Module:zlw-ocs-noun

local export = {}

--[=[

Authorship: Zhnka

]=]

--[=[

TERMINOLOGY:

-- "slot" = A particular combination of case/number. Example slot names for nouns are "gen_s" (genitive singular) and "voc_p" (vocative plural). Each slot is filled with zero or more forms.

-- "form" = The declined Czech form representing the value of a given slot.

-- "lemma" = The dictionary form of a given Czech term. Generally the nominative masculine singular, but may occasionally be another form if the nominative masculine singular is missing. ]=]

local lang = require("Module:languages").getByCode("zlw-ocs") local m_table = require("Module:table") local m_links = require("Module:links") local m_string_utilities = require("Module:string utilities") local iut = require("Module:inflection utilities") local m_para = require("Module:parameters") local com = require("Module:zlw-ocs-common")

local current_title = mw.title.getCurrentTitle local NAMESPACE = current_title.nsText local PAGENAME = current_title.text

local u = mw.ustring.char local rsplit = mw.text.split local rfind = mw.ustring.find local rmatch = mw.ustring.match local rgmatch = mw.ustring.gmatch local rsubn = mw.ustring.gsub local ulen = mw.ustring.len local usub = mw.ustring.sub local uupper = mw.ustring.upper local ulower = mw.ustring.lower

local force_cat = false -- set to true to make categories appear in non-mainspace pages, for testing

-- version of rsubn that discards all but the first return value local function rsub(term, foo, bar) local retval = rsubn(term, foo, bar) return retval end

-- version of rsubn that returns a 2nd argument boolean indicating whether -- a substitution was made. local function rsubb(term, foo, bar) local retval, nsubs = rsubn(term, foo, bar) return retval, nsubs > 0 end

local function track(track_id) require("Module:debug/track")("zlw-ocs-noun/" .. track_id) return true end

local output_noun_slots = { nom_s = "nom|s", gen_s = "gen|s", dat_s = "dat|s", acc_s = "acc|s", voc_s = "voc|s", loc_s = "loc|s", ins_s = "ins|s", nom_d = "nom|d", gen_d = "gen|d", dat_d = "dat|d", acc_d = "acc|d", voc_d = "voc|d", loc_d = "loc|d", ins_d = "ins|d", nom_p = "nom|p", gen_p = "gen|p", dat_p = "dat|p", acc_p = "acc|p", voc_p = "voc|p", loc_p = "loc|p", ins_p = "ins|p", }

local function get_output_noun_slots(alternant_multiword_spec) -- FIXME: To save memory we modify the table in-place. This won't work if we ever end up with multiple calls to -- this module in the same Lua invocation, and we would need to clone the table. if alternant_multiword_spec.actual_number ~= "allthree" then for slot, accel_form in pairs(output_noun_slots) do			output_noun_slots[slot] = accel_form:gsub("|[sp]$", "") end end return output_noun_slots end

local potential_lemma_slots = {"nom_s", "nom_p", "gen_s"}

local cases = { nom = true, gen = true, dat = true, acc = true, voc = true, loc = true, ins = true, }

local clitic_cases = { gen = true, dat = true, acc = true, }

local function dereduce(base, stem) local dereduced_stem = com.dereduce(base, stem) if not dereduced_stem then error("Unable to dereduce stem '" .. stem .. "'") end return dereduced_stem end

local function apply_special_cases(base, slot, stem, ending) local palatalize_voc if base.c_as_k and rfind(ending, "^[aouyáóúůý]") then local k_stem = rsub(stem, "c$", "k") stem = {stem, k_stem} elseif slot == "voc_s" and ending == "e" and base.palatalize_voc and not base["-velar"] then local palstem = com.apply_first_palatalization(stem) if base.animacy == "inan" and rfind(stem, com.cons_c .. "r$") and not rfind(stem, "rr$") then -- optional r -> ř stem = {stem, palstem} else stem = palstem end elseif rfind(ending, "^[ěií]") or slot == "loc_s" and ending == "e" then -- loc_s of hard masculines is sometimes -e/ě; the user might indicate this as -e, which we should handle -- correctly stem = com.apply_second_palatalization(stem) end return stem, ending end

local function skip_slot(number, slot) return number == "sg" and rfind(slot, "_p$") or		number == "pl" and rfind(slot, "_s$") end

-- Basic function to combine stem(s) and ending(s) and insert the result into the appropriate slot. `stems` is either -- the `stems` object passed into the declension functions (containing the various stems; see below) or a string to -- override the stem. (NOTE: If you pass a string in as `stems`, you should pass the value of `stems.footnotes` as the -- value of `footnotes` as it will be lost otherwise. If you need to supply your own footnote in addition, use -- iut.combine_footnotes to combine any user-specified footnote(s) with your footnote(s).) `endings` is either a -- string specifying a single ending or a list of endings. If `endings` is nil, no forms are inserted. If an ending is -- "-", the value of `stems` is ignored and the lemma is used instead as the stem; this is important in case the user -- used `decllemma:` to specify a declension lemma different from the actual lemma, or specified '.foreign' (which has -- a similar effect). local function add(base, slot, stems, endings, footnotes) if not endings then return end -- Call skip_slot based on the declined number; if the actual number is different, we correct this in -- decline_noun at the end. if skip_slot(base.number, slot) then return end local stems_footnotes = type(stems) == "table" and stems.footnotes or nil footnotes = iut.combine_footnotes(iut.combine_footnotes(base.footnotes, stems_footnotes), footnotes) if type(endings) == "string" then endings = {endings} end for _, ending in ipairs(endings) do -- Compute the stem. If ending is "-", use the lemma regardless. Otherwise if `stems` is a string, use it. -- Otherwise `stems` is an object containing four stems (vowel-vs-non-vowel cross regular-vs-oblique); -- compute the appropriate stem based on the slot and whether the ending begins with a vowel. local stem if ending == "-" then stem = base.actual_lemma ending = "" elseif type(stems) == "string" then stem = stems else local is_vowel_ending = rfind(ending, "^" .. com.vowel_c) if stems.oblique_slots == "all" or				(stems.oblique_slots == "gen_p" or stems.oblique_slots == "all-oblique") and slot == "gen_p" or				stems.oblique_slots == "all-oblique" and (slot == "ins_s" or slot == "dat_p" or slot == "loc_p" or slot == "ins_p") then if is_vowel_ending then stem = stems.oblique_vowel_stem else stem = stems.oblique_nonvowel_stem end elseif is_vowel_ending then stem = stems.vowel_stem else stem = stems.nonvowel_stem end end -- Maybe apply the first or second Slavic palatalization. stem, ending = apply_special_cases(base, slot, stem, ending) ending = iut.combine_form_and_footnotes(ending, footnotes) local function combine_stem_ending(stem, ending) return com.combine_stem_ending(base, slot, stem, ending) end iut.add_forms(base.forms, slot, stem, ending, combine_stem_ending) end end

local function process_slot_overrides(base, do_slot) for slot, overrides in pairs(base.overrides) do		-- Call skip_slot based on the declined number; if the actual number is different, we correct this in -- decline_noun at the end. if skip_slot(base.number, slot) then error("Override specified for invalid slot '" .. slot .. "' due to '" .. base.number .. "' number restriction") end if do_slot(slot) then base.slot_overridden[slot] = true base.forms[slot] = nil for _, override in ipairs(overrides) do				for _, value in ipairs(override.values) do					local form = value.form local combined_notes = iut.combine_footnotes(base.footnotes, value.footnotes) if override.full then if form ~= "" then iut.insert_form(base.forms, slot, {form = form, footnotes = combined_notes}) end else -- Convert a null ending to "-" in the acc/voc sg slots so that e.g. Kerberos declared as						--  works correctly and generates accusative 'Kerberos/Kerbera' not -- #'Kerber/Kerbera'. if (slot == "acc_s" or slot == "voc_s") and form == "" then form = "-" end for _, stems in ipairs(base.stem_sets) do							add(base, slot, stems, form, combined_notes) end end end end end end end

local function add_decl(base, stems,	gen_s, dat_s, acc_s, voc_s, loc_s, ins_s,	nom_d, gen_d, dat_d, 	nom_p, gen_p, dat_p, acc_p, loc_p, ins_p, nom_s, footnotes ) add(base, "nom_s", stems, "-", footnotes) add(base, "gen_s", stems, gen_s, footnotes) add(base, "dat_s", stems, dat_s, footnotes) add(base, "acc_s", stems, acc_s, footnotes) add(base, "voc_s", stems, voc_s, footnotes) add(base, "loc_s", stems, loc_s, footnotes) add(base, "ins_s", stems, ins_s, footnotes) add(base, "nom_d", stems, nom_d, footnotes) add(base, "gen_d", stems, gen_d, footnotes) add(base, "dat_d", stems, dat_d, footnotes) if base.number == "pl" then -- If this is a plurale tantum noun and we're processing the nominative plural, use the user-specified lemma -- rather than generating the plural from the synthesized singular, which may not match the specified lemma -- (e.g. tvargle "Olomouc cheese" using  would try to generate 'tvargle/tvargly', and peníze		-- "money" using  would try to generate 'peněze'). local acc_p_like_nom = m_table.deepEquals(nom_p, acc_p) nom_p = "-" if acc_p_like_nom then acc_p = "-" end end add(base, "nom_p", stems, nom_p, footnotes) add(base, "gen_p", stems, gen_p, footnotes) add(base, "dat_p", stems, dat_p, footnotes) add(base, "acc_p", stems, acc_p, footnotes) add(base, "loc_p", stems, loc_p, footnotes) add(base, "ins_p", stems, ins_p, footnotes) add(base, "nom_s", stems, nom_s, footnotes) end

local function add_sg_decl(base, stems,	gen_s, dat_s, acc_s, voc_s, loc_s, ins_s, footnotes ) add_decl(base, stems, gen_s, dat_s, acc_s, voc_s, loc_s, ins_s,		nil, nil, nil,		nil, nil, nil, nil, nil, nil, footnotes) end

local function add_du_only_decl(base, stems,	gen_d, dat_d, footnotes ) add_decl(base, stems, nil, nil, nil, nil, nil, nil, 		"-", gen_d, dat_d,		nil, nil, nil, nil, nil, nil, footnotes) end

local function add_pl_only_decl(base, stems,	gen_p, dat_p, acc_p, loc_p, ins_p, footnotes ) add_decl(base, stems, nil, nil, nil, nil, nil, nil, 		nil, nil, nil,		"-", gen_p, dat_p, acc_p, loc_p, ins_p, footnotes) end

local function handle_derived_slots_and_overrides(base) local function is_non_derived_slot(slot) return slot ~= "voc_p" and slot ~= "acc_s" and slot ~= "clitic_acc_s" end

local function is_derived_slot(slot) return not is_non_derived_slot(slot) end

base.slot_overridden = {} -- Handle overrides for the non-derived slots. Do this before generating the derived -- slots so overrides of the source slots (e.g. nom_p) propagate to the derived slots. process_slot_overrides(base, is_non_derived_slot)

-- Generate the remaining slots that are derived from other slots. if not base.pron and not base.det then -- Pronouns don't have a vocative (singular or plural). iut.insert_forms(base.forms, "voc_p", base.forms.nom_p) end if not base.forms.acc_s and not base.slot_overridden.acc_s then iut.insert_forms(base.forms, "acc_s", base.forms[base.animacy == "inan" and "nom_s" or base.animacy == "pr" and "gen_s"]) end if not base.forms.clitic_acc_s and not base.slot_overridden.clitic_acc_s then iut.insert_forms(base.forms, "clitic_acc_s", base.forms[base.animacy == "inan" and "nom_s" or "clitic_gen_s"]) end

-- Handle overrides for derived slots, to allow them to be overridden. process_slot_overrides(base, is_derived_slot)

-- Compute linked versions of potential lemma slots, for use in. -- We substitute the original lemma (before removing links) for forms that -- are the same as the lemma, if the original lemma has links. for _, slot in ipairs(potential_lemma_slots) do iut.insert_forms(base.forms, slot .. "_linked", iut.map_forms(base.forms[slot], function(form) if form == base.orig_lemma_no_links and rfind(base.orig_lemma, "%[%[") then return base.orig_lemma else return form end end)) end end

-- Table mapping declension types to functions to decline the noun. The function takes two arguments, `base` and -- `stems`; the latter specifies the computed stems (vowel vs. non-vowel, singular vs. plural) and whether the noun -- is reducible and/or has vowel alternations in the stem. Most of the specifics of determining which stem to use -- and how to modify it for the given ending are handled in add_decl; the declension functions just need to generate -- the appropriate endings. local decls = {} -- Table specifying additional properties for declension types. Every declension type must have such a table, which -- specifies which category or categories to add and what annotation to show in the title bar of the declension table. -- -- * Only the `cat` property of this table is mandatory; there is also a `desc` property to specify the annotation, but --  this can be omitted and the annotation will then be computed from the `cat` property. The `cat` property is either --  a string, a list of strings or a function (of two arguments, `base` and `stems` as above) returning a string or --   list of strings. The string can contain the keywords GENDER to substitute the gender (and animacy for masculine --  nouns) and POS (to substitute the pluralized part of speech). The keyword GENPOS is equivalent to 'GENDER POS'. If --  no keyword is present, ' GENPOS' is added onto the end. If only GENDER is present, ' POS' is added onto the end. --  In all cases, the language name is added onto the beginning to form the full category name. -- * The `desc` property is of the same form as the `cat` property and specifies the annotation to display in the title --  bar (which may have the same format as the category minus the part of speech, or may be abbreviated). The value --  may not be a list of strings, as only one annotation is displayed. If omitted, it is derived from the category --  spec(s) by taking the last category (if more than one is given) and removing ' POS' before keyword substitution. local declprops = {}

decls["hard-m"] = function(base, stems) base.palatalize_voc = not rfind(stems.vowel_stem, "c$") local gen_s = base.animacy == "pr" and "a" or {"a", "u"} local dat_s = base.animacy == "inan" and {"u"} or {"u", "ovi"} local loc_s = base.animacy == "inan" and rfind(base.lemma, "l$") and {"e", "u"} or base.animacy == "inan" and {"ě", "u"} or base.animacy == "anml" and {"ě", "u", "ovi"} or {"u", "ovi"} local voc_s = velar and {"e", "u"} or "e" local loc_p = rfind(base.lemma, "l$") and "éch" or "iech" local acc_s = base.animacy == "anml" and {"", "a"} add_decl(base, stems, gen_s, dat_s, acc_s, voc_s, loc_s, "em",		"y", "ú", "oma",		{"i", "ové"}, "óv", "óm", "y", loc_p, "y") end

declprops["hard-m"] = { desc = function(base, stems) return "hard o-stem" end, cat = function(base, stems) return "hard masculine o-stem" end }

decls["soft-m"] = function(base, stems) if rfind(base.lemma, "l$") then local dat_s = base.animacy == "inan" and "u" or {"u", "ovi"} local loc_s = base.animacy == "inan" and {"i", "u"} or base.animacy == "anml" and {"i", "u", "ovi"} or {"u", "ovi"} local nom_p = rfind(base.lemma, "tel") and {"é", "i"} or {"i"} local gen_p = rfind(base.lemma, "tel") and "" or nil local acc_s = base.animacy == "anml" and {"", "ě"} add_decl(base, stems, "e", {}, {}, {}, {}, "em",		"e", {}, {},		nom_p, gen_p, {}, "e", "ích", "i") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, dat_s, nil, "u", loc_s, nil,		nil, "ú", "oma",		"ové", "óv", "óm"		) elseif rfind(base.lemma, "z$") then local dat_s = base.animacy == "inan" and "u" or {"u", "ovi"} local loc_s = base.animacy == "inan" and {"i", "u"} or {"u", "ovi"} add_decl(base, stems, "ě", dat_s, nil, {}, loc_s, "em",		"ě", "ú", "oma",		{"i", "ové"}, "óv", "óm", "ě", "ích", "i") add_decl(base, com.apply_first_palatalization(stem, "is soft"), {}, {}, {}, "e") else base.palatalize_voc = true local dat_s = base.animacy == "inan" and "u" or {"u", "ovi"} local loc_s = base.animacy == "inan" and {"i", "u"} or base.animacy == "anml" and {"i", "u", "ovi"} or {"u", "ovi"} local voc_s = base.animacy == "pr" and rfind(base.lemma, "ec$") and stems.reducible and "e" or base.animacy == "pr" and rfind(base.lemma, "z$") and "e" or "u" local nom_p = rfind(base.lemma, "ař") and {"é", "ové"} or {"i", "ové"} local acc_s = base.animacy == "anml" and {"", "ě"} add_decl(base, stems, "ě", dat_s, acc_s, voc_s, loc_s, "em",		"ě", "ú", "oma",		nom_p, "óv", "óm", "ě", "ích", "i") end end

declprops["soft-m"] = { desc = function(base, stems) return "soft o-stem" end, cat = function(base, stems) return "soft masculine o-stem" end }

decls["ěnín-m"] = function(base, stems) add_decl(base, stems, "ěnína", {"ěnínu", "ěnínovi"}, "ěnína", "ěníne", {"ěnínu", "ěnínovi"}, "ěnínem",		"ěníny", "ěnínú", "ěnínoma",		{"ěné", "anové"}, {"an", "anóv"}, "anóm", "any", "ěniech", "any") end

declprops["ěnín-m"] = { desc = function(base, stems) return "hard o-stem" end, cat = function(base, stems) return "hard masculine o-stem" end }

decls["enín-m"] = function(base, stems) add_decl(base, stems, "enína", {"enínu", "enínovi"}, "enína", "eníne", {"enínu", "enínovi"}, "enínem",		"eníny", "enínú", "enínoma",		{"ené", "anové"}, {"an", "anóv"}, "anóm", "any", "eniech", "any") end

declprops["enín-m"] = { desc = function(base, stems) return "hard o-stem" end, cat = function(base, stems) return "hard masculine o-stem" end }

decls["u-m"] = function(base, stems) local gen_s = base.animacy == "inan" and "u" or "a" local acc_s = base.animacy == "inan" and "-" or "a" local loc_p = rfind(base.lemma, "l$") and "éch" or "iech" add_decl(base, stems, gen_s, {"ovi", "u"}, acc_s, "e", "u", "em",		"y", "ú", "oma",		{"ové", "i"}, "óv", "óm", "y", loc_p, "y") end

declprops["u-m"] = { desc = function(base, stems) return "u-stem" end, cat = function(base, stems) return "u-stem" end }

decls["a-m"] = function(base, stems) local it_ist = rfind(stems.vowel_stem, "is?t$") or rfind(stems.vowel_stem, "ast$") add_decl(base, stems, "y", "ě", "u", "o", "ě", "ú",		"ě", "ú", "ama",		"y", "", "ám", "y", "ách", "ami") end

declprops["a-m"] = { desc = function(base, stems) return "hard a-stem" end, cat = function(base, stems) return "hard masculine a-stem" end }

decls["ě-m"] = function(base, stems) add_decl(base, stems, "ě", "i", {}, {}, "i", {},		"i", {}, "ěma",		"ě", "í", "iem", "ě", "iech", "ěmi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "u", "e", nil, "ú", nil, "ú") end

declprops["ě-m"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft masculine a-stem" end }

decls["ijo-m"] = function(base, stems) if rfind(base.lemma, "lí$") then local acc_s = base.animacy == "inan" and "í" or {"é", "í"} add_decl(base, stems, "é", {}, acc_s, {}, "í", "ím",		"é", {}, "íma",		"í", "í", "ím", "é", "ích", "ími") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, "ú", nil, "ú", nil, nil, nil, "ú") else local acc_s = base.animacy == "inan" and "í" or {"ie", "í"} add_decl(base, stems, "ie", {}, acc_s, {}, "í", "ím",		"ie", {}, "íma",		"í", "í", "ím", "ie", "ích", "ími") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, "ú", nil, "ú", nil, nil, nil, "ú") end end

declprops["ijo-m"] = { desc = function(base, stems) return "soft o-stem" end, cat = function(base, stems) return "soft masculine o-stem" end }

decls["ija-m"] = function(base, stems) if rfind(base.lemma, "l[íé]") then add_decl(base, stems, "é", "í", {}, "í", "í", "ím",		"í", {}, "éma",		"é", "í", "ém", "é", "éch", "iemi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "ú", nil, nil, nil, nil, "ú") else add_decl(base, stems, "ie", "í", {}, "í", "í", "ím",		"í", {}, "iema",		"ie", "í", "iem", "ie", "iech", "iemi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "ú", nil, nil, nil, nil, "ú") end end

declprops["ija-m"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft masculine a-stem" end }

decls["istem-m"] = function(base, stems) add_decl(base, stems, "i", "i", {"", "i"}, "i", "i", "em",		"i", {}, "ma",		"ie", "í", "em", "i", "ech", "mi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, nil, nil, nil, nil, nil, "ú") end

declprops["istem-m"] = { desc = function(base, stems) return "i-stem" end, cat = function(base, stems) return "masculine i-stem" end }

decls["n-m"] = function(base, stems) add_decl(base, stems, "e", "i", "-", "i", "i", "em",		"y", "ú", "oma",		{"i", "ové"}, "óv", "óm", "y", "iech", "y") end

declprops["n-m"] = { desc = function(base, stems) return "n-stem" end, cat = function(base, stems) return "masculine n-stem" end } decls["tstem-m"] = function(base, stems) add_decl(base, stems, "te", "ti", "et", "te", "ti", "tem",		"ty", "tú", "toma",		"ty", "tóv", "tóm", "ty", "tiech", "ty") end

declprops["tstem-m"] = { desc = function(base, stems) return "t-stem" end, cat = function(base, stems) return "masculine t-stem" end }

decls["hard-f"] = function(base, stems) base.no_palatalize_c = true if rfind(base.lemma, "la$") then add_decl(base, stems, "y", "e", "u", "o", "e", "ú",		"e", "ú", "ama",		"y", "", "ám", "y", "ách", "ami") else add_decl(base, stems, "y", "ě", "u", "o", "ě", "ú",		"ě", "ú", "ama",		"y", "", "ám", "y", "ách", "ami") end end

declprops["hard-f"] = { cat = "hard" }

declprops["hard-f"] = { desc = function(base, stems) return "hard a-stem" end, cat = function(base, stems) return "hard feminine a-stem" end }

decls["soft-f"] = function(base, stems) local gen_p = rfind(base.lemma, "ice$") and nil or "í" add_decl(base, stems, "ě", "i", {}, {}, "i", {},		"i", {}, "ěma",		"ě", gen_p, "iem", "ě", "iech", "ěmi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "u", "e", nil, "ú", nil, "ú") add_decl(base, com.onlyndt(stems.oblique_vowel_stem), nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, "") end

declprops["soft-f"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft feminine a-stem" end }

decls["e-f"] = function(base, stems) add_decl(base, stems, "e", "i", {}, "e", "i", {},		"i", {}, "ema",		"e", {"í", ""}, "ém", "e", "éch", "emi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "u", nil, nil, "ú", nil, "ú") end

declprops["e-f"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft feminine a-stem" end }

decls["cons-f"] = function(base, stems) if rfind(base.lemma, "l$") then add_decl(base, stems, "e", "i", {}, {}, "i", {},		"i", {}, "ema",		"e", "í", "ém", "e", "ích", "emi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "u", "e", nil, "ú",		nil, "ú") else add_decl(base, stems, "ě", "i", "u", "e", "i", "ú",		"i", "ú", "ěma",		"ě", "í", "iem", "ě", "ích", "ěmi") end end

declprops["cons-f"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft feminine a-stem" end }

decls["v-f"] = function(base, stems) add_decl(base, stems, "ve", "vi", "ev", "vi", "vi", "v́ú",		"vi", "v́ú", "vema",		"ve", "ví", "vem", "ve", "vech", "vemi") end

declprops["v-f"] = { desc = function(base, stems) return "v-stem" end, cat = function(base, stems) return "v-stem" end }

decls["r-f"] = function(base, stems) if base.lemma == "máti" then stem = "mat" end add_decl(base, stem, "eře", "eři", "eř", "-", "eři", "eřú",		"eři", "eřú", "eřma",		"ery", "er", "erám", "ery", "erách", "erami") end

declprops["r-f"] = { desc = function(base, stems) return "r-stem" end, cat = function(base, stems) return "r-stem" end }

decls["istem-f"] = function(base, stems) local acc_s = rfind(base.lemma, "i$") and "i" or "-" add_decl(base, stems, "i", "i", acc_s, "i", "i", {},		"i", {}, "ma",		"i", "í", "em", "i", "ech", "mi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, nil, nil, nil, "ú", nil, "ú") end

declprops["istem-f"] = { desc = function(base, stems) return "i-stem" end, cat = function(base, stems) return "feminine i-stem" end }

decls["i-f"] = function(base, stems) add_decl(base, stems, "ě", "i", {}, {}, "i", {},		"i", {}, "ěma",		"ě", {}, "iem", "ě", "iech", "ěmi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "u", "e", nil, "ú", nil, "ú", nil, nil, "") end

declprops["i-f"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft feminine a-stem" end }

decls["í-f"] = function(base, stems) if rfind(stem, "l$") then add_decl(base, stems, "é", "í", {}, "í", "í", {},		"í", {}, "éma",		"é", "í", "ém", "é", "éch", "émi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "ú", nil, nil, "ú", nil, "ú") else local voc_s = rfind(base.lemma, "ie$") and "ie" or "í" add_decl(base, stems, "ie", "í", {}, voc_s, "í", {},		"í", {}, "iema",		"ie", "í", "iem", "ie", "iech", "iemi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "ú", nil, nil, "ú", nil, "ú") end end

declprops["í-f"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft feminine a-stem" end }

decls["é-f"] = function(base, stems) add_decl(base, stems, "é", "í", {}, "í", "í", {},		"í", {}, "éma",		"é", "í", "ém", "ie", "éch", "émi") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, nil, "ú", nil, nil, "ú", nil, "ú") end

declprops["é-f"] = { desc = function(base, stems) return "soft a-stem" end, cat = function(base, stems) return "soft feminine a-stem" end } decls["hard-n"] = function(base, stems) if rfind(base.lemma, "lo$") then add_decl(base, stems, "a", "u", "-", "-", {"e", "u"}, "em",		"e", "ú", "oma",		"a", "", "óm", "a", "éch", "y") else add_decl(base, stems, "a", "u", "-", "-", {"ě", "u"}, "em",		"ě", "ú", "oma",		"a", "", "óm", "a", "iech", "y") end end

declprops["hard-n"] = { desc = function(base, stems) return "hard o-stem" end, cat = function(base, stems) return "hard neuter o-stem" end }

decls["soft-n"] = function(base, stems) if rfind(base.lemma, "le$") then add_decl(base, stems, "e", {}, "-", "-", "i", "em",		"i", {}, {},		"e", "í", {}, "e", "ích", "i") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, "u", nil, nil, "u", nil, nil, "ú", "oma", nil, nil, "óm") else add_decl(base, stems, "ě", {}, "-", "-", "i", "em",		"i", {}, {},		"ě", "í", {}, "ě", "ích", "i") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, "u", nil, nil, "u", nil, nil, "ú", "oma", nil, nil, "óm") end end

declprops["soft-n"] = { desc = function(base, stems) return "soft o-stem" end, cat = function(base, stems) return "soft neuter o-stem" end }

decls["ie-n"] = function(base, stems) add_decl(base, stems, "ie", {}, "ie", "ie", "í", "ím",		"í", {}, "íma",		"ie", "í", "ím", "ie", "ích", "ími") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, "ú", nil, nil, "ú", nil, nil, "ú") end

declprops["ie-n"] = { desc = function(base, stems) return "soft o-stem" end, cat = function(base, stems) return "soft neuter o-stem" end }

decls["é-n"] = function(base, stems) add_decl(base, stems, "é", {}, "é", "é", "í", "ím",		"í", {}, "íma",		"é", "í", "ím", "é", "ích", "ími") add_decl(base, com.convert_paired_plain_to_palatal(stems.oblique_vowel_stem), nil, "ú", nil, nil, "ú", nil, nil, "ú") end

declprops["é-n"] = { desc = function(base, stems) return "soft o-stem" end, cat = function(base, stems) return "soft neuter o-stem" end }

decls["n-n"] = function(base, stems) add_decl(base, stems, "ene", "eni", "-", "-", "eni", "enem",		"eně", "enú", "enoma",		"ena", "en", "enóm", "ena", "eniech", "eny") end

declprops["n-n"] = { desc = function(base, stems) return "n-stem" end, cat = function(base, stems) return "neuter n-stem" end }

decls["tstem-n"] = function(base, stems) add_decl(base, stems, "ěte", "ěti", "-", "-", "ěti", "ětem",		"ětě", "atú", "atma",		"ata", "at", "atóm", "ata", "ětech", "aty") end

declprops["tstem-n"] = { desc = function(base, stems) return "t-stem" end, cat = function(base, stems) return "neuter t-stem" end }

decls["e-tstem-n"] = function(base, stems) add_decl(base, stems, "ete", "eti", "-", "-", "eti", "etem",		"etě", "atú", "atma",		"ata", "at", "atóm", "ata", "etech", "aty") end

declprops["e-tstem-n"] = { desc = function(base, stems) return "t-stem" end, cat = function(base, stems) return "neuter t-stem" end }

decls["adj"] = function(base, stems) local props = {} local propspec = table.concat(props, ".") if propspec ~= "" then propspec = "<" .. propspec .. ">"	end local adj_alternant_multiword_spec = require("Module:zlw-ocs-adjective").do_generate_forms({base.lemma .. propspec}) local function copy(from_slot, to_slot) base.forms[to_slot] = adj_alternant_multiword_spec.forms[from_slot] end if base.number ~= "pl" then if base.gender == "m" then copy("nom_m", "nom_s") copy("gen_mn", "gen_s") copy("dat_mn", "dat_s") copy("loc_mn", "loc_s") copy("ins_mn", "ins_s") elseif base.gender == "f" then copy("nom_f", "nom_s") copy("gen_f", "gen_s") copy("dat_f", "dat_s") copy("acc_f", "acc_s") copy("loc_f", "loc_s") copy("ins_f", "ins_s") else copy("nom_n", "nom_s") copy("gen_mn", "gen_s") copy("dat_mn", "dat_s") copy("acc_n", "acc_s") copy("loc_mn", "loc_s") copy("ins_mn", "ins_s") end if not base.forms.voc_s then iut.insert_forms(base.forms, "voc_s", base.forms.nom_s) end end if base.number ~= "sg" then if base.gender == "m" then copy("nom_mp", "nom_p") copy("acc_mfp", "acc_p") copy("nom_md", "nom_d") elseif base.gender == "f" then copy("nom_fp", "nom_p") copy("acc_mfp", "acc_p") copy("nom_fnd", "nom_d") else copy("nom_np", "nom_p") copy("acc_np", "acc_p") copy("nom_fnd", "nom_d") end copy("gen_p", "gen_p") copy("dat_p", "dat_p") copy("ins_p", "ins_p") copy("loc_p", "loc_p") copy("gen_d", "gen_d") copy("dat_d", "dat_d") end end

local function get_stemtype(base) if rfind(base.lemma, "ý$") then return "hard" elseif rfind(base.lemma, "í$") then return "soft" else return "possessive" end end

declprops["adj"] = { cat = function(base, stems) return {"adjectival POS", get_stemtype(base) .. " GENDER adjectival POS"} end, }

decls["indecl"] = function(base, stems) -- Indeclinable. Note that fully indeclinable nouns should not have a table at all rather than one all of whose forms -- are the same; but having an indeclinable declension is useful for nouns that may or may not be indeclinable, e.g. -- desatero "group of ten" or the plural of peso, which may be indeclinable 'pesos'. add_decl(base, stems, "-", "-", "-", "-", "-", "-",		"-", "-", "-", "-", "-", "-") end

declprops["indecl"] = { cat = function(base, stems) if base.adj then return {"adjectival POS", "indeclinable adjectival POS", "indeclinable GENDER adjectival POS"} else return {"indeclinable POS", "indeclinable GENPOS"} end end }

decls["manual"] = function(base, stems) -- Anything declined manually using overrides. We don't set any declensions except the nom_s (or nom_p if plurale	-- tantum). add(base, base.number == "pl" and "nom_p" or "nom_s", stems, "-") end

declprops["manual"] = { desc = "GENDER", cat = {}, }

local function set_pron_defaults(base) if base.gender or base.lemma ~= "ona" and base.number or base.animacy then error("Can't specify gender, number or animacy for pronouns") end

local function pron_props -- Return values are GENDER, NUMBER, ANIMACY, HAS_CLITIC. if base.lemma == "kto" then return "none", "sg", "pr", false elseif base.lemma == "čso" then return "none", "sg", "inan", false else error(("Unrecognized pronoun '%s'"):format(base.lemma)) end end

local gender, number, animacy, has_clitic = pron_props base.gender = gender base.actual_gender = gender base.number = number base.actual_number = number base.animacy = animacy base.actual_animacy = animacy base.has_clitic = has_clitic end

local function determine_pronoun_stems(base) if base.stem_sets then error("Reducible and vowel alternation specs cannot be given with pronouns") end base.stem_sets = base.decl = "pron" end

decls["pron"] = function(base, stems) if base.lemma == "kto" then add_decl(base, stems, "koho", "komu", nil, nil, "kom", "kým") elseif base.lemma == "čso" then add_decl(base, stems, {"čeho", "čso"}, "čemu", nil, nil, "čem", "čím") else error(("Internal error: Unrecognized pronoun lemma '%s'"):format(base.lemma)) end end

declprops["pron"] = { desc = "GENDER pronoun", cat = {}, }

local function set_num_defaults(base) if base.gender or base.animacy then error("Can't specify gender, number or animacy for numeral") end

local function num_props -- Return values are GENDER, NUMBER, ANIMACY, HAS_CLITIC. return "none", "pl", "none", false end

local gender, number, animacy, has_clitic = num_props base.gender = gender base.actual_gender = gender base.number = number base.actual_number = number base.animacy = animacy base.actual_animacy = animacy base.has_clitic = has_clitic end

local function determine_numeral_stems(base) if base.stem_sets then error("Reducible and vowel alternation specs cannot be given with numerals") end local stem = rmatch(base.lemma, "^(.*)" .. com.vowel_c .. "$") or base.lemma base.stem_sets = base.decl = "num" end

decls["num"] = function(base, stems) local after_prep_footnote =	"[after a preposition]" if base.lemma == "jeden" then add_pl_only_decl(base, "", "jednoho", "jednomu", "jeden", "jednom", "jedniem") elseif base.lemma == "dva" or base.lemma == "dvě" then -- in compound numbers; stem is dv- add_pl_only_decl(base, stems, "ú", "ěma", "-", "ú", "ěma") elseif base.lemma == "tři" or base.lemma == "čtyři" then local is_three = base.lemma == "tři" add_pl_only_decl(base, stems, is_three and "í" or "", "em", "-", "ech", "mi") elseif base.lemma == "devět" then add_pl_only_decl(base, "", {"devieti", "devěti"}, {"devieti", "devěti"}, "-", {"devieti", "devěti"}, {"devieti", "devěti"}, stems.footnotes) elseif base.lemma == "desět" then add_pl_only_decl(base, "", {"desieti", "desěti"}, {"desieti", "desěti"}, "-", {"desieti", "desěti"}, {"desieti", "desěti"}, stems.footnotes) elseif base.lemma == "sta" or base.lemma == "stě" or base.lemma == "set" then add_pl_only_decl(base, "", "set", "stóm", "-", "stech", "sty", stems.footnotes) elseif rfind(base.lemma, "[cs]et$") then -- deset and all numbers ending in -cet (dvacet, třicet, čtyřicet and inverted compound		-- numerals such as pětadvacet "25" and dvaatřicet "32") local begin = rmatch(base.lemma, "^(.*)et$") add_pl_only_decl(base, stems, "i", "i", "-", "i", "i") add_pl_only_decl(base, begin, "íti", "íti", "-", "íti", "íti", stems.footnotes) elseif rfind(base.lemma, "oje$") then -- dvoje, troje -- stem is without -e add_pl_only_decl(base, stems, "ích", "ím", "-", "ích", "ími") elseif rfind(base.lemma, "ery$") then -- čtvery, patery, šestery, sedmery, osmery, devatery, desatery -- stem is without -y add_pl_only_decl(base, stems, "ých", "ým", "-", "ých", "ými") else add_pl_only_decl(base, stems, "i", "i", "-", "i", "i") end end

declprops["num"] = { desc = "GENDER numeral", cat = {}, }

local function set_det_defaults(base) if base.gender or base.number or base.animacy then error("Can't specify gender, number or animacy for determiner") end

local function det_props -- Return values are GENDER, NUMBER, ANIMACY, HAS_CLITIC. return "none", "none", "none", false end

local gender, number, animacy, has_clitic = det_props base.gender = gender base.actual_gender = gender base.number = number base.actual_number = number base.animacy = animacy base.actual_animacy = animacy base.has_clitic = has_clitic end

local function determine_determiner_stems(base) if base.stem_sets then error("Reducible and vowel alternation specs cannot be given with determiners") end local stem = rmatch(base.lemma, "^(.*)" .. com.vowel_c .. "$") or base.lemma base.stem_sets = base.decl = "det" end

decls["det"] = function(base, stems) add_sg_decl(base, stems, "a", "a", "-", nil, "a", "a") end

declprops["det"] = { desc = "GENDER determiner", cat = {}, }

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

--[=[ Parse a single override spec (e.g. 'nomplé:ové' or 'ins:autodráhou:autodrahou[rare]') and return two values: the slot(s) the override applies to, and an object describing the override spec. The input is actually a list where the footnotes have been separated out; for example, given the spec 'inspl:čobotami:čobotámi[rare]:čobitmi[archaic]', the input will be a list {"inspl:čobotami:čobotámi", "[rare]", ":čobitmi", "[archaic]", ""}. The object returned for 'ins:autodráhou:autodrahou[rare]' looks like this:

{ full = true, values = { {	 form = "autodráhou" },	{	 form = "autodrahou", footnotes = {"[rare]"} } } }

The object returned for 'nomplé:ové' looks like this:

{ values = { {	 form = "é", },	{	 form = "ové", } } } ]=] local function parse_override(segments) local retval = {values = {}} local part = segments[1] local slots = {} while true do		local case = usub(part, 1, 3) if cases[case] then -- ok		else error(("Unrecognized case '%s' in override: '%s'"):format(case, table.concat(segments))) end part = usub(part, 4) local slot if rfind(part, "^pl") then part = usub(part, 3) slot = case .. "_p" elseif rfind(part, "^du") then part = usub(part, 3) slot = case .. "_d" else slot = case .. "_s" end table.insert(slots, slot) if rfind(part, "^%+") then part = usub(part, 2) else break end end if rfind(part, "^:") then retval.full = true part = usub(part, 2) end segments[1] = part local colon_separated_groups = iut.split_alternating_runs_and_strip_spaces(segments, ":") for i, colon_separated_group in ipairs(colon_separated_groups) do		local value = {} local form = colon_separated_group[1] if form == "" then error(("Use - to indicate an empty ending for slot%s '%s': '%s'"):format(#slots > 1 and "s" or "", table.concat(slots), table.concat(segments))) elseif form == "-" then value.form = "" else value.form = form end value.footnotes = fetch_footnotes(colon_separated_group) table.insert(retval.values, value) end return slots, retval end

--[=[ Parse an indicator spec (text consisting of angle brackets and zero or more dot-separated indicators within them). Return value is an object of the form

{ overrides = { SLOT = {OVERRIDE, OVERRIDE, ...}, -- as returned by parse_override ... },  forms = {}, -- forms for a single spec alternant; see `forms` below footnotes = {"FOOTNOTE", "FOOTNOTE", ...}, -- may be missing stems = { -- may be missing {	 reducible = TRUE_OR_FALSE, footnotes = {"FOOTNOTE", "FOOTNOTE", ...}, -- may be missing -- The following fields are filled in by determine_stems vowel_stem = "STEM", nonvowel_stem = "STEM", oblique_slots = one of {nil, "gen_p", "all", "all-oblique"}, oblique_vowel_stem = "STEM" or nil (only needs to be set if oblique_slots is non-nil), oblique_nonvowel_stem = "STEM" or nil (only needs to be set if oblique_slots is non-nil), },	... },  gender = "GENDER", -- "m", "f", "n" number = "NUMBER", -- "sg", "pl"; may be missing animacy = "ANIMACY", -- "inan", "an"; may be missing hard = true, -- may be missing soft = true, -- may be missing mixed = true, -- may be missing surname = true, -- may be missing istem = true, -- may be missing ["-istem"] = true, -- may be missing tstem = true, -- may be missing nstem = true, -- may be missing tech = true, -- may be missing foreign = true, -- may be missing mostlyindecl = true, -- may be missing indecl = true, -- may be missing manual = true, -- may be missing adj = true, -- may be missing decllemma = "DECLENSION-LEMMA", -- may be missing declgender = "DECLENSION-GENDER", -- may be missing declnumber = "DECLENSION-NUMBER", -- may be missing

-- The following additional fields are added by other functions: orig_lemma = "ORIGINAL-LEMMA", -- as given by the user orig_lemma_no_links = "ORIGINAL-LEMMA-NO-LINKS", -- links removed lemma = "LEMMA", -- `orig_lemma_no_links`, converted to singular form if plural and lowercase if all-uppercase forms = { SLOT = { {		form = "FORM", footnotes = {"FOOTNOTE", "FOOTNOTE", ...} -- may be missing },	 ...	},	...  },  decl = "DECL", -- declension, e.g. "hard-m" vowel_stem = "VOWEL-STEM", -- derived from vowel-ending lemmas nonvowel_stem = "NONVOWEL-STEM", -- derived from non-vowel-ending lemmas } ]=] local function parse_indicator_spec(angle_bracket_spec) local inside = rmatch(angle_bracket_spec, "^<(.*)>$") assert(inside) local base = {overrides = {}, forms = {}} if inside ~= "" then local segments = iut.parse_balanced_segment_run(inside, "[", "]") local dot_separated_groups = iut.split_alternating_runs_and_strip_spaces(segments, "%.") for i, dot_separated_group in ipairs(dot_separated_groups) do			local part = dot_separated_group[1] local case_prefix = usub(part, 1, 3) if cases[case_prefix] then local slots, override = parse_override(dot_separated_group) for _, slot in ipairs(slots) do					if base.overrides[slot] then error(("Two overrides specified for slot '%s'"):format(slot)) else base.overrides[slot] = {override} end end elseif part == "" then if #dot_separated_group == 1 then error("Blank indicator: '" .. inside .. "'") end base.footnotes = fetch_footnotes(dot_separated_group) elseif rfind(part, "^[-*#ě]*$") or rfind(part, "^[-*#ě]*,") then if base.stem_sets then error("Can't specify reducible/vowel-alternant indicator twice: '" .. inside .. "'") end local comma_separated_groups = iut.split_alternating_runs_and_strip_spaces(dot_separated_group, ",") local stem_sets = {} for i, comma_separated_group in ipairs(comma_separated_groups) do					local pattern = comma_separated_group[1] local orig_pattern = pattern local reducible, vowelalt, oblique_slots if pattern == "-" then -- default reducible, no vowel alt else local before, after before, reducible, after = rmatch(pattern, "^(.-)(%-?%*)(.-)$") if before then pattern = before .. after reducible = reducible == "*" end if pattern ~= "" then if not rfind(pattern, "^##?ě?$") then error("Unrecognized vowel-alternation pattern '" .. pattern .. "', should be one of #, ##, #ě or ##ě: '" .. inside .. "'") end if pattern == "#ě" or pattern == "##ě" then vowelalt = "quant-ě" else vowelalt = "quant" end -- `oblique_slots` will be later changed to "all" if the lemma ends in a consonant. if pattern == "##" or pattern == "##ě" then oblique_slots = "all-oblique" else oblique_slots = "gen_p" end end end table.insert(stem_sets, {						reducible = reducible,						vowelalt = vowelalt,						oblique_slots = oblique_slots,						footnotes = fetch_footnotes(comma_separated_group)					}) end base.stem_sets = stem_sets elseif #dot_separated_group > 1 then error("Footnotes only allowed with slot overrides, reducible or vowel alternation specs or by themselves: '" .. table.concat(dot_separated_group) .. "'") elseif part == "m" or part == "f" or part == "n" then if base.gender then error("Can't specify gender twice: '" .. inside .. "'") end base.gender = part elseif part == "sg" or part == "du" or part == "pl" then if base.number then error("Can't specify number twice: '" .. inside .. "'") end base.number = part elseif part == "pr" or part == "anml" or part == "inan" then if base.animacy then error("Can't specify animacy twice: '" .. inside .. "'") end base.animacy = part elseif part == "hard" or part == "soft" or part == "mixed" or part == "ija" or part == "ustem" or part == "vstem" or part == "rstem" or part == "istem" or				part == "-istem" or part == "tstem" or part == "nstem" or part == "tech" or part == "foreign" or				part == "indecl" or part == "pron" or part == "det" or part == "num" or				-- Use 'velar' with words like petanque and Braque that end with a pronounced velar (and hence are declined				-- like velars) but not with a spelled velar; use '-velar' with words like hadíth that end with a spelled but -- silent velar. part == "collapse_ee" or part == "persname" or part == "c_as_k" or part == "velar" or part == "-velar" then if base[part] then error("Can't specify '" .. part .. "' twice: '" .. inside .. "'") end base[part] = true elseif part == "+" then if base.adj then error("Can't specify '+' twice: '" .. inside .. "'") end base.adj = true elseif part == "!" then if base.manual then error("Can't specify '!' twice: '" .. inside .. "'") end base.manual = true elseif rfind(part, "^mixedistem:") then if base.mixedistem then error("Can't specify 'mixedistem:' twice: '" .. inside .. "'") end base.mixedistem = rsub(part, "^mixedistem:", "") elseif rfind(part, "^decllemma:") then if base.decllemma then error("Can't specify 'decllemma:' twice: '" .. inside .. "'") end base.decllemma = rsub(part, "^decllemma:", "") elseif rfind(part, "^declgender:") then if base.declgender then error("Can't specify 'declgender:' twice: '" .. inside .. "'") end base.declgender = rsub(part, "^declgender:", "") elseif rfind(part, "^declnumber:") then if base.declnumber then error("Can't specify 'declnumber:' twice: '" .. inside .. "'") end base.declnumber = rsub(part, "^declnumber:", "") else error("Unrecognized indicator '" .. part .. "': '" .. inside .. "'") end end end return base end

local function is_regular_noun(base) return not base.adj and not base.pron and not base.det and not base.num end

local function process_declnumber(base) base.actual_number = base.number if base.declnumber then if base.declnumber == "sg" or base.declnumber == "du" or base.declnumber == "pl" then base.number = base.declnumber else error(("Unrecognized value '%s' for 'declnumber', should be 'sg' or 'pl'"):format(base.declnumber)) end end end

local function set_defaults_and_check_bad_indicators(base) -- Set default values. local regular_noun = is_regular_noun(base) if base.pron then set_pron_defaults(base) elseif base.det then set_det_defaults(base) elseif base.num then set_num_defaults(base) elseif not base.adj then if not base.gender then if base.manual then base.gender = "none" else error("For nouns, gender must be specified") end end base.number = base.number or "allthree" process_declnumber(base) base.animacy = base.animacy or "inan" base.actual_gender = base.gender base.actual_animacy = base.animacy if base.declgender then if base.declgender == "m-an" then base.gender = "m" base.animacy = "pr" elseif base.declgender == "m-in" then base.gender = "m" base.animacy = "inan" elseif base.declgender == "f" or base.declgender == "n" then base.gender = base.declgender else error(("Unrecognized value '%s' for 'declgender', should be 'm-an', 'm-in', 'f' or 'n'"):format(base.declgender)) end end end -- Check for bad indicator combinations. if (base.hard and 1 or 0) + (base.soft and 1 or 0) + (base.mixed and 1 or 0) > 1 then error("At most one of 'hard', 'soft' and 'mixed' can be specified") end if base.istem and base["-istem"] then error("'istem' and '-istem' cannot be specified together") end if (base.istem or base["-istem"]) then if not regular_noun then error("'istem' and '-istem' can only be specified with regular nouns") end end if base.declgender and not regular_noun then error("'declgender' can only be specified with regular nouns") end end

local function set_all_defaults_and_check_bad_indicators(alternant_multiword_spec) local is_multiword = #alternant_multiword_spec.alternant_or_word_specs > 1 iut.map_word_specs(alternant_multiword_spec, function(base)		set_defaults_and_check_bad_indicators(base)		base.multiword = is_multiword -- FIXME: not currently used; consider deleting		alternant_multiword_spec.has_clitic = alternant_multiword_spec.has_clitic or base.has_clitic		if base.pron then			alternant_multiword_spec.saw_pron = true		else			alternant_multiword_spec.saw_non_pron = true		end		if base.det then			alternant_multiword_spec.saw_det = true		else			alternant_multiword_spec.saw_non_det = true		end		if base.num then			alternant_multiword_spec.saw_num = true		else			alternant_multiword_spec.saw_non_num = true		end	end) end

local function undo_second_palatalization(base, word, is_adjective) local function try(from, to) local stem = rmatch(word, "^(.*)" .. from .. "$") if stem then return stem .. to		end return nil end return is_adjective and try("št", "sk") or		is_adjective and try("čt", "ck") or		try("c", "k") or -- FIXME, this could be wrong and c correct try("ř", "r") or		try("z", "h") or -- FIXME, this could be wrong and z or g correct try("š", "ch") or		word end

-- For a plural-only lemma, synthesize a likely singular lemma. It doesn't have to be -- theoretically correct as long as it generates all the correct plural forms. local function synthesize_singular_lemma(base) if not base.stem_sets then base.stem_sets = – end

local lemma_determined -- Loop over all stem sets in case the user specified multiple ones (e.g. '*,-*'). If we try to reconstruct -- different lemmas for different stem sets, we'll throw an error below. for _, stems in ipairs(base.stem_sets) do		local stem, lemma while true do			if base.indecl then -- If specified as indeclinable, leave it alone; e.g. 'pesos' indeclinable plural of peso. lemma = base.lemma break elseif base.gender == "m" then stem = rmatch(base.lemma, "^(.*)i$") if stem then if base.soft then -- Blíženci "Gemini" -- Since the nominative singular has no ending. lemma = com.convert_paired_plain_to_palatal(stem, ending) else lemma = undo_second_palatalization(base, stem) end else stem = rmatch(base.lemma, "^(.*)ové$") or rmatch(base.lemma, "^(.*)[éyě]$") or rmatch(base.lemma, "^(.*)ie$") if stem then -- manželé "married couple", Velšané "Welsh people" lemma = stem else error(("Masculine plural-only lemma '%s' should end in -i, -ové or -é"):format(base.lemma)) end end if stems.reducible == nil then if rfind(lemma, com.cons_c .. "[ck]$") and not com.is_monosyllabic(base.lemma) then stems.reducible = true end if stems.reducible then lemma = dereduce(base, lemma) end end break elseif base.gender == "f" then stem = rmatch(base.lemma, "^(.*)y$") if stem then lemma = stem .. "a" break end stem = rmatch(base.lemma, "^(.*)[eě]$") if stem then lemma = base.lemma break end stem = rmatch(base.lemma, "^(.*)i$") if stem then -- i-stems. lemma = stem base.istem = true break end error(("Feminine plural-only lemma '%s' should end in -y, -ě, -e or -i"):format(base.lemma)) elseif base.gender == "n" then stem = rmatch(base.lemma, "^(.*)a$") if stem then lemma = stem .. "o" break end stem = rmatch(base.lemma, "^(.*)[eěí]$") if stem then -- singular lemma also in -e, -ě or -í; e.g. věčná loviště "happy hunting ground" lemma = base.lemma break end error(("Neuter plural-only lemma '%s' should end in -a, -í, -ě or -e"):format(base.lemma)) else error(("Internal error: Unrecognized gender '%s'"):format(base.gender)) end end if lemma_determined and lemma_determined ~= lemma then error(("Attempt to set two different singular lemmas '%s' and '%s'"):format(lemma_determined, lemma)) end lemma_determined = lemma end base.lemma = lemma_determined end

-- For an adjectival lemma, synthesize the masc singular form. local function synthesize_adj_lemma(base) local stem if base.indecl then base.decl = "indecl" stem = base.lemma else local gender, number local function sub_ov(stem) stem = stem:gsub("ov$", "óv") return stem end while true do			if base.number == "pl" then if base.gender == "m" then stem = rmatch(base.lemma, "^(.*)í$") if stem then if base.soft then -- nothing to do						else if base.animacy ~= "pr" then error(("Masculine plural-only adjectival lemma '%s' ending in -í can only be animate unless '.soft' is specified"):									format(base.lemma)) end base.lemma = undo_second_palatalization(base, stem, "is adjective") .. "ý" end break end stem = rmatch(base.lemma, "^(.*)é$") if stem then if base.animacy == "pr" then error(("Masculine plural-only adjectival lemma '%s' ending in -é must be inanimate"):								format(base.lemma)) end base.lemma = stem .. "ý" break end stem = rmatch(base.lemma, "^(.*ov)i$") or rmatch(base.lemma, "^(.*in)i$") if stem then if base.animacy ~= "pr" then error(("Masculine plural-only possessive adjectival lemma '%s' ending in -i must be animate"):								format(base.lemma)) end base.lemma = sub_ov(stem) break end stem = rmatch(base.lemma, "^(.*ov)y$") or rmatch(base.lemma, "^(.*in)y$") if stem then if base.animacy == "pr" then error(("Masculine plural-only possessive adjectival lemma '%s' ending in -y must be inanimate"):								format(base.lemma)) end base.lemma = sub_ov(stem) break end if base.animacy == "pr" then error(("Animate masculine plural-only adjectival lemma '%s' should end in -í, -ovi or -ini"):							format(base.lemma)) elseif base.soft then error(("Soft masculine plural-only adjectival lemma '%s' should end in -í"):format(base.lemma)) else error(("Inanimate masculine plural-only adjectival lemma '%s' should end in -é, -ovy or -iny"):							format(base.lemma)) end elseif base.gender == "f" then stem = rmatch(base.lemma, "^(.*)é$") -- hard adjective if stem then base.lemma = stem .. "ý" break end stem = rmatch(base.lemma, "^(.*)í$") -- soft adjective if stem then break end stem = rmatch(base.lemma, "^(.*ov)y$") or rmatch(base.lemma, "^(.*in)y$") -- possessive adjective if stem then base.lemma = sub_ov(stem) break end error(("Feminine plural-only adjectival lemma '%s' should end in -é, -í, -ovy or -iny"):format(base.lemma)) else stem = rmatch(base.lemma, "^(.*)á$") -- hard adjective if stem then base.lemma = stem .. "ý" break end stem = rmatch(base.lemma, "^(.*)í$") -- soft adjective if stem then break end stem = rmatch(base.lemma, "^(.*ov)a$") or rmatch(base.lemma, "^(.*in)a$") -- possessive adjective if stem then base.lemma = sub_ov(stem) break end error(("Neuter plural-only adjectival lemma '%s' should end in -á, -í, -ova or -ina"):format(base.lemma)) end else if base.gender == "m" then stem = rmatch(base.lemma, "^(.*)[ýí]$") or rmatch(base.lemma, "^(.*)óv$") or rmatch(base.lemma, "^(.*)in$") if stem then break end error(("Masculine adjectival lemma '%s' should end in -ý, -í, -óv or -in"):format(base.lemma)) elseif base.gender == "f" then stem = rmatch(base.lemma, "^(.*)á$") if stem then base.lemma = stem .. "ý" break end stem = rmatch(base.lemma, "^(.*)í$") if stem then break end stem = rmatch(base.lemma, "^(.*ov)a$") or rmatch(base.lemma, "^(.*in)a$") if stem then base.lemma = sub_ov(stem) break end error(("Feminine adjectival lemma '%s' should end in -á, -í, -ova or -ina"):format(base.lemma)) else stem = rmatch(base.lemma, "^(.*)í$") if stem then break end stem = rmatch(base.lemma, "^(.*ov)o$") or rmatch(base.lemma, "^(.*in)o$") if stem then base.lemma = sub_ov(stem) break end error(("Neuter adjectival lemma '%s' should end in -é, -í, -ovo or -ino"):format(base.lemma)) end end end base.decl = "adj" end

-- Now set the stem sets if not given. -- Now set the stem sets if not given. if not base.stem_sets then base.stem_sets = end for _, stems in ipairs(base.stem_sets) do -- Set the stems. stems.vowel_stem = stem stems.nonvowel_stem = stem end end

-- Determine the declension based on the lemma, gender and number. The declension is set in base.decl. In the process, -- we set either base.vowel_stem (if the lemma ends in a vowel) or base.nonvowel_stem (if the lemma does not end in a -- vowel), which is used by determine_stems. In some cases (specifically with certain foreign nouns), we set -- base.lemma to a new value; this is as if the user specified 'decllemma:'. local function determine_declension(base) if base.indecl then base.decl = "indecl" base.nonvowel_stem = base.lemma return end -- Determine declension stem = rmatch(base.lemma, "^(.*)a$") if stem then if base.gender == "m" then if base.animacy ~= "pr" then error("Masculine lemma in -a must be animate") end base.decl = "a-m" elseif base.gender == "f" then if base.hard then -- e.g. doňa, which seems not to have soft alternates as piraňa does (despite IJP; but see the note at the				-- bottom) base.decl = "hard-f" elseif rfind(stem, "e$") then -- idea, diarea (subtype '.tech'), Korea, etc.				base.decl = "ea-f" elseif rfind(stem, "i$") then -- signoria, sinfonia, paranoia, etc.				base.decl = "ia-f" elseif rfind(stem, "[ou]$") then -- stoa, kongrua, Samoa, Nikaragua, etc.				base.decl = "oa-f" elseif not base.persname and rfind(stem, "^.*[ňj]$") then -- maracuja, papája, sója; piraňa etc. Also Keňa, Troja/Trója, Amudarja. -- Not Táňa, Darja, which decline like gejša, skica, etc. (subtype of hard feminines). base.decl = "mixed-f" else base.decl = "hard-f" end elseif base.gender == "n" then if rfind(stem, "m$") then base.decl = "ma-n" else error("Lemma ending in -a and neuter must end in -ma") end end base.vowel_stem = stem return end local ending stem, ending = rmatch(base.lemma, "^(.*)ěnín$") if stem then if base.gender == "m" then base.decl = "ěnín-m" end base.vowel_stem = stem return end stem, ending = rmatch(base.lemma, "^(.*)enín$") if stem then if base.gender == "m" then base.decl = "enín-m" end base.vowel_stem = stem return end stem, ending = rmatch(base.lemma, "^(.*)ie$") if stem then if base.gender == "m" then base.decl = "ija-m" elseif base.gender == "f" then base.decl = "í-f" elseif base.gender == "n" then base.decl = "ie-n" end base.vowel_stem = stem return end

stem, ending = rmatch(base.lemma, "^(.*)e$") if stem then if base.gender == "m" then if base.hard then -- -e be damned; e.g. Sofokles with hard stem 'Sofokle-' (genitive 'Sofoklea', dative 'Sofokleovi', etc.) base.nonvowel_stem = base.lemma base.decl = "hard-m" return elseif rfind(stem, "i") then -- zombie, hippie, yuppie, rowdie base.decl = "ie-m" return end elseif base.gender == "f" then base.decl = "e-f" else if base.tstem then base.decl = "e-tstem-n" else base.decl = "soft-n" end end base.vowel_stem = stem return end stem, ending = rmatch(base.lemma, "^(.*)ě$") if stem then if base.gender == "m" then base.decl = "ě-m" elseif base.gender == "f" then base.decl = "soft-f" else if base.nstem then base.decl = "n-n" else stem = com.onlyndt(stem) base.decl = "tstem-n" end end base.vowel_stem = stem return end stem = rmatch(base.lemma, "^(.*)o$") if stem then if base.hard then base.decl = "hard-n" elseif rfind(stem, "[aeiuy]$") then -- These have gen pl in -í and often other soft plural endings. base.decl = "semisoft-n" else base.decl = "hard-n" end base.vowel_stem = stem return end stem = rmatch(base.lemma, "^(.*[iy])$") if stem then if base.gender == "m" then base.decl = "i-m" elseif base.gender == "f" then if base.soft then -- Uruguay, Paraguay base.decl = "soft-f" else -- máti, pramáti; note also indeclinable tsunami/cunami, okapi if base.istem then base.decl = "istem-f" elseif base.rstem then base.decl = "r-f" else base.decl = "i-f" end if stem:find("i$") then stem = stem:gsub("i$", "") else error("Feminine nouns in -y are either soft or indeclinable; use '.soft' or '.indecl' as needed") end end else error("Neuter nouns in -i are indeclinable; use '.indecl' if needed") end base.vowel_stem = stem return end stem = rmatch(base.lemma, "^(.*u)$") if stem then if base.gender == "m" then -- Cf. emu, guru, etc.			base.decl = "u-m" elseif base.gender == "f" then -- Only one I know is budižkničemu, which is indeclinable in the singular and declines in the plural as -- if written 'budižkničema'. error("Feminine nouns in -u are indeclinable; use '.indecl' if needed") else error("Neuter nouns in -u are indeclinable; use '.indecl' if needed") end base.vowel_stem = stem return end stem = rmatch(base.lemma, "^(.*)í$") if stem then if base.gender == "m" then if base.ija then base.decl = "ija-m" else base.decl = "ijo-m" end elseif base.gender == "f" then base.decl = "í-f" else base.decl = "í-n" end base.vowel_stem = stem return end stem = rmatch(base.lemma, "^(.*)é$") if stem then if base.gender == "m" then base.decl = "é-m" elseif base.gender == "f" then base.decl = "é-f" else base.decl = "é-n" end base.vowel_stem = stem return end stem = rmatch(base.lemma, "^(.*" .. com.cons_c .. ")$")	if stem then if base.gender == "m" then if base.tstem then stem = rmatch(base.lemma, "^(.*)et$") base.decl = "tstem-m" elseif base.ustem then base.decl = "u-m" elseif base.nstem then base.decl = "n-m" elseif base.hard then base.decl = "hard-m" elseif base.soft then base.decl = "soft-m" elseif base.mixed then base.decl = "mixed-m" elseif rfind(base.lemma, com.inherently_soft_c .. "$") or rfind(base.lemma, "tel$") then base.decl = "soft-m" elseif base.istem then base.decl = "istem-m" else base.decl = "hard-m" end elseif base.gender == "f" then if base.soft then base.decl = "cons-f" elseif base.vstem or rfind(base.lemma, "ev$") then base.decl = "v-f" stem = rmatch(stem, "(.*)ev$") else base.decl = "istem-f" end elseif base.gender == "n" then if base.foreign then stem = rmatch(base.lemma, "^(.*)um$") or rmatch(base.lemma, "^(.*)on$") if not stem then error("Unrecognized neuter foreign ending, should be -um or -on") end base.decl = "hard-n" -- set the lemma here as if decllemma: were given base.lemma = stem .. "o" base.vowel_stem = stem return else error("Neuter nouns ending in a consonant should use '.foreign' or '.decllemma:...'") end end base.nonvowel_stem = stem return end error("Unrecognized ending for lemma: '" .. base.lemma .. "'") end

-- Determine the default value for the 'reducible' flag. local function determine_default_reducible(base) -- Nouns in vowels other than -a/o as well as masculine nouns ending in all vowels don't have null endings so not -- reducible. Note, we are never called on adjectival nouns. if rfind(base.lemma, "[iyuíeě]$") or base.gender == "m" and rfind(base.lemma, "[ao]$") or base.tstem then base.default_reducible = false return end

local stem stem = rmatch(base.lemma, "^(.*" .. com.cons_c .. ")$")	if stem then -- When analyzing existing manual declensions in -ec and -ek, 290 were reducible vs. 23 non-reducible. Of these -- 23, 15 were monosyllabic (and none of the 290 reducible nouns were monosyllabic) -- and two of these were -- actually reducible but irregularly: švec "shoemaker" (gen sg 'ševce') and žnec "reaper (person)" -- (gen sg. 'žence'). Of the remaining 8 multisyllabic non-reducible words, two were actually reducible but -- irregularly: stařec "old man" (gen sg 'starce') and tkadlec "weaver" (gen sg 'tkalce'). The remaining -- six consisted of 5 compounds of monosyllabic words: dotek, oblek, kramflek, pucflek, -- pokec, plus česnek, which should be reducible but would lead to an impossible consonant cluster. if base.gender == "m" and rfind(stem, "e[ck]$") and not com.is_monosyllabic(stem) then base.default_reducible = true elseif base.gender == "f" and rfind(stem, "eň$") then -- pochodeň "torch", píseň "leather", žeň "harvest"; not reveň "rhubarb" or dřeň "pulp", -- which need an override. base.default_reducible = true else base.default_reducible = false end return end if base.number == "sg" then base.default_reducible = false return end if rfind(base.lemma, "isko$") then -- e.g. středisko base.default_reducible = "mixed" return end stem = rmatch(base.lemma, "^(.*)" .. com.vowel_c .. "$") if not stem then error(("Internal error: Something wrong, lemma '%s' doesn't end in consonant or vowel"):format(base.lemma)) end -- Substitute 'ch' with a single character to make the following code simpler. stem = stem:gsub("ch", com.TEMP_CH) if rfind(stem, com.cons_c .. "[lr]" .. com.cons_c .. "$") then -- vrba, vlha; not reducible. (But note jablko, reducible; needs override.) base.default_reducible = false elseif not base.foreign and (rfind(stem, com.cons_c .. "[bkhlrmnv]$") or base.c_as_k and rfind(stem, com.cons_c .. "c$")) then -- ayahuasca has gen pl 'ayahuasek' base.default_reducible = true elseif base.foreign and rfind(stem, com.cons_c .. "r$") then -- Foreign nouns in -CCum seem generally non-reducible in the gen pl except for those in -Crum like centrum, -- Examples: album, verbum, signum, interregnum, sternum. infernum has gen pl 'infern/inferen'. base.default_reducible = true else base.default_reducible = false end end

-- Determine the stems to use for each stem set: vowel and nonvowel stems, for singular -- and plural. We assume that one of base.vowel_stem or base.nonvowel_stem has been -- set in determine_declension, depending on whether the lemma ends in -- a vowel. We construct all the rest given the reducibility, vowel alternation spec and -- any explicit stems given. We store the determined stems inside of the stem-set objects -- in `base.stem_sets`, meaning that if the user gave multiple reducible or vowel-alternation -- patterns, we will compute multiple sets of stems. The reason is that the stems may vary -- depending on the reducibility and vowel alternation. local function determine_stems(base) if not base.stem_sets then base.stem_sets = – end -- Set default reducible and check for default mixed reducible, which needs to be expanded into two entries. local default_mixed_reducible = false for _, stems in ipairs(base.stem_sets) do		if stems.reducible == nil then stems.reducible = base.default_reducible end if stems.reducible == "mixed" then default_mixed_reducible = true end end if default_mixed_reducible then local new_stem_sets = {} for _, stems in ipairs(base.stem_sets) do			if stems.reducible == "mixed" then local non_reducible_copy = m_table.shallowcopy(stems) non_reducible_copy.reducible = false stems.reducible = true table.insert(new_stem_sets, stems) table.insert(new_stem_sets, non_reducible_copy) else table.insert(new_stem_sets, stems) end end base.stem_sets = new_stem_sets end

-- Now determine all the stems for each stem set. for _, stems in ipairs(base.stem_sets) do		local lemma_is_vowel_stem = not not base.vowel_stem if base.vowel_stem then stems.vowel_stem = base.vowel_stem stems.nonvowel_stem = stems.vowel_stem -- Apply vowel alternation first in cases like jádro -> jader; apply_vowel_alternation will throw an error -- if the vowel being modified isn't the last vowel in the stem. stems.oblique_nonvowel_stem = com.apply_vowel_alternation(stems.vowelalt, stems.nonvowel_stem) if stems.reducible then stems.nonvowel_stem = dereduce(base, stems.nonvowel_stem) stems.oblique_nonvowel_stem = dereduce(base, stems.oblique_nonvowel_stem) end else stems.nonvowel_stem = base.nonvowel_stem -- The user specified #, #ě, ## or ##ě and we're dealing with a term like masculine bóh or feminine -- sól that ends in a consonant. In this case, all slots except the nom_s and maybe acc_s have vowel -- alternation. if stems.oblique_slots then stems.oblique_slots = "all" end stems.oblique_nonvowel_stem = com.apply_vowel_alternation(stems.vowelalt, stems.nonvowel_stem) if stems.reducible then stems.vowel_stem = com.reduce(base.nonvowel_stem) if not stems.vowel_stem then error("Unable to reduce stem '" .. base.nonvowel_stem .. "'") end else stems.vowel_stem = base.nonvowel_stem end end stems.oblique_vowel_stem = com.apply_vowel_alternation(stems.vowelalt, stems.vowel_stem) end end

local function detect_indicator_spec(base) if base.pron then determine_pronoun_stems(base) elseif base.det then determine_determiner_stems(base) elseif base.num then determine_numeral_stems(base) elseif base.adj then process_declnumber(base) synthesize_adj_lemma(base) elseif base.manual then if base.stem_sets then -- FIXME, maybe this should be allowed? error("Reducible and vowel alternation specs cannot be given with manual declensions") end base.stem_sets = base.decl = "manual" else if base.number == "pl" then synthesize_singular_lemma(base) end determine_declension(base) determine_default_reducible(base) determine_stems(base) end end

local function detect_all_indicator_specs(alternant_multiword_spec) -- Keep track of all genders seen in the singular and plural so we can determine whether to add the term to -- Category:Czech nouns that change gender in the plural. alternant_multiword_spec.sg_genders = {} alternant_multiword_spec.pl_genders = {} iut.map_word_specs(alternant_multiword_spec, function(base)		detect_indicator_spec(base)		if base.number ~= "pl" then			alternant_multiword_spec.sg_genders[base.actual_gender] = true		end		if base.number ~= "sg" then			-- All t-stem masculines are neuter in the plural.			local plgender				plgender = base.actual_gender			alternant_multiword_spec.pl_genders[plgender] = true		end	end) if (alternant_multiword_spec.saw_pron and 1 or 0) + (alternant_multiword_spec.saw_det and 1 or 0) + (alternant_multiword_spec.saw_num and 1 or 0) > 1 then error("Can't combine pronouns, determiners and/or numerals") end end

local propagate_multiword_properties

local function propagate_alternant_properties(alternant_spec, property, mixed_value, nouns_only) local seen_property for _, multiword_spec in ipairs(alternant_spec.alternants) do		propagate_multiword_properties(multiword_spec, property, mixed_value, nouns_only) if seen_property == nil then seen_property = multiword_spec[property] elseif multiword_spec[property] and seen_property ~= multiword_spec[property] then seen_property = mixed_value end end alternant_spec[property] = seen_property end

propagate_multiword_properties = function(multiword_spec, property, mixed_value, nouns_only) local seen_property = nil local last_seen_nounal_pos = 0 local word_specs = multiword_spec.alternant_or_word_specs or multiword_spec.word_specs for i = 1, #word_specs do		local is_nounal if word_specs[i].alternants then propagate_alternant_properties(word_specs[i], property, mixed_value) is_nounal = not not word_specs[i][property] elseif nouns_only then is_nounal = is_regular_noun(word_specs[i]) else is_nounal = not not word_specs[i][property] end if is_nounal then if not word_specs[i][property] then error("Internal error: noun-type word spec without " .. property .. " set") end for j = last_seen_nounal_pos + 1, i - 1 do				word_specs[j][property] = word_specs[j][property] or word_specs[i][property] end last_seen_nounal_pos = i			if seen_property == nil then seen_property = word_specs[i][property] elseif seen_property ~= word_specs[i][property] then seen_property = mixed_value end end end if last_seen_nounal_pos > 0 then for i = last_seen_nounal_pos + 1, #word_specs do			word_specs[i][property] = word_specs[i][property] or word_specs[last_seen_nounal_pos][property] end end multiword_spec[property] = seen_property end

local function propagate_properties_downward(alternant_multiword_spec, property, default_propval) local function set_and_fetch(obj, default) local retval if obj[property] then retval = obj[property] else obj[property] = default retval = default end if not obj["actual_" .. property] then obj["actual_" .. property] = retval end return retval end local propval1 = set_and_fetch(alternant_multiword_spec, default_propval) for _, alternant_or_word_spec in ipairs(alternant_multiword_spec.alternant_or_word_specs) do		local propval2 = set_and_fetch(alternant_or_word_spec, propval1) if alternant_or_word_spec.alternants then for _, multiword_spec in ipairs(alternant_or_word_spec.alternants) do				local propval3 = set_and_fetch(multiword_spec, propval2) for _, word_spec in ipairs(multiword_spec.word_specs) do					local propval4 = set_and_fetch(word_spec, propval3) if propval4 == "mixed" then -- FIXME, use clearer error message. error("Attempt to assign mixed " .. property .. " to word") end set_and_fetch(word_spec, propval4) end end else if propval2 == "mixed" then -- FIXME, use clearer error message. error("Attempt to assign mixed " .. property .. " to word") end set_and_fetch(alternant_or_word_spec, propval2) end end end

--[=[ Propagate `property` (one of "animacy", "gender" or "number") from nouns to adjacent adjectives. We proceed as follows: 1. We assume the properties in question are already set on all nouns. This should happen in  set_defaults_and_check_bad_indicators. 2. We first propagate properties upwards and sideways. We recurse downwards from the top. When we encounter a multiword spec, we proceed left to right looking for a noun. When we find a noun, we fetch its property (recursing if the noun  is an alternant), and propagate it to any adjectives to its left, up to the next noun to the left. When we have processed the last noun, we also propagate its property value to any adjectives to the right (to handle e.g.  anděl strážný "guardian angel", where the adjective strážný should inherit the 'masculine' and 'animate'   properties of anděl). Finally, we set the property value for the multiword spec itself by combining all the non-nil properties of the individual elements. If all non-nil properties have the same value, the result is that value, otherwise it is `mixed_value` (which is "mixed" for animacy and gender, but "allthree" for number). 3. When we encounter an alternant spec in this process, we recursively process each alternant (which is a multiword  spec) using the previous step, and combine any non-nil properties we encounter the same way as for multiword specs. 4. The effect of steps 2 and 3 is to set the property of each alternant and multiword spec based on its children or its neighbors. ]=] local function propagate_properties(alternant_multiword_spec, property, default_propval, mixed_value) propagate_multiword_properties(alternant_multiword_spec, property, mixed_value, "nouns only") propagate_multiword_properties(alternant_multiword_spec, property, mixed_value, false) propagate_properties_downward(alternant_multiword_spec, property, default_propval) end

local function determine_noun_status(alternant_multiword_spec) for i, alternant_or_word_spec in ipairs(alternant_multiword_spec.alternant_or_word_specs) do		if alternant_or_word_spec.alternants then local is_noun = false for _, multiword_spec in ipairs(alternant_or_word_spec.alternants) do				for j, word_spec in ipairs(multiword_spec.word_specs) do					if is_regular_noun(word_spec) then multiword_spec.first_noun = j						is_noun = true break end end end if is_noun then alternant_multiword_spec.first_noun = i			end elseif is_regular_noun(alternant_or_word_spec) then alternant_multiword_spec.first_noun = i			return end end end

-- Set the part of speech based on properties of the individual words. local function set_pos(alternant_multiword_spec) if alternant_multiword_spec.args.pos then alternant_multiword_spec.pos = alternant_multiword_spec.args.pos elseif alternant_multiword_spec.saw_pron and not alternant_multiword_spec.saw_non_pron then alternant_multiword_spec.pos = "pronoun" elseif alternant_multiword_spec.saw_det and not alternant_multiword_spec.saw_non_det then alternant_multiword_spec.pos = "determiner" elseif alternant_multiword_spec.saw_num and not alternant_multiword_spec.saw_non_num then alternant_multiword_spec.pos = "numeral" else alternant_multiword_spec.pos = "noun" end alternant_multiword_spec.plpos = require("Module:string utilities").pluralize(alternant_multiword_spec.pos) end

local function normalize_all_lemmas(alternant_multiword_spec, pagename) iut.map_word_specs(alternant_multiword_spec, function(base)		if base.lemma == "" then			base.lemma = pagename		end		base.orig_lemma = base.lemma		base.orig_lemma_no_links = m_links.remove_links(base.lemma)		local lemma = base.orig_lemma_no_links		-- If the lemma is all-uppercase, lowercase it but note this, so that later in combine_stem_ending we convert it		-- back to uppercase. This allows us to handle all-uppercase acronyms without a lot of extra complexity.		-- FIXME: This may not make sense at all.		if uupper(lemma) == lemma then			base.all_uppercase = true			lemma = ulower(lemma)		end		base.actual_lemma = lemma		base.lemma = base.decllemma or lemma	end) end

local function decline_noun(base) for _, stems in ipairs(base.stem_sets) do		if not decls[base.decl] then error("Internal error: Unrecognized declension type '" .. base.decl .. "'") end decls[base.decl](base, stems) end handle_derived_slots_and_overrides(base) local function copy(from_slot, to_slot) base.forms[to_slot] = base.forms[from_slot] end copy("nom_d", "acc_d") copy("nom_d", "voc_d") copy("gen_d", "loc_d") copy("dat_d", "ins_d") if base.actual_number ~= base.number then local source_num = base.number == "sg" and "_s" or base.number == "du" and "_d" or "_p" local dest_num = base.number == "sg" and {"_p", "_d"} or base.number == "du" and {"_s", "_p"} or {"_s", "_d"} for case, _ in pairs(cases) do copy(case .. source_num, case .. dest_num) copy("nom" .. source_num .. "_linked", "nom" .. dest_num .. "_linked") end if base.actual_number ~= "allthree" then local erase_num = base.actual_number == "sg" and {"_d", "_p"} or base.actual_number == "du" and {"_s", "_p"} or {"_s", "_d"} for case, _ in pairs(cases) do base.forms[case .. erase_num] = nil end base.forms["nom" .. erase_num .. "_linked"] = nil end end end

local function get_variants(form) return nil --[=[	FIXME return form:find(com.VAR1) and "var1" or		form:find(com.VAR2) and "var2" or		form:find(com.VAR3) and "var3" or		nil ]=] end

-- Compute the categories to add the noun to, as well as the annotation to display in the -- declension title bar. We combine the code to do these functions as both categories and -- title bar contain similar information. local function compute_categories_and_annotation(alternant_multiword_spec) local all_cats = {} local function insert(cattype) m_table.insertIfNot(all_cats, "Old Czech " .. cattype) end if alternant_multiword_spec.pos == "noun" then if alternant_multiword_spec.actual_number == "sg" then insert("uncountable nouns") elseif alternant_multiword_spec.actual_number == "du" then insert("dualia tantum") elseif alternant_multiword_spec.actual_number == "pl" then insert("pluralia tantum") end end local annotation local annparts = {} local decldescs = {} local vowelalts = {} local foreign = {} local irregs = {} local stemspecs = {} local reducible = nil local function get_genanim(gender, animacy) local gender_code_to_desc = { m = "masculine", f = "feminine", n = "neuter", none = nil, }		local animacy_code_to_desc = { pr = "personal", anml = "animal", inan = "inanimate", none = nil, }		local descs = {} table.insert(descs, gender_code_to_desc[gender]) if gender ~= "f" and gender ~= "n" then -- masculine or "none" (e.g. certain pronouns and numerals) table.insert(descs, animacy_code_to_desc[animacy]) end return table.concat(descs, " ") end

local function trim(text) text = text:gsub(" +", " ") return mw.text.trim(text) end

local function do_word_spec(base) local actual_genanim = get_genanim(base.actual_gender, base.actual_animacy) local declined_genanim = get_genanim(base.gender, base.animacy) local genanim if accatual_genanim ~= declined_genanim then genanim = ("%s (declined as %s)"):format(actual_genanim, declined_genanim) insert("nouns with actual gender different from declined gender") else genanim = actual_genanim end if base.actual_gender == "m" then -- Insert a category for 'Czech masculine animate nouns' or 'Czech masculine inanimate nouns'; the base categories -- Category:Czech masculine nouns, Czech animate nouns are auto-inserted. insert(actual_genanim .. " " .. alternant_multiword_spec.plpos) end for _, stems in ipairs(base.stem_sets) do			local props = declprops[base.decl] local cats = props.cat if type(cats) == "function" then cats = cats(base, stems) end if type(cats) == "string" then cats = {cats} end local default_desc for i, cat in ipairs(cats) do				if not cat:find("GENDER") and not cat:find("GENPOS") and not cat:find("POS") then cat = cat end cat = cat:gsub("GENPOS", "GENDER POS") if not cat:find("POS") then cat = cat .. " POS" end if i == #cats then default_desc = cat:gsub(" POS", "") end cat = cat:gsub("GENDER", actual_genanim) cat = cat:gsub("POS", alternant_multiword_spec.plpos) -- Need to trim `cat` because actual_genanim may be an empty string. insert(trim(cat)) end

local desc = props.desc if type(desc) == "function" then desc = desc(base, stems) end desc = desc or default_desc desc = desc:gsub("GENDER", genanim) -- Need to trim `desc` because genanim may be an empty string. m_table.insertIfNot(decldescs, trim(desc))

local vowelalt if stems.vowelalt == "quant" then vowelalt = "quant-alt" insert("nouns with quantitative vowel alternation") elseif stems.vowelalt == "quant-ě" then vowelalt = "í-ě-alt" insert("nouns with í-ě alternation") end if vowelalt then m_table.insertIfNot(vowelalts, vowelalt) end if reducible == nil then reducible = stems.reducible elseif reducible ~= stems.reducible then reducible = "mixed" end if stems.reducible then insert("nouns with reducible stem") end if base.foreign then m_table.insertIfNot(foreign, "foreign") if not base.decllemma then -- NOTE: there are nouns that use both 'foreign' and 'decllemma', e.g. Zeus. insert("nouns with regular foreign declension") end end -- User-specified 'decllemma:' indicates irregular stem. Don't consider foreign nouns in -us/-os/-es, -um/-on or -- silent -e (e.g. software) where this ending is simply dropped in oblique and plural forms as irregular; -- there are too many of these and they are already categorized above as 'nouns with regular foreign declension'. if base.decllemma then m_table.insertIfNot(irregs, "irreg-stem") insert("nouns with irregular stem") end m_table.insertIfNot(stemspecs, stems.vowel_stem) end end local key_entry = alternant_multiword_spec.first_noun or 1 if #alternant_multiword_spec.alternant_or_word_specs >= key_entry then local alternant_or_word_spec = alternant_multiword_spec.alternant_or_word_specs[key_entry] if alternant_or_word_spec.alternants then for _, multiword_spec in ipairs(alternant_or_word_spec.alternants) do				key_entry = multiword_spec.first_noun or 1 if #multiword_spec.word_specs >= key_entry then do_word_spec(multiword_spec.word_specs[key_entry]) end end else do_word_spec(alternant_or_word_spec) end end if alternant_multiword_spec.actual_number == "sg" or alternant_multiword_spec.actual_number == "pl" or alternant_multiword_spec.actual_number == "du" then -- not "allthree" or "none" (for sebe) table.insert(annparts, alternant_multiword_spec.actual_number == "sg" and "sg-only" or alternant_multiword_spec.actual_number == "du" and "du-only" or "pl-only") end if #decldescs == 0 then table.insert(annparts, "indecl") else table.insert(annparts, table.concat(decldescs, " // ")) end if #vowelalts > 0 then table.insert(annparts, table.concat(vowelalts, "/")) end if reducible == "mixed" then table.insert(annparts, "mixed-reducible") elseif reducible then table.insert(annparts, "reducible") end if #foreign > 0 then table.insert(annparts, table.concat(foreign, " // ")) end if #irregs > 0 then table.insert(annparts, table.concat(irregs, " // ")) end alternant_multiword_spec.annotation = table.concat(annparts, " ") if #stemspecs > 1 then insert("nouns with multiple stems") end if alternant_multiword_spec.actual_number == "allthree" and not m_table.deepEquals(alternant_multiword_spec.sg_genders, alternant_multiword_spec.pl_genders) then insert("nouns that change gender in the plural") end alternant_multiword_spec.categories = all_cats end

local function show_forms(alternant_multiword_spec) local lemmas = {} for _, slot in ipairs(potential_lemma_slots) do		if alternant_multiword_spec.forms[slot] then for _, formobj in ipairs(alternant_multiword_spec.forms[slot]) do -- FIXME, now can support footnotes as qualifiers in headwords? table.insert(lemmas, formobj.form) end break end end local props = { lemmas = lemmas, slot_table = alternant_multiword_spec.output_noun_slots, lang = lang, canonicalize = function(form) -- return com.remove_variant_codes(form) return form end, }	iut.show_forms(alternant_multiword_spec.forms, props) end

local function make_table(alternant_multiword_spec) local forms = alternant_multiword_spec.forms

local function template_prelude(min_width) return rsub([=[ {title}{annotation} {\op}| style="background:#F9F9F9;text-align:center;min-width:MINWIDTHem" class="inflection-table" ]=], "MINWIDTH", min_width) end

local function template_postlude return [=[ This table shows the most common forms around the 13th century. See also Appendix:Old Czech nouns and Appendix:Old Czech pronunciation.''  ]=] end
 * {\cl}{notes_clause}

local table_spec_allthree = template_prelude("45") .. [=[ ! style="width:33%;background:#d9ebff" | ! style="background:#d9ebff" | singular ! style="background:#d9ebff" | dual ! style="background:#d9ebff" | plural !style="background:#eff7ff"|nominative !style="background:#eff7ff"|genitive !style="background:#eff7ff"|dative !style="background:#eff7ff"|accusative !style="background:#eff7ff"|vocative !style="background:#eff7ff"|locative !style="background:#eff7ff"|instrumental ]=] .. template_postlude
 * {nom_s}
 * {nom_d}
 * {nom_p}
 * {gen_s}
 * {gen_d}
 * {gen_p}
 * {dat_s}
 * {dat_d}
 * {dat_p}
 * {acc_s}
 * {acc_d}
 * {acc_p}
 * {voc_s}
 * {voc_d}
 * {voc_p}
 * {loc_s}
 * {loc_d}
 * {loc_p}
 * {ins_s}
 * {ins_d}
 * {ins_p}

local function get_table_spec_one_number(number, numcode) local table_spec_one_number = [=[ ! style="width:33%;background:#d9ebff" | ! style="background:#d9ebff" | NUMBER !style="background:#eff7ff"|nominative !style="background:#eff7ff"|genitive !style="background:#eff7ff"|dative !style="background:#eff7ff"|accusative !style="background:#eff7ff"|vocative !style="background:#eff7ff"|locative !style="background:#eff7ff"|instrumental ]=]		return template_prelude("30") .. table_spec_one_number:gsub("NUMBER", number):gsub("CODE", numcode) .. template_postlude end
 * {nom_CODE}
 * {gen_CODE}
 * {dat_CODE}
 * {acc_CODE}
 * {voc_CODE}
 * {loc_CODE}
 * {ins_CODE}

local function get_table_spec_one_number_clitic(number, numcode) local table_spec_one_number_clitic = [=[ ! rowspan=2 style="width:33%;background:#d9ebff"| ! colspan=2 style="background:#d9ebff" | NUMBER ! style="width:33%;background:#d9ebff" | stressed ! style="background:#d9ebff" | clitic !style="background:#eff7ff"|nominative !style="background:#eff7ff"|genitive !style="background:#eff7ff"|dative !style="background:#eff7ff"|accusative !style="background:#eff7ff"|vocative !style="background:#eff7ff"|locative !style="background:#eff7ff"|instrumental ]=]		return template_prelude("40") .. table_spec_one_number_clitic:gsub("NUMBER", number):gsub("CODE", numcode) .. template_postlude end
 * colspan=2 | {nom_CODE}
 * {gen_CODE}
 * {clitic_gen_CODE}
 * {dat_CODE}
 * {clitic_dat_CODE}
 * {acc_CODE}
 * {clitic_acc_CODE}
 * colspan=2 | {voc_CODE}
 * colspan=2 | {loc_CODE}
 * colspan=2 | {ins_CODE}

local notes_template = [=[ {footnote} ]=]

if alternant_multiword_spec.title then forms.title = alternant_multiword_spec.title else forms.title = 'Declension of ' .. forms.lemma .. '' end

local annotation = alternant_multiword_spec.annotation if annotation == "" then forms.annotation = "" else forms.annotation = " (" .. annotation .. " )" end

local number, numcode if alternant_multiword_spec.actual_number == "sg" then number, numcode = "singular", "s" elseif alternant_multiword_spec.actual_number == "du" then number, numcode = "dual", "d" elseif alternant_multiword_spec.actual_number == "pl" then number, numcode = "plural", "p" elseif alternant_multiword_spec.actual_number == "none" then -- used for sebe number, numcode = "", "s" end

local table_spec = alternant_multiword_spec.actual_number == "allthree" and table_spec_allthree or		alternant_multiword_spec.has_clitic and get_table_spec_one_number_clitic(number, numcode) or		get_table_spec_one_number(number, numcode) forms.notes_clause = forms.footnote ~= "" and m_string_utilities.format(notes_template, forms) or "" return m_string_utilities.format(table_spec, forms) end

local function compute_headword_genders(alternant_multiword_spec) local genders = {} local number if alternant_multiword_spec.actual_number == "pl" then number = "-p" elseif alternant_multiword_spec.actual_number == "du" then number = "-d" else number = "" end iut.map_word_specs(alternant_multiword_spec, function(base)		local animacy = base.animacy		if animacy == "inan" then			animacy = "in"		end		m_table.insertIfNot(genders, base.gender .. "-" .. animacy .. number)	end) return genders end

-- Externally callable function to parse and decline a noun given user-specified arguments. -- Return value is ALTERNANT_MULTIWORD_SPEC, an object where the declined forms are in -- `ALTERNANT_MULTIWORD_SPEC.forms` for each slot. If there are no values for a slot, the -- slot key will be missing. The value for a given slot is a list of objects -- {form=FORM, footnotes=FOOTNOTES}. function export.do_generate_forms(parent_args, from_headword) local params = { [1] = {required = true, default = "bóh"}, title = {}, pagename = {}, json = {type = "boolean"}, pos = {}, }

if from_headword then params["head"] = {list = true} params["lemma"] = {list = true} params["g"] = {list = true} params["f"] = {list = true} params["m"] = {list = true} params["adj"] = {list = true} params["dim"] = {list = true} params["id"] = {} end

local args = m_para.process(parent_args, params) local parse_props = { parse_indicator_spec = parse_indicator_spec, angle_brackets_omittable = true, allow_blank_lemma = true, }	local alternant_multiword_spec = iut.parse_inflected_text(args[1], parse_props) alternant_multiword_spec.title = args.title alternant_multiword_spec.args = args local pagename = args.pagename or from_headword and args.head[1] or mw.title.getCurrentTitle.subpageText normalize_all_lemmas(alternant_multiword_spec, pagename) set_all_defaults_and_check_bad_indicators(alternant_multiword_spec) -- These need to happen before detect_all_indicator_specs so that adjectives get their genders and numbers set -- appropriately, which are needed to correctly synthesize the adjective lemma. propagate_properties(alternant_multiword_spec, "animacy", "inan", "mixed") propagate_properties(alternant_multiword_spec, "number", "allthree", "allthree") -- FIXME, the default value (third param) used to be 'm' with a comment indicating that this applied only to	-- plural adjectives, where it didn't matter; but in Czech, plural adjectives are distinguished for gender and -- animacy. Make sure 'mixed' works. propagate_properties(alternant_multiword_spec, "gender", "mixed", "mixed") detect_all_indicator_specs(alternant_multiword_spec) -- Propagate 'actual_number' after calling detect_all_indicator_specs, which sets 'actual_number' for adjectives. propagate_properties(alternant_multiword_spec, "actual_number", "allthree", "allthree") determine_noun_status(alternant_multiword_spec) set_pos(alternant_multiword_spec) alternant_multiword_spec.output_noun_slots = get_output_noun_slots(alternant_multiword_spec) local inflect_props = { skip_slot = function(slot) return skip_slot(alternant_multiword_spec.actual_number, slot) end, slot_table = alternant_multiword_spec.output_noun_slots, get_variants = get_variants, inflect_word_spec = decline_noun, }	iut.inflect_multiword_or_alternant_multiword_spec(alternant_multiword_spec, inflect_props) compute_categories_and_annotation(alternant_multiword_spec) alternant_multiword_spec.genders = compute_headword_genders(alternant_multiword_spec) if args.json then alternant_multiword_spec.args = nil return require("Module:JSON").toJSON(alternant_multiword_spec) end return alternant_multiword_spec end

-- Entry point for. Template-callable function to parse and decline a noun given -- user-specified arguments and generate a displayable table of the declined forms. function export.show(frame) local parent_args = frame:getParent.args local alternant_multiword_spec = export.do_generate_forms(parent_args) if type(alternant_multiword_spec) == "string" then -- JSON return value return alternant_multiword_spec end show_forms(alternant_multiword_spec) return make_table(alternant_multiword_spec) .. require("Module:utilities").format_categories(alternant_multiword_spec.categories, lang, nil, nil, force_cat) end

return export