Module:si-headword

පහත දැක්වෙන උපදෙස්, Module:si-headword/documentation හි පිහිටා ඇත. ^{[සංස්කරණය]}

ප්‍රයෝජනවත් සබැඳි: උප පිටු ලැයිස්තුව • සබැඳි • transclusions • testcases • sandbox

This module is used for Sinhalese headword-line templates. This module currently implements {{si-noun}}, {{si-proper noun}}, {{si-verb}}, {{si-adj}}, {{si-adv}}, {{si-intj}} and {{si-con}} (for conjunctions). See the documentation of those templates for more information. Other Sinhalese headword templates are in the process of being converted to use this module.

The module is always invoked the same way, by passing a single parameter to the "show" function. This parameter is the name of the part of speech, but in plural (examples given are for nouns, and for adjective forms respectively):

{{#invoke:si-headword|show|nouns}}
{{#invoke:si-headword|show|adjective forms}}

The template will, by default, accept the following parameters (specific parts of speech may accept or require others):

|head=, |head2=, |head3=, ...

Override the headword display; used to add links to individual words in a multiword term.

|id=

Sense ID for linking to this headword. See {{senseid}} for more information.

|nolink=1 or |nolinkhead=1

Don't link individual words in the headword of a multiword term. Useful for foreign or otherwise unanalyzable terms like a posteriori and yabba dabba doo where the expression functions as a whole in Sinhalese but the individual parts are not Sinhalese words.

|splithyph=1

Indicate that automatic splitting and linking of words should split on hyphens in multiword expressions with spaces in them, even if the hyphenated component would normally be linked as-is or with hyphens converted to spaces. See #Autosplitting below.

|nosplithyph=1

Indicate that automatic splitting and linking of words should not split on hyphens in multiword expressions with spaces in them, even if this would normally happen. See #Autosplitting below.

|hyphspace=1

Indicate that hyphenated components should be linked as a whole using the space-separated equivalent, even if this would not normally happen (i.e. because the space-separated equivalent is not defined as an Sinhalese term). See #Autosplitting below.

|nosuffix=1

Prevent terms beginning with a hyphen from being interpreted as suffixes. See #Suffix handling below.

|nomultiwordcat=1

Prevent multiword terms (those with spaces or with hyphens in the middle) from being added to Category:Sinhalese multiword terms.

|pagename=

Override the page name used to compute default values of various sorts. Useful when testing, for documentation pages, etc.

|sort=

Sort key. Rarely needs to be specified, as it is normally automatically generated.

Autosplitting

All templates using this module use an intelligent autosplitting algorithm to link portions of multipart and multiword expressions, as follows:

• If there are spaces in the term but no apostrophes or hyphens, the module will automatically split and link distinct space-separated words, similarly to {{head}}; hence, absent without leave will be linked as [[absent]] [[without]] [[leave]].
• If there are spaces and apostrophes but no hyphens, the module will likewise split and link distinct space-separated words, but may also split up words with apostrophes in them. Specifically:
  1. If a word ends in 's, the part before the 's will be linked as a word, and the 's will be linked separately to -'s, on the assumption that the 's is functioning as a possessive. For example, Abel's impossibility theorem will be linked as [[Abel]][[-'s|'s]] [[impossibility]] [[theorem]]. (An exception is made for one's, someone's, he's, she's and it's, which are linked as-is without splitting.)
  2. If a word ends in ', the apostrophe will be linked to -' (on the assumption that the ' is functioning as a plural possessive, similarly to above), and the part before will be separately linked. If the part before ends in an s, the module converts it to its singular equivalent and looks that up to see if it exists and has a definition as an Sinhalese term. If so, the term is linked to the singular form; otherwise, it is linked to the plural form. (Converting to the singular means that -ies becomes -y; -es is dropped after sh, ch and x; and otherwise s is dropped.) For example, flies' graveyard will be linked as [[fly|flies]][[-'|']] [[graveyard]] because fly exists as an Sinhalese term, but Achilles' heel will be linked as [[Achilles]][[-'|']] [[heel]] because Achille does not exist as an Sinhalese term.
  3. All other terms containing apostrophes are linked unsplit.
• If there are hyphens in the term but no spaces or apostrophes, the hyphenated components will be linked individually. For example, beggar-thy-neighbor will be linked as [[beggar]]-[[thy]]-[[neighbor]].
  • An exception to this occurs with certain recognized prefixes, which are linked with the hyphen included in the prefix. For example, Afro-American is linked as [[Afro-]][[American]] and co-occurrence is linked as [[co-]][[occurrence]], because Afro- and co- are in the list of recognized prefixes. (For the full list, see below.)
• If there are hyphens and apostrophes but no spaces, the effect is similar to the situation with spaces and apostrophes. For example, beggar's-lice is linked as [[beggar]][[-'s|'s]]-[[lice]].
• If there are both hyphens and spaces, the space-separated components that do not have hyphens will be linked separately, as above. Any hyphen-separated components may be linked in one of three ways:
  1. If |hyphspace=1 is specified or the hyphen-separated component exists as an Sinhalese term when the hyphens are converted to spaces, it will be linked to that term. For example, closed-circuit television will be linked as [[closed circuit|closed-circuit]] [[television]] because closed circuit exists as an Sinhalese term. (In this case, closed-circuit also exists but is approximately a soft redirect to closed circuit, as is often the case with such attributive compounds. This is why we prefer the space-separated variant.)
  2. If |nosplithyph=1 is specified or the hyphen-separated component exists as an Sinhalese term in its unmodified form but not when the hyphens are converted to spaces, it will be linked as an unmodified whole. For example, coin-operated laundry will be linked as [[coin-operated]] [[laundry]] because coin-operated exists as an Sinhalese term but coin operated does not. (An example that requires |nosplithyph=1 is close-up lens, where the default algorithm would incorrectly link the first component to close up. Here, close up [a verb] and close-up [an adjective] both exist but refer to different things.)
  3. If |splithyph=1 is specified or the hyphen-separated component does not exist as an Sinhalese term (either unmodified or when the hyphens are converted to spaces), each hyphenated component is linked separately. Examples where this happens are adult-onset diabetes (linked as [[adult]]-[[onset]] [[diabetes]]) and Bombieri-Friedlander-Iwaniec theorem linked as [[Bombieri]]-[[Friedlander]]-[[Iwaniec]] [[theorem]]). Note that when separately linking hyphenated components, prefixes are recognized and handled specially, as documented below.

Special prefix handling

As described above, when splitting hyphenated components, if a component is not the last component and looks like one of the following prefixes, the following hyphen will be included inside of the link.

Specially handled prefixes

• acro
• acousto
• Afro
• agro
• anarcho
• angio
• Anglo
• ante
• anti
• arch
• auto
• bi
• bio
• cis
• co
• cryo
• crypto
• de
• demi
• eco
• electro
• Euro
• ex
• Greco
• hemi
• hydro
• hyper
• hypo
• infra
• Indo
• inter
• intra
• Judeo
• macro
• meta
• micro
• mini
• multi
• neo
• neuro
• non
• para
• peri
• post
• pre
• pro
• proto
• pseudo
• re
• semi
• sub
• super
• trans
• un
• vice

Suffix handling

If the term begins with a hyphen (-), it is assumed to be a suffix rather than a base form, and is categorized into Category:Sinhalese suffixes and Category:Sinhalese POS-forming suffixes rather than Category:Sinhalese POSs (e.g. Category:Sinhalese noun-forming suffixes rather than Category:Sinhalese nouns). This can be overridden using |nosuffix=1. (An example where this is necessary is -ussification, which refers to a linguistic process of blending words with the suffix -ussy but is not itself a suffix.)

Link modifications

The default behavior described above under #Autosplitting is sufficient in most circumstances, but some multiword terms need special linking behavior to handle things like inflected terms (e.g. those ending in -ing or -s), capitalized terms, multiword subexpressions, etc. One way to handle that is to use |head= and spell out the entire headword, appropriately linked, effectively ignoring the default linking behavior. But this can be awkward for long multiword terms. For cases like this, a shortcut syntax is provided to apply link modifications on top of the autolinked term. To enable this, put a tilde (~) at the beginning of the value specified to |head=, followed by the changes to individual words.

For example, for the term acute necrotising ulcerative gingivitis, we would like to link necrotising to necrotise. This can be done as follows:

• {{si-noun|head=~necrotising:necrotise}}

or more compactely as

• {{si-noun|head=~necrotis[ing:e]}}

This is equivalent to writing {{si-noun|head=[[acute]] [[necrotise|necrotising]] [[ulcerative]] [[gingivitis]]}}, but shorter. In general, syntax of the form prefix[from:to] is equivalent to writing prefixfrom:prefixto, and says to replace prefixfrom with prefixto in the default output produced by the #Autosplitting mechanism described above.

The same syntax works on the beginning of a word, which is especially useful when linking to the lowercase equivalent of a capitalized term. For example, for admiral of the Swiss Navy, use the following to link Navy to navy:

• {{si-noun|head=~[N:n]avy}}

This is equivalent to writing {{si-noun|head=[[admiral]] [[of]] [[the]] [[Swiss]] [[navy|Navy]]}} but shorter.

Modifications need to match full words, but can be applied to multiple words. A ~ on the right-hand side is a shortcut that stands for the left-hand side, which is especially useful when multiple words are given on the left-hand side, and causes the words to be linked together. For example, for acute respiratory distress syndrome, to link respiratory distress as a single entity, use the following:

• {{si-noun|head=~respiratory distress:~}}

which is equivalent to {{si-noun|head=[[acute]] [[respiratory distress]] [[syndrome]]}}. The right-hand side need not consist solely of a tilde, but can contain other surrounding text. For example, for Charlie Brown Christmas tree, use the following to link to the Wikipedia entry for Charlie Brown:

• {{si-noun|head=~Charlie Brown:w:~}}

This is equivalent to writing {{si-noun|head=[[w:Charlie Brown|Charlie Brown]] [[Christmas]] [[tree]]}}.

Multiple modifications can be specified, separated by a semicolon (optionally with surrounding spaces). For example, for Admiral of the Fleet, use:

• {{si-noun|head=~[A:a]dmiral; [F:f]leet}}

This is equivalent to writing {{si-noun|head=[[admiral|Admiral]] [[of]] [[the]] [[fleet|Fleet]]}}.

---

local export = {}
local pos_functions = {}

local lang = require("Module:languages").getByCode("si")
local PAGENAME = mw.title.getCurrentTitle().text

local function glossary_link(entry, text)
	text = text or entry
	return "[[Appendix:Glossary#" .. entry .. "|" .. text .. "]]"
end

local function track(page)
	require("Module:debug").track("si-headword/" .. page)
	return true
end

-- The main entry point.
-- This is the only function that can be invoked from a template.
function export.show(frame)
	
	local poscat = frame.args[1] or error("Part of speech has not been specified. Please pass parameter 1 to the module invocation.")
	
	if poscat == "verbs" then
		poscat = "ක්‍රියා පද"
	elseif poscat == "nouns" then
		poscat = "නාම පද"
	end
	
	local params = {
		["head"] = {list = true},
		["suff"] = {type = "boolean"},
		["sort"] = {},
		["pagename"] = {}, -- for testing
	}
	
	local pos_data = pos_functions[poscat]
	if pos_data then
		for key, val in pairs(pos_data.params) do
			params[key] = val
		end
	end
	
	local args, unknown_args = require("Module:parameters").process(frame:getParent().args, params, pos_data.return_unknown)
	
	if unknown_args and next(unknown_args) then
		track("unknown args")
		track("unknown args/POS/" .. tostring(poscat))
		for parameter, value in pairs(unknown_args) do
			track("unknown args/param/" .. tostring(parameter))
			mw.log("unknown parameter in [[Module:headword]]: |" .. tostring(parameter) .. "=" .. tostring(value))
		end
	end
	
	local data = {lang = lang, pos_category = poscat, categories = {}, heads = args["head"], inflections = {}}

	if #data.heads == 0 and args.pagename then
		table.insert(data.heads, args.pagename)
	end

	if args["suff"] then
		data.pos_category = "suffixes"
		
		if poscat == "adjectives" or poscat == "adverbs" or poscat == "නාම පද" or poscat == "ක්‍රියා පද" then
			table.insert(data.categories, ("%s %s-forming suffixes")
				:format(lang:getCanonicalName(), poscat:gsub("s$", "")))
		else
			error("No category exists for suffixes forming " .. poscat .. ".")
		end
	end
	
	if pos_data then
		pos_data.func(args, data)
	end
	
	local extra_categories = {}
	local pagename = args.pagename or PAGENAME
	if pagename:find("[Qq][^Uu]") or pagename:find("[Qq]$") then
		table.insert(data.categories, lang:getCanonicalName() .. " words containing Q not followed by U")
	end
	if pagename:find("([A-Za-z])%1%1") then
		table.insert(data.categories, lang:getCanonicalName() .. " words containing three consecutive instances of the same letter")
	end
	if pagename:find("([A-Za-z])%1%1%1") then
		table.insert(data.categories, lang:getCanonicalName() .. " words containing four consecutive instances of the same letter")
	end
	-- mw.ustring.toNFD performs decomposition, so letters that decompose
	-- to an ASCII vowel and a diacritic, such as é, are counted as vowels and
	-- do not need to be included in the pattern.
	if not mw.ustring.find(mw.ustring.lower(mw.ustring.toNFD(pagename)), "[aeiouyæœø]") then
		table.insert(data.categories, lang:getCanonicalName() .. " words without vowels")
	end
	if pagename:find("yre$") then
		table.insert(data.categories, lang:getCanonicalName() .. ' words ending in "-yre"')
	end
	if not pagename:find(" ") and mw.ustring.len(pagename) >= 25 then
		table.insert(extra_categories, "Long " .. lang:getCanonicalName() .. ' words')
	end
	if pagename:find("^[^aeiou ]*a[^aeiou ]*e[^aeiou ]*i[^aeiou ]*o[^aeiou ]*u[^aeiou ]*$") then
		table.insert(data.categories, lang:getCanonicalName() .. ' words that use all vowels in alphabetical order')
	end
	data.sort_key = args.sort
	return require("Module:headword").full_headword(data)
		.. (#extra_categories > 0
			and require("Module:utilities").format_categories(extra_categories, lang, args.sort)
			or "")
end

-- This function does the common work between adjectives and adverbs
local function make_comparatives(params, data, pagename)
	local comp_parts = {label = glossary_link("comparative"), accel = {form = "comparative"}}
	local sup_parts = {label = glossary_link("superlative"), accel = {form = "superlative"}}
	
	if #params == 0 then
		table.insert(params, {"more"})
	end
	
	-- To form the stem, replace -(e)y with -i and remove a final -e.
	local stem = pagename:gsub("([^aeiou])e?y$", "%1i"):gsub("e$", "")
	
	-- Go over each parameter given and create a comparative and superlative form
	for i, val in ipairs(params) do
		local comp = val[1]
		local comp_qual = val[2]
		local sup = val[3]
		local sup_qual = val[4]
		local comp_part, sup_part
		
		if comp == "more" and pagename ~= "many" and pagename ~= "much" then
			comp_part = "[[more]] " .. pagename
			sup_part = "[[most]] " .. pagename
		elseif comp == "further" and pagename ~= "far" then
			comp_part = "[[further]] " .. pagename
			sup_part = "[[furthest]] " .. pagename
		elseif comp == "er" then
			comp_part = stem .. "er"
			sup_part = stem .. "est"
		elseif comp == "-" or sup == "-" then
			-- Allowing '-' makes it more flexible to not have some forms
			if comp ~= "-" then
				comp_part = comp
			end
			if sup ~= "-" then
				sup_part = sup
			end
		else
			-- If the full comparative was given, but no superlative, then
			-- create it by replacing the ending -er with -est.
			if not sup then
				if comp:find("er$") then
					sup = comp:gsub("er$", "est")
				else
					error("The superlative of \"" .. comp .. "\" cannot be generated automatically. Please provide it with the \"sup" .. (i == 1 and "" or i) .. "=\" parameter.")
				end
			end
			
			comp_part = comp
			sup_part = sup
		end

		if comp_part then
			table.insert(comp_parts, {term = comp_part, qualifiers = {comp_qual}})
		end
		if sup_part then
			table.insert(sup_parts, {term = sup_part, qualifiers = {sup_qual}})
		end
	end
	
	table.insert(data.inflections, comp_parts)
	table.insert(data.inflections, sup_parts)
end

pos_functions["adjectives"] = {
	params = {
		[1] = {list = true, allow_holes = true},
		["comp_qual"] = {list = "comp=_qual", allow_holes = true},
		["sup"] = {list = true, allow_holes = true},
		["sup_qual"] = {list = "sup=_qual", allow_holes = true},
		},
	func = function(args, data)
		local pagename = args.pagename or PAGENAME
		local shift = 0
		local is_not_comparable = false
		local is_comparative_only = false
		
		-- If the first parameter is ?, then don't show anything, just return.
		if args[1][1] == "?" then
			return
		-- If the first parameter is -, then move all parameters up one position.
		elseif args[1][1] == "-" then
			shift = 1
			is_not_comparable = true
		-- If the only argument is +, then remember this and clear parameters
		elseif args[1][1] == "+" and args[1].maxindex == 1 then
			shift = 1
			is_comparative_only = true
		end
		
		-- Gather all the comparative and superlative parameters.
		local params = {}
		
		for i = 1, args[1].maxindex - shift do
			local comp = args[1][i + shift]
			local comp_qual = args["comp_qual"][i + shift]
			local sup = args["sup"][i]
			local sup_qual = args["sup_qual"][i + shift]
			
			if comp or sup then
				table.insert(params, {comp, comp_qual, sup, sup_qual})
			end
		end
		
		if shift == 1 then
			-- If the first parameter is "-" but there are no parameters,
			-- then show "not comparable" only and return.
			-- If there are parameters, then show "not generally comparable"
			-- before the forms.
			if #params == 0 then
				if is_not_comparable then
					table.insert(data.inflections, {label = "not " .. glossary_link("comparable")})
					table.insert(data.categories, lang:getCanonicalName() .. " uncomparable adjectives")
					return
				end
				if is_comparative_only then
					table.insert(data.inflections, {label = glossary_link("comparative") .. " form only"})
					table.insert(data.categories, lang:getCanonicalName() .. " comparative-only adjectives")
					return
				end
			else
				table.insert(data.inflections, {label = "not generally " .. glossary_link("comparable")})
			end
		end
		
		-- Process the parameters
		make_comparatives(params, data, pagename)
	end
}

pos_functions["adverbs"] = {
	params = {
		[1] = {list = true, allow_holes = true},
		["comp_qual"] = {list = "comp=_qual", allow_holes = true},
		["sup"] = {list = true, allow_holes = true},
		["sup_qual"] = {list = "sup=_qual", allow_holes = true},
		},
	func = function(args, data)
		local pagename = args.pagename or PAGENAME
		local shift = 0
		
		-- If the first parameter is ?, then don't show anything, just return.
		if args[1][1] == "?" then
			return
		-- If the first parameter is -, then move all parameters up one position.
		elseif args[1][1] == "-" then
			shift = 1
		end
		
		-- Gather all the comparative and superlative parameters.
		local params = {}
		
		for i = 1, args[1].maxindex - shift do
			local comp = args[1][i + shift]
			local comp_qual = args["comp_qual"][i + shift]
			local sup = args["sup"][i]
			local sup_qual = args["sup_qual"][i + shift]
			
			if comp or sup then
				table.insert(params, {comp, comp_qual, sup, sup_qual})
			end
		end
		
		if shift == 1 then
			-- If the first parameter is "-" but there are no parameters,
			-- then show "not comparable" only and return. If there are parameters,
			-- then show "not generally comparable" before the forms.
			if #params == 0 then
				table.insert(data.inflections, {label = "not " .. glossary_link("comparable")})
				table.insert(data.categories, lang:getCanonicalName() .. " uncomparable adverbs")
				return
			else
				table.insert(data.inflections, {label = "not generally " .. glossary_link("comparable")})
			end
		end
		
		-- Process the parameters
		make_comparatives(params, data, pagename)
	end
}

pos_functions["conjunctions"] = {
	params = {
		[1] = { alias_of = "head" },
	},
	return_unknown = true,
	func = function (args, data)
		
	end,
}

pos_functions["interjections"] = {
	params = {
		[1] = { alias_of = "head" },
	},
	return_unknown = true,
	func = function (args, data)
		
	end,
}

local function default_plural(noun)
	local new_pl
	if noun:find("[sxz]$") or noun:find("[cs]h$") then
		new_pl = noun .. "es"
	elseif noun:find("[^aeiou]y$") then
		new_pl = noun:gsub("y$", "i") .. "es"
	else
		new_pl = noun .. "s"
	end
	return new_pl
end

local function canonicalize_plural(pl, stem, pagename)
	if pl == "s" then
		return stem .. "s"
	elseif pl == "es" then
		return stem .. "es"
	elseif pl == "+" then
		return default_plural(pagename)
	else
		return nil
	end
end

pos_functions["නාම පද"] = {
	params = {
		[1] = {list = true, allow_holes = true},
		["pl=qual"] = { list = true, allow_holes = true },
		},
	func = function(args, data)
		local pagename = args.pagename or PAGENAME
		-- Gather all the plural parameters from the numbered parameters.
		local plurals = {}
		
		for i = 1, args[1].maxindex do
			local pl = args[1][i]
			
			if pl then
				local qual = args["plqual"][i]
				
				if qual then
					table.insert(plurals, {term = pl, qualifiers = {qual}})
				else
					table.insert(plurals, pl)
				end
			end
		end
		
		local need_default_plural = true
		if plurals[1] == "-" then
			-- Uncountable noun; may occasionally have a plural
			table.remove(plurals, 1)  -- Remove the "-"
			table.insert(data.categories, lang:getCanonicalName() .. " uncountable nouns")
			
			-- If plural forms were given explicitly, then show "usually"
			if #plurals > 0 then
				table.insert(data.inflections, {label = "usually " .. glossary_link("uncountable")})
				table.insert(data.categories, lang:getCanonicalName() .. " countable nouns")
			else
				table.insert(data.inflections, {label = glossary_link("uncountable")})
			end
			need_default_plural = false
		elseif plurals[1] == "~" then
			-- Mixed countable/uncountable noun, always has a plural
			table.remove(plurals, 1)  -- Remove the "~"
			table.insert(data.inflections, {label = glossary_link("countable") .. " and " .. glossary_link("uncountable")})
			table.insert(data.categories, lang:getCanonicalName() .. " uncountable nouns")
			table.insert(data.categories, lang:getCanonicalName() .. " countable nouns")
			
			-- If no plural was given, add a default one now
			if #plurals == 0 then
				plurals = {default_plural(pagename)}
			end
		else
			-- The default (countable noun), always has a plural
			table.insert(data.categories, lang:getCanonicalName() .. " countable nouns")
		end
		-- Plural is unknown
		if plurals[1] == "?" then
			table.remove(plurals, 1)  -- Remove the "?"
			-- Not desired; see [[Wiktionary:Tea_room/2021/August#"Plural unknown or uncertain"]]
			-- table.insert(data.inflections, {label = "plural unknown or uncertain"})
			table.insert(data.categories, lang:getCanonicalName() .. " nouns with unknown or uncertain plurals")
			if #plurals > 0 then
				error("Can't specify explicit plurals along with '?' for unknown/uncertain plural")
			end
			return
		end
		-- Plural is not attested
		if plurals[1] == "!" then
			table.remove(plurals, 1)  -- Remove the "!"
			table.insert(data.inflections, {label = "plural not attested"})
			table.insert(data.categories, lang:getCanonicalName() .. " nouns with unattested plurals")
			if #plurals > 0 then
				error("Can't specify explicit plurals along with '!' for unattested plural")
			end
			return
		end

		-- If no plural was given, maybe add a default one, otherwise (when "-" was given) return
		if #plurals == 0 then
			if need_default_plural then
				plurals = {default_plural(pagename)}
			else
				return
			end
		end
		
		-- There are plural forms to show, so show them
		local pl_parts = {label = "බහුවචනය", accel = {form = "p"}}
		
		local function check_ies(pl, stem)
			local newplural, nummatches = stem:gsub("([^aeiou])y$","%1ies")
			return nummatches > 0 and pl == newplural
		end
		local stem = pagename
		local irregular = false
		for i, pl in ipairs(plurals) do
			local canon_pl = canonicalize_plural(pl, stem, pagename)
			if canon_pl then
				table.insert(pl_parts, canon_pl)
			elseif type(pl) == "table" then
				canon_pl = canonicalize_plural(pl.term, stem, pagename)
				if canon_pl then
					table.insert(pl_parts, {term=canon_pl, qualifiers=pl.qualifiers})
				end
			end
			if not canon_pl then
				table.insert(pl_parts, pl)
				if type(pl) == "table" then
					pl = pl.term
				end
				if not stem:find(" ") and not (pl == stem .. "s" or pl == stem .. "es" or check_ies(pl, stem)) then
					irregular = true
					if pl == stem then
						table.insert(data.categories, lang:getCanonicalName() .. " indeclinable nouns")
					end
				end
			end
		end
		if irregular then
			table.insert(data.categories, lang:getCanonicalName() .. " nouns with irregular plurals")
		end
		
		table.insert(data.inflections, pl_parts)
	end
}

pos_functions["proper nouns"] = {
	params = {
		[1] = {list = true},
		},
	func = function(args, data)
		local pagename = args.pagename or PAGENAME
		local plurals = args[1]
		
		-- Decide what to do next...
		local mode = nil
		
		if plurals[1] == "?" or plurals[1] == "!" or plurals[1] == "-" or plurals[1] == "~" then
			mode = plurals[1]
			table.remove(plurals, 1)  -- Remove the mode parameter
		end
		
		-- Plural is unknown
		if mode == "?" then
			table.insert(data.categories, lang:getCanonicalName() .. " proper nouns with unknown or uncertain plurals")
			return
		-- Plural is not attested
		elseif mode == "!" then
			table.insert(data.inflections, {label = "plural not attested"})
			table.insert(data.categories, lang:getCanonicalName() .. " proper nouns with unattested plurals")
			return
		-- Uncountable noun; may occasionally have a plural
		elseif mode == "-" then
			-- If plural forms were given explicitly, then show "usually"
			if #plurals > 0 then
				table.insert(data.inflections, {label = "usually " .. glossary_link("uncountable")})
				table.insert(data.categories, lang:getCanonicalName() .. " countable proper nouns")
			else
				table.insert(data.inflections, {label = glossary_link("uncountable")})
			end
		-- Mixed countable/uncountable noun, always has a plural
		elseif mode == "~" then
			table.insert(data.inflections, {label = glossary_link("countable") .. " and " .. glossary_link("uncountable")})
			table.insert(data.categories, lang:getCanonicalName() .. " countable proper nouns")
			
			-- If no plural was given, add a default one now
			if #plurals == 0 then
				plurals = {"s"}
			end
		elseif #plurals > 0 then
			table.insert(data.categories, lang:getCanonicalName() .. " countable proper nouns")
		end
		
		-- If there are no plurals to show, return now
		if #plurals == 0 then
			return
		end
		
		-- There are plural forms to show, so show them
		local pl_parts = {label = "plural", accel = {form = "p"}}
		
		local stem = pagename
		
		for i, pl in ipairs(plurals) do
			if pl == "s" then
				table.insert(pl_parts, stem .. "s")
			elseif pl == "es" then
				table.insert(pl_parts, stem .. "es")
			else
				table.insert(pl_parts, pl)
			end
	
		end
		
		table.insert(data.inflections, pl_parts)
	end
}

local function base_default_verb_forms(verb)
	local s_form = default_plural(verb)
	local ing_form, ed_form
	local vowel = "aeiouáéíóúàèìòùâêîôûäëïöüæœø"
	local ulvowel = vowel .. "AEIOUÁÉÍÓÚÀÈÌÒÙÂÊÎÔÛÄËÏÖÜÆŒØ"

	-- (1) Check for C*VC verbs.
	--
	-- flip -> flipping/flipped, strum -> strumming/strummed, nag -> nagging/nagged, etc.
	-- Do not include words with final -y, e.g. 'stay' (staying/stayed), 'toy' (toying/toyed),
	-- or with final -w, e.g. 'flow' (flowing/flowed), or with final -h, e.g. 'ah' (ahing/ahed),
	-- or with final -x, e.g. 'box' (boxing/boxed), or ending in an uppercase consonant,
	-- e.g. 'XOR' (XORing/XORed), 'OK' (OKing/OKed). Check specially for initial y- as a consonant,
	-- e.g. 'yip' (yipping/yipped), otherwise treat y as a vowel, so we don't trigger on 'hyphen'
	-- but do trigger on 'gyp'.
	local last_cons = mw.ustring.match(verb, "^[Yy][" .. vowel .. "y]([^A-Z" .. vowel .. "ywxh])$")
	if not last_cons then
		last_cons = mw.ustring.match(verb, "^[^" .. ulvowel .. "yY]*[" .. ulvowel .. "yY]([^A-Z" .. vowel .. "ywxh])$")
	end
	if last_cons then
		ing_form = verb .. last_cons .. "ing"
		ed_form = verb .. last_cons .. "ed"
	else
		-- (2) Generate -ing form.
		-- (2a) lie -> lying, untie -> untying, etc.
		local stem = verb:match("^(.*)ie$")
		if stem then
			ing_form = stem .. "ying"
		else
			-- (2b) argue -> arguing, sprue -> spruing, dialogue -> dialoguing, etc.
			stem = verb:match("^(.*)ue$")
			if stem then
				ing_form = stem .. "uing"
			else
				stem = mw.ustring.match(verb, "^(.*[" .. ulvowel .. "yY][^" .. vowel .. "y]+)e$")
				if stem then
					-- (2c) baptize -> baptizing, rake -> raking, type -> typing, parse -> parsing, etc.
					-- (ending in vowel + consonant(s) + -e); but not referee -> refereeing,
					-- backhoe -> backhoeing, redye -> redyeing (ending in some other vowel + -e or in -ye);
					-- and not be -> being (no vowel before the consonant preceding the -e)
					ing_form = stem .. "ing"
				else
					-- (2d) regular verbs
					ing_form = verb .. "ing"
				end
			end
		end

		-- (3) Generate -ed form.
		if verb:find("e$") then
			-- (3a) baptize -> baptized, rake -> raked, parse -> parsed, free -> freed, hoe -> hoed
			ed_form = verb .. "d"
		else
			stem = mw.ustring.match(verb, "^(.*[^" .. ulvowel .. "yY])y$")
			if stem then
				-- (3b) marry -> married, levy -> levied, try -> tried, etc.; but not toy -> toyed
				ed_form = stem .. "ied"
			else
				-- (3c) regular verbs
				ed_form = verb .. "ed"
			end
		end
	end

	return s_form, ing_form, ed_form
end


local function default_verb_forms(verb)
	local full_s_form, full_ing_form, full_ed_form = base_default_verb_forms(verb)
	if verb:find(" ") then
		local first, rest = verb:match("^(.-)( .*)$")
		local first_s_form, first_ing_form, first_ed_form = base_default_verb_forms(first)
		return full_s_form, full_ing_form, full_ed_form, first_s_form .. rest, first_ing_form .. rest, first_ed_form .. rest
	else
		return full_s_form, full_ing_form, full_ed_form, nil, nil, nil
	end
end


pos_functions["ක්‍රියා පද"] = {
	params = {
		[1] = {list = "pres_3sg", allow_holes = true},
		["pres_3sg_qual"] = {list = "pres_3sg=_qual", allow_holes = true},
		[2] = {list = "pres_ptc", allow_holes = true},
		["pres_ptc_qual"] = {list = "pres_ptc=_qual", allow_holes = true},
		[3] = {list = "past", allow_holes = true},
		["past_qual"] = {list = "past=_qual", allow_holes = true},
		[4] = {list = "past_ptc", allow_holes = true},
		["past_ptc_qual"] = {list = "past_ptc=_qual", allow_holes = true},
		},
	func = function(args, data)
		-- Get parameters
		local par1 = args[1][1]
		local par2 = args[2][1]
		local par3 = args[3][1]
		local par4 = args[4][1]
		
		local pres_3sgs, pres_ptcs, pasts, past_ptcs

		local pagename = args.pagename or PAGENAME

		------------------------------------------- UTILITY FUNCTIONS #1 ------------------------------------------

		-- These functions are used directly in the <> format as well as in the utility functions #2 below.

		local function compute_double_last_cons_stem(verb)
			local last_cons = verb:match("([bcdfghjklmnpqrstvwxyzBCDFGHJKLMNPQRSTVWXYZ])$")
			if not last_cons then
				error("Verb stem '" .. verb .. "' must end in a consonant to use ++")
			end
			return verb .. last_cons
		end

		local function compute_plusplus_s_form(verb, default_s_form)
			if verb:find("[sz]$") then
				-- regas -> regasses, derez -> derezzes
				return compute_double_last_cons_stem(verb) .. "es"
			else
				return default_s_form
			end
		end

		------------------------------------------- UTILITY FUNCTIONS #2 ------------------------------------------

		-- These functions are used in both in the separate-parameter format and in the override params such as past_ptc2=. 

		local new_default_s, new_default_ing, new_default_ed, split_default_s, split_default_ing, split_default_ed =
			default_verb_forms(pagename)

		local function compute_double_last_cons_stem_of_split_verb(verb, ending)
			local first, rest = verb:match("^(.-)( .*)$")
			if not first then
				error("Verb '" .. verb .. "' must have a space in it to use ++*")
			end
			local last_cons = first:match("([bcdfghjklmnpqrstvwxyzBCDFGHJKLMNPQRSTVWXYZ])$")
			if not last_cons then
				error("First word '" .. first .. "' must end in a consonant to use ++*")
			end
			return first .. last_cons .. ending .. rest
		end

		local function check_non_nil_star_form(form)
			if form == nil then
				error("Verb '" .. pagename .. "' must have a space in it to use * or ++*")
			end
			return form
		end

		local function sub_tilde(form)
			if not form then
				return nil
			end
			local retval = form:gsub("~", pagename) -- discard second return value
			return retval
		end

		local function canonicalize_s_form(form)
			if form == "+" then
				return new_default_s
			elseif form == "*" then
				return check_non_nil_star_form(split_default_s)
			elseif form == "++" then
				return compute_plusplus_s_form(pagename, new_default_s)
			elseif form == "++*" then
				if pagename:find("^[^ ]*[sz] ") then
					return compute_double_last_cons_stem_of_split_verb(pagename, "es")
				else
					return check_non_nil_star_form(split_default_s)
				end
			else
				return sub_tilde(form)
			end
		end
		
		local function canonicalize_ing_form(form)
			if form == "+" then
				return new_default_ing
			elseif form == "*" then
				return check_non_nil_star_form(split_default_ing)
			elseif form == "++" then
				return compute_double_last_cons_stem(pagename) .. "ing"
			elseif form == "++*" then
				return compute_double_last_cons_stem_of_split_verb(pagename, "ing")
			else
				return sub_tilde(form)
			end
		end
			
		local function canonicalize_ed_form(form)
			if form == "+" then
				return new_default_ed
			elseif form == "*" then
				return check_non_nil_star_form(split_default_ed)
			elseif form == "++" then
				return compute_double_last_cons_stem(pagename) .. "ed"
			elseif form == "++*" then
				return compute_double_last_cons_stem_of_split_verb(pagename, "ed")
			else
				return sub_tilde(form)
			end
		end

		--------------------------------- MAIN PARSING/CONJUGATING CODE --------------------------------

		local past_ptcs_given

		if par1 and par1:find("<") then

			-------------------------- ANGLE-BRACKET FORMAT --------------------------

			if par2 or par3 or par4 then
				error("Can't specify 2=, 3= or 4= when 1= contains angle brackets: " .. par1)
			end
			-- In the angle bracket format, we always copy the full past tense specs to the past participle
			-- specs if none of the latter are given, so act as if the past participle is always given.
			-- There is a separate check to see if the past tense and past participle are identical, in any case.
			past_ptcs_given = true
			local iut = require("Module:inflection utilities")

			-- (1) Parse the indicator specs inside of angle brackets.

			local function parse_indicator_spec(angle_bracket_spec)
				local inside = angle_bracket_spec:match("^<(.*)>$")
				assert(inside)
				local segments = iut.parse_balanced_segment_run(inside, "[", "]")
				local comma_separated_groups = iut.split_alternating_runs(segments, ",")
				if #comma_separated_groups > 4 then
					error("Too many comma-separated parts in indicator spec: " .. angle_bracket_spec)
				end

				local function fetch_qualifiers(separated_group)
					local qualifiers
					for j = 2, #separated_group - 1, 2 do
						if separated_group[j + 1] ~= "" then
							error("Extraneous text after bracketed qualifiers: '" .. table.concat(separated_group) .. "'")
						end
						if not qualifiers then
							qualifiers = {}
						end
						table.insert(qualifiers, separated_group[j])
					end
					return qualifiers
				end

				local function fetch_specs(comma_separated_group)
					if not comma_separated_group then
						return {{}}
					end
					local specs = {}
					
					local colon_separated_groups = iut.split_alternating_runs(comma_separated_group, ":")
					for _, colon_separated_group in ipairs(colon_separated_groups) do
						local form = colon_separated_group[1]
						if form == "*" or form == "++*" then
							error("* and ++* not allowed inside of indicator specs: " .. angle_bracket_spec)
						end
						if form == "" then
							form = nil
						end
						table.insert(specs, {form = form, qualifiers = fetch_qualifiers(colon_separated_group)})
					end
					return specs
				end

				local s_specs = fetch_specs(comma_separated_groups[1])
				local ing_specs = fetch_specs(comma_separated_groups[2])
				local ed_specs = fetch_specs(comma_separated_groups[3])
				local en_specs = fetch_specs(comma_separated_groups[4])
				for _, spec in ipairs(s_specs) do
					if spec.form == "++" and #ing_specs == 1 and not ing_specs[1].form and not ing_specs[1].qualifiers
						and #ed_specs == 1 and not ed_specs[1].form and not ed_specs[1].qualifiers then
						ing_specs[1].form = "++"
						ed_specs[1].form = "++"
						break
					end
				end

				return {
					forms = {},
					s_specs = s_specs,
					ing_specs = ing_specs,
					ed_specs = ed_specs,
					en_specs = en_specs,
				}
			end

			local parse_props = {
				parse_indicator_spec = parse_indicator_spec,
			}
			local alternant_multiword_spec = iut.parse_inflected_text(par1, parse_props)

			-- (2) Remove any links from the lemma, but remember the original form
			--     so we can use it below in the 'lemma_linked' form.

			iut.map_word_specs(alternant_multiword_spec, function(base)
				if base.lemma == "" then
					base.lemma = pagename
				end
				base.orig_lemma = base.lemma
				base.lemma = require("Module:links").remove_links(base.lemma)
			end)

			-- (3) Conjugate the verbs according to the indicator specs parsed above.

			local all_verb_slots = {
				lemma = "infinitive",
				lemma_linked = "infinitive",
				s_form = "3|s|pres",
				ing_form = "pres|ptcp",
				ed_form = "past",
				en_form = "past|ptcp",
			}
			local function conjugate_verb(base)
				local def_s_form, def_ing_form, def_ed_form = base_default_verb_forms(base.lemma)

				local function process_specs(slot, specs, default_form, canonicalize_plusplus)
					for _, spec in ipairs(specs) do
						local form = spec.form
						if not form or form == "+" then
							form = default_form
						elseif form == "++" then
							form = canonicalize_plusplus()
						end
						-- If there's a ~ in the form, substitute it with the lemma,
						-- but make sure to first replace % in the lemma with %% so that
						-- it doesn't get interpreted as a capture replace expression.
						if form:find("~") then
							-- Assign to a var because gsub returns multiple values.
							local subbed_lemma = base.lemma:gsub("%%", "%%%%")
							form = form:gsub("~", subbed_lemma)
						end
						-- If the form is -, don't insert any forms, which will result
						-- in there being no overall forms (in fact it will be nil).
						-- We check for that down below and substitute a single "-" as
						-- the form, which in turn gets turned into special labels like
						-- "no present participle".
						if form ~= "-" then
							iut.insert_form(base.forms, slot, {form = form, footnotes = spec.qualifiers})
						end
					end
				end

				process_specs("s_form", base.s_specs, def_s_form,
					function() return compute_plusplus_s_form(base.lemma, def_s_form) end)
				process_specs("ing_form", base.ing_specs, def_ing_form,
					function() return compute_double_last_cons_stem(base.lemma) .. "ing" end)
				process_specs("ed_form", base.ed_specs, def_ed_form,
					function() return compute_double_last_cons_stem(base.lemma) .. "ed" end)

				-- If the -en spec is completely missing, substitute the -ed spec in its entirely.
				-- Otherwise, if individual -en forms are missing or use +, we will substitute the
				-- default -ed form, as with the -ed spec.
				local en_specs = base.en_specs
				if #en_specs == 1 and not en_specs[1].form and not en_specs[1].qualifiers then
					en_specs = base.ed_specs
				end

				process_specs("en_form", en_specs, def_ed_form,
					function() return compute_double_last_cons_stem(base.lemma) .. "ed" end)

				iut.insert_form(base.forms, "lemma", {form = base.lemma})
				-- Add linked version of lemma for use in head=. We write this in a general fashion in case
				-- there are multiple lemma forms (which isn't possible currently at this level, although it's
				-- possible overall using the ((...,...)) notation).
				iut.insert_forms(base.forms, "lemma_linked", iut.map_forms(base.forms.lemma, function(form)
					if form == base.lemma and base.orig_lemma:find("%[%[") then
						return base.orig_lemma
					else
						return form
					end
				end))
			end

			local inflect_props = {
				slot_table = all_verb_slots,
				inflect_word_spec = conjugate_verb,
			}
			iut.inflect_multiword_or_alternant_multiword_spec(alternant_multiword_spec, inflect_props)

			-- (4) Fetch the forms and put the conjugated lemmas in data.heads if not explicitly given.

			local function fetch_forms(slot)
				local forms = alternant_multiword_spec.forms[slot]
				-- See above. This should only occur if the user explicitly used -
				-- for a spec.
				if not forms or #forms == 0 then
					forms = {{form = "-"}}
				end
				return forms
			end

			pres_3sgs = fetch_forms("s_form")
			pres_ptcs = fetch_forms("ing_form")
			pasts = fetch_forms("ed_form")
			past_ptcs = fetch_forms("en_form")
			-- Use the "linked" form of the lemma as the head if no head= explicitly given.
			-- If no links in this form and it has multiple words, autolink the individual words.
			-- The user can override this using head=.
			if #data.heads == 0 then
				for _, lemma_obj in ipairs(alternant_multiword_spec.forms.lemma_linked) do
					local lemma = lemma_obj.form
					if not lemma:find("%[%[") then
						local m_headword = require("Module:headword")
						if m_headword.head_is_multiword(lemma) then
							lemma = m_headword.add_multiword_links(lemma)
						end
					end
					table.insert(data.heads, lemma)
				end
			end
		else
			-------------------------- SEPARATE-PARAM FORMAT --------------------------

			local pres_3sg, pres_ptc, past

			if par1 and not par2 and not par3 then
				-- Use of a single parameter other than "++", "*" or "++*" is now the "legacy" format,
				-- and no longer supported.
				if par1 == "es" or par1 == "ies" or par1 == "d" then
					error("Legacy parameter 1=es/ies/d no longer supported, just use 'si-verb' without params")
				elseif par1 == "++" or par1 == "*" or par1 == "++*" then
					pres_3sg = canonicalize_s_form(par1)
					pres_ptc = canonicalize_ing_form(par1)
					past = canonicalize_ed_form(par1)
				else
					error("Legacy parameter 1=STEM no longer supported, just use 'si-verb' without params")
				end
			else
				if par3 then
					track("xxx3")
				elseif par2 then
					track("xxx2")
				end
			end

			if not pres_3sg or not pres_ptc or not past then
				-- Either all three should be set above, or none of them.
				assert(not pres_3sg and not pres_ptc and not past)

				if par1 then
					pres_3sg = canonicalize_s_form(par1)
				else
					pres_3sg = new_default_s
				end

				if par2 then
					pres_ptc = canonicalize_ing_form(par2)
				else
					pres_ptc = new_default_ing
				end

				if par3 then
					past = canonicalize_ed_form(par3)
				else
					past = new_default_ed
				end
			end

			if par4 then
				past_ptcs_given = true
				past_ptc = canonicalize_ed_form(par4)
			else
				past_ptc = past
			end

			pres_3sgs = {{form = pres_3sg}}
			pres_ptcs = {{form = pres_ptc}}
			pasts = {{form = past}}
			past_ptcs = {{form = past_ptc}}
		end

		------------------------------------------- HANDLE OVERRIDES ------------------------------------------

		local pres_3sg_infls, pres_ptc_infls, past_infls, past_ptc_infls

		local function strip_brackets(qualifiers)
			if not qualifiers then
				return nil
			end
			local stripped_qualifiers = {}
			for _, qualifier in ipairs(qualifiers) do
				local stripped_qualifier = qualifier:match("^%[(.*)%]$")
				if not stripped_qualifier then
					error("Internal error: Qualifier should be surrounded by brackets at this stage: " .. qualifier)
				end
				table.insert(stripped_qualifiers, stripped_qualifier)
			end
			return stripped_qualifiers
		end

		local function collect_forms(label, accel_form, defaults, overrides, override_qualifiers, canonicalize)
			if defaults[1].form == "-" then
				return {label = "no " .. label}
			else
				local into_table = {label = label, accel = {form = accel_form}}
				local maxindex = math.max(#defaults, overrides.maxindex)
				local qualifiers = override_qualifiers[1] and {override_qualifiers[1]} or strip_brackets(defaults[1].footnotes)
				table.insert(into_table, {term = defaults[1].form, qualifiers = qualifiers})

				-- Present 3rd singular
				for i = 2, maxindex do
					local override_form = canonicalize(overrides[i])
		
					if override_form then
						-- If there is an override such as past_ptc2=..., only use the qualifier specified
						-- using an override (past_ptc2_qual=...), if any; it doesn't make sense to combine
						-- an override form with a qualifier specified inside of angle brackets.
						table.insert(into_table, {term = override_form, qualifiers = {override_qualifiers[i]}})
					elseif defaults[i] then
						-- If the form comes from inside angle brackets, allow any override qualifier
						-- (past_ptc2_qual=...) to override any qualifier specified inside of angle brackets.
						-- FIXME: Maybe we should throw an error here if both exist.
						local qualifiers = override_qualifiers[i] and {override_qualifiers[i]} or strip_brackets(defaults[i].footnotes)
						table.insert(into_table, {term = defaults[i].form, qualifiers = qualifiers})
					end
				end

				return into_table
			end
		end

		local pres_3sg_infls = collect_forms("third-person singular simple present", "3|s|pres",
			pres_3sgs, args[1], args.pres_3sg_qual, canonicalize_s_form)
		local pres_ptc_infls = collect_forms("present participle", "pres|ptcp",
			pres_ptcs, args[2], args.pres_ptc_qual, canonicalize_ing_form)
		local past_infls = collect_forms("simple past", "past",
			pasts, args[3], args.past_qual, canonicalize_ed_form)
		local past_ptc_infls = collect_forms("past participle", "past|ptcp",
			past_ptcs, args[4], args.past_ptc_qual, canonicalize_ed_form)
		
		-- Are the past forms identical to the past participle forms? If so, we use a single
		-- combined "simple past and past participle" label on the past tense forms.
		-- We check for two conditions: Either no past participle forms were given at all, or
		-- they were given but are identical in every way (all forms and qualifiers) to the past
		-- tense forms. The former "no explicit past participle forms" check is important in the
		-- "separate-parameter" format; if past tense overrides are given and no past participle
		-- forms given, the past tense overrides should apply to the past participle as well.
		-- In the angle-bracket format, it's expected that all forms and qualifiers are specified
		-- using that format, and we explicitly copy past tense forms and qualifiers to past
		-- participle ones if the latter are omitted, so we disable to "no explicit past participle
		-- forms" check.
		if args[4].maxindex > 0 or args.past_ptc_qual.maxindex > 0 then
			past_ptcs_given = true
		end

		local identical = true

		-- For the past and past participle to be identical, there must be
		-- the same number of inflections, and each inflection must match
		-- in term and qualifiers.
		if #past_infls ~= #past_ptc_infls then
			identical = false
		else
			for key, val in ipairs(past_infls) do
				if past_ptc_infls[key].term ~= val.term then
					identical = false
					break
				else
					local quals1 = past_ptc_infls[key].qualifiers
					local quals2 = val.qualifiers
					if (not not quals1) ~= (not not quals2) then
						-- one is nil, the other is not
						identical = false
					elseif quals1 and quals2 then
						-- qualifiers present in both; each qualifier must match
						if #quals1 ~= #quals2 then
							identical = false
						else
							for k, v in ipairs(quals1) do
								if v ~= quals2[k] then
									identical = false
									break
								end
							end
						end
					end
					if not identical then
						break
					end
				end
			end
		end
		
		-- Insert the forms
		table.insert(data.inflections, pres_3sg_infls)
		table.insert(data.inflections, pres_ptc_infls)
		
		if not past_ptcs_given or identical then
			if past_ptcs[1].form == "-" then
				past_infls.label = "no simple past or past participle"
			else
				past_infls.label = "simple past and past participle"
				past_infls.accel = {form = "past|and|past|ptcp"}
			end
			table.insert(data.inflections, past_infls)
		else
			table.insert(data.inflections, past_infls)
			table.insert(data.inflections, past_ptc_infls)
		end
	end
}

return export