Module:User:JeffDoozan/es-verb

local export = {}

--[=[

Authorship: Ben Wing

]=]

--[=[

TERMINOLOGY:

-- "slot" = A particular combination of tense/mood/person/number/etc. Example slot names for verbs are "pres_1s" (present indicative first-person singular), "pres_sub_2sv" (present	 subjunctive second-person singular voseo form) "impf_sub_ra_3p" (imperfect subjunctive -ra form third-person	 plural), "imp_1p_comb_lo" (imperative first-person plural combined with clitic lo). Each slot is filled with zero or more forms.

-- "form" = The conjugated Spanish form representing the value of a given slot.

-- "lemma" = The dictionary form of a given Spanish term. For Spanish, always the infinitive. ]=]

--[=[

FIXME:

1. Implement no_pres_stressed for aterir, garantir. (NOTE: Per RAE, garantir used in all forms in Argentina/Uruguay.) [DONE] 2. Support concluyo. [DONE] 3. Fixes for veo -> ve vs. preveo -> prevé. [DONE] 4. Various more irregular verbs, e.g. predecir, redecir, bendecir, maldecir. [DONE] 5. Raising of e -> i, o -> u before -iendo, -ió, etc. occurs only in -ir verbs. [DONE] 6. Raising of e -> i, o -> u happens before subjunctive -amos, -áis in -ir verbs. [DONE] 7. Implement reflexive verbs. [DONE] 8. Implement categories. [DONE] 9. Implement show_forms. [DONE] 10. Reconcile stems.vowel_alt from irregular verbs with vowel_alt from indicators. May require moving the irregular-verb handling code in construct_stems into detect_indicator_spec. [DONE] 11. Implement make_table. [DONE] 12. Vowel alternation should show u-ue (jugar), i-ie (adquirir), e-í (reír) alternations specially. [DONE] 13. Handle linking of multiword forms as is done in Module:es-headword. [DONE] 14. Implement comparison against previous module. [DONE] 15. Implement categorization of irregularities for individual tenses. 16. Support nocomb=1. [DONE] 17. (Possibly) display irregular forms in a different color, as with the old module. 18. (Possibly) display a "rule" description indicating the types of alternations. 19. Implement replace_reflexive_indicators. 20. Implement verbs with attached clitics e.g. pasarlo, corrérsela. [DONE] 21. When footnote + tú/vos notation, add a space before tú/vos. 22. Fix erguir so ie-i vowel alternation produces ye- at beginning of word, similarly for errar. Also allow multiple vowel alternation specs in irregular verbs, for errar. Finally, ie should show as e-ye for errar and as e-ye-i for erguir. [DONE] 23. Figure out why red links in combined forms show up as black not red. 24. Consider including alternative superseded forms of verbs like ciar (e.g. pret_3s = cio, ció with footnote). 25. Allow conjugation of suffixes e.g. -ir, -ecer; need to fix in Module:inflection utilities. [DONE] 26. Allow specification of stems esp. so that footnotes can be hung off them; use + for the default. 27. Don't remove monosyllabic accents when conjugating suffixes. [DONE] 28. If multiword expression with no <>, add <> after first word, as with Module:es-headword. [DONE] 29. (Possibly) link the parts of a reflexive or cliticized infinitive, as done in Module:es-headword. [DONE] 30. Final fixes to allow Module:es-headword to use this module. [DONE] --]=]

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

local force_cat = false -- set to true for debugging local check_for_red_links = false -- set to false for debugging

local rfind = mw.ustring.find local rmatch = mw.ustring.match local rsplit = mw.text.split local rsub = com.rsub

local function link_term(term) return m_links.full_link({ lang = lang, term = term }, "term") end

local V = com.V -- vowel regex class local AV = com.AV -- accented vowel regex class local C = com.C -- consonant regex class

local fut_sub_note = "[mostly obsolete, now mainly used in legal language]" local pres_sub_voseo_note = "[Argentine and Uruguayan " .. link_term("voseo") .. " prefers the " .. link_term("tú") .. " form for the present subjunctive]"

local vowel_alternants = m_table.listToSet({"ie", "ie-i", "ye", "ye-i", "ue", "ue-u", "hue", "i", "í", "ú", "+"}) local vowel_alternant_to_desc = { ["ie"] = "e-ie", ["ie-i"] = "e-ie-i", ["ye"] = "e-ye", ["ye-i"] = "e-ye-i", ["ue"] = "o-ue", ["ue-u"] = "o-ue-u", ["hue"] = "o-hue", ["i"] = "e-i", ["í"] = "i-í", ["ú"] = "u-ú", }

local raise_vowel = {["e"] = "i", ["o"] = "u"}

local all_persons_numbers = { ["1s"] = "1|s", ["2s"] = "2|s", ["2sv"] = "2|s|voseo", ["3s"] = "3|s", ["1p"] = "1|p", ["2p"] = "2|p", ["3p"] = "3|p", }

local person_number_list_basic = {"1s", "2s", "3s", "1p", "2p", "3p"} local person_number_list_voseo = {"1s", "2s", "2sv", "3s", "1p", "2p", "3p"} local imp_person_number_list = {"2s", "2sv", "3s", "1p", "2p", "3p"} local neg_imp_person_number_list = {"2s", "3s", "1p", "2p", "3p"}

person_number_to_reflexive_pronoun = { ["1s"] = "me", ["2s"] = "te", ["2sv"] = "te", ["3s"] = "se", ["1p"] = "nos", ["2p"] = "os", ["3p"] = "se", }

-- Initialize all the slots for which we generate forms. The particular slots may depend on whether we're generating -- combined slots (`not alternant_multiword_spec.nocomb`, which is always false if we're dealing with a verb with an -- attached clitic, such as hincarla, or a reflexive or partly-reflexive verb, where a partly-reflexive verb is -- a conjoined term made up of two or more verbs, where some but not all are reflexive). It may also depend on whether -- we're being requested to generate some double-combined forms, such as llevándoselo; see the comment below for -- `verb_slot_double_combined_rows`. local function add_slots(alternant_multiword_spec) -- "Basic" slots: All slots that go into the regular table (not the combined-form table). alternant_multiword_spec.verb_slots_basic = { {"infinitive", "inf"}, {"infinitive_linked", "inf"}, {"gerund", "ger"}, {"pp_ms", "m|s|past|part"}, {"pp_fs", "f|s|past|part"}, {"pp_mp", "m|p|past|part"}, {"pp_fp", "f|p|past|part"}, }

-- Slots that go into the combined-form table, along with double-combined slots (e.g. llevándoselo) that are -- requested for use with. alternant_multiword_spec.verb_slots_combined = {}

-- Special slots used to handle non-reflexive parts of reflexive verbs in. -- For example, for a reflexive-only verb like jambarse, we want to be able to use on	-- jambe (which should mention that it is a part of 'me jambe', first-person singular present subjunctive, and	-- 'se jambe', third-person singular present subjunctive) or on jambamos (which should mention that it is a	-- part of 'nos jambamos', first-person plural present indicative or preterite). Similarly, we want to use -- on jambando (which should mention that it is a part of 'se ... jambando', syntactic	-- variant of jambándose, which is the gerund of jambarse). To do this, we need to be able to map -- non-reflexive parts like jambe, jambamos, jambando, etc. to their reflexive equivalent(s), to the -- tag(s) of the equivalent(s), and, in the case of forms like jambando, jambar and imperatives, to the -- separated syntactic variant of the verb+clitic combination. We do this by creating slots for the non-reflexive -- part equivalent of each basic reflexive slot, and for the separated syntactic-variant equivalent of each basic -- reflexive slot that is formed of verb+clitic. We use slots in this way to deal with multiword lemmas. Note that -- we run into difficulties mapping between reflexive verbs, non-reflexive part equivalents, and separated syntactic -- variants if a slot contains more than one form. To handle this, if there are the same number of forms in two -- slots we're trying to match up, we assume the forms match one-to-one; otherwise we don't match up the two slots -- (which means won't work in this case, but such a case is extremely rare and not worth	-- worrying about). Alternatives that handle this "properly" are significantly more complicated and require -- non-trivial modifications to Module:inflection utilities. local need_special_verb_form_of_slots = alternant_multiword_spec.from_verb_form_of and alternant_multiword_spec.refl

if need_special_verb_form_of_slots then alternant_multiword_spec.verb_slots_reflexive_verb_form_of = { {"infinitive_non_reflexive", "-"}, {"infinitive_variant", "-"}, {"gerund_non_reflexive", "-"}, {"gerund_variant", "-"}, }	else alternant_multiword_spec.verb_slots_reflexive_verb_form_of = {} end

-- For generating combined forms, i.e. combinations of a basic form (specifically, infinitive, gerund or an	-- imperative form) with a clitic (or in some cases, two clitics). This is a list of lists of the form -- {BASIC_SLOT, CLITICS} where BASIC_SLOT is the slot to add the clitic pronouns to (e.g. "gerund" or "imp_2s") -- and CLITICS is a list of the clitic pronouns to add. alternant_multiword_spec.verb_slot_combined_rows = {}

-- For generating double combined forms (e.g. llevándoselo or dámela). This is used by -- when it detects that it is being requested to find the inflection tags for a double-combined form. The number of	-- double-combined forms is relatively large, so to optimize this, Module:es-inflections (which implements	-- ) detects which two clitics are involved, and we only generate double-combined forms -- involving those two clitics; this is specified using `double_combined_forms_to_include`, passed into -- do_generate_forms. The value of this field is a list of lists of the form {SINGLE_COMB_SLOT, CLITICS} where -- SINGLE_COMB_SLOT is the single-combined slot to add the object clitic pronouns to (e.g. "gerund_comb_se" or	-- "imp_2s_comb_me") and CLITICS is a list of the clitic pronouns to add. CLITICS will normally be a length-one -- list whose value is one of {"lo", "la", "le", "los", "las", "les"}. alternant_multiword_spec.verb_slot_double_combined_rows = {}

-- Add entries for a slot with person/number variants. -- `verb_slots` is the table to add to. -- `slot_prefix` is the prefix of the slot, typically specifying the tense/aspect. -- `tag_suffix` is a string listing the set of inflection tags to add after the person/number tags. -- `person_number_list` is a list of the person/number slot suffixes to add to `slot_prefix`. local function add_personal_slot(verb_slots, slot_prefix, tag_suffix, person_number_list) for _, persnum in ipairs(person_number_list) do			local persnum_tag = all_persons_numbers[persnum] local slot = slot_prefix .. "_" .. persnum local accel = persnum_tag .. "|" .. tag_suffix table.insert(verb_slots, {slot, accel}) end end

-- Add a personal slot (i.e. a slot with person/number variants) to `verb_slots_basic`. local function add_basic_personal_slot(slot_prefix, tag_suffix, person_number_list, no_special_verb_form_of_slot,		need_variant_slot) add_personal_slot(alternant_multiword_spec.verb_slots_basic, slot_prefix, tag_suffix, person_number_list) -- Add special slots for handling non-reflexive parts of reflexive verbs in. -- See comment above in `need_special_verb_form_of_slots`. if need_special_verb_form_of_slots and not no_special_verb_form_of_slot then for _, persnum in ipairs(person_number_list) do				local persnum_tag = all_persons_numbers[persnum] local basic_slot = slot_prefix .. "_" .. persnum local accel = persnum_tag .. "|" .. tag_suffix table.insert(alternant_multiword_spec.verb_slots_reflexive_verb_form_of, {basic_slot .. "_non_reflexive", "-"}) if need_variant_slot then table.insert(alternant_multiword_spec.verb_slots_reflexive_verb_form_of, {basic_slot .. "_variant", "-"}) end end end end

add_basic_personal_slot("pres", "pres|ind", person_number_list_voseo) add_basic_personal_slot("impf", "impf|ind", person_number_list_basic) add_basic_personal_slot("pret", "pret|ind", person_number_list_basic) add_basic_personal_slot("fut", "fut|ind", person_number_list_basic) add_basic_personal_slot("cond", "cond", person_number_list_basic) add_basic_personal_slot("pres_sub", "pres|sub", person_number_list_voseo) add_basic_personal_slot("impf_sub_ra", "impf|sub", person_number_list_basic) add_basic_personal_slot("impf_sub_se", "impf|sub", person_number_list_basic) add_basic_personal_slot("fut_sub", "fut|sub", person_number_list_basic) -- Need variant slots because the imperative clitics are suffixed. add_basic_personal_slot("imp", "imp", imp_person_number_list, nil, "need variant slot") -- Don't need special non-reflexive-part slots because the negative imperative is multiword, of which the -- individual words are 'no' + subjunctive. add_basic_personal_slot("neg_imp", "neg|imp", neg_imp_person_number_list, "no special verb form of") -- Don't need special non-reflexive-part slots because we don't want jambando mapping to jambándome -- (only jambándose) or jambar mapping to jambarme (only jambarse). add_basic_personal_slot("infinitive", "inf", person_number_list_basic, "no special verb form of") add_basic_personal_slot("gerund", "ger", person_number_list_basic, "no special verb form of")

local third_person_object_clitics = {"lo", "la", "le", "los", "las", "les"}

-- Add combined-form slots. if not alternant_multiword_spec.nocomb then -- Add a row of slots representing the combination of a basic slot with a clitic. `basic_slot` is the basic slot -- descriptor, `tag_prefix` is a string describing the inflection tags of the basic slot, and `personal_clitics` -- is a list of the personal clitics ("me", "te", "se", "nos" or "os") to add to the basic slot. local function add_combined_slot_row(basic_slot, tag_prefix, personal_clitics) -- First, add each individual combined slot to `verb_slots_combined`. local clitics_with_object = m_table.append(personal_clitics, third_person_object_clitics) for _, clitic in ipairs(clitics_with_object) do local slot = basic_slot .. "_comb_" .. clitic -- You have to pass this through full_link to get a Spanish-specific link local accel = tag_prefix .. "|combined with " .. clitic .. "" table.insert(alternant_multiword_spec.verb_slots_combined, {slot, accel}) end

-- Also, add the row to `verb_slot_combined_rows`. table.insert(alternant_multiword_spec.verb_slot_combined_rows, {basic_slot, clitics_with_object})

-- Also do double-combined forms for a specific set of clitics, if requested. See the comment above -- `verb_slot_double_combined_rows` above. if alternant_multiword_spec.double_combined_forms_to_include then for _, personal_clitic in ipairs(personal_clitics) do					for _, object_clitic in ipairs(third_person_object_clitics) do						for _, form_to_include in ipairs(alternant_multiword_spec.double_combined_forms_to_include) do							local to_include_personal_clitic, to_include_object_clitic = unpack(form_to_include) if personal_clitic == to_include_personal_clitic and object_clitic == to_include_object_clitic then local single_comb_slot = basic_slot .. "_comb_" .. personal_clitic local slot = single_comb_slot .. "_" .. object_clitic local accel = tag_prefix .. "|combined with " .. personal_clitic .. " and " ..									object_clitic .. "" table.insert(alternant_multiword_spec.verb_slots_combined, {slot, accel}) table.insert(alternant_multiword_spec.verb_slot_double_combined_rows,									{single_comb_slot, {object_clitic}}) break end end end end end end

add_combined_slot_row("infinitive", "inf", {"me", "te", "se", "nos", "os"}) add_combined_slot_row("gerund", "gerund", {"me", "te", "se", "nos", "os"})

local function add_combined_imp_slot_row(persnum, personal_clitics) add_combined_slot_row("imp_" .. persnum, all_persons_numbers[persnum] .. "|imp", personal_clitics) end add_combined_imp_slot_row("2s", {"me", "te", "nos"}) add_combined_imp_slot_row("2sv", {"me", "te", "nos"}) add_combined_imp_slot_row("3s", {"me", "se", "nos"}) add_combined_imp_slot_row("1p", {"te", "nos", "os"}) add_combined_imp_slot_row("2p", {"me", "nos", "os"}) add_combined_imp_slot_row("3p", {"me", "se", "nos"}) end

-- Generate the list of all slots. alternant_multiword_spec.all_verb_slots = {} for _, slot_and_accel in ipairs(alternant_multiword_spec.verb_slots_basic) do		table.insert(alternant_multiword_spec.all_verb_slots, slot_and_accel) end for _, slot_and_accel in ipairs(alternant_multiword_spec.verb_slots_combined) do		table.insert(alternant_multiword_spec.all_verb_slots, slot_and_accel) end for _, slot_and_accel in ipairs(alternant_multiword_spec.verb_slots_reflexive_verb_form_of) do		table.insert(alternant_multiword_spec.all_verb_slots, slot_and_accel) end

alternant_multiword_spec.verb_slots_basic_map = {} for _, slotaccel in ipairs(alternant_multiword_spec.verb_slots_basic) do		local slot, accel = unpack(slotaccel) alternant_multiword_spec.verb_slots_basic_map[slot] = accel end

alternant_multiword_spec.verb_slots_combined_map = {} for _, slotaccel in ipairs(alternant_multiword_spec.verb_slots_combined) do		local slot, accel = unpack(slotaccel) alternant_multiword_spec.verb_slots_combined_map[slot] = accel end end

local function match_against_verbs(ref_verb, prefixes) return function(verb) for _, prefix in ipairs(prefixes) do if verb == prefix .. ref_verb then return prefix, ref_verb end end return nil end end

--[=[

Special cases for verbs:

diluviar, atardecer: impersonal; all finite non-3s forms are nonexistent or hypothetical. Handle using '.only3s'.

empecer, atañer, concernir: all finite non-third-person forms are nonexistent or hypothetical. Handle using '.only3sp'.

desnacer: Former module claimed an irregular past participle 'desnato'. Verb is not in RAE at all and barely exists; unlikely to have irregular past participle.

desposeer: Former module claimed an irregular past participle 'desposeso'. Not per RAE.

valer: Former module claimed an irregular imperative 'val'. Not per RAE.

manumitir: Former module claimed an irregular past participle 'manumiso'. Not per RAE.

raer: Former module claimed a pres 1sg rao. Not per RAE.

rehuir: Handle using +ú.

sustituir: Former module claimed an irregular past participle 'sustituto'. Not per RAE.

venir: Former module claimed an irregular clitic combination ven + nos -> venos. Not per native speakers.

-

Verbs to check: rehuir (+ú), prohibir (+í), reunir (+ú)

-

-ir verbs:

There are several types of vowel alternations:

1. No alternation. Includes some verbs in -e- and -o-, e.g. aterir(se) (no_pres_stressed), tra(n)sgredir (pres_stressed forms rare), abolir (pres_stressed forms rare), colorir (no_pres_stressed), sumergir, divergir, convergir, arrecir (no_pres_stressed), rostir, polir (obsolete), condir (obsolete), possibly ascondir (obsolete), atordir (obsolete), sacodir (obsolete), sobrevendir (possible misspelling), empedernir (no_pres_stressed), decebir (obsolete; possibly actually like concebir, i.e. decibo not decebo), premir (obsolete), expremir (obsolete), exir (obsolete), escreuir (obsolete; fix conjugation), escrebir (obsolete), agredir; sometimes the stressed forms are rare or disused. (Also embaír, desvaír are no_pres_stressed.) 2a. ie: Infinitive has -e-, changing to -ie- when stressed. No raising before i+V. Only hendir, cernir, discernir, concernir (only3sp); discernir -> discierno, discerniendo, discernió, discernamos. 2b. ye: Infinitive has -e-, changing to -ye- when stressed. No raising before i+V. Does not occur (cf. errar). 3a. ie-i: Infinitive has -e- or -i-, changing to -ie- when stressed. Raising before i+V and 1p/2p pres subjunctive: sentir -> siento, sintiendo, sintió, sintamos. adquirir -> adquiero, adquiriendo, adquirió, adquiramos. 3b. ye-i: Infinitive has -e-, changing to -ye- when stressed. Raising before i+V and 1p/2p pres subjunctive. Only erguir: erguir -> yergo, irguiendo, irguió, irgamos. 4. i: Infinitive has -e-, changing to -i- when stressed. Raising before i+V and 1p/2p pres subjunctive: vestir -> visto, vistiendo, vistió, vistamos. Variant: ceñir -> ciño, ciñendo, ciñó, ciñamos. NOTE: preterir (no_pres_stressed). 5. ue-u: Infinitive has -o-, changing to -ue- when stressed. Raising before i+V and 1p/2p pres subjunctive: Only dormir, morir and compounds. dormir -> duermo, durmiendo, durmió, durmamos. 6. ue: This type would be parallel to 'ie' but doesn't appear to exist.

-

Verbs to fix (extra forms need to be excised or deleted): The above verbs under type (1) -ir vowel alternations; neviscar (impersonal), acaecer (third-person only), acontecer (third-person only), cellisquear (impersonal), pintear (impersonal? other meaning "to play hookey" given, not in RAE), diluviar (impersonal). desabrir (not like abrir, rare in pres_stressed/sub forms), jabrir (not like abrir).

Verbs with existing errors:
 * abeldar: (missing ) [FORMS TO DELETE]
 * acaecer: (used only in 3rd person) [DELETE ALL FIRST AND SECOND PERSON FORMS]
 * acontecer: (used only in 3rd person) [DELETE ALL FIRST AND SECOND PERSON FORMS]
 * anticuar (conjugated without ú) [FORMS TO DELETE]
 * antojar: (used only in 3rd person) [DELETE ALL FIRST AND SECOND PERSON FORMS]
 * aerografiar: (conjugated without í) [FORMS TO DELETE]
 * afiliar: (conjugated with í/+ should be only +) [FORMS TO DELETE]
 * agraviar: (wrongly has í) [FORMS TO DELETE]
 * arrecir: (not given as no_pres_stressed) [FORMS TO DELETE]
 * aserrar: (missing ) [FORMS TO DELETE]
 * aspaventar: (missing ) [OK]
 * atesar: (wrongly has ie) [FORMS TO DELETE]
 * avalentar: (wrongly has ie) [FORMS TO DELETE]
 * autorregularse (conjugated as non-reflexive) [OK]
 * auxiliar: (conjugated with í/+ should be only +) [FORMS TO DELETE]
 * balbucir (pres1_and_sub nonexistent) [DELETE PRES1_AND_SUB FORMS]
 * caçar (conjugated as -zar) [OK]
 * calefacer (extra form caleface) [FIX]
 * cellisquear: (used only in 3rd person singular) [DELETE ALL FIRST AND SECOND PERSON FORMS, ALL 3RD PLURAL FORMS AND ALL PP NON-MS FORMS]
 * chilenizar (conjugated with no cons alternation) [OK]
 * colorir: (not given as no_pres_stressed) [FORMS TO DELETE]
 * comisariar (conjugated without í) [FORMS TO DELETE]
 * complacer (missing complega in imperative 3s) [OK]
 * comprehender (conjugated as comprender) [OK]
 * decebir (was conjugated without ) [FORMS TO DELETE]
 * desafiliar: (conjugated with í/+ should be only +) [FORMS TO DELETE]
 * desagregar (conjugated with no cons alternation) [FORMS TO DELETE]
 * desairar: (conjugated with í should be only +) [FORMS TO DELETE]
 * descordar: (missing ) [FORMS TO DELETE]
 * deseleccionar (conjugated as desseleccionar) [OK]
 * desestacionalizar (conjugated with no cons alternation) [FORMS TO DELETE]
 * desgonzar (conjugated with no cons alternation) [FORMS TO DELETE]
 * desguinzar (conjugated with no cons alternation) [FORMS TO DELETE]
 * deshacer (extra form deshace, desháceme etc.) [OK]
 * desnacer (extra form desnato etc.) [DELETE PP FORMS]
 * desposeer (extra form desposeso etc.) [DELETE PP FORMS]
 * dezmar: (wrongly has ie) [FORMS TO DELETE]
 * diluviar: (used only in 3rd person singular) [DELETE ALL FIRST AND SECOND PERSON FORMS, ALL 3RD PLURAL FORMS AND ALL PP NON-MS FORMS]
 * draftear (extraneous param lang=es) [REMOVE PARAM]
 * ejemplarizar (extraneous param compound=1) [compound -> combined]
 * ejercitar (extraneous param compound=1) [compound -> combined]
 * empecer (empezca has imperatives, 1s in its verb form entry) [FIX]
 * encentar: (wrongly has ie) [OK]
 * encubertar: (wrongly has ie) [FORMS TO DELETE]
 * entrechocar (conjugated with no cons alternation) [FORMS TO DELETE]
 * entredecir (imp_2s has entredice should be entredí; extra future/cond forms entredeciré etc.) [FORMS TO DELETE/FIX]
 * escenografiar: (conjugated without í) [FORMS TO DELETE]
 * estar (incorrect combined forms, e.g. éstela instead of estela) [OK]
 * extasiar: (conjugated with í/+ should be only í) [FORMS TO DELETE]
 * facer (extra form face, fáceme, etc.) [FIX]
 * ferrar: (missing ) [FORMS TO DELETE]
 * gloriar: (conjugated with í/+ should be only í) [FORMS TO DELETE]
 * hacer (extra form hace, háceme, etc.) [DELETE hácete]
 * homogeneizar (extra form homogeneízo) [DELETE homogeneízo, homogeneízas, homogeneíza, homogeneízan, homogeneícen, homogeneícemos]
 * incensar: (missing ) [FORMS TO DELETE]
 * ir (extraneous param aux=ser) [REMOVE PARAM]
 * jacer (extra form jace, jáceme, etc.) [FIX]
 * jarrear (extraneous param impersonal=yes)
 * litografiar: (conjugated without í) [FORMS TO DELETE]
 * manumitir (extra form manumiso, etc.) [FIX; is an adjective]
 * mecanografiar: (conjugated without í) [FORMS TO DELETE]
 * mordiscar (conjugated with no cons alternation) [FORMS TO DELETE]
 * neviscar: (used only in 3rd person singular) [DELETE ALL FIRST AND SECOND PERSON FORMS, ALL 3RD PLURAL FORMS AND ALL PP NON-MS FORMS]
 * obsipar (conjugated as obispar) [DELETE VERB]
 * obstar: (used only in 3rd person) [DELETE ALL FIRST AND SECOND PERSON FORMS]
 * orificar (conjugated as orificiar) [OK]
 * complacer, placer (missing plega in imperative 3s, etc.) [OK]
 * prevaler (extra form preval, incorrect form prévalme etc.) [DELETE preval]
 * preterir: (not given as no_pres_stressed) [FORMS TO DELETE]
 * raer (extra form rao) [DELETE rao]
 * rebordar (extraneous param lang=es) [REMOVE PARAM]
 * redecir (incorrect imp_2s redice instead of redí) [FIX]
 * reinstitucionalizar (conjugated with no cons alternation) [FORMS TO DELETE]
 * reirse [DELETE]
 * serigrafiar: (conjugated without í) [FORMS TO DELETE]
 * sobresalir (incorrect combined forms e.g. sobrésalme) [OK]
 * superpoblar (missing ) [FORMS TO DELETE]
 * sustituir (extra form sustituto, etc.) [OK]
 * usucapir: (used only in inf and pp) [DELETE ALL FORMS BUT PP]
 * valefacer (extra form valeface) [FIX]
 * valer (extra form val, etc.) [OK]

Example: afeitar (afeítate, afeítese, afeítense). Need a script to find them.
 * LOTS OF COMBINED FORMS OF VERBS IN -iar, -uar, -ai-, -au-, -ei-, -eu-, etc.


 * acostar: normally  but in meaning "arrive at the coast", <>
 * adecuar: <+,ú>
 * aerografiar: <í>
 * aferrar: <+,ie[obsolete]>
 * afiliar: <>
 * aforar: in meaning "to gauge, to measure": <>; in meaning "otorgar fuero": 
 * arrecir: 
 * agraviar: <>
 * agriar: <í,+>
 * aserrar: 
 * asolar: in meaning "destroy": ; in meaning "to dry up": <>
 * aspaventar: 
 * atentar: in meaning "to commit a crime" does not have vowel alt
 * atesar: <>; possibly  in obsolete meaning "atiesar"
 * atestar: in meaning "to pack": ; in meaning "testify": <>
 * autoevacuarse: <+,ú>
 * auxiliar: <>
 * avalentar: <>
 * cimentar: <ie,+>
 * colar: most meanings <ue> but "canonically confer (an ecclesiastical benefit)" <>
 * colorir: <no_pres_stressed>
 * comisariar: <í>
 * desafiliar: <>
 * desaforar: "to deprive of fuero": <ue,+[less common]>
 * desairar: <>
 * desolar: <ue,+>
 * desmembrar: <ie,+>
 * dezmar: <>
 * ejecutoriar: <í,+>
 * emparentar: <ie,+>
 * encentar: <>
 * encubertar: <>
 * engrosar: <ue,+>
 * escenografiar: <í>
 * evacuar: <+,ú>
 * expatriar: <í,+>
 * extasiar: <í>
 * ferrar: <ie>
 * follar: <ue>
 * gloriar: <í>
 * hibernar: no vowel alt or e-ie; e-ie not in RAE, ask about it
 * historiar: <í,+>
 * incensar: <ie>
 * invernar: <ie,+>
 * licuar: <+,ú>
 * litografiar: <í>
 * mecanografiar: <í>
 * paliar: <í,+>
 * preterir: <i.no_pres_stressed>
 * promiscuar: <+,ú>
 * readecuar: <+,ú>
 * repatriar: <í,+>
 * retrocar: <ue,+>
 * serigrafiar: <í>
 * soterrar: <ie,+>
 * superpoblar: <ue>
 * templar: <+,ie[in some parts of Latin America]>
 * vidriar: <í,+>

Second round of verbs to fix:


 * apropriar: forms point to reflexive [FORMS TO FIX]
 * autogobernarse: (missing <ie>) [FORMS TO DELETE]
 * autorreproducirse: (wrongly had regular preterite and impf/fut sub) [FORMS TO DELETE]
 * aventarse: (missing <ie>) [FORMS TO DELETE]
 * aventar: forms point to reflexive [FORMS TO FIX]
 * culiar: yo culio or culío? [VERIFY]
 * desnacer: delete dasnatos
 * poseer: delete posesa, posesos, posesas
 * reproducir: (wrongly had regular preterite and impf/fut sub) [FORMS TO DELETE]
 * trasgredir: (wrongly had <i>) [FORMS TO DELETE]

-

Irregular conjugations.

Each entry is processed in turn and consists of an object with two fields: 1. match=: Specifies the irregular verbs that match this object. 2. forms=: Specifies the irregular stems and forms for these verbs.

The value of match= is either a string beginning with "^" (match only the specified verb), a string not beginning with "^" (match any verb ending in that string), or a function that is passed in the verb and should return the prefix of the verb if it matches, otherwise nil. The function match_against_verbs is provided to facilitate matching a set of verbs with a common ending and specific prefixes (e.g. andar and desandar but not mandar, etc.).

The value of forms= is a table specifying stems and individual override forms. Each key of the table names either a stem (e.g. `pres_stressed`), a stem property (e.g. `vowel_alt`) or an individual override form (e.g. `pres_1s`). Each value of a stem can either be a string (a single stem), a list of strings, or a list of objects of the form {form = STEM, footnotes = {FOONOTES}}. Each value of an individual override should be of exactly the same form except that the strings specify full forms rather than stems. The values of a stem property depend on the specific property but are generally strings or booleans.

In order to understand how the stem specifications work, it's important to understand the phonetic modifications done by combine_stem_ending. In general, the complexities of predictable stem and ending modifications are all handled in this function. In particular:

1. Spelling-based modifications (c/z, g/gu, gu/gü, g/j) occur automatically as appropriate for the ending. 2. Raising of e -> i, o -> u in -ir verbs before an ending beginning with i + vowel, as well as in the 1p/2p forms of  the present subjunctive (dormir -> durmiendo, durmió, durmamos), are handled here. Raising happens only for -ir verbs and only when the stem setting `raising_conj` is true (which is normally set to true when vowel alternations  `ie-i`, `ye-i`, `ue-u`, `i`, `í` or `ú` are specified). 3. Numerous modifications are automatically made before an ending beginning with i + vowel. These include: a. final -i of stem absorbed: sonreír -> sonrió, sonriera, sonriendo; b. in the preterite of irregular verbs (likewise for other tenses derived from the preterite stem, i.e. imperfect     and and future subjunctive), initial i absorbed after j and u (dijeron not #dijieron, likewise for condujeron,	  trajeron; also fueron not #fuyeron). This happens only when stem setting `pret_conj` == "irreg"; this must be set explicitly by irregular verbs. Does not apply everywhere because of cases like regular tejer (tejieron not	 #tejeron), regular concluir (concluyeron not #conclueron). c. initial i of ending -> y after vowel and word-initially: poseer -> poseyó, poseyera, poseyendo; ir -> yendo; d. initial i of ending -> y after gü, which becomes gu: argüir -> arguyó, arguyera, arguyendo; e. initial i of ending absorbed after ñ, ll, y: tañer -> tañó, tañera, tañendo; bullir -> bulló, bullera, bullendo 4. If the ending begins with (h)i, it gets an accent after a/e/i/o to prevent the two merging into a diphthong: caer -> caíste, caímos; reír -> reíste, reímos (pres and pret). This does not apply after u, e.g.  concluir -> concluiste, concluimos. 5. In -uir verbs (i.e. -ir verbs with stem ending in -u), a y is added before endings beginning with a/e/o: concluir -> concluyo, concluyen, concluya, concluyamos. Note that preterite concluyó, gerund concluyendo, etc. are handled by a different rule above (3b).

The following stems are recognized:

-- pres_unstressed: The present indicative unstressed stem (2s voseo, 1p, 2p). Also controls the imperative 2p and gerund. Defaults to the infinitive stem. -- pres_stressed: The present indicative stressed stem (1s, 2s, 3s, 3p). Also controls the imperative 2s. Default is empty if indicator `no_pres_stressed`, else a vowel alternation if such an indicator is given (e.g. `ue`, `ì`), else the infinitive stem. -- pres1_and_sub: Overriding stem for 1s present indicative and the entire subjunctive. Only set by irregular verbs and by the indicator `no_pres_stressed` (since verbs of this sort, e.g. aterir, are missing the entire	 subjunctive as well as the forms with stressed root). Used by many irregular verbs, e.g. caer, roer, salir, tener, valer, venir, etc. Some verbs set this and then supply an override for the pres_1sg if it's irregular, e.g. saber, with irregular subjunctive stem "sep-" and special 1s present indicative "sé*" (the * indicates that the monosyllabic accent should not be removed). -- pres1: Special stem for 1s present indicative. Normally, do not set this explicitly. If you need to specify an    irregular 1s present indicative, use the form override pres_1s= to specify the entire form. Defaults to pres1_and_sub if given, else pres_stressed. -- pres_sub_unstressed: The present subjunctive unstressed stem (1p, 2p, also 2s voseo for -ar verbs). Defaults to    pres1_and_sub if given, else the infinitive stem. -- pres_sub_stressed: The present subjunctive stressed stem (1s, 2s, 3s, 1p, also 2s voseo for -er/-ir verbs). Defaults to pres1. -- impf: The imperfect stem. Defaults to the infinitive stem. -- pret: The preterite stem. Defaults to the infinitive stem. -- pret_conj: Determines the set of endings used in the preterite. Should be one of "ar", "er", "ir" or "irreg". Defaults to the conjugation as determined from the infinitive. -- fut: The future stem. Defaults to the infinitive stem. -- cond: The conditional stem. Defaults to fut. -- impf_sub_ra: The imperfect subjunctive -ra stem. Defaults to the preterite stem. -- impf_sub_se: The imperfect subjunctive -se stem. Defaults to the preterite stem. -- fut_sub: The future subjunctive stem. Defaults to the preterite stem. -- pp: The past participle stem. Default is based on the verb conjugation: infinitive stem + "ad" for -ar verbs, otherwise infinitive stem + "id". ]=]

local irreg_conjugations = { {		-- andar, desandar -- we don't want to match e.g. mandar. match = match_against_verbs("andar", {"", "des"}), forms = {pret = "anduv", pret_conj = "irreg"} },	{		-- asir, desasir match = "asir", -- use 'asgu' because we're in a front environment; if we use 'asg', we'll get '#asjo' forms = {pres1_and_sub = "asgu"} },	{		-- abrir, cubrir and compounds match = function(verb) local prefix, base_verb = rmatch(verb, "^(.*)(brir)$") -- Only match abrir, cubrir and compounds, and don't match desabrir/jabrir if not prefix then return nil elseif not prefix:find("a$") and not prefix:find("cu$") then return nil elseif prefix == "desa" or prefix == "ja" then return nil else return prefix, base_verb end end, forms = {pp = "biert"} },	{		match = "caber", forms = {pres1_and_sub = "quep", pret = "cup", pret_conj = "irreg", fut = "cabr"} },	{		-- caer, decaer, descaer, recaer match = "caer", -- use 'caigu' because we're in a front environment; if we use 'caig', we'll get '#caijo' forms = {pres1_and_sub = "caigu"} },	{		-- cocer, escocer, precocer, etc.		match = "cocer", -- override cons_alt, otherwise the verb would be categorized as a c-zc alternating verb forms = {vowel_alt = "ue", pres1 = "cuez", pres_sub_unstressed = "coz", cons_alt = "c-z"}, -- not cozco, as would normally be generated },	{		-- dar, desdar match = match_against_verbs("dar", {"", "des"}), forms = { -- we need to override various present indicative forms and add an accent for the compounds; -- not needed for the simplex and in fact the accents will be removed in that case pres_1s = "doy", pres_2s = "dás", pres_3s = "dá", pres_2p = "dáis", pres_3p = "dán", pret = "d", pret_conj = "er", pres_sub_1s = "dé*", -- * signals that the monosyllabic accent must remain pres_sub_2s = "dés", pres_sub_3s = "dé*", pres_sub_2p = "déis", pres_sub_3p = "dén", imp_2s = "dá", }	},	{		-- decir, redecir, entredecir match = match_against_verbs("decir", {"", "re", "entre"}), forms = { -- for this and variant verbs in -decir, we set cons_alt to false because we don't want the -- verb categorized as a c-zc alternating verb, which would happen by default -- use 'digu' because we're in a front environment; if we use 'dig', we'll get '#dijo' pres1_and_sub = "digu", vowel_alt = "i", cons_alt = false, pret = "dij", pret_conj = "irreg", pp = "dich", fut = "dir", imp_2s = "dí" -- need the accent for the compounds; it will be removed in the simplex }	},	{		-- antedecir, interdecir match = match_against_verbs("decir", {"ante", "inter"}), forms = { pres1_and_sub = "digu", vowel_alt = "i", cons_alt = false, pret = "dij", pret_conj = "irreg", pp = "dich", fut = "dir" -- imp_2s regular }	},	{		-- bendecir, maldecir match = match_against_verbs("decir", {"ben", "mal"}), forms = { pres1_and_sub = "digu", vowel_alt = "i", cons_alt = false, pret = "dij", pret_conj = "irreg", pp = {"decid", "dit"} -- imp_2s regular, fut regular }	},	{		-- condecir, contradecir, desdecir, predecir, others? match = "decir", forms = { pres1_and_sub = "digu", vowel_alt = "i", cons_alt = false, pret = "dij", pret_conj = "irreg", pp = "dich", fut = {"decir", "dir"} -- imp_2s regular }	},	{		-- conducir, producir, reducir, traducir, etc.		match = "ducir", forms = {pret = "duj", pret_conj = "irreg"} },	{		-- elegir, reelegir; not preelegir, per RAE match = match_against_verbs("elegir", {"", "re"}), forms = {vowel_alt = "i", pp = {"elegid", "elect"}} },	{		-- erguir match = "^erguir", forms = { vowel_alt = {"i", "ye-i"}, -- We would not need to make this verb irregular except for the alternative unstressed present subjunctive forms -- in yerg-. We can't use pres_sub_unstressed = {"irgu", "yergu"} because the "i" and "ye-i" vowel alternations -- are e->i raising in the unstressed present subjunctive and we wrongly get irgamos/yirgamos etc.			pres_sub_2sv = {"irgás", "yergás"}, pres_sub_1p = {"irgamos", "yergamos"}, pres_sub_2p = {"irgáis", "yergáis"}, },	},	{		match = "^estar", forms = { pres_1s = "estoy", pres_2s = "estás", pres_2sv = "estás", pres_3s = "está", pres_3p = "están", pret = "estuv", pret_conj = "irreg", pres_sub_1s = "esté", pres_sub_2s = "estés", pres_sub_2sv = "estés", pres_sub_3s = "esté", pres_sub_3p = "estén", imp_2s = "está", imp_2sv = "está", }	},	{		-- freír, refreír match = "freír", forms = {vowel_alt = "í", pp = {"freíd", "frit"}} },	{		match = "garantir", forms = { pres_stressed = , pres1_and_sub = , }	},	{		match = "^haber", forms = { pres_1s = "he", pres_2s = "has", pres_2sv = "has", pres_3s = {"ha", {form = "hay", footnotes = {"[used impersonally]"}}}, pres_1p = "hemos", pres_3p = "han", pres1_and_sub = "hay", -- only for subjunctive as we override pres_1s pret = "hub", pret_conj = "irreg", imp_2s = {"habe", "he"}, imp_2sv = {"habe", "he"}, fut = "habr", }	},	{		match = "satisfacer", forms = { -- see below for cons_alt setting and pres1_and_sub setting pres1_and_sub = "satisfagu", cons_alt = false, pret = "satisfic", pret_conj = "irreg", pp = "satisfech", fut = "satisfar", imp_2s = {"satisface", "satisfaz"} }	},	{		match = match_against_verbs("hacer", {"contra", "re"}), -- contrahacer/rehacer require an extra accent in the preterite (rehíce, rehízo). forms = { -- see below for cons_alt setting and pres1_and_sub setting pres1_and_sub = "hagu", cons_alt = false, pret = "hic", pret_1s = "híce", pret_3s = "hízo", pret_conj = "irreg", pp = "hech", fut = "har", imp_2s = "haz" }	},	{		-- hacer, deshacer, contrahacer, rehacer, facer, desfacer, jacer match = function(verb) return rmatch(verb, "^(.*[hjf])(acer)$") end, forms = { -- for these verbs, we set cons_alt to false because we don't want the verb categorized as a			-- c-zc alternating verb, which would happen by default -- use 'agu' because we're in a front environment; if we use 'ag', we'll get '#hajo' pres1_and_sub = "agu", cons_alt = false, pret = "ic", pret_conj = "irreg", pp = "ech", fut = "ar", imp_2s = "az" }	},	{		-- imprimir, reimprimir match = "imprimir", forms = {pp = {"imprimid", "impres"}} },	{		-- infecir match = "infecir", -- override cons_alt, otherwise the verb would be categorized as a c-zc alternating verb forms = {vowel_alt = "i", pres1 = "infiz", pres_sub_unstressed = "infez", cons_alt = "c-z"}, -- not infizco, as would normally be generated },	{		match = "infecir", -- override cons_alt, otherwise the verb would be categorized as a c-zc alternating verb forms = {pres1_and_sub = "infez", cons_alt = "c-z"}, -- not mezco, as would normally be generated },	{		match = "^ir", forms = { pres_1s = "voy", pres_2s = "vas", pres_2sv = "vas", pres_3s = "va", pres_1p = "vamos", pres_2p = "vais", pres_3p = "van", pres1_and_sub = "vay", -- only for subjunctive as we override pres_1s full_impf = "ib", impf_1p = "íbamos", pret = "fu", pret_conj = "irreg", -- this signals that fu + -ieron -> fueron not fuyeron pret_1s = "fui", pret_3s = "fue", imp_2s = "ve", imp_2sv = "andá", imp_1p = {"vamos", "vayamos"}, refl_imp_2p = {"idos", "iros"}, imp_2p_comb_os = {"idos", "iros"}, }	},	{		-- mecer, remecer -- we don't want to match e.g. adormecer, estremecer match = match_against_verbs("mecer", {"re", ""}), -- override cons_alt, otherwise the verb would be categorized as a c-zc alternating verb forms = {pres1_and_sub = "mez", cons_alt = "c-z"}, -- not mezco, as would normally be generated },	{		-- morir, desmorir, premorir match = "morir", forms = {vowel_alt = "ue-u", pp = "muert"}, },	{		-- oír, desoír, entreoír, trasoír match = "oír", -- use 'oigu' because we're in a front environment; if we use 'oig', we'll get '#oijo' forms = {pres1_and_sub = "oigu"} },	{		match = "olver", -- solver, volver, bolver and derivatives forms = {vowel_alt = "ue", pp = "uelt"} },	{		-- placer, desplacer match = "placer", forms = { pret_3s = {"plació", {form = "plugo", footnotes = {"[archaic]"}}}, pret_3p = {"placieron", {form = "pluguieron", footnotes = {"[archaic]"}}}, pres_sub_3s = {"plazca", {form = "plega", footnotes = {"[archaic]"}}, {form = "plegue", footnotes = {"[archaic]"}}}, impf_sub_ra_3s = {"placiera", {form = "pluguiera", footnotes = {"[archaic]"}}}, impf_sub_ra_3p = {"placieran", {form = "pluguieran", footnotes = {"[archaic]"}}}, impf_sub_se_3s = {"placiese", {form = "pluguiese", footnotes = {"[archaic]"}}}, impf_sub_se_3p = {"placiesen", {form = "pluguiesen", footnotes = {"[archaic]"}}}, fut_sub_3s = {"placiere", {form = "pluguiere", footnotes = {"[archaic]"}}}, fut_sub_3p = {"placieren", {form = "pluguieren", footnotes = {"[archaic]"}}}, }	},	{		match = "poder", forms = {vowel_alt = "ue", pret = "pud", pret_conj = "irreg", fut = "podr", gerund = "pudiendo"} },	{		-- poner, componer, deponer, imponer, oponer, suponer, many others match = "poner", forms = { -- use 'pongu' because we're in a front environment; if we use 'pong', we'll get '#ponjo' pres1_and_sub = "pongu", pret = "pus", pret_conj = "irreg", fut = "pondr", pp = "puest", imp_2s = "pón" -- need the accent for the compounds; it will be removed in the simplex }	},	{		-- proveer, desproveer match = "proveer", forms = {pp = {"proveíd", "provist"}}, },	{		match = "pudrir", forms = {pp = "podrid"} },	{		-- querer, desquerer, malquerer match = "querer", forms = {vowel_alt = "ie", pret = "quis", pret_conj = "irreg", fut = "querr"} },	{		match = "^raer", -- use 'raigu' because we're in a front environment; if we use 'raig', we'll get '#raijo' forms = {pres1_and_sub = {"raigu", "ray"}} },	{		-- roer, corroer match = "roer", -- use 'roigu' because we're in a front environment; if we use 'roig', we'll get '#roijo' forms = {pres1_and_sub = {"ro", "roigu", "roy"}} },	{		-- romper, entrerromper, arromper, derromper; not corromper; FIXME: not sure about interromper (obsolete) match = function(verb) local prefix, base_verb = rmatch(verb, "^(.*)(romper)$") -- Don't match corromper if prefix == "cor" then return nil else return prefix, base_verb end end, forms = {pp = "rot"} },	{		-- saber, resaber match = "saber", forms = { pres_1s = "sé*", -- * signals that the monosyllabic accent must remain pres1_and_sub = "sep", -- only for subjunctive as we override pres_1s pret = "sup", pret_conj = "irreg", fut = "sabr", }	},	{		match = "salir", forms = { -- use 'salgu' because we're in a front environment; if we use 'salg', we'll get '#saljo' pres1_and_sub = "salgu", fut = "saldr", imp_2s = "sal", -- These don't exist per the RAE. imp_2s_comb_lo = {}, imp_2s_comb_los = {}, imp_2s_comb_la = {}, imp_2s_comb_las = {}, imp_2s_comb_le = {}, imp_2s_comb_les = {}, },	},	{		match = "scribir", -- escribir, describir, proscribir, etc.		forms = {pp = {"scrit", {form = "script", footnotes = {"[Argentina and Uruguay]"}}}} },	{		match = "^ser", forms = { pres_1s = "soy", pres_2s = "eres", pres_2sv = "sos", pres_3s = "es", pres_1p = "somos", pres_2p = "sois", pres_3p = "son", pres1_and_sub = "se", -- only for subjunctive as we override pres_1s full_impf = "er", impf_1p = "éramos", pret = "fu", pret_conj = "irreg", -- this signals that fu + -ieron -> fueron not fuyeron pret_1s = "fui", pret_3s = "fue", fut = "ser", imp_2s = "sé*", -- * signals that the monosyllabic accent must remain imp_2sv = "sé*", }	},	{		match = "^soler", forms = { vowel_alt = "ue", fut = , fut_sub = , pp = , }	},	{		-- tener, abstener, contener, detener, obtener, sostener, and many others match = "tener", forms = { -- use 'tengu' because we're in a front environment; if we use 'teng', we'll get '#tenjo' pres1_and_sub = "tengu", vowel_alt = "ie", pret = "tuv", pret_conj = "irreg", fut = "tendr", imp_2s = "tén" -- need the accent for the compounds; it will be removed in the simplex }	},	{		-- traer, atraer, detraer, distraer, extraer, sustraer, and many others match = "traer", -- use 'traigu' because we're in a front environment; if we use 'traig', we'll get '#traijo' forms = {pres1_and_sub = "traigu", pret = "traj", pret_conj = "irreg"} },	{		-- valer, equivaler, prevaler match = "valer", -- use 'valgu' because we're in a front environment; if we use 'valg', we'll get '#valjo' forms = {pres1_and_sub = "valgu", fut = "valdr"} },	{		match = "venir", forms = { -- use 'vengu' because we're in a front environment; if we use 'veng', we'll get '#venjo' pres1_and_sub = "vengu", vowel_alt = "ie-i", pret = "vin", pret_conj = "irreg", -- uniquely for this verb, pres sub 2sv/1p/2p do not raise the vowel even though we are an			-- e-ie-i verb (contrast sentir -> sintamos/sintáis) pres_sub_2sv = "vengás", pres_sub_1p = "vengamos", pres_sub_2p = "vengáis", fut = "vendr", imp_2s = "vén" -- need the accent for the compounds; it will be removed in the simplex }	},	{		-- We want to match antever etc. but not atrever etc. No way to avoid listing each verb. match = match_against_verbs("ver", {"ante", "entre", "pre", "re", ""}), forms = { -- we need to override various present indicative forms and add an accent for the compounds; -- not needed for the simplex and in fact the accents will be removed in that case pres_2s = "vés", pres_2sv = "vés", pres_3s = "vé", pres_2p = "véis", pres_3p = "vén", pres1_and_sub = "ve", impf = "ve", pp = "vist", imp_2s = "vé" -- need the accent for the compounds; it will be removed in the simplex }	},	{		-- yacer, adyacer, subyacer match = "yacer", -- use 'yazqu/yazgu/yagu' because we're in a front environment; see 'decir' above forms = {pres1_and_sub = {"yazqu", "yazgu", "yagu"}, imp_2s = {"yace", "yaz"}} }, }

local reflexive_masc_forms = { ["su"] = {"mi", "tu", "su", "nuestro", "vuestro", "su"}, ["sus"] = {"mis", "tus", "sus", "nuestros", "vuestros", "sus"}, ["sí"] = {"mí", "ti", "sí", "nosotros", "vosotros", "sí"}, ["consigo"] = {"conmigo", "contigo", "consigo", "con nosotros", "con vosotros", "consigo"}, }

local reflexive_fem_forms = { ["su"] = {"mi", "tu", "su", "nuestra", "vuestra", "su"}, ["sus"] = {"mis", "tus", "sus", "nuestras", "vuestras", "sus"}, ["sí"] = {"mí", "ti", "sí", "nosotras", "vosotras", "sí"}, ["consigo"] = {"conmigo", "contigo", "consigo", "con nosotras", "con vosotras", "consigo"}, }

local reflexive_forms = { ["se"] = {"me", "te", "se", "nos", "os", "se"}, ["suyo"] = {"mío", "tuyo", "suyo", "nuestro", "vuestro", "suyo"}, ["suya"] = {"mía", "tuya", "suya", "nuestra", "vuestra", "suya"}, ["suyos"] = {"míos", "tuyos", "suyos", "nuestros", "vuestros", "suyos"}, ["suyas"] = {"mías", "tuyas", "suyas", "nuestras", "vuestras", "suyas"}, }

local function skip_slot(base, slot, allow_overrides) if not allow_overrides and (base.basic_overrides[slot] or base.combined_overrides[slot] or		base.refl and base.basic_reflexive_only_overrides[slot]) then -- Skip any slots for which there are overrides. return true end

if base.only3s and (slot:find("^pp_f") or slot:find("^pp_mp")) then -- diluviar, atardecer, neviscar; impersonal verbs have only masc sing pp		return true end

if not slot:find("[123]") then -- Don't skip non-personal slots. return false end

if base.nofinite then return true end

if (base.only3s or base.only3sp or base.only3p) and (slot:find("^imp_") or slot:find("^neg_imp_")) then return true end

if base.only3s and not slot:find("3s") then -- diluviar, atardecer, neviscar return true end

if base.only3sp and not slot:find("3[sp]") then -- atañer, concernir return true end

if base.only3p and not slot:find("3p") then -- caer cuatro gotas, caer chuzos de punta, entrarle los siete males return true end

return false end

local function escape_reflexive_indicators(arg1) if not arg1:find("pron>") then return arg1 end local segments = iut.parse_balanced_segment_run(arg1, "<", ">") -- Loop over every other segment. The even-numbered segments are angle-bracket specs while -- the odd-numbered segments are the text between them. for i = 2, #segments - 1, 2 do		if segments[i] == " " then segments[i] = "⦃⦃mpron⦄⦄" elseif segments[i] == " " then segments[i] = "⦃⦃fpron⦄⦄" elseif segments[i] == " " then segments[i] = "⦃⦃pron⦄⦄" end end return table.concat(segments) end

local function undo_escape_form(form) -- assign to var to throw away second value local newform = form:gsub("⦃⦃", "<"):gsub("⦄⦄", ">") return newform end

local function remove_reflexive_indicators(form) -- assign to var to throw away second value local newform = form:gsub("⦃⦃.-⦄⦄", "") return newform end

local function replace_reflexive_indicators(slot, form) if not form:find("⦃") then return form end error("Internal error: replace_reflexive_indicators not implemented yet") end

-- Add the `stem` to the `ending` for the given `slot` and apply any phonetic modifications. -- `is_combining_ending` is true if `ending` is actually the ending (this function is also -- called to combine prefix + stem). WARNING: This function is written very carefully; changes -- to it can easily have unintended consequences. local function combine_stem_ending(base, slot, stem, ending, is_combining_ending) if not is_combining_ending then return stem .. ending end

if base.stems.raising_conj and (rfind(ending, "^i" .. V) or		slot == "pres_sub_1p" or slot == "pres_sub_2p" or slot == "pres_sub_2sv") then -- need to raise e -> i, o -> u: dormir -> durmió, durmiera, durmiendo, durmamos stem = rsub(stem, "([eo])(" .. C .. "*)$", function(vowel, rest) return raise_vowel[vowel] .. rest end) -- also with stem ending in -gu or -qu (e.g. erguir -> irguió, irguiera, irguiendo, irgamos) stem = rsub(stem, "([eo])(" .. C .. "*[gq]u)$", function(vowel, rest) return raise_vowel[vowel] .. rest end) end

-- Lots of sound changes involving endings beginning with i + vowel if rfind(ending, "^i" .. V) then -- (1) final -i of stem absorbed: sonreír -> sonrió, sonriera, sonriendo; note that this rule may be fed -- by the preceding one (stem sonre- raised to sonri-, then final i absorbed) stem = stem:gsub("i$", "")

-- (2) In the preterite of irregular verbs (likewise for other tenses derived from the preterite stem, i.e.		--    imperfect and future subjunctive), initial i absorbed after j (dijeron not #dijieron, likewise for		--     condujeron, trajeron) and u (fueron not #fuyeron). Does not apply in regular verb tejer (tejieron not		--    #tejeron) and concluir (concluyeron not #conclueron). if base.stems.pret_conj == "irreg" and rfind(stem, "[ju]$") then ending = ending:gsub("^i", "") end

-- (3) initial i -> y after vowel and word-initially: poseer -> poseyó, poseyera, poseyendo; -- concluir -> concluyó, concluyera, concluyendo; ir -> yendo; but not conseguir/delinquir if stem == "" or (rfind(stem, V .. "$") and not rfind(stem, "[gq]u$")) then ending = ending:gsub("^i", "y") end

-- (4) -gü + ie- -> -guye-: argüir -> arguyó, arguyera, arguyendo if stem:find("gü$") then -- transfer the y to the stem to avoid gü -> gu below in front/back conversions stem = stem:gsub("ü$", "uy") ending = ending:gsub("^i", "") end

-- (5) initial i absorbed after ñ, ll, y: tañer -> tañó, tañera, tañendo; bullir -> bulló, bullera, bullendo if rfind(stem, "[ñy]$") or rfind(stem, "ll$") then ending = ending:gsub("^i", "") end end

-- If ending begins with i, it must get an accent after a/e/i/o to prevent the two merging into a diphthong: -- caer -> caíste, caímos; reír -> reíste, reímos (pres and pret). This does not apply after u, e.g. -- concluir -> concluiste, concluimos. if ending:find("^i") and stem:find("[aeio]$") then ending = ending:gsub("^i", "í") end

-- If -oír/-uir (i.e. -ir with stem ending in -o/u, e.g. oír, concluir), a y must be added before endings -- beginning with a/e/o. Check for base.stems.pret_conj == "irreg" to exclude stem fu- of ir. if base.conj == "ir" and rfind(ending, "^[aeoáéó]") and base.stems.pret_conj ~= "irreg" then if rfind(stem, "[oú]$") then -- oír -> oye, rehuir -> rehúyo/rehúye (with indicator 'ú') stem = stem .. "y" elseif rfind(stem, "[^gq]u$") then -- concluir, but not conseguir or delinquir stem = stem .. "y" elseif stem:find("ü$") then -- argüir -> arguyendo stem = stem:gsub("ü$", "uy") end end

-- Spelling changes in the stem; it depends on whether the stem given is the pre-front-vowel or -- pre-back-vowel variant, as indicated by `frontback`. We want these front-back spelling changes to happen -- between stem and ending, not between prefix and stem; the prefix may not have the same "front/backness" -- as the stem. local is_front = rfind(ending, "^[eiéí]") if base.frontback == "front" and not is_front then -- parecer -> parezco, conducir -> conduzco; use zqu to avoid triggering the following gsub; -- the third line will replace zqu -> zc		if slot ~= "pret_3s" then -- exclude hice -> hizo (not #hizco) stem = rsub(stem, "(" .. V .. ")c$", "%1zqu") end stem = stem:gsub("sc$", "squ") -- evanescer -> evanesco, fosforescer -> fosforesco stem = stem:gsub("c$", "z") -- ejercer -> ejerzo, uncir -> unzo stem = stem:gsub("qu$", "c") -- delinquir -> delinco, parecer -> parezqu- -> parezco stem = stem:gsub("g$", "j") -- coger -> cojo, afligir -> aflijo stem = stem:gsub("gu$", "g") -- distinguir -> distingo stem = stem:gsub("gü$", "gu") -- may not occur; argüir -> arguyo handled above elseif base.frontback == "back" and is_front then stem = stem:gsub("gu$", "gü") -- averiguar -> averigüé stem = stem:gsub("g$", "gu") -- cargar -> cargué stem = stem:gsub("c$", "qu") -- marcar -> marqué stem = rsub(stem, "[çz]$", "c") -- aderezar/adereçar -> aderecé end

return stem .. ending end

local function add(base, slot, stems, endings, is_combining_ending, allow_overrides) if skip_slot(base, slot, allow_overrides) then return end local function do_combine_stem_ending(stem, ending) return combine_stem_ending(base, slot, stem, ending, is_combining_ending) end iut.add_forms(base.forms, slot, stems, endings, do_combine_stem_ending, nil, nil, base.all_footnotes) end

local function add3(base, slot, prefix, stems, endings, allow_overrides) if prefix == "" then return add(base, slot, stems, endings, "is combining ending", allow_overrides) end

if skip_slot(base, slot, allow_overrides) then return end

local is_combining_ending = false

local function do_combine_stem_ending(stem, ending) return combine_stem_ending(base, slot, stem, ending, is_combining_ending) end

-- Have to reimplement add_multiple_forms ourselves due to the is_combining_ending -- flag, which needs to be different when adding prefix to stems vs. stems to ending. -- Otherwise we get e.g. #reímpreso instead of reimpreso. local tempdest = {} iut.add_forms(tempdest, slot, prefix, stems, do_combine_stem_ending) is_combining_ending = true iut.add_forms(base.forms, slot, tempdest[slot], endings, do_combine_stem_ending) end

local function insert_form(base, slot, form) if not skip_slot(base, slot) then iut.insert_form(base.forms, slot, form) end end

local function insert_forms(base, slot, forms) if not skip_slot(base, slot) then iut.insert_forms(base.forms, slot, forms) end end

local function add_single_stem_tense(base, slot_pref, stems, s1, s2, s3, p1, p2, p3) local function addit(slot, ending) add3(base, slot_pref .. "_" .. slot, base.prefix, stems, ending) end addit("1s", s1) addit("2s", s2) addit("3s", s3) addit("1p", p1) addit("2p", p2) addit("3p", p3) end

local function add_present_indic(base) local function addit(slot, stems, ending) add3(base, "pres_" .. slot, base.prefix, stems, ending) end local s2, s2v, s3, p1, p2, p3	if base.conj == "ar" then s2, s2v, s3, p1, p2, p3 = "as", "ás", "a", "amos", "áis", "an" elseif base.conj == "er" then s2, s2v, s3, p1, p2, p3 = "es", "és", "e", "emos", "éis", "en" elseif base.conj == "ir" then s2, s2v, s3, p1, p2, p3 = "es", "ís", "e", "imos", "ís", "en" else error("Internal error: Unrecognized conjugation " .. base.conj) end

addit("1s", base.stems.pres1, "o") addit("2s", base.stems.pres_stressed, s2) addit("2sv", base.stems.pres_unstressed, s2v) addit("3s", base.stems.pres_stressed, s3) addit("1p", base.stems.pres_unstressed, p1) addit("2p", base.stems.pres_unstressed, p2) addit("3p", base.stems.pres_stressed, p3) end

local function add_present_subj(base) local function addit(slot, stems, ending) add3(base, "pres_sub_" .. slot, base.prefix, stems, ending) end local s1, s2, s2v, s3, p1, p2, p3, voseo_stem if base.conj == "ar" then s1, s2, s2v, s3, p1, p2, p3 = "e", "es", "és", "e", "emos", "éis", "en" else s1, s2, s2v, s3, p1, p2, p3 = "a", "as", "ás", "a", "amos", "áis", "an" end

addit("1s", base.stems.pres_sub_stressed, s1) addit("2s", base.stems.pres_sub_stressed, s2) addit("2sv", base.stems.pres_sub_unstressed, s2v) addit("3s", base.stems.pres_sub_stressed, s3) addit("1p", base.stems.pres_sub_unstressed, p1) addit("2p", base.stems.pres_sub_unstressed, p2) addit("3p", base.stems.pres_sub_stressed, p3) end

local function add_imper(base) local function addit(slot, stems, ending) add3(base, "imp_" .. slot, base.prefix, stems, ending) end if base.conj == "ar" then addit("2s", base.stems.pres_stressed, "a") addit("2sv", base.stems.pres_unstressed, "á") addit("2p", base.stems.pres_unstressed, "ad") elseif base.conj == "er" then addit("2s", base.stems.pres_stressed, "e") addit("2sv", base.stems.pres_unstressed, "é") addit("2p", base.stems.pres_unstressed, "ed") elseif base.conj == "ir" then addit("2s", base.stems.pres_stressed, "e") addit("2sv", base.stems.pres_unstressed, "í") addit("2p", base.stems.pres_unstressed, "id") else error("Internal error: Unrecognized conjugation " .. base.conj) end end

local function add_finite_non_present(base) local function add_tense(slot, stem, s1, s2, s3, p1, p2, p3) add_single_stem_tense(base, slot, stem, s1, s2, s3, p1, p2, p3) end

local stems = base.stems

if stems.full_impf then -- An override needs to be supplied for the impf_1p due to the accent on the stem. add_tense("impf", stems.full_impf, "a", "as", "a", {}, "ais", "an") elseif base.conj == "ar" then add_tense("impf", stems.impf, "aba", "abas", "aba", "ábamos", "abais", "aban") else add_tense("impf", stems.impf, "ía", "ías", "ía", "íamos", "íais", "ían") end

if stems.pret_conj == "irreg" then add_tense("pret", stems.pret, "e", "iste", "o", "imos", "isteis", "ieron") elseif stems.pret_conj == "ar" then add_tense("pret", stems.pret, "é", "aste", "ó", "amos", "asteis", "aron") else add_tense("pret", stems.pret, "í", "iste", "ió", "imos", "isteis", "ieron") end

if stems.pret_conj == "ar" then add_tense("impf_sub_ra", stems.impf_sub_ra, "ara", "aras", "ara", "áramos", "arais", "aran") add_tense("impf_sub_se", stems.impf_sub_se, "ase", "ases", "ase", "ásemos", "aseis", "asen") add_tense("fut_sub", stems.fut_sub, "are", "ares", "are", "áremos", "areis", "aren") else add_tense("impf_sub_ra", stems.impf_sub_ra, "iera", "ieras", "iera", "iéramos", "ierais", "ieran") add_tense("impf_sub_se", stems.impf_sub_se, "iese", "ieses", "iese", "iésemos", "ieseis", "iesen") add_tense("fut_sub", stems.fut_sub, "iere", "ieres", "iere", "iéremos", "iereis", "ieren") end

add_tense("fut", stems.fut, "é", "ás", "á", "emos", "éis", "án") add_tense("cond", stems.cond, "ía", "ías", "ía", "íamos", "íais", "ían") end

local function add_non_finite_forms(base) local stems = base.stems local function addit(slot, stems, ending) add3(base, slot, base.prefix, stems, ending) end insert_form(base, "infinitive", {form = base.verb}) -- Also insert "infinitive + reflexive pronoun" combinations if we're handling a reflexive verb. See comment below for -- "gerund + reflexive pronoun" combinations. if base.refl then for _, persnum in ipairs(person_number_list_basic) do insert_form(base, "infinitive_" .. persnum, {form = base.verb}) end end local ger_ending = base.conj == "ar" and "ando" or "iendo" addit("gerund", stems.pres_unstressed, ger_ending) -- Also insert "gerund + reflexive pronoun" combinations if we're handling a reflexive verb. We insert exactly the same -- form as for the bare gerund; later on in add_reflexive_or_fixed_clitic_to_forms, we add the appropriate clitic -- pronouns. It's important not to do this for non-reflexive verbs, because in that case, the clitic pronouns won't be	-- added, and will wrongly consider all these combinations as possible inflections of the bare -- gerund. Thanks to User:JeffDoozan for this bug fix. if base.refl then for _, persnum in ipairs(person_number_list_basic) do addit("gerund_" .. persnum, stems.pres_unstressed, ger_ending) end end addit("pp_ms", stems.pp, "o") addit("pp_fs", stems.pp, "a") addit("pp_mp", stems.pp, "os") addit("pp_fp", stems.pp, "as") end

local function copy_subjunctives_to_imperatives(base) -- Copy subjunctives to imperatives, unless there's an override for the given slot (as with the imp_1p of ir). for _, persnum in ipairs({"3s", "1p", "3p"}) do local from = "pres_sub_" .. persnum local to = "imp_" .. persnum insert_forms(base, to, iut.map_forms(base.forms[from], function(form) return form end)) end end

-- Remove monosyllabic accents (e.g. the 3sg preterite of fiar is fio not #fió). Note that there are a -- few monosyllabic verb forms that intentionally have an accent, to distinguish them from other words -- with the same pronunciation. These are as follows: -- (1) sé 1sg present indicative of saber; -- (2) sé 2sg imperative of ser; -- (3) dé 1sg and 3sg present subjunctive of dar. -- For these, a * is added, which indicates that the accent needs to remain. If we see such a *, we remove -- it but otherwise leave the form alone. local function remove_monosyllabic_accents(base) for _, slotaccel in ipairs(base.alternant_multiword_spec.verb_slots_basic) do		local slot, accel = unpack(slotaccel) if base.forms[slot] then for _, form in ipairs(base.forms[slot]) do				if form.form:find("%*") then -- * means leave alone any accented vowel form.form = form.form:gsub("%*", "") elseif not rfind(form.form, "^%-") and rfind(form.form, AV) and not rfind(form.form, V .. C .. V) then -- Has an accented vowel and no VCV sequence and not a suffix; may be monosyllabic, in which -- case we need to remove the accent. Check # of syllables and remove accent if only 1. Note -- that the checks for accented vowel and VCV sequence are not strictly needed, but are -- optimizations to avoid running the whole syllabification algorithm on every verb form. local syllables = com.syllabify(form.form) if #syllables == 1 then form.form = com.remove_accent_from_syllable(syllables[1]) end end end end end end

-- Add the appropriate clitic pronouns in `clitics` to the forms in `base_slot`. `store_cliticized_form` is a function -- of three arguments (clitic, formobj, cliticized_form) and should store the cliticized form for the specified clitic -- and form object. local function add_forms_with_clitic(base, base_slot, clitics, store_cliticized_form) if not base.forms[base_slot] then -- This can happen, e.g. in only3s/only3sp/only3p verbs. return end for _, formobj in ipairs(base.forms[base_slot]) do -- Figure out the correct accenting of the verb when a clitic pronoun is attached to it. We may need to		-- add or remove an accent mark: -- (1) No accent mark currently, none needed: infinitive sentar -> sentarlo; imperative singular ten -> tenlo; -- (2) Accent mark currently, still needed: infinitive oír -> oírlo; -- (3) No accent mark currently, accent needed: imperative singular siente -> siéntelo; -- (4) Accent mark currently, not needed: imperative singular está -> estalo, sé -> selo. local syllables = com.syllabify(formobj.form) local sylno = com.stressed_syllable(syllables) table.insert(syllables, "lo") local needs_accent = com.accent_needed(syllables, sylno) if needs_accent then syllables[sylno] = com.add_accent_to_syllable(syllables[sylno]) else syllables[sylno] = com.remove_accent_from_syllable(syllables[sylno]) end table.remove(syllables) -- remove added clitic pronoun local reaccented_form = table.concat(syllables) for _, clitic in ipairs(clitics) do			local cliticized_form -- Some further special cases. if base_slot == "imp_1p" and (clitic == "nos" or clitic == "os") then -- Final -s disappears: sintamos + nos -> sintámonos, sintamos + os -> sintámoos cliticized_form = reaccented_form:gsub("s$", "") .. clitic elseif base_slot == "imp_2p" and clitic == "os" then -- Final -d disappears, which may cause an accent to be required: -- haced + os -> haceos, sentid + os -> sentíos if reaccented_form:find("id$") then cliticized_form = reaccented_form:gsub("id$", "íos") else cliticized_form = reaccented_form:gsub("d$", "os") end else cliticized_form = reaccented_form .. clitic end store_cliticized_form(clitic, formobj, cliticized_form) end end end

-- Generate the combinations of verb form (infinitive, gerund or various imperatives) + clitic pronoun. local function add_combined_forms(base) for _, base_slot_and_clitics in ipairs(base.alternant_multiword_spec.verb_slot_combined_rows) do		local base_slot, clitics = unpack(base_slot_and_clitics) add_forms_with_clitic(base, base_slot, clitics,			function(clitic, formobj, cliticized_form)				insert_form(base, base_slot .. "_comb_" .. clitic, {form = cliticized_form, footnotes = formobj.footnotes})			end		) end for _, single_comb_slot_and_clitics in ipairs(base.alternant_multiword_spec.verb_slot_double_combined_rows) do		local single_comb_slot, clitics = unpack(single_comb_slot_and_clitics) add_forms_with_clitic(base, single_comb_slot, clitics,			function(clitic, formobj, cliticized_form)				insert_form(base, single_comb_slot .. "_" .. clitic, {form = cliticized_form, footnotes = formobj.footnotes})			end		) end end

local function process_slot_overrides(base, do_basic, reflexive_only) local overrides = reflexive_only and base.basic_reflexive_only_overrides or		do_basic and base.basic_overrides or base.combined_overrides for slot, forms in pairs(overrides) do		add(base, slot, base.prefix, forms, false, "allow overrides") end end

-- Prefix `form` with `clitic`, adding fixed text `between` between them. Add links as appropriate unless the user -- requested no links. Check whether form already has brackets (as will be the case if the form has a fixed clitic). local function add_clitic_to_form(base, clitic, between, form) if base.alternant_multiword_spec.args.noautolinkverb then return clitic .. between .. form else local clitic_pref = "" .. clitic .. "" .. between if form:find("%[%[") then return clitic_pref .. form else return clitic_pref .. "" .. form .. "" end end end

-- Add a reflexive pronoun or fixed clitic, e.g. lo, as appropriate to the base form that were generated. -- `do_joined` means to do only the forms where the pronoun is joined to the end of the form; otherwise, do only the -- forms where it is not joined and precedes the form. local function add_reflexive_or_fixed_clitic_to_forms(base, do_reflexive, do_joined) for _, slotaccel in ipairs(base.alternant_multiword_spec.verb_slots_basic) do		local slot, accel = unpack(slotaccel) local clitic if not do_reflexive then clitic = base.clitic elseif slot:find("[123]") then local persnum = slot:match("^.*_(.-)$") clitic = person_number_to_reflexive_pronoun[persnum] else clitic = "se" end if base.forms[slot] then if do_reflexive and slot:find("^pp_") or slot == "infinitive_linked" then -- do nothing with reflexive past participles or with infinitive linked (handled at the end) elseif slot:find("^neg_imp_") then error("Internal error: Should not have forms set for negative imperative at this stage") else local slot_has_suffixed_clitic = slot:find("infinitive") or slot:find("gerund") or slot:find("^imp_") -- Maybe generate non-reflexive parts and separated syntactic variants for use in. -- See comment in add_slots above `need_special_verb_form_of_slots`. Check for do_joined so we only -- run this code once. if do_reflexive and do_joined and base.alternant_multiword_spec.from_verb_form_of and -- Skip personal variants of infinitives and gerunds so we don't think jambando is a -- non-reflexive equivalent of jambándome. not slot:find("infinitive_") and not slot:find("gerund_") then -- Clone the forms because we will be destructively modifying them just below, adding the reflexive -- pronoun. insert_forms(base, slot .. "_non_reflexive", mw.clone(base.forms[slot])) if slot_has_suffixed_clitic then insert_forms(base, slot .. "_variant", iut.map_forms(base.forms[slot], function(form) return add_clitic_to_form(base, clitic, " ... ", form) end)) end end if slot_has_suffixed_clitic then if do_joined then add_forms_with_clitic(base, slot, {clitic},							function(clitic, formobj, cliticized_form)								formobj.form = cliticized_form							end						) end elseif not do_joined then -- Add clitic as separate word before all other forms. for _, form in ipairs(base.forms[slot]) do						form.form = add_clitic_to_form(base, clitic, " ", form.form) end end end end end end

local function handle_infinitive_linked(base) -- 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({"infinitive"}) do insert_forms(base, slot .. "_linked", iut.map_forms(base.forms[slot], function(form) if form == base.lemma and rfind(base.linked_lemma, "%[%[") then return base.linked_lemma else return form end end)) end end

local function generate_negative_imperatives(base) -- Copy subjunctives to negative imperatives, preceded by "no". for _, persnum in ipairs(neg_imp_person_number_list) do local from = "pres_sub_" .. persnum local to = "neg_imp_" .. persnum insert_forms(base, to, iut.map_forms(base.forms[from], function(form) if base.alternant_multiword_spec.args.noautolinkverb then return "no " .. form elseif form:find("%[%[") then -- already linked, e.g. when reflexive return "no " .. form else return "no " .. form .. "" end end)) end end

local function add_missing_links_to_forms(base) -- Any forms without links should get them now. Redundant ones will be stripped later. for slot, forms in pairs(base.forms) do		for _, form in ipairs(forms) do			if not form.form:find("%[%[") then form.form = "" .. form.form .. "" end end end end

local function conjugate_verb(base) add_present_indic(base) add_present_subj(base) add_imper(base) add_finite_non_present(base) add_non_finite_forms(base) -- This should happen before add_combined_forms so overrides of basic forms end up part of the combined forms. process_slot_overrides(base, "do basic") -- do basic slot overrides -- This should happen after process_slot_overrides in case a derived slot is based on an override (as with the	-- imp_3s of dar, estar). copy_subjunctives_to_imperatives(base) -- This should happen after process_slot_overrides because overrides may have accents in them that need to be -- removed. (This happens e.g. for most present indicative forms of ver, which have accents in them for the	-- prefixed derived verbs, but the accents shouldn't be present in the base verb.) remove_monosyllabic_accents(base) if not base.nocomb then add_combined_forms(base) end -- We need to add joined reflexives, then joined and non-joined clitics, then non-joined reflexives, so we get -- házmelo but no me lo haga. if base.refl then -- This should happen after remove_monosyllabic_accents so the * marking the preservation of monosyllabic -- accents doesn't end up in the middle of a word. add_reflexive_or_fixed_clitic_to_forms(base, "do reflexive", "do joined") process_slot_overrides(base, "do basic", "do reflexive") -- do reflexive-only basic slot overrides end if base.clitic then -- This should happen after reflexives are added. add_reflexive_or_fixed_clitic_to_forms(base, false, "do joined") add_reflexive_or_fixed_clitic_to_forms(base, false, false) end if base.refl then add_reflexive_or_fixed_clitic_to_forms(base, "do reflexive", false) end -- This should happen after add_reflexive_or_fixed_clitic_to_forms so negative imperatives get the reflexive pronoun -- and clitic in them. generate_negative_imperatives(base) if not base.nocomb then process_slot_overrides(base, false) -- do combined slot overrides end -- This should happen before add_missing_links_to_forms so that the comparison `form == base.lemma` -- in handle_infinitive_linked works correctly and compares unlinked forms to unlinked forms. handle_infinitive_linked(base) if not base.alternant_multiword_spec.args.noautolinkverb then add_missing_links_to_forms(base) end end

local function parse_indicator_spec(angle_bracket_spec) -- Store the original angle bracket spec so we can reconstruct the overall conj spec with the lemma(s) in them. local base = {angle_bracket_spec = angle_bracket_spec} local function parse_err(msg) error(msg .. ": " .. angle_bracket_spec) end local function fetch_footnotes(separated_group) local footnotes for j = 2, #separated_group - 1, 2 do			if separated_group[j + 1] ~= "" then parse_err("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

local inside = angle_bracket_spec:match("^<(.*)>$") assert(inside) if inside == "" then return base end local segments = iut.parse_balanced_segment_run(inside, "[", "]") local dot_separated_groups = iut.split_alternating_runs(segments, "%.") for i, dot_separated_group in ipairs(dot_separated_groups) do		local comma_separated_groups = iut.split_alternating_runs(dot_separated_group, "%s*,%s*") local first_element = comma_separated_groups[1][1] if vowel_alternants[first_element] then for j = 1, #comma_separated_groups do				local alt = comma_separated_groups[j][1] if not vowel_alternants[alt] then parse_err("Unrecognized vowel alternant '" .. alt .. "'") end if base.vowel_alt then for _, existing_alt in ipairs(base.vowel_alt) do						if existing_alt.form == alt then parse_err("Vowel alternant '" .. alt .. "' specified twice") end end else base.vowel_alt = {} end table.insert(base.vowel_alt, {form = alt, footnotes = fetch_footnotes(comma_separated_groups[j])}) end elseif (first_element == "no_pres_stressed" or first_element == "no_pres1_and_sub" or				first_element == "only3s" or first_element == "only3sp" or first_element == "only3p") then if #comma_separated_groups[1] > 1 then parse_err("No footnotes allowed with '" .. first_element .. "' spec") end if base[first_element] then parse_err("Spec '" .. first_element .. "' specified twice") end base[first_element] = true else parse_err("Unrecognized spec '" .. comma_separated_groups[1][1] .. "'") end end

return base end

-- Reconstruct the overall verb spec from the output of iut.parse_inflected_text, so we can use it in -- Module:accel/es. function export.reconstruct_verb_spec(alternant_multiword_spec) local parts = {}

for _, alternant_or_word_spec in ipairs(alternant_multiword_spec.alternant_or_word_specs) do		table.insert(parts, alternant_or_word_spec.user_specified_before_text) if alternant_or_word_spec.alternants then table.insert(parts, "((") for i, multiword_spec in ipairs(alternant_or_word_spec.alternants) do				if i > 1 then table.insert(parts, ",") end for _, word_spec in ipairs(multiword_spec.word_specs) do					table.insert(parts, word_spec.user_specified_before_text) table.insert(parts, word_spec.user_specified_lemma) table.insert(parts, word_spec.angle_bracket_spec) end table.insert(parts, multiword_spec.user_specified_post_text) end table.insert(parts, "))") else table.insert(parts, alternant_or_word_spec.user_specified_lemma) table.insert(parts, alternant_or_word_spec.angle_bracket_spec) end end table.insert(parts, alternant_multiword_spec.user_specified_post_text)

-- As a special case, if we see e.g. "amar<>", remove the <>. Don't do this if there are spaces, hyphens or -- alternants. local retval = table.concat(parts) if not retval:find("[ %-]") and not retval:find("%(%(") then local retval_no_angle_brackets = retval:match("^(.*)<>$") if retval_no_angle_brackets then return retval_no_angle_brackets end end return retval end

-- Normalize all lemmas, substituting the pagename for blank lemmas and adding links to multiword lemmas. local function normalize_all_lemmas(alternant_multiword_spec, pagename)

-- (1) Add links to all before and after text. Remember the original text so we can reconstruct the verb spec later. if not alternant_multiword_spec.args.noautolinktext then for _, alternant_or_word_spec in ipairs(alternant_multiword_spec.alternant_or_word_specs) do			alternant_or_word_spec.user_specified_before_text = alternant_or_word_spec.before_text alternant_or_word_spec.before_text = com.add_links(alternant_or_word_spec.before_text) if alternant_or_word_spec.alternants then for _, multiword_spec in ipairs(alternant_or_word_spec.alternants) do					for _, word_spec in ipairs(multiword_spec.word_specs) do						word_spec.user_specified_before_text = word_spec.before_text word_spec.before_text = com.add_links(word_spec.before_text) end multiword_spec.user_specified_post_text = multiword_spec.post_text multiword_spec.post_text = com.add_links(multiword_spec.post_text) end end end alternant_multiword_spec.user_specified_post_text = alternant_multiword_spec.post_text alternant_multiword_spec.post_text = com.add_links(alternant_multiword_spec.post_text) end

-- (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.user_specified_lemma = base.lemma

base.lemma = m_links.remove_links(base.lemma) local refl_verb, clitic = rmatch(base.lemma, "^(.-)(l[aeo]s?)$") if not refl_verb then refl_verb, clitic = base.lemma, nil end local verb, refl = rmatch(refl_verb, "^(.-)(se)$") if not verb then verb, refl = refl_verb, nil end base.user_specified_verb = verb base.refl = refl base.clitic = clitic

if base.refl and base.clitic then -- We have to parse the verb suffix to see how to construct the base verb; e.g. -- abrírsela -> abrir but oírsela -> oír. We parse the verb suffix again in all cases -- in detect_indicator_spec, after splitting off the prefix of irrregular verbs. local actual_verb local inf_stem, suffix = rmatch(base.user_specified_verb, "^(.*)([aáeéií]r)$") if not inf_stem then error("Unrecognized infinitive: " .. base.user_specified_verb) end if suffix == "ír" and inf_stem:find("[aeo]$") then -- accent on suffix should remain base.verb = base.user_specified_verb else base.verb = inf_stem .. com.remove_accent_from_syllable(suffix) end else base.verb = base.user_specified_verb end

local linked_lemma if alternant_multiword_spec.args.noautolinkverb or base.user_specified_lemma:find("%[%[") then linked_lemma = base.user_specified_lemma elseif base.refl or base.clitic then -- Reconstruct the linked lemma with separate links around base verb, reflexive pronoun and clitic. linked_lemma = base.user_specified_verb == base.verb and "" .. base.user_specified_verb .. "" or "" .. base.user_specified_verb .. "" linked_lemma = linked_lemma .. (refl and "" .. refl .. "" or "") .. (clitic and "" .. clitic .. "" or "") else -- Add links to the lemma so the user doesn't specifically need to, since we preserve -- links in multiword lemmas and include links in non-lemma forms rather than allowing -- the entire form to be a link. linked_lemma = com.add_links(base.user_specified_lemma) end base.linked_lemma = linked_lemma end) end

local function construct_stems(base) local stems = base.stems stems.pres_unstressed = stems.pres_unstressed or base.inf_stem stems.pres_stressed = stems.pres_stressed or -- If no_pres_stressed given, pres_stressed stem should be empty so no forms are generated. base.no_pres_stressed and {} or		base.vowel_alt or		base.inf_stem stems.pres1_and_sub = stems.pres1_and_sub or -- If no_pres_stressed given, the entire subjunctive is missing. base.no_pres_stressed and {} or -- If no_pres1_and_sub given, pres1 and entire subjunctive are missing. base.no_pres1_and_sub and {} or		nil stems.pres1 = stems.pres1 or stems.pres1_and_sub or stems.pres_stressed stems.impf = stems.impf or base.inf_stem stems.pret = stems.pret or base.inf_stem stems.pret_conj = stems.pret_conj or base.conj stems.fut = stems.fut or base.inf_stem .. base.conj stems.cond = stems.cond or stems.fut stems.pres_sub_stressed = stems.pres_sub_stressed or stems.pres1 stems.pres_sub_unstressed = stems.pres_sub_unstressed or stems.pres1_and_sub or stems.pres_unstressed stems.impf_sub_ra = stems.impf_sub_ra or stems.pret stems.impf_sub_se = stems.impf_sub_se or stems.pret stems.fut_sub = stems.fut_sub or stems.pret stems.pp = stems.pp or base.conj == "ar" and combine_stem_ending(base, "pp_ms", base.inf_stem, "ad", "is combining ending") or -- use combine_stem_ending esp. so we get reído, caído, etc.		combine_stem_ending(base, "pp_ms", base.inf_stem, "id", "is combining ending") end

local function detect_indicator_spec(base) base.forms = {} base.non_reflexive_forms = {} base.stems = {}

if (base.only3s and 1 or 0) + (base.only3sp and 1 or 0) + (base.only3p and 1 or 0) > 1 then error("Only one of 'only3s', 'only3sp' and 'only3p' can be specified") end

base.basic_overrides = {} base.basic_reflexive_only_overrides = {} base.combined_overrides = {} for _, irreg_conj in ipairs(irreg_conjugations) do		if type(irreg_conj.match) == "function" then base.prefix, base.non_prefixed_verb = irreg_conj.match(base.verb) elseif irreg_conj.match:find("^%^") and rsub(irreg_conj.match, "^%^", "") == base.verb then -- begins with ^, for exact match, and matches base.prefix, base.non_prefixed_verb = "", base.verb else base.prefix, base.non_prefixed_verb = rmatch(base.verb, "^(.*)(" .. irreg_conj.match .. ")$")		end if base.prefix then -- we found an irregular verb base.irreg_verb = true for stem, forms in pairs(irreg_conj.forms) do				if stem:find("^refl_") then stem = stem:gsub("^refl_", "") if not base.alternant_multiword_spec.verb_slots_basic_map[stem] then error("Internal error: setting for 'refl_" .. stem .. "' does not refer to a basic verb slot") end base.basic_reflexive_only_overrides[stem] = forms elseif base.alternant_multiword_spec.verb_slots_basic_map[stem] then -- an individual form override of a basic form base.basic_overrides[stem] = forms elseif base.alternant_multiword_spec.verb_slots_combined_map[stem] then -- an individual form override of a combined form base.combined_overrides[stem] = forms else base.stems[stem] = forms end end break end end base.prefix = base.prefix or "" base.non_prefixed_verb = base.non_prefixed_verb or base.verb local inf_stem, suffix = rmatch(base.non_prefixed_verb, "^(.*)([aeií]r)$") if not inf_stem then error("Unrecognized infinitive: " .. base.verb) end base.inf_stem = inf_stem suffix = suffix == "ír" and "ir" or suffix base.conj = suffix base.frontback = suffix == "ar" and "back" or "front"

if base.stems.vowel_alt then -- irregular verb with specified vowel alternation if base.vowel_alt then error(base.verb .. " is a recognized irregular verb, and should not have vowel alternations specified with it") end base.vowel_alt = iut.convert_to_general_list_form(base.stems.vowel_alt) end

-- Convert vowel alternation indicators into stems. if base.vowel_alt then for _, altform in ipairs(base.vowel_alt) do			altform.alt = altform.form -- save original indicator local alt = altform.alt if base.conj == "ir" then local raising = (					alt == "ie-i" or alt == "ye-i" or alt == "ue-u" or alt == "i" or alt == "í" or alt == "ú"				) if base.stems.raising_conj == nil then base.stems.raising_conj = raising elseif base.stems.raising_conj ~= raising then error("Can't currently support a mixture of raising (e.g. 'ie-i') and non-raising (e.g. 'ie') vowel alternations in -ir verbs") end end if alt == "+" then altform.form = base.inf_stem else local normalized_alt = alt if alt == "ie-i" or alt == "ye-i" or alt == "ue-u" then if base.conj ~= "ir" then error("Vowel alternation '" .. alt .. "' only supported with -ir verbs") end -- ie-i is like i except for the vowel raising before i+V, similarly for ye-i, ue-u, -- so convert appropriately. normalized_alt = alt == "ie-i" and "ie" or alt == "ye-i" and "ye" or "ue" end local ret = com.apply_vowel_alternation(base.inf_stem, normalized_alt) if ret.err then error("To use '" .. alt .. "', present stem '" .. base.inf_stem .. "' " .. ret.err) end altform.form = ret.ret end end end end

local function detect_all_indicator_specs(alternant_multiword_spec) -- Propagate some settings up or down. iut.map_word_specs(alternant_multiword_spec, function(base)		for _, prop in ipairs { "refl", "clitic", "only3s", "only3sp", "only3p" } do			if base[prop] then				alternant_multiword_spec[prop] = true			end		end		base.alternant_multiword_spec = alternant_multiword_spec		-- If reflexive or fixed clitic, don't include combined forms.		alternant_multiword_spec.nocomb = alternant_multiword_spec.nocomb or base.clitic or base.refl	end)

add_slots(alternant_multiword_spec)

iut.map_word_specs(alternant_multiword_spec, function(base)		base.nocomb = alternant_multiword_spec.args.nocomb		detect_indicator_spec(base)		construct_stems(base)	end) end

local function add_categories_and_annotation(alternant_multiword_spec, base, multiword_lemma) local function insert_ann(anntype, value) m_table.insertIfNot(alternant_multiword_spec.annotation[anntype], value) end

local function insert_cat(cat, also_when_multiword) -- Don't place multiword terms in categories like 'Spanish verbs ending in -ar' to avoid spamming the -- categories with such terms. if also_when_multiword or not multiword_lemma then m_table.insertIfNot(alternant_multiword_spec.categories, "Spanish " .. cat) end end

if check_for_red_links and not alternant_multiword_spec.from_headword and not alternant_multiword_spec.from_verb_form_of and multiword_lemma then for _, slot_and_accel in ipairs(alternant_multiword_spec.all_verb_slots) do			local slot = slot_and_accel[1] local forms = base.forms[slot] local must_break = false if forms then for _, form in ipairs(forms) do					if not form.form:find("%[%[") then local title = mw.title.new(form.form) if title and not title.exists then insert_cat("verbs with red links in their inflection tables") must_break = true break end end end end if must_break then break end end end

insert_cat("verbs ending in -" .. base.conj)

if base.irreg_verb then insert_ann("irreg", "irregular") insert_cat("irregular verbs") else insert_ann("irreg", "regular") end

if base.only3s then insert_ann("defective", "impersonal") insert_cat("impersonal verbs") elseif base.only3sp then insert_ann("defective", "third-person only") insert_cat("third-person-only verbs") elseif base.only3p then insert_ann("defective", "third-person plural only") insert_cat("third-person-plural-only verbs") elseif base.no_pres_stressed or base.no_pres1_and_sub then insert_ann("defective", "defective") insert_cat("defective verbs") else insert_ann("defective", "regular") end

if base.clitic then insert_cat("verbs with lexical clitics") end

if base.refl then insert_cat("reflexive verbs") end

if not base.vowel_alt then insert_ann("vowel_alt", "non-alternating") else local inf_stem = base.inf_stem:gsub("[gq]u$", "x") for _, alt in ipairs(base.vowel_alt) do			if alt.alt == "+" then insert_ann("vowel_alt", "non-alternating") else local desc if alt.alt == "ue" and rfind(inf_stem, "u" .. C .. "*$") then desc = "u-ue alternation" -- jugar elseif alt.alt == "ie" and rfind(inf_stem, "i" .. C .. "*$") then desc = "i-ie alternation" -- adquirir elseif alt.alt == "í" and rfind(inf_stem, "e" .. C .. "*$") then desc = "e-í alternation" -- reír, freír, etc.				else desc = vowel_alternant_to_desc[alt.alt] .. " alternation" end insert_ann("vowel_alt", desc) insert_cat("verbs with " .. desc) end end end

local cons_alt = base.stems.cons_alt if cons_alt == nil then if base.conj == "ar" then if base.inf_stem:find("z$") then cons_alt = "c-z" elseif base.inf_stem:find("ç$") then cons_alt = "c-ç" elseif base.inf_stem:find("c$") then cons_alt = "c-qu" elseif base.inf_stem:find("g$") then cons_alt = "g-gu" elseif base.inf_stem:find("gu$") then cons_alt = "gu-gü" end else if base.no_pres_stressed or base.no_pres1_and_sub then cons_alt = nil -- no c-zc alternation in balbucir or arrecir elseif rfind(base.inf_stem, V .. "c$") then cons_alt = "c-zc" elseif base.inf_stem:find("sc$") then cons_alt = "hard-soft" elseif base.inf_stem:find("c$") then cons_alt = "c-z" elseif base.inf_stem:find("qu$") then cons_alt = "c-qu" elseif base.inf_stem:find("g$") then cons_alt = "g-j" elseif base.inf_stem:find("gu$") then cons_alt = "g-gu" elseif base.inf_stem:find("gü$") then cons_alt = "gu-gü" end end end

if cons_alt then local desc = cons_alt .. " alternation" insert_ann("cons_alt", desc) insert_cat("verbs with " .. desc) else insert_ann("cons_alt", "non-alternating") end end

-- Compute the categories to add the verb to, as well as the annotation to display in the -- conjugation 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) alternant_multiword_spec.categories = {} local ann = {} alternant_multiword_spec.annotation = ann ann.irreg = {} ann.defective = {} ann.vowel_alt = {} ann.cons_alt = {}

local multiword_lemma = false for _, form in ipairs(alternant_multiword_spec.forms.infinitive) do		if form.form:find(" ") then multiword_lemma = true break end end

iut.map_word_specs(alternant_multiword_spec, function(base)		add_categories_and_annotation(alternant_multiword_spec, base, multiword_lemma)	end) local ann_parts = {} local irreg = table.concat(ann.irreg, " or ") if irreg ~= "" and irreg ~= "regular" then table.insert(ann_parts, irreg) end local defective = table.concat(ann.defective, " or ") if defective ~= "" and defective ~= "regular" then table.insert(ann_parts, defective) end local vowel_alt = table.concat(ann.vowel_alt, " or ") if vowel_alt ~= "" and vowel_alt ~= "non-alternating" then table.insert(ann_parts, vowel_alt) end local cons_alt = table.concat(ann.cons_alt, " or ") if cons_alt ~= "" and cons_alt ~= "non-alternating" then table.insert(ann_parts, cons_alt) end alternant_multiword_spec.annotation = table.concat(ann_parts, "; ") end

local function show_forms(alternant_multiword_spec) local lemmas = iut.map_forms(alternant_multiword_spec.forms.infinitive, remove_reflexive_indicators) alternant_multiword_spec.lemmas = lemmas -- save for later use in make_table

local reconstructed_verb_spec = export.reconstruct_verb_spec(alternant_multiword_spec)

-- Initialize the footnotes with those for the future subjunctive and maybe the pres subjunctive -- voseo usage. In the latter case, we only do it if there is a distinct pres subjunctive voseo form. local function create_footnote_obj local obj = iut.create_footnote_obj iut.get_footnote_text({footnotes = {fut_sub_note}}, obj) -- Compute whether the tú and voseo variants are different, for each voseo variant. -- We use this later in make_table. for _, slot in ipairs({"pres_2s", "pres_sub_2s", "imp_2s"}) do alternant_multiword_spec["separate_" .. slot .. "v"] = false iut.map_word_specs(alternant_multiword_spec, function(base)				if not m_table.deepEquals(base.forms[slot], base.forms[slot .. "v"]) then					alternant_multiword_spec["separate_" .. slot .. "v"] = true				end			end) end if alternant_multiword_spec.separate_pres_sub_2sv then iut.get_footnote_text({footnotes = {pres_sub_voseo_note}}, obj) end return obj end

local function transform_accel_obj(slot, formobj, accel_obj) -- No accelerators for negative imperatives, which are always multiword and derived directly from the -- present subjunctive. if slot:find("^neg_imp") then return nil end if accel_obj then accel_obj.form = "verb-form-" .. reconstructed_verb_spec end return accel_obj end

local props = { lang = lang, lemmas = lemmas, create_footnote_obj = create_footnote_obj, transform_accel_obj = transform_accel_obj, }	props.slot_list = alternant_multiword_spec.verb_slots_basic iut.show_forms(alternant_multiword_spec.forms, props) alternant_multiword_spec.footnote_basic = alternant_multiword_spec.forms.footnote props.create_footnote_obj = nil props.slot_list = alternant_multiword_spec.verb_slots_combined iut.show_forms(alternant_multiword_spec.forms, props) alternant_multiword_spec.footnote_combined = alternant_multiword_spec.forms.footnote end

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

local pres_2sv_template = 'tú {pres_2sv}vos ' local pres_sub_2sv_template = 'tú {pres_sub_2sv}vos<sup style="color:red">2 ' local imp_2sv_template = 'tú {imp_2sv}vos '

local basic_table = [=[ {description} Conjugation of {title} (See Appendix:Spanish verbs) {\op}| style="background:#F9F9F9;text-align:center;width:100%" ! colspan="3" style="background:#e2e4c0" | infinitive ! colspan="3" style="background:#e2e4c0" | gerund ! rowspan="3" colspan="2" style="background:#e2e4c0" | <span title="participio (pasado)">past participle ! colspan="2" style="background:#e2e4c0" | masculine ! colspan="2" style="background:#e2e4c0" | feminine ! colspan="2" style="background:#e2e4c0" | singular ! colspan="2" style="background:#e2e4c0" | plural ! colspan="2" rowspan="2" style="background:#DEDEDE" | ! colspan="3" style="background:#DEDEDE" | singular ! colspan="3" style="background:#DEDEDE" | plural ! style="background:#DEDEDE" | 1st person ! style="background:#DEDEDE" | 2nd person ! style="background:#DEDEDE" | 3rd person ! style="background:#DEDEDE" | 1st person ! style="background:#DEDEDE" | 2nd person ! style="background:#DEDEDE" | 3rd person ! rowspan="6" style="background:#c0cfe4" | indicative ! style="background:#ECECEC;width:12.5%" | ! style="background:#ECECEC;width:12.5%" | yo ! style="background:#ECECEC;width:12.5%" | tú vos ! style="background:#ECECEC;width:12.5%" | él/ella/ello usted ! style="background:#ECECEC;width:12.5%" | nosotros nosotras ! style="background:#ECECEC;width:12.5%" | vosotros vosotras ! style="background:#ECECEC;width:12.5%" | ellos/ellas ustedes ! style="height:3em;background:#ECECEC" | present ! style="height:3em;background:#ECECEC" | <span title="pretérito imperfecto (copréterito)">imperfect ! style="height:3em;background:#ECECEC" | <span title="pretérito perfecto simple (pretérito indefinido)">preterite ! style="height:3em;background:#ECECEC" | <span title="futuro simple (futuro imperfecto)">future ! style="height:3em;background:#ECECEC" | <span title="condicional simple (pospretérito de modo indicativo)">conditional ! style="background:#DEDEDE;height:.75em" colspan="8" | ! rowspan="5" style="background:#c0e4c0" | subjunctive ! style="background:#ECECEC" | ! style="background:#ECECEC" | yo ! style="background:#ECECEC" | tú vos ! style="background:#ECECEC" | él/ella/ello usted ! style="background:#ECECEC" | nosotros nosotras ! style="background:#ECECEC" | vosotros vosotras ! style="background:#ECECEC" | ellos/ellas ustedes ! style="height:3em;background:#ECECEC" | present ! style="height:3em;background:#ECECEC" | <span title="pretérito imperfecto de subjuntivo">imperfect (ra) ! style="height:3em;background:#ECECEC" | <span title="pretérito imperfecto de subjuntivo">imperfect (se) ! style="height:3em;background:#ECECEC" | <span title="futuro simple de subjuntivo (futuro de subjuntivo)">future <sup style="color:red">1 ! style="background:#DEDEDE;height:.75em" colspan="8" | ! rowspan="6" style="background:#e4d4c0" | imperative ! style="background:#ECECEC" | ! style="background:#ECECEC" | — ! style="background:#ECECEC" | tú vos ! style="background:#ECECEC" | usted ! style="background:#ECECEC" | nosotros nosotras ! style="background:#ECECEC" | vosotros vosotras ! style="background:#ECECEC" | ustedes ! style="height:3em;background:#ECECEC" | affirmative ! style="height:3em;background:#ECECEC" | negative ]=]
 * colspan="5" | {infinitive}
 * colspan="5" | {gerund}
 * colspan="2" style="background:#e2e4c0" |
 * colspan="2" | {pp_ms}
 * colspan="2" | {pp_fs}
 * colspan="2" | {pp_mp}
 * colspan="2" | {pp_fp}
 * -{reflexive_non_finite_clause}
 * {pres_1s}
 * {pres_2s}{pres_2sv_text}
 * {pres_3s}
 * {pres_1p}
 * {pres_2p}
 * {pres_3p}
 * {impf_1s}
 * {impf_2s}
 * {impf_3s}
 * {impf_1p}
 * {impf_2p}
 * {impf_3p}
 * {pret_1s}
 * {pret_2s}
 * {pret_3s}
 * {pret_1p}
 * {pret_2p}
 * {pret_3p}
 * {fut_1s}
 * {fut_2s}
 * {fut_3s}
 * {fut_1p}
 * {fut_2p}
 * {fut_3p}
 * {cond_1s}
 * {cond_2s}
 * {cond_3s}
 * {cond_1p}
 * {cond_2p}
 * {cond_3p}
 * {pres_sub_1s}
 * {pres_sub_2s}{pres_sub_2sv_text}
 * {pres_sub_3s}
 * {pres_sub_1p}
 * {pres_sub_2p}
 * {pres_sub_3p}
 * {impf_sub_ra_1s}
 * {impf_sub_ra_2s}
 * {impf_sub_ra_3s}
 * {impf_sub_ra_1p}
 * {impf_sub_ra_2p}
 * {impf_sub_ra_3p}
 * {impf_sub_se_1s}
 * {impf_sub_se_2s}
 * {impf_sub_se_3s}
 * {impf_sub_se_1p}
 * {impf_sub_se_2p}
 * {impf_sub_se_3p}
 * {fut_sub_1s}
 * {fut_sub_2s}
 * {fut_sub_3s}
 * {fut_sub_1p}
 * {fut_sub_2p}
 * {fut_sub_3p}
 * {imp_2s}{imp_2sv_text}
 * {imp_3s}
 * {imp_1p}
 * {imp_2p}
 * {imp_3p}
 * {imp_3p}
 * {neg_imp_2s}
 * {neg_imp_3s}
 * {neg_imp_1p}
 * {neg_imp_2p}
 * {neg_imp_3p}
 * {\cl}{notes_clause}
 * {\cl}{notes_clause}

local reflexive_non_finite_template = [=[

! rowspan="3" style="background:#e2e4c0" | personal non-finite ! style="background:#ECECEC;width:12.5%" | ! style="background:#ECECEC;width:12.5%" | yo ! style="background:#ECECEC;width:12.5%" | tú vos ! style="background:#ECECEC;width:12.5%" | él/ella/ello usted ! style="background:#ECECEC;width:12.5%" | nosotros nosotras ! style="background:#ECECEC;width:12.5%" | vosotros vosotras ! style="background:#ECECEC;width:12.5%" | ellos/ellas ustedes ! style="height:3em;background:#ECECEC" | infinitive ! style="height:3em;background:#ECECEC" | gerund ! style="background:#DEDEDE;height:.75em" colspan="8" |
 * {infinitive_1s}
 * {infinitive_2s}
 * {infinitive_3s}
 * {infinitive_1p}
 * {infinitive_2p}
 * {infinitive_3p}
 * {gerund_1s}
 * {gerund_2s}
 * {gerund_3s}
 * {gerund_1p}
 * {gerund_2p}
 * {gerund_3p}
 * -]=]

local combined_form_combined_tu_vos_template = [=[

! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#f2caa4" | with informal second-person singular tú/vos imperative {imp_2s} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative
 * {imp_2s_comb_me}
 * {imp_2s_comb_te}
 * {imp_2s_comb_le}
 * {imp_2s_comb_nos}
 * not used
 * {imp_2s_comb_les}
 * {imp_2s_comb_me}
 * {imp_2s_comb_te}
 * {imp_2s_comb_lo}, {imp_2s_comb_la}
 * {imp_2s_comb_nos}
 * not used
 * {imp_2s_comb_los}, {imp_2s_comb_las}
 * -]=]

local combined_form_separate_tu_vos_template = [=[

! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#f2caa4" | with informal second-person singular tú imperative {imp_2s} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative ! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#f2caa4" | with informal second-person singular vos imperative {imp_2sv} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative
 * {imp_2s_comb_me}
 * {imp_2s_comb_te}
 * {imp_2s_comb_le}
 * {imp_2s_comb_nos}
 * not used
 * {imp_2s_comb_les}
 * {imp_2s_comb_me}
 * {imp_2s_comb_te}
 * {imp_2s_comb_lo}, {imp_2s_comb_la}
 * {imp_2s_comb_nos}
 * not used
 * {imp_2s_comb_los}, {imp_2s_comb_las}
 * {imp_2sv_comb_me}
 * {imp_2sv_comb_te}
 * {imp_2sv_comb_le}
 * {imp_2sv_comb_nos}
 * not used
 * {imp_2sv_comb_les}
 * {imp_2sv_comb_me}
 * {imp_2sv_comb_te}
 * {imp_2sv_comb_lo}, {imp_2sv_comb_la}
 * {imp_2sv_comb_nos}
 * not used
 * {imp_2sv_comb_los}, {imp_2sv_comb_las}
 * -]=]

local combined_form_table = [=[ {description} Selected combined forms of {title} These forms are generated automatically and may not actually be used. Pronoun usage varies by region. {\op}| class="inflection-table" style="background:#F9F9F9;text-align:center;width:100%" ! colspan="2" rowspan="2" style="background:#DEDEDE" | ! colspan="3" style="background:#DEDEDE" | singular ! colspan="3" style="background:#DEDEDE" | plural ! style="background:#DEDEDE" | 1st person ! style="background:#DEDEDE" | 2nd person ! style="background:#DEDEDE" | 3rd person ! style="background:#DEDEDE" | 1st person ! style="background:#DEDEDE" | 2nd person ! style="background:#DEDEDE" | 3rd person ! rowspan="3" style="background:#c0cfe4" | with infinitive {infinitive} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative ! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#d0cfa4" | with gerund {gerund} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative ! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#f2caa4" | with formal second-person singular imperative {imp_3s} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative ! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#f2caa4" | with first-person plural imperative {imp_1p} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative ! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#f2caa4" | with informal second-person plural imperative {imp_2p} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative ! style="background:#DEDEDE;height:.35em" colspan="8" | ! rowspan="3" style="background:#f2caa4" | with formal second-person plural imperative {imp_3p} ! style="height:3em;background:#ECECEC" | dative ! style="height:3em;background:#ECECEC" | accusative ]=]
 * {infinitive_comb_me}
 * {infinitive_comb_te}
 * {infinitive_comb_le}, {infinitive_comb_se}
 * {infinitive_comb_nos}
 * {infinitive_comb_os}
 * {infinitive_comb_les}, {infinitive_comb_se}
 * {infinitive_comb_me}
 * {infinitive_comb_te}
 * {infinitive_comb_lo}, {infinitive_comb_la}, {infinitive_comb_se}
 * {infinitive_comb_nos}
 * {infinitive_comb_os}
 * {infinitive_comb_los}, {infinitive_comb_las}, {infinitive_comb_se}
 * {gerund_comb_me}
 * {gerund_comb_te}
 * {gerund_comb_le}, {gerund_comb_se}
 * {gerund_comb_nos}
 * {gerund_comb_os}
 * {gerund_comb_les}, {gerund_comb_se}
 * {gerund_comb_me}
 * {gerund_comb_te}
 * {gerund_comb_lo}, {gerund_comb_la}, {gerund_comb_se}
 * {gerund_comb_nos}
 * {gerund_comb_os}
 * {gerund_comb_los}, {gerund_comb_las}, {gerund_comb_se}
 * -{tu_vos_clause}
 * {imp_3s_comb_me}
 * not used
 * {imp_3s_comb_le}, {imp_3s_comb_se}
 * {imp_3s_comb_nos}
 * not used
 * {imp_3s_comb_les}
 * {imp_3s_comb_me}
 * not used
 * {imp_3s_comb_lo}, {imp_3s_comb_la}, {imp_3s_comb_se}
 * {imp_3s_comb_nos}
 * not used
 * {imp_3s_comb_los}, {imp_3s_comb_las}
 * not used
 * {imp_1p_comb_te}
 * {imp_1p_comb_le}
 * {imp_1p_comb_nos}
 * {imp_1p_comb_os}
 * {imp_1p_comb_les}
 * not used
 * {imp_1p_comb_te}
 * {imp_1p_comb_lo}, {imp_1p_comb_la}
 * {imp_1p_comb_nos}
 * {imp_1p_comb_os}
 * {imp_1p_comb_los}, {imp_1p_comb_las}
 * {imp_2p_comb_me}
 * not used
 * {imp_2p_comb_le}
 * {imp_2p_comb_nos}
 * {imp_2p_comb_os}
 * {imp_2p_comb_les}
 * {imp_2p_comb_me}
 * not used
 * {imp_2p_comb_lo}, {imp_2p_comb_la}
 * {imp_2p_comb_nos}
 * {imp_2p_comb_os}
 * {imp_2p_comb_los}, {imp_2p_comb_las}
 * {imp_3p_comb_me}
 * not used
 * {imp_3p_comb_le}
 * {imp_3p_comb_nos}
 * not used
 * {imp_3p_comb_les}, {imp_3p_comb_se}
 * {imp_3p_comb_me}
 * not used
 * {imp_3p_comb_lo}, {imp_3p_comb_la}
 * {imp_3p_comb_nos}
 * not used
 * {imp_3p_comb_los}, {imp_3p_comb_las}, {imp_3p_comb_se}
 * {\cl}{notes_clause}

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

forms.title = link_term(alternant_multiword_spec.lemmas[1].form) if alternant_multiword_spec.annotation ~= "" then forms.title = forms.title .. " (" .. alternant_multiword_spec.annotation .. ")" end forms.description = ""

-- Format the basic table. forms.footnote = alternant_multiword_spec.footnote_basic forms.reflexive_non_finite_clause = alternant_multiword_spec.refl and m_string_utilities.format(reflexive_non_finite_template, forms) or "" forms.notes_clause = forms.footnote ~= "" and m_string_utilities.format(notes_template, forms) or "" -- The separate_* values are computed in show_forms. forms.pres_2sv_text = alternant_multiword_spec.separate_pres_2sv and m_string_utilities.format(pres_2sv_template, forms) or "" forms.pres_sub_2sv_text = alternant_multiword_spec.separate_pres_sub_2sv and m_string_utilities.format(pres_sub_2sv_template, forms) or "" forms.imp_2sv_text = alternant_multiword_spec.separate_imp_2sv and m_string_utilities.format(imp_2sv_template, forms) or "" local formatted_basic_table = m_string_utilities.format(basic_table, forms)

-- Format the combined table. local formatted_combined_table if alternant_multiword_spec.refl or alternant_multiword_spec.args.nocomb or alternant_multiword_spec.clitic then formatted_combined_table = "" else forms.footnote = alternant_multiword_spec.footnote_combined forms.notes_clause = forms.footnote ~= "" and m_string_utilities.format(notes_template, forms) or "" -- separate_imp_2sv is computed in show_forms. local tu_vos_template = alternant_multiword_spec.separate_imp_2sv and combined_form_separate_tu_vos_template or			combined_form_combined_tu_vos_template forms.tu_vos_clause = m_string_utilities.format(tu_vos_template, forms) formatted_combined_table = m_string_utilities.format(combined_form_table, forms) end

-- Paste them together. return formatted_basic_table .. formatted_combined_table end

-- Externally callable function to parse and conjugate a verb given user-specified arguments. -- Return value is WORD_SPEC, an object where the conjugated forms are in `WORD_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, from_verb_form_of, double_combined_forms_to_include) local params = { [1] = {required = from_verb_form_of}, ["nocomb"] = {type = "boolean"}, ["noautolinktext"] = {type = "boolean"}, ["noautolinkverb"] = {type = "boolean"}, ["pagename"] = {}, -- for testing/documentation pages ["json"] = {type = "boolean"}, -- for bot use }

if from_headword then params["head"] = {list = true} params["pres"] = {list = true} --present params["pres_qual"] = {list = "pres=_qual", allow_holes = true} params["pret"] = {list = true} --preterite params["pret_qual"] = {list = "pret=_qual", allow_holes = true} params["part"] = {list = true} --participle params["part_qual"] = {list = "part=_qual", allow_holes = true} params["attn"] = {type = "boolean"} params["id"] = {} end

local args = require("Module:parameters").process(parent_args, params) local PAGENAME = mw.title.getCurrentTitle.text local function in_template_space return mw.title.getCurrentTitle.nsText == "Template" end

-- Determine the verb spec we're being asked to generate the conjugation of. This may be taken from the -- current page title or the value of |pagename=; but not when called from, where the -- page title is a non-lemma form. Note that the verb spec may omit the infinitive; e.g. it may be "<ue>". -- For this reason, we use the value of `pagename` computed here down below, when calling normalize_all_lemmas. local pagename = not from_verb_form_of and args.pagename or from_headword and args.head[1] or PAGENAME local arg1 = args[1] if not arg1 then if (pagename == "es-conj" or pagename == "es-verb") and in_template_space then arg1 = "licuar<+,ú>" elseif pagename == "es-verb form of" and in_template_space then arg1 = "amar" else arg1 = pagename end end

-- When called from, determine the non-lemma form whose inflections we're being asked to -- determine. This normally comes from the page title or the value of |pagename=. local verb_form_of_form if from_verb_form_of then verb_form_of_form = args.pagename if not verb_form_of_form then if PAGENAME == "es-verb form of" and in_template_space then verb_form_of_form = "ame" else verb_form_of_form = PAGENAME end end end

if arg1:find(" ") and not arg1:find("<") then -- If multiword lemma without <> already, try to add it after the first word.

local need_explicit_angle_brackets = false if arg1:find("%(%(") then need_explicit_angle_brackets = true else local refl_clitic_verb, orig_refl_clitic_verb, post

-- Try to preserve the brackets in the part after the verb, but don't do it			-- if there aren't the same number of left and right brackets in the verb -- (which means the verb was linked as part of a larger expression). refl_clitic_verb, post = rmatch(arg1, "^(.-)( .*)$") local left_brackets = rsub(refl_clitic_verb, "[^%[]", "") local right_brackets = rsub(refl_clitic_verb, "[^%]]", "") if #left_brackets == #right_brackets then arg1 = refl_clitic_verb .. "<>" .. post else need_explicit_angle_brackets = true end end

if need_explicit_angle_brackets then error("Multiword argument without <> and with alternants, a multiword linked verb or unbalanced brackets; please include <> explicitly: " .. arg1) end end

local parse_props = { parse_indicator_spec = parse_indicator_spec, allow_default_indicator = true, allow_blank_lemma = true, }	local escaped_arg1 = escape_reflexive_indicators(arg1) local alternant_multiword_spec = iut.parse_inflected_text(escaped_arg1, parse_props) alternant_multiword_spec.pos = pos or "verbs" alternant_multiword_spec.args = args alternant_multiword_spec.from_headword = from_headword alternant_multiword_spec.from_verb_form_of = from_verb_form_of alternant_multiword_spec.verb_form_of_form = verb_form_of_form

-- Now determine if we need to generate any double-combined forms, and if so, which clitics are involved. -- See the comment above the initialization of `verb_slot_double_combined_rows` above in add_slots. if verb_form_of_form and rfind(verb_form_of_form, AV) then -- All double-clitic forms have an explicit accent, so we check for this. In addition, all double-clitic forms -- are of the form "(me|te|se|nos|os)(lo|la|le)s$". We have no alternations in Lua patterns, but we can exploit -- the similarity of the clitics in question. local single_comb_form, object_clitic = rmatch(verb_form_of_form, "^(.*)(l[aeo]s?)$") if single_comb_form then local personal_clitic = rmatch(single_comb_form, "^.*([mts]e)$") if not personal_clitic then personal_clitic = rmatch(single_comb_form, "^.-(n?os)$") end if personal_clitic then if personal_clitic == "nos" then -- "os" is a substring of "nos"; conceivably we could have a form ending in -n + os, and we don't -- know whether to interpret as -n + os or - + nos. alternant_multiword_spec.double_combined_forms_to_include = {{"nos", object_clitic}, {"os", object_clitic}} else alternant_multiword_spec.double_combined_forms_to_include = end end end end

normalize_all_lemmas(alternant_multiword_spec, pagename) detect_all_indicator_specs(alternant_multiword_spec) local inflect_props = { slot_list = alternant_multiword_spec.all_verb_slots, inflect_word_spec = conjugate_verb, -- We add links around the generated verbal forms rather than allow the entire multiword -- expression to be a link, so ensure that user-specified links get included as well. include_user_specified_links = true, }	iut.inflect_multiword_or_alternant_multiword_spec(alternant_multiword_spec, inflect_props)

-- Remove redundant brackets around entire forms. for slot, forms in pairs(alternant_multiword_spec.forms) do		for _, form in ipairs(forms) do			form.form = com.strip_redundant_links(form.form) end end

compute_categories_and_annotation(alternant_multiword_spec) if args.json and not from_headword and not from_verb_form_of then return require("Module:JSON").toJSON({forms = alternant_multiword_spec.forms}) end return alternant_multiword_spec end

-- Entry point for. Template-callable function to parse and conjugate a verb given -- user-specified arguments and generate a displayable table of the conjugated forms. function export.show(frame) local parent_args = frame:getParent.args local alternant_multiword_spec = export.do_generate_forms(parent_args) 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