Changelog¶

[Unreleased]¶

[2.2.0] - 2025-07-13¶

Added¶

join() — join words into a natural-language list with configurable conjunction and separators
ordinal() — convert a number to its ordinal string representation (1 → “1st”, 2 → “2nd”, 11 → “11th”)
template() — interpolate pluralize/singularize into string templates with count-aware support
pluralio.utils module with full docstrings and type hints
90 tests covering all utility functions, including property-based tests with hypothesis
Sphinx documentation page for utility functions
Cookbook recipes for join, ordinal, and template

[2.1.2] - 2025-07-13¶

Fixed¶

restore() now clears the lru_cache via lazy import of _clear_regex_cache, preventing stale results after registry restoration (critical for test isolation in conftest.py)
_match_case() no longer lowercases target when source contains no uppercase letters (e.g. digit-only sources like "123")
register_language() removed redundant _clear_regex_cache() call — register() already handles cache clearing

Added¶

Regression tests for all three bug fixes

[2.1.1] - 2025-07-13¶

Fixed¶

Esperanto: pluralize("librojn") no longer produces "librojjn" — added no-op rules for -jn and -j to prevent double pluralization of already-plural forms
register() now clears the lru_cache via lazy import of _clear_regex_cache, preventing stale results when language rules are replaced at runtime
_match_case() no longer raises IndexError when the target string is empty (e.g. add_irregular("Test", ""))

Added¶

Regression tests for all three bug fixes

[2.1.0] - 2025-07-13¶

Added¶

Esperanto (eo) language support — simplest pluralization system: -j suffix for plural, -jn for accusative plural. Zero irregulars, zero exceptions.
Esperanto uncountables: pronouns (mi, vi, li, ŝi, ĝi, ni, ili, oni, si), correlatives (kio, tio, ĉio, nenio, iom), particles (kaj, aŭ, sed, la, etc.)
Tests: test_eo_plurals.py, test_eo_singles.py, round-trip tests

Changed¶

pyproject.toml description updated to include EO
Keywords updated with esperanto

[2.0.2] - 2025-07-13¶

Changed¶

is_plural / is_singular simplified: removed redundant irregular checks, delegate to singularize / pluralize
README: coverage badge corrected to 97%, Python 3.14 added to feature list
release.yml build Python bumped from 3.12 to 3.14

[2.0.1] - 2025-07-13¶

Fixed¶

mypy CI: warn_unused_ignores disabled for tests.test_property (hypothesis decorators vary across Python versions)
tests/__init__.py added for proper module resolution

Changed¶

Deduplicated _pluralize_hyphenated / _singularize_hyphenated into shared _transform_hyphenated
CONTRIBUTING.md updated with new rules/ subpackage structure and all test files
supported_languages() docstring now lists all 5 languages
Python 3.14 added to CI matrix

[2.0.0] - 2025-07-13¶

Breaking¶

Language rule modules moved from pluralio/rules_en.py to pluralio/rules/en.py (same for es, fr, it, pt). Use import pluralio as before — the public API is unchanged. If you imported rule modules directly (e.g. from pluralio import rules_en), update to from pluralio.rules import en.

Changed¶

pluralio/rules/ is now a subpackage with its own __init__.py that triggers registration of all built-in languages
_match_case optimized with islower() short-circuit (~5% faster on lowercase inputs)

[1.8.3] - 2025-07-13¶

Fixed¶

_match_case now guards against empty strings, preventing IndexError on edge-case inputs
is_singular now uses pluralize(x) != x instead of singularize(x) == x, fixing inconsistency with is_plural

Added¶

__all__ exports in core.py and registry.py for explicit public API surface
snapshot() and restore() functions in registry.py for test-safe registry isolation
mypy type checking now covers tests/ directory in CI
ruff lint rules expanded: B (bugbear), C4 (comprehensions), SIM (simplifications), PT (pytest style)
.hypothesis/ added to .gitignore

Changed¶

Extracted _transform() in core.py to deduplicate pluralize/singularize logic
conftest.py now uses public snapshot/restore API instead of accessing private _REGISTRY
test_edge_cases.py (2948 lines) split into 6 per-language files: test_cross_edge_cases.py, test_en_edge_cases.py, test_es_edge_cases.py, test_pt_edge_cases.py, test_fr_edge_cases.py, test_it_edge_cases.py
__version__ fallback reads from pyproject.toml when package metadata is unavailable
Module docstring documents side-effect import requirement for language registration
LanguageRules docstring improved with .. warning:: block about intentional container mutability

[1.8.2] - 2025-07-13¶

Fixed¶

Italian -tà abstract nouns (verità, felicità, libertà) now invariable via new à$ plural rule
Italian -ese nationality words (inglese, francese, giapponese, cinese, portoghese, olandese, svedese, norvegese, finlandese, danese) now round-trip correctly via new esi$ → ese singular rule
Italian round-trip for 27 common -e words: piede, fine, margine, bottone, gigante, santone, difensore, attaccante, esemplare, necessario, sufficiente, importante, evidente, intelligente, potente, urgente, simile, facile, difficile, utile, inutile, possibile, impossibile, probabile, improbabile, portatile, turbine
sangue, peggio, week-end added to uncountables
Months (gennaio–dicembre), seasons (primavera, estate, autunno, inverno), and numbers (due, tre, cinque, sei, sette, nove, dieci, mille) added to uncountables

Added¶

New plural rule: à$ → invariable (for abstract nouns ending in -tà)
New singular rule: esi$ → ese (for nationality words)
27 new Italian irregular plurals for common -e words
27 new Italian uncountables (months, seasons, numbers, invariable nouns)

Changed¶

Italian irregulars expanded: 239 (was 212)
Italian uncountables expanded: 144 (was 117)
Italian plural rules: 19 (was 18)
Italian singular rules: 12 (was 11)

[1.8.1] - 2025-07-13¶

Fixed¶

Italian round-trip for bicchiere, giornale, mare, padre, madre, re (masculine -e words that singularized to -o instead of -e)
Italian round-trip for zio → zii (was singularizing to zo)
società added to uncountables (invariable noun, was being pluralized incorrectly)

Added¶

13 new Italian irregular plurals: bicchiere, giornale, mare, padre, madre, re, zio, zia, fratello, nonno, nonna, cugino, cugina

Changed¶

Italian irregulars expanded: 212 (was 199)
Italian uncountables expanded: 117 (was 116)

[1.8.0] - 2025-07-13¶

Added¶

54 new Italian irregular plurals: Greek-origin masculine -a (schema, dogma, emblema, idioma, fantasma), profession nouns in -ista (artista, autista, giornalista, linguista, turista, astronauta), masculine -e nouns (fiume, ponte, cliente, residente, presidente, generale, animale, fossile, fucile, cortile), adjectives in -are (volgare, regolare, singolare, particolare), feminine -a nouns (borsa, coppia, porta, torta, tazza, sala, ruota, penna, palla, tela, sorella, frusta), body part irregulars (braccio → braccia, lenzuolo → lenzuola, ciglio → ciglia), -cio words (laccio, straccio, sacrificio, abbraccio), -go piana (obbligo → obblighi), and feminine -e nouns (torre, corte, sorte, morte, canzone)
60 new Italian uncountables: tech loanwords (streaming, download, upload, hashtag, follower, like, share, tweet, link, click, login, logout, reset, backup, input, output, format, record, report, budget), media terms (show, preview, trailer, remake, sequel, break, flop, sketch, spot), sports (tennis, golf, hockey, rugby, match, round, ring, volley, sprint, set), accented invariables (città, virtù, tè, perché, cioè, sé), days of the week (lunedì–domenica), and misc (shock, stop, start, loop, cross, short, staff)

Changed¶

Italian irregulars expanded: 199 (was 145)
Italian uncountables expanded: 116 (was 56)

[1.7.2] - 2025-07-13¶

Fixed¶

Italian masculine -e words (cuore, studente, dente, nome, sole, colore, valore, dottore, signore, attore, professore, imperatore, scultore, pittore, scrittore) now round-trip correctly
Italian feminine -zione/-sione words (nazione, stazione, lezione, funzione, etc.) now round-trip correctly via -ioni → -ione singular rule
Italian masculine -tore words now singularize correctly via -tori → -tore regex rule
Italian feminine -cia/-gia words after consonant (pioggia → piogge, roccia → rocce, fascia → fasce, ascia → asce) now pluralize correctly via consonant-aware regex
Italian feminine -cia/-gia words after vowel (valigia → valigie, camicia → camicie) now pluralize correctly via vowel-aware regex
Italian -ie feminine words (superficie → superfici, effigie → effigi) now pluralize correctly via -ie → -i regex rule
Italian Greek-origin masculine -a words (problema → problemi, sistema → sistemi, clima → climi, dramma → drammi, programma → programmi, etc.) now pluralize correctly as irregulars
Italian masculine -io words (studio → studi, esercizio → esercizi) now round-trip correctly as irregulars
Italian feminine -a plurals (sedie, chiavi, isole, stelle, barche, etc.) now idempotent as irregulars
is_singular/is_plural now correct for masculine -e words (cuore, studente, etc.)
Removed 19 loanwords from _IRREGULAR_PLURALS (already in _UNCOUNTABLE, no duplication needed)

Changed¶

Italian regex rules expanded: 18 plural rules (was 15), 11 singular rules (was 8)
Italian irregulars expanded: 145 (was 121)

[1.7.1] - 2025-07-13¶

Fixed¶

Italian feminine -e singulars (notte, mente, fronte, sede, parte, classe, noce, fede, pace, croce, gente, nave, valle, parete, radice) now round-trip correctly
Italian masculine -io/-cio/-gio words (bacio, spazio, ufficio, socio, cambio, esempio, principio, stadio, negozio, formaggio, viaggio, raggio, coraggio, passaggio, messaggio) now round-trip correctly
Italian feminine -a plurals (case, scuole, banane, gatte, paste, piante, famiglie, squadre, feste, donne, ragazze, macchine, piazze, pizze, bambine) now idempotent
voce moved from uncountable to irregular plurals (voce → voci)
noir added to uncountable for film-noir compound
Removed 18 redundant _EXTRA_SINGLES (all auto-generated from inverse mapping)
is_singular/is_plural now correct for feminine -e words

[1.7.0] - 2025-07-13¶

Added¶

Italian (it) language support with 15 plural rules and 8 singular rules
121 irregular plurals covering sdrucciola -co → -ci (amico → amici), sdrucciola -go → -gi (asparago → asparagi), piana -go → -ghi explicit mappings (lago → laghi), completely irregular words (uomo → uomini, dio → dei, uovo → uova), foreign loanwords (film, bar, computer), common -e/-io words for round-trip correctness (cane → cani, vizio → vizi), feminine -e singulars (notte → notti, mente → menti), masculine -io/-cio/-gio words (bacio → baci, formaggio → formaggi, viaggio → viaggi), and feminine -a words for plural idempotency (casa → case, scuola → scuole)
56 uncountable/invariable words including foreign loanwords, Greek-origin forms in -i (analisi, crisi), pluralia tantum (occhiali, forbici), truncated forms (foto, auto), accented-vowel words (caffè), and invariable loanwords (noir)
Italian test suite: test_it_plurals.py, test_it_singles.py, and Italian edge cases in test_edge_cases.py

[1.6.2] - 2025-07-13¶

Added¶

9 pluralia tantum words: ciseaux, lunettes, jumelles, pincettes, arrérages, ambages, fraîtures, mœurs, condoléances, frais, gens
2 -ou → -oux irregulars: gnou → gnoux, bayou → bayoux
4 -al → -als exceptions: étal → étals, val → vals, gal → gals, recal → recals
cour → cours irregular mapping for round-trip correctness
7 accentless variant mappings: generaux → general, hopitaux → hopital, metaux → metal, signaux → signal, regaux → regal, recitaux → recital, voeux → voeu
aux → au extra single for compound singularization (pots-aux-feux → pot-au-feu)
French hyphenated compound pluralization: all noun segments are now pluralized/singularized (e.g. café-théâtre → cafés-théâtres, chou-fleur → choux-fleurs, pot-au-feu → pots-aux-feux)

Changed¶

French hyphenated compound handling in core.py: pluralizes/singularizes all noun segments, skipping function words (de, en, et, etc.) and fixed parts (vie, ciel)
Removed cours from uncountable (now handled via cour → cours irregular)
Updated supported languages table with accurate rule counts (104 irregulars, 27 extra singles, 81 uncountables)

[1.6.1] - 2025-07-13¶

Fixed¶

bleu, pneu, émeu now pluralize to bleus, pneus, émeus instead of bleux, pneux, émeux (were incorrectly caught by eu$ → eux regex; added as irregulars with +s)
tuyau, noyau, boyau, sarrau now pluralize to tuyaux, noyaux, boyaux, sarraux instead of tuyaus, noyaus, etc. (added au$ → aux regex rule and irregular mappings)
tuyaux, noyaux, boyaux, sarraux now singularize back to tuyau, noyau, boyau, sarrau instead of tuyal, noyal, etc. (added to _EXTRA_SINGLES)
bail now pluralizes to baux instead of bails (was missing from -ail → -aux irregulars)
endroit removed from uncountable — now correctly pluralizes to endroits
paris removed from uncountable — now correctly singularizes to pari
fils added to uncountable (pluralia tantum) — no longer singularizes to fil
ideaux, émaux (accentless variants) now singularize to ideal, email via _EXTRA_SINGLES

Added¶

au$ → aux pluralization regex rule (between eau$ and eu$)
Accentless variant mappings for ideaux → ideal and emaux → email
Tests for all new irregulars and fixed cases

[1.6.0] - 2025-07-13¶

Added¶

French (fr) language support: 5 regex rules (-al → -aux, -eau → -eaux, -eu → -eux, -s/-x/-z → invariable, default +s), 90 irregular plurals (cheval → chevaux, travail → travaux, bijou → bijoux, œil → yeux, monsieur → messieurs, etc.), 20 extra singles for singularization restoration (travaux → travail, yeux → œil, etc.), 60 uncountable/invariable words (information, temps, riz, croix, nez, etc.)
French edge case tests: case preservation (title, all caps), mixed case, hyphenated words, idempotency, round-trip (50+ words), count-aware pluralization, whitespace preservation, single letters, uncountable consistency (70+ words), is_singular/is_plural checks

Changed¶

Updated pyproject.toml description to include French (EN/ES/PT/FR)
Added french to keywords in pyproject.toml
Updated README to reflect French language support

[1.5.5] - 2025-07-12¶

Fixed¶

dinamarqueses now singularizes to dinamarqués (was dinamarques — missing accent restoration in _EXTRA_SINGLES)
fraces now singularizes to frac (was fraz — incorrectly caught by ces$ → z singularization rule; added frac → fraces to irregular plurals)

Added¶

452 Spanish edge case tests: case preservation (title, all caps), mixed case, hyphenated words (noun+noun, leading/double hyphen), idempotency (45+ already-plural words), round-trip (45+ words), count-aware pluralization, whitespace preservation, single letters, uncountable consistency (70+ words), is_singular/is_plural checks
150 English edge case tests: count-aware pluralization (18 words × 3 counts), mixed case, single letters, uncountable consistency (40+ words), is_singular/is_plural checks, hyphenated compound round-trip (13 compounds)

[1.5.4] - 2025-07-12¶

Fixed¶

Portuguese verb+noun hyphenated compounds now pluralize the last segment (noun) instead of the first (verb): quebra-cabeça → quebra-cabeças (was quebras-cabeça), guarda-chuva → guarda-chuvas, beija-flor → beija-flores, arranha-céu → arranha-céus, passa-tempo → passa-tempos, guarda-roupa → guarda-roupas
Added 28 Portuguese verb prefixes to _LAST_SEGMENT_PLURAL_FIRST_WORDS: quebra, guarda, arranha, limpa, mata, saca, abre, corta, lança, trava, cata, chupa, espanta, passa, pisa, salta, adivinha, roí, cai, sobe, desce, para, pára, bota, tira, pega, leva, traz, beija

Changed¶

Moved 3 Portuguese compound words from uncountable to pluralizable: guarda-chuva, beija-flor, passa-tempo — these are verb+noun compounds that pluralize the noun, not invariable words
Portuguese uncountable count adjusted from 91 to 88

Added¶

725 Portuguese edge case tests: case preservation (title, all caps, mixed), hyphenated words (noun+noun, verb+noun), idempotency (140+ already-plural words), round-trip (140+ words), NFD normalization, is_plural/is_singular checks, count-aware pluralization, whitespace preservation, single letters, uncountable consistency

[1.5.3] - 2025-07-12¶

Added¶

291 additional Portuguese irregular plurals (97 → 388): expanded -ão → -ões group with ~80 common words (criação, emoção, região, questão, lição, etc.), added -ês → -eses gentilicios (freguês, português, japonês, inglês, francês, holandês, dinamarquês, chinês, norueguês, polonês), added monosyllables (som, flor, cor, mar, paz, luz, cruz, rapaz, arroz), added -el → -éis (anel, pincel, painel, pastel, coronel), added -ol → -óis (girassol), added -ães group (catalão, guardião, sotão, caimão, tecelão), and ~130 tech/business loanwords (framework, endpoint, callback, middleware, hash, bucket, pipeline, build, ticket, socket, fixture, mock, diff, commit, driver, buffer, proxy, header, footer, script, backlog, kanban, scrum, review, merge, branch, fork, push, pull, tag, log, bug, hack, patch, release, deploy, rollback, backup, snapshot, dashboard, plugin, addon, snippet, template, theme, skin, layout, form, input, output, flag, switch, toggle, hook, trigger, handler, listener, observer, wrapper, adapter, parser, lexer, compiler, debugger, profiler, linter, runner, worker, master, slave, leader, follower, node, host, peer, client, broker, shard, replica, pod, volume, image, registry, chart, graph, test, suite, case, stub, spy, coverage, report, alert, event, message, webhook, payload, request, response, session, cookie, query, cursor, field, schema, migration, seed, job, task, queue, stack, heap, pool, cache, stream, pipe, port, channel, signal, beacon, sensor, device, badge, card, menu, tab, icon, button, label, filter, sort, block, section, item, element, post, comment, user, account, profile, role, group, team, project, issue, plan, tier, quota, limit, invoice, payment, charge, refund, license, subscription, monitor, scanner, manager, browser, printer, computer, sender, receiver, editor, visitor, sponsor, partner, provider, supplier, investor, founder, developer)
56 additional Portuguese extra singles (32 → 88): accent restoration for all new -ões, -ães, -éis, -óis, -eses, -ãos, and -eis plural forms

Changed¶

Moved 14 tech loanwords from uncountable to irregular plurals (framework, pipeline, build, socket, commit, proxy, deploy, backup, switch, node, host, cookie, cache, post) — these are pluralizable with +s, not invariable
Portuguese uncountable count adjusted from 92 to 91

[1.5.2] - 2025-07-12¶

Changed¶

Optimized _split_whitespace: fast path for words without surrounding whitespace (common case), skips 3 string operations
Skip .lower() allocation for already-lowercase strings via islower() check
Inlined unicodedata.normalize call to avoid redundant variable assignment
Performance: ~30% faster across all languages (EN/ES/PT pluralize and singularize)

[1.5.1] - 2025-07-12¶

Added¶

62 additional Portuguese uncountables (30 → 92): -x invariables (fax, complex, suplex, etc.), anatomical -ps (bíceps, tríceps, fórceps), Greek/biblical -is (oásis, gênesis), -s invariables (pires, ourives, cosmos, seis), adverbs (menos, jamais), music genres (blues, soul, funk, reggae, folk, metal), sports (rugby, skate, poker, darts), and tech/foreign loanwords (deploy, commit, cloud, kernel, streaming, etc.)

Changed¶

Test suite: 3,116 → 3,188 tests
100% coverage maintained

[1.5.0] - 2025-07-12¶

Added¶

Portuguese (pt) pluralization and singularization support
Portuguese regex rules: -ão→-ões, -m→-ns, -il→-is, -l→-is, -r→-res, -z→-zes, -x→invariable, vowel→+s
Portuguese singularization rules: -ões→-ão, -ães→-ão, -zes→-z, -ns→-m, consonant+-is→-il, -is→-l, double consonant+-es→strip s, -tes/-des/-mes/-les→strip s, -res→-r, default strip s
97 Portuguese irregular plurals covering -ão→-ões/-ães/-ãos splits, accent shifts (-el→-éis, -ol→-óis), -il paroxítona→-eis, and foreign loanwords
32 Portuguese extra singles for accent restoration (papéis→papel, sóis→sol, méis→mel, fiéis→fiel, etc.)
92 Portuguese uncountables including -x invariables (tórax, látex, fax, complex), anatomical -ps (bíceps, tríceps, fórceps), Greek/biblical -is (oásis, gênesis), -s invariables (lápis, vírus, óculos, férias, pires, ourives), pronouns (nós, vós), adverbs (menos, jamais), compound words, music genres (blues, soul, funk, reggae), sports (rugby, skate, poker), and foreign/tech loanwords (software, deploy, commit, cloud, kernel)
Portuguese test suite: pluralization, singularization, and round-trip tests

Fixed¶

cantil→cantiis (should be cantis): added il$→is rule before l$→is
leis→lel, reis→rel, pais→pal: added consonant+is→il rule and extra singles for vowel+is words
mel→meis (should be méis), fiel→fieis (should be fiéis): added to irregulars
árvores→árvor (should be árvore): added to extra singles
chaves→chav, chefes→chef, peixes→peix: replaced overly aggressive consonant+es rule with specific -res→-r rule
grão→grões (should be grãos), são→sões (should be sãos): added to irregulars
nós→nó, vós→vó: added to uncountables

Changed¶

Test suite: 2,859 → 3,116 tests
100% coverage maintained across all modules

[1.4.2] - 2025-07-12¶

Changed¶

Optimized _match_case: single-pass case detection instead of 3 passes, with fast path for all-lowercase words
Skip unicodedata.normalize for ASCII-only strings (most English words)
Eliminated redundant .lower() call in _apply_rules — caller passes pre-computed lowercase
Simplified _apply_rules signature (removed unused irregulars and rule_list params)
Streamlined _split_whitespace to avoid redundant string operations
Performance: ~145K → ~285K ops/sec (~96% faster)

[1.4.1] - 2025-07-12¶

Added¶

19 English uncountable -ics disciplines: politics, ethics, gymnastics, linguistics, athletics, civics, statistics, informatics, classics, mechanics, dynamics, genetics, obstetrics, pediatrics, psychiatrics, orthopaedics, academics, logistics, hysterics
4 English uncountable games: billiards, darts, checkers, craps
5 Spanish compound uncountables: quitamanchas, matasanos, guardabosques, guardacostas
Spanish invariable adjective gris to uncountables

Fixed¶

annex round-trip: annexes singularized to annexe (British) instead of annex (American English)
finesse round-trip: finesses singularized to finess instead of finesse
bellyache round-trip: bellyaches singularized to bellyach instead of bellyache
peine round-trip: peines singularized to pein instead of peine (vowel+consonant+e pattern)
rol round-trip: roles singularized to role (English loanword) instead of rol (Spanish singular)

Changed¶

Test suite: 2,817 → 2,859 tests
Deep validation: 1,141 English words + 466 Spanish words, 0 failures

[1.4.0] - 2025-07-11¶

Added¶

compás → compases irregular plural (was incorrectly handled by regex)
17 Spanish loanwords ending in -r with -s pluralization: monitor, scanner, manager, browser, printer, computer, editor, visitor, sponsor, partner, provider, supplier, investor, founder, developer, sender, receiver
9 Spanish words ending in -é as explicit irregulars: café, té, bebé, puré, cliché, paté, rosé, bidé, frappé
10 Spanish accent restoration mappings in _EXTRA_SINGLES: caimanes→caimán, guardianes→guardián, sotanes→sotán, comodines→comodín, edenes→edén, limones→limón, melones→melón, dragones→dragón, campeones→campeón
20 Spanish uncountables: medical/scientific (parálisis, tuberculosis, psoriasis, elefantiasis, pediculosis, rabies, mumps), biblical (génesis, apocalipsis), anatomical (biceps, triceps, cuádriceps, forceps), compound words (lavacoches, sacamuelas, cortaplumas, abrelatas, parachoques, rompecorazones, sacaorchos)
5 Spanish -y words: ley, rey, buey, hoy, convoy
4 non-tech loanwords: sandwich, whisky, ponche, parche

Fixed¶

Idempotency bug: pluralize(pluralize("café", lang="es"), lang="es") returned "cafeses" instead of "cafés". The regex rule -és$ → -eses was incorrectly matching already-plural forms like cafés, tés, bebés. Fixed by adding these words as explicit irregulars.

Changed¶

Spanish irregular plurals: 61 → 351
Spanish extra singles: 34 → 80
Spanish uncountables: 62 → 87
Test suite: 2,799 → 2,817 tests
Updated comparison table in README to include pluralsingular as competitor
Updated roadmap to reflect actual release history

[1.3.0] - 2025-07-11¶

Added¶

~130 Latin/Greek classical irregular plurals (Fase 1): -um→-a, -a→-ata, -a→-ae, -en→-ina, -is→-ides, -itis→-itides, -on→-a, -ix/ex→-ices, Hebrew -im, Arabic -i, and other irregulars
~40 English uncountables (Fase 2): herd animals, fish, and other invariable nouns
~300 explicit inverse singularizations (Fase 3): -uses→-use, -eses→-esis, -oes→-oe, -xes→-xe, -zzes→-zz, -ois→-oi
~98 foreign -o words with -os pluralization (Fase 4): musical terms, food, animals, and general loanwords
~130 proper noun irregulars with case preservation (Fase 5): names ending in -ie/-y→-ies
~22 singular -s words with -es pluralization (Fase 6): acropolis, aegis, alias, cosmos, ethos, etc.
~52 suffix-based uncountables (Fase 7): demonyms in -ese, fish names, pox diseases
rhinoceros to _UNCOUNTABLE
octopi, hippopotami, platypi to _IRREGULAR_SINGLES for correct singularization of Latin plural forms

Changed¶

Roadmap Fase 8 (verbs/pronouns) evaluated and explicitly marked as out of scope

[1.2.0] - 2025-07-11¶

Changed¶

pluralize() and singularize() now preserve leading and trailing whitespace around the input word
English -che words handled as explicit irregulars instead of a regex rule, fixing singularization for words like caches, niches, quiches
Expanded compound prefix detection for hyphenated words: meta, post, re, pre, anti, pro, non, sub, co, ex, inter, intra, multi, semi, pseudo, proto, neo

Added¶

20+ new English -che irregulars (ache, cliche, quiche, brioche, pastiche, douche, gouache, cloche, barouche, cartouche, caliche, huarache, seiche, troche, microfiche, attache, synecdoche, etc.)
6 compound -ache words (headache, earache, toothache, backache, heartache, stomachache)
13 demonyms ending in -ese as uncountables (japanese, chinese, vietnamese, burmese, lebanese, portuguese, javanese, sundanese, senegalese, congolese, sudanese, maltese, siamese)
ref/ROADMAP-INFLECT.md — roadmap to reach feature parity with inflect in English word coverage
Comprehensive edge-case tests for whitespace preservation, -che singularization, and compound prefixes

[0.1.0] - 2025-07-10¶

Added¶

pluralize() and singularize() for English and Spanish
count parameter for count-aware pluralization
Case preservation (title case, all caps)
Hyphenated word support (mother-in-law → mothers-in-law)
Extensibility API: add_irregular, add_plural, add_singular, add_uncountable, add_plural_rule, add_singular_rule, register_language
supported_languages() to list registered languages
LanguageRules dataclass for custom language registration
py.typed marker for PEP 561
CI workflow (lint + type check + test on Python 3.10–3.13)
Release workflow (trusted publishing to PyPI)
95% coverage requirement enforced in CI