Module:User:Vriullop
Jump to navigation
Jump to search
- This module sandbox lacks a documentation subpage. You may create it.
- Useful links: subpage list • links • transclusions • testcases • user page • user talk page • userspace
This is a private module sandbox of Vriullop, for their own experimentation. Items in this module may be added and removed at Vriullop's discretion; do not rely on this module's stability.
-- Sandbox: Module:headword
local export = {}
local lemmas = {
["abbreviations"] = true,
["acronyms"] = true,
["adjectives"] = true,
["adnominals"] = true,
["adpositions"] = true,
["adverbs"] = true,
["affixes"] = true,
["articles"] = true,
["circumfixes"] = true,
["circumpositions"] = true,
["classifiers"] = true,
["cmavo"] = true,
["cmavo clusters"] = true,
["cmene"] = true,
["comparative adjectives"] = true,
["comparative adverbs"] = true,
["conjunctions"] = true,
["contractions"] = true,
["counters"] = true,
["determiners"] = true,
["diacritical marks"] = true,
["fu'ivla"] = true,
["gismu"] = true,
["Han tu"] = true,
["hanzi"] = true,
["hanja"] = true,
["idioms"] = true,
["infixes"] = true,
["interfixes"] = true,
["initialisms"] = true,
["interjections"] = true,
["kanji"] = true,
["letters"] = true,
["ligatures"] = true,
["lujvo"] = true,
["morphemes"] = true,
["non-constituents"] = true,
["nouns"] = true,
["numbers"] = true,
["numeral symbols"] = true,
["numerals"] = true,
["particles"] = true,
["phrases"] = true,
["postpositions"] = true,
["predicatives"] = true,
["prefixes"] = true,
["prepositions"] = true,
["prepositional phrases"] = true,
["preverbs"] = true,
["pronominal adverbs"] = true,
["pronouns"] = true,
["proverbs"] = true,
["proper nouns"] = true,
["punctuation marks"] = true,
["relatives"] = true,
["roots"] = true,
["suffixes"] = true,
["superlative adjectives"] = true,
["superlative adverbs"] = true,
["syllables"] = true,
["symbols"] = true,
["verbs"] = true,
}
local nonlemmas = {
["active participles"] = true,
["adjectival participles"] = true,
["adjective forms"] = true,
["adjective comparative forms"] = true,
["adjective superlative forms"] = true,
["adverb forms"] = true,
["adverb comparative forms"] = true,
["adverb superlative forms"] = true,
["adverbial participles"] = true,
["agent participles"] = true,
["article forms"] = true,
["circumfix forms"] = true,
["combined forms"] = true,
["determiner comparative forms"] = true,
["determiner forms"] = true,
["determiner superlative forms"] = true,
["diminutive nouns"] = true,
["gerunds"] = true,
["infinitives"] = true,
["jyutping"] = true,
["kanji readings"] = true,
["misspellings"] = true,
["negative participles"] = true,
["nominal participles"] = true,
["noun case forms"] = true,
["noun forms"] = true,
["noun plural forms"] = true,
["numeral forms"] = true,
["participles"] = true,
["participle forms"] = true,
["particle forms"] = true,
["passive participles"] = true,
["past active participles"] = true,
["past participles"] = true,
["past participle forms"] = true,
["past passive participles"] = true,
["perfect active participles"] = true,
["perfect passive participles"] = true,
["pinyin"] = true,
["plurals"] = true,
["postposition forms"] = true,
["prefix forms"] = true,
["preposition contractions"] = true,
["preposition forms"] = true,
["prepositional pronouns"] = true,
["present active participles"] = true,
["present participles"] = true,
["present passive participles"] = true,
["pronoun forms"] = true,
["proper noun forms"] = true,
["rafsi"] = true,
["romanizations"] = true,
["romaji"] = true,
["singulatives"] = true,
["suffix forms"] = true,
["verb forms"] = true,
["verbal nouns"] = true,
}
function export.full_headword(lang, sc, heads, tr, genders, inflections, categories, sort_key)
local tracking_categories = {}
heads = heads or {""}
categories = categories or {}
if #categories == 0 then
if lang:getCode() ~= "und" then
table.insert(tracking_categories, "head tracking/no pos")
require("Module:debug").track("head tracking/no pos")
require("Module:debug").track("head tracking/no pos/lang/" .. lang:getCode())
end
else
for _, cat in ipairs(categories) do
if mw.ustring.sub(cat, 1, mw.ustring.len(lang:getCanonicalName()) + 1) ~= lang:getCanonicalName() .. " " then
table.insert(tracking_categories, "head tracking/no lang category")
require("Module:debug").track("head tracking/no lang category")
require("Module:debug").track("head tracking/no lang category/lang/" .. lang:getCode())
end
end
if mw.ustring.sub(categories[1], 1, mw.ustring.len(lang:getCanonicalName()) + 1) == lang:getCanonicalName() .. " " then
local pos = mw.ustring.sub(categories[1], mw.ustring.len(lang:getCanonicalName()) + 2)
if lemmas[pos] or lemmas[pos:gsub("^reconstructed ", "")] then
table.insert(categories, 1, lang:getCanonicalName() .. " lemmas")
elseif nonlemmas[pos] or nonlemmas[pos:gsub("^reconstructed ", "")] or lemmas[pos:gsub("^mutated ", "")] or nonlemmas[pos:gsub("^mutated ", "")] then
table.insert(categories, lang:getCanonicalName() .. " non-lemma forms")
else
table.insert(tracking_categories, "head tracking/unrecognized pos")
require("Module:debug").track("head tracking/unrecognized pos")
require("Module:debug").track("head tracking/unrecognized pos/lang/" .. lang:getCode())
require("Module:debug").track("head tracking/unrecognized pos/pos/" .. pos)
end
end
end
-- Try to detect the script if it was not provided
if not sc then
sc = require("Module:scripts").findBestScript(heads[1] ~= "" and heads[1] or mw.title.getCurrentTitle().subpageText, lang)
end
-- Try to generate a transliteration if necessary
-- Generate it if the script is not Latn or similar, and if no transliteration was provided
if tr == "-" then
tr = nil
elseif not tr and not ((sc:getCode():find("Latn", nil, true)) or sc:getCode() == "Latinx") then
-- TODO: This should handle transliterations for each headword separately
tr = lang:transliterate(heads[1] ~= "" and require("Module:links").remove_links(heads[1]) or mw.title.getCurrentTitle().subpageText, sc)
end
-- Format and return all the gathered information
return
format_headword(lang, sc, heads) ..
format_transliteration(lang, sc, tr) ..
format_genders(lang, sc, genders) ..
format_inflections(lang, sc, inflections) ..
require("Module:utilities").format_categories(categories, lang, sort_key) ..
require("Module:utilities").format_categories(tracking_categories, lang, sort_key)
end
-- Format a headword
function format_headword(lang, sc, heads)
local m_links = require("Module:links")
local m_scriptutils = require("Module:script utilities")
if type(heads) ~= "table" then
heads = {heads}
end
-- Create a default headword.
local default_head = mw.title.getCurrentTitle().subpageText
-- Add links to multi-word page names when appropriate
if default_head:find(" ", nil, false) then
default_head = mw.text.split(default_head, " ", true)
default_head = "[[" .. table.concat(default_head, "]] [[") .. "]]"
end
if lang:getType() == "reconstructed" or (lang:getType() ~= "appendix-constructed" and mw.title.getCurrentTitle().nsText == "Appendix") then
default_head = "*" .. default_head
end
-- If the list is empty, fill it with the default headword
if #heads == 0 then
heads = {default_head}
end
-- Try to detect the script if it was not provided
if not sc then
sc = require("Module:scripts").findBestScript(heads[1], lang)
end
for i, head in ipairs(heads) do
-- If the head is the empty string "", then replace it with the default
if head == "" then
head = default_head
end
-- Apply processing to the headword, for formatting links and such
if head:find("[[", nil, true) then
head = m_links.language_link(head, nil, lang)
end
-- Add language and script wrapper
head = m_scriptutils.tag_text(head, lang, sc, "head")
heads[i] = head
end
return table.concat(heads, " ''or'' ")
end
-- Format transliteration
function format_transliteration(lang, sc, tr)
if tr then
local ret = " (<span class=\"tr\" lang=\"\">" .. tr .. "</span>)"
if lang and mw.title.new(lang:getCanonicalName() .. " transliteration", "Wiktionary").exists then
ret = " [[Wiktionary:" .. lang:getCanonicalName() .. " transliteration|•]]" .. ret
end
return ret
else
return ""
end
end
function format_genders(lang, sc, genders)
if genders and #genders > 0 then
local gen = require("Module:gender and number")
return " " .. gen.format_list(genders, lang)
else
return ""
end
end
-- Format the inflections following the headword
function format_inflections(lang, sc, inflections)
if inflections and #inflections > 0 then
-- Format each inflection individually
for key, infl in ipairs(inflections) do
-- If this inflection is a table, it contains parts
-- consisting of alternating label-term pairs. Format them too.
if type(infl) == "table" then
inflections[key] = format_parts(lang, sc, infl)
end
end
return " (" .. table.concat(inflections, ", ") .. ")"
else
return ""
end
end
function format_parts(lang, sc, parts)
local m_links = require("Module:links")
for key, part in ipairs(parts) do
if type(part) ~= "table" then
part = {term = part}
end
local qualifiers = ""
if part.qualifiers and #part.qualifiers > 0 then
qualifiers = mw.getCurrentFrame():expandTemplate{title = "qualifier", args = part.qualifiers} .. " "
end
-- Convert the term into a full link
-- Don't show a transliteration here, the consensus seems to be not to
-- show them in headword lines to avoid clutter.
local nolink = parts.nolink or (part.term == mw.title.getCurrentTitle().text)
part = m_links.full_link(not nolink and part.term or nil, part.alt or (nolink and part.term or nil), lang, part.sc or sc, "bold", part.id, {genders = part.genders, tr = "-"}, mw.title.getCurrentTitle().prefixedText)
if parts.accel then
part = "<span class=\"form-of lang-" .. lang:getCode() .. " " .. parts.accel .. "\">" .. part .. "</span>"
end
part = qualifiers .. part
parts[key] = part
end
return "''" .. parts.label .. "''" .. (#parts > 0 and " " .. table.concat(parts, " ''or'' ") or "")
end
return export