args.number end
if args.indent and args.indent ~=""
then indent = args.indent end
if indent or number then
if not indent then indent = "4" end --default value
interlinear_wrapper:css("margin-left", indent .. 'em')
if number then
interlinear_wrapper:tag("div")
:addClass("number")
:wikitext(args.number)
end
end
--lines to display above the interlinear block
if args.top and args.top ~= "" then
interlinear_wrapper:tag("div")
:addClass("top")
:wikitext(args.top)
end
-- Producing the interlinear block
local blocks_wrapper = interlinear_wrapper:tag("div")
:addClass("block_wrapper")
-- non-standard spacing
local _spacing = tonumber(args.spacing)
if _spacing and _spacing <= 20 then
blocks_wrapper:css('column-gap', _spacing .. 'em')
end
for wi = 1, number_of_words do
local block = blocks_wrapper:tag("div")
:addClass("word_block")
for i,_ in ipairs (line) do
if line[i].whole ~= "" then -- skipping empty lines
local p = block:tag("p")
p:attr(line[i].attr)
if line[i].class then
p:addClass(line[i].class)
end
local _text = line[i].words[wi]
if _text == "" or _text == " "
then _text = " " end
-- <p> elements without content mess up the interlinear display
p:wikitext(_text)
end
end
end
--- "comments", added at the end of each line
if line.hasComments then
local comment_block = blocks_wrapper:tag("div")
:addClass("comment_block")
for i,_ in ipairs (line) do
local p = comment_block:tag("p")
if line[i].c then
p:wikitext(line[i].c)
else p:wikitext(" ")
end
end
end
--Add hidden lines containing the content of each line of interlinear text
-- this is for accessibility
for i,v in ipairs(line) do
local hidden_line = interlinear_wrapper:tag("p")
hidden_line:addClass("hidden_text")
:wikitext(v.whole)
end
-- Free translation
local ft_line = interlinear_wrapper:tag("p")
if free_translation and free_translation ~= "" then
ft_line:addClass("free_translation")
ft_line:wikitext(free_translation)
end
ft_line:node(msg:print_errors()) -- for error messages
-- bottom
if args.bottom and args.bottom ~= "" then
local bottom = interlinear_wrapper:tag('p')
:addClass('bottom')
:wikitext(args.bottom)
end
return interlinear_wrapper
end
--------------------
-- The following is the function called by Template:Interlinear.
-- It processes the template arguments, then calls parse() to split the input lines into words
-- and it then builds the output html.
--------------------
function p.interlinearise(f)
---------------------
-- Prepare arguments
---------------------
frame = f
local if_auto_translit = false
local args = getArgs(frame, { -- configuration for Module:Arguments
trim = true,
removeBlanks = false,
parentFirst = true,
-- wrappers = {
-- 'Template:Interlinear', 'Template:Fs interlinear',
-- 'Template:Interlinear/sandbox', 'Template:Fs interlinear/sandbox'
-- },
})
local template_name = frame:getParent():getTitle()
if template_name == 'Template:Fs interlinear/sandbox' then
args.italics1 = args.italics1 or "no"
args.italics2 = args.italics2 or "yes"
args.glossing3 = args.glossing3 or "yes"
if args.lang and not args.lang2 then args.lang2 = args.lang .."-Latn" end
if args.transl and not args.transl2 then args.transl2 = args.transl end
if_auto_translit = true
end
-- if args.wordseparator and (args.wordseparator ~= "") then
-- conf.WordSeparator = conf.WordSeparator .. args.wordseparator
-- end
if args.glossing then
glossing_type = set_glossing_type(args.glossing)
end
-- messages will be displayed only in preview mode
if frame:preprocess("{{REVISIONID}}") == "" then
if not args['display-messages']
or yesno(args['display-messages']) then
displaying_messages = true
end
end
msg = UserMessages
local line = {}
local function set_italics(n)
line[n].class = "italics"
line[n].tone_sup = true -- single digits are assumed to be tone markers and will hence be superscripted
if args['tone-superscripting'] and not yesno(args['tone-superscripting'])
then line[n].tone_sup = false end
end
local offset, last_line = 0, 0
for j,v in ipairs(args) do -- iterates over the unnamed parameters from the template
last_line = last_line +1
if is_empty(v) then
offset = offset + 1
else
local i = j - offset
line[i] = {}
-- normalise
if args.wordseparator and (args.wordseparator ~= "") then
v = mw.ustring.gsub(v,"([" .. args.wordseparator .. "]+)","%1 ")
end
v = mw.ustring.gsub(v,"[" .. conf.WordSeparator .. "]+"," ")
line[i].whole = v
line[i].length = mw.ustring.len(v)
local _c = args["c" .. i]
if _c and _c ~= "" then
line.hasComments = true
line[i].c = _c
end
---prepare style arguments----
line[i].class = ""
local _style = args["style" .. i]
if not _style then
_style = ""
else
_style = tidyCss(_style)
end
--line[i].attr holds the attributes for the <p> elements that enclose the words in line i
line[i].attr = { style = _style }
local _lang = args["lang" .. i]
if _lang and #_lang > 1 then
line[i].lang = _lang
else _lang = args.lang
if _lang and #_lang > 1 and i == 1 then -- if a lang= parameter is supplied, it's assumed to apply to line 1
line[i].lang = _lang
end
end
line[i].attr.lang = line[i].lang
if yesno(args["italics" .. i]) then
set_italics(i)
end
local _gloss = yesno(args["glossing" .. i]);
if _gloss then
line[i].glossing = _gloss
end
if yesno(args['ipa' .. i]) then
line[i].ipa = true
end
local _wrapper = args['wrapper' .. i]
if _wrapper and line[i].words then
line[i].wrapper = _wrapper
end
if yesno(args["smallcaps" .. i]) then
line[i].class = line[i].class .. " smallcaps"
line[i].glossing = false
end
local _transl = args["transl" .. i]
if _transl and #_transl > 1 then
_transl = mw.ustring.lower(_transl)
local _lookup = lang_data.translit_title_table[_transl]
if _lookup then
if _lang and _lookup[_lang] then
_transl = _lookup[_lang]
else _transl = _lookup.default
end
if _transl then
line[i].attr.title = _transl
end
else msg:add("error", "Transliteration scheme '" .. _transl .. "' not recognised")
end
end
local _class = args['class' .. i]
if _class then
line[i].class = line[i].class .. " " .. _class
end
if line[i].class == "" then
line[i].class = nil
end
end -- ends the first if-statement in the loop
end -- ends the FOR cycle
local line_count = #line
if line_count == 0 then
msg:add("error", template_name .. ": no lines supplied.")
return msg:print_errors()
elseif line_count == 1 then
msg:add("error", template_name .. ": only 1 line supplied.")
return msg:print_errors()
end
if line_count > 1 then
local _italics = args.italics
local n = tonumber(_italics)
if n and n > 0 then
set_italics(n)
elseif not (_italics and not yesno(_italics))
and not (args["italics1"] and not yesno(args["italics1"])) then
set_italics(1) -- by default, the first line will get italicised, unless italics=no or italics1=no
end
-- the last unnamed parameter is assumed to be the free translation:
free_translation = args[last_line]
if not is_empty(free_translation) then
line [line_count] = nil
end --... and is thus excluded from interlinearising
end
---------------------
-- Segment lines into words
---------------------
for i,v in ipairs(line) do
local wc, n = 1, 1
line[i].words = {}
while n <= line[i].length do
buffer = ""
n = parse(line[i], n, 0, true) + 2
line[i].words[wc] = buffer
wc = wc + 1
end
end
----Check for mismatches in number of words across lines----
local number_of_words, mismatch_found = 0, false
for i,v in ipairs(line) do -- find the maximum number of words in any line
local wc = #line[i].words
if wc ~= number_of_words then
if i ~= 1 and wc ~= 0 then
mismatch_found = true
end
if wc > number_of_words then
number_of_words = wc
end
end
end
----Deal with mismatches---
if mismatch_found then
local error_text = "Mismatch in the number of words between lines: "
for i,v in ipairs(line) do
local wc = #line[i].words
error_text = error_text .. wc .. " word(s) in line " .. i .. ", "
if wc ~= number_of_words then
for current_word = wc+1, number_of_words do
line[i].words[current_word] = " "
end
end
end
if string.sub(error_text, -2) == ", " then
error_text = string.sub(error_text, 1, #error_text - 2) .. " "
end
error_text = error_text .. help_link("mismatch")
UserMessages:add("error", error_text)
end
-- Wrap in first line of {{Fs interlinear}} in {{Script}}
--if template_name == 'Template:Fs interlinear/sandbox'
if args.script and line[1].words then
line[1].words = template_to_line(
line[1].words, number_of_words,
'Script', args.script)
if args.lang and line[2].words then
line[2].words = template_to_line(
line[2].words, number_of_words,
'transliteration', args.lang)
end
elseif args.lang and line[1].words then
line[1].words = template_to_line(
line[1].words, number_of_words,
'lang', args.lang)
end
---------------------
-- Build the HTML
---------------------
local divHtml = build_interlinear_html(args, number_of_words, line)
local temp_track = ""
if last_line == 2 then
temp_track = "[[Category:Pages with interlinear glosses using two unnamed parameters]]"
end
if last_line > 3 and template_name ~= 'Template:Fs interlinear' then
temp_track = "[[Category:Pages with interlinear glosses using more than three unnamed parameters]]"
end
return tostring(divHtml) .. temp_track .. msg:print_warnings()
end
return p
