Add the Czech, Slovak, Slovenian and Croatian packs (3.0, 3f)

README.rst

@@ -82,7 +82,7 @@ Quick Start Example
     >>> russian.parse("Сидоров Иван Петрович").family
     'Сидоров'
     >>> locales.available()
-    ('ja', 'ne', 'ru', 'tr_az', 'zh')
+    ('cs', 'hr', 'ja', 'ne', 'ru', 'sk', 'sl', 'tr_az', 'zh')
 
 Learn more
 ----------

docs/design/decisions/M3-marker-led-clause.md

@@ -11,7 +11,7 @@
 - A rule that extends where a vocabulary claim acts re-asks C-i of the whole set it reads; M3 made the marker vocabulary act on the first word of any bracketed clause, and the re-ask removed `roz` (`Rosalind Roz Smith` reads no maiden name, middle 'Roz' under the default order) — 2026-08-26 #335, M3 makes.
 - Declaring `maiden_delimiters` is a declaration about the data, still needed for markerless and one-word clauses: `Cherice J. (Johnson) Williams` reads nickname 'Johnson' by default — 2026-08-26 #335, what configuring.
 
-Elsewhere: a marker followed by a comma, standing alone or glued to either word, is no marker (`Jane (née , Jones) Smith` → nickname 'née , Jones') — decisions/N1's 2026-10-12 #653 entry; the delimiters trumping every reading inside the clause (`Jane Doe (nee Smith MA)` → maiden 'Smith MA') — decisions/M2's 2026-09-19 (Derek), #533 review — THE DELIMITERS TRUMP entry; the bare take's first-word floor, which reads `Jane Smith née V` as maiden 'V' — decisions/M2's 2026-10-04 (Derek), #601 entry; `roz`'s exclusion, now on the word (`uv run python tools/vocab/query.py --word roz`) — decisions/vocabulary-exclusions' Excluded (MAIDEN_MARKERS) block.
+Elsewhere: a marker followed by a comma, standing alone or glued to either word, is no marker (`Jane (née , Jones) Smith` → nickname 'née , Jones') — decisions/N1's 2026-10-12 #653 entry; the delimiters trumping every reading inside the clause (`Jane Doe (nee Smith MA)` → maiden 'Smith MA') — decisions/M2's 2026-09-19 (Derek), #533 review — THE DELIMITERS TRUMP entry; the bare take's first-word floor, which reads `Jane Smith née V` as maiden 'V' — decisions/M2's 2026-10-04 (Derek), #601 entry; `roz`, refused as a default marker and since 3.0 a reading off by default that the Czech pack turns on (`uv run python tools/vocab/query.py --word roz`) — decisions/vocabulary-exclusions' Excluded (MAIDEN_MARKERS) block.
 
 Open: the fullwidth-colon marker (`旧姓：佐藤` arriving as one word) is decisions/M2's Open line for #317.
 

docs/design/decisions/deviates-registry.md

@@ -2,7 +2,7 @@
 
 ## Standing
 
-- A locale pack declares its deviations from the defaults as a module-level `DEVIATES` callable, acceptance metadata kept off `Locale`, whose value equality and pickling a callable field would break; `test_registry_is_the_pack_contract` fails unless every registered pack ships `DEVIATES` and a rotator list, which is what lets mechanisms.md#LOCALE-PACKS-PURE-DATA promise packs no code paths of their own. A field on `Locale` is to be revisited only if a runtime consumer of the declaration, such as an `explain()`, appears, and none has — 2026-07-18. A vocabulary pack's deviations are derivable instead, from the reading index (`_reading_index.changed_words`), which tests/v2/test_locales.py's choice gate reads; the gate ors the derived surface into every registered pack with `languages`, so a hand `DEVIATES` declares only what a fragment or a patch beside a choice changes; the contract test refuses a `DEVIATES` on a pack that is nothing but a choice of languages (`locales.NE`, the first) and requires one of every other pack — 2026-10-29 #753, phase 3b; 2026-10-30 #753, phase 3d.
+- A locale pack declares its deviations from the defaults as a module-level `DEVIATES` callable, acceptance metadata kept off `Locale`, whose value equality and pickling a callable field would break; `test_registry_is_the_pack_contract` fails unless every registered policy or segmentation pack ships `DEVIATES`, and every registered pack a rotator list, which is what lets mechanisms.md#LOCALE-PACKS-PURE-DATA promise packs no code paths of their own. A field on `Locale` is to be revisited only if a runtime consumer of the declaration, such as an `explain()`, appears, and none has — 2026-07-18. A vocabulary pack's deviations are derivable instead, from the reading index (`_reading_index.changed_words`), which tests/v2/test_locales.py's choice gate reads; the gate ors the derived surface into every registered pack with `languages`, so a hand `DEVIATES` declares only what a fragment or a patch beside a choice changes; the contract test refuses a `DEVIATES` on a pack that is nothing but a choice of languages (`locales.NE`, the first, and `CS`, `SK`, `SL` and `HR`), and a choice pack whose choice changes no word, and requires one of every other pack — 2026-10-29 #753, phase 3b; 2026-10-30 #753, phase 3d; 2026-10-30 #753, phase 3f.
 
 Open: none in this file.
 
@@ -11,3 +11,4 @@ Open: none in this file.
 - 2026-07-18 (d4aaafa; the DEVIATES design note) — a `deviates` field ON Locale was rejected: callables break value equality and pickle-by-reference, and DEVIATES is acceptance metadata, not runtime behavior (revisit only if a runtime consumer like explain() appears). The REGISTRY became the contract instead: test_registry_is_the_pack_contract fails structurally unless every registered pack ships DEVIATES and its rotator list. This is why LOCALE-PACKS-PURE-DATA can promise "no code paths of their own" and mean it.
 - 2026-10-29 #753, phase 3b — A VOCABULARY PACK'S DEVIATIONS ARE DERIVED, NOT DECLARED. `Locale.languages` selects among the default's readings, so the words a choice can move are computable: `_reading_index.changed_words(chosen)` is every word joining or leaving a compiled constant under the choice, and tests/v2/test_locales.py's `test_non_interference_each_choice` requires the default's reading for every corpus name holding none of them, with a negative control (`test_the_choice_gate_fails_when_its_surface_misses_a_word`). No pack ships in 3b, so the gate runs over choices of the files' tags; the registry contract test still asks every REGISTERED pack for `DEVIATES` and a rotator list, and learns to read a derived surface for a pack with `languages` in the PR that registers the first one (3d). The policy and segmentation packs (JA, RU, TR_AZ, ZH) keep theirs, decisions/vocabulary-model's 2026-10-27 #753 entry, item 7.
 - 2026-10-30 #753, phase 3d — THE CONTRACT READS A DERIVED SURFACE. `locales.NE` registers the first pack whose content is a choice of languages and nothing else (`Locale.lexicon` empty, `Locale.policy` the empty patch). tests/v2/test_locales.py's `_deviates(code)` gives the gate, the rotator test and the combined gate every pack's predicate: the surface derived from its `languages` through `changed_words` and the 3b choice gate's `_surface_deviates`, or'd with its `DEVIATES`, so a pack with a fragment or a patch beside a choice needs to declare only that part (`test_a_pack_with_a_choice_and_a_patch_declares_only_the_patch`; nothing refuses a hand copy of the choice's words there); `test_registry_is_the_pack_contract` refuses a `DEVIATES` on a pack that is nothing but its choice, so a hand copy cannot drift beside the derived one, and still requires one of every other pack. Its rotator list stays required, as the gate's positive side. decisions/vocabulary-model's 2026-10-30 phase 3d entry, item 2.
+- 2026-10-30 #753, phase 3f — A CHOICE PACK MUST CHANGE A WORD. `locales.CS`, `SK`, `SL` and `HR` register as choice packs under phase 3d's contract, with no `DEVIATES`. `test_registry_is_the_pack_contract` adds one refusal: a choice pack whose choice changes no word (an empty `changed_words`), since its gate would pass on nothing; and `test_choice_rotators_cover_every_changed_word` holds each choice pack's rotator list to a name per changed word, so the gate's positive side reaches every word. The combined gate reads the surface of the union of every registered pack's languages, since `parser_for` applies one choice, not one per pack, or'd with every pack's hand `DEVIATES` whatever the pack, since a mixed pack carries both. decisions/vocabulary-model's 2026-10-30 phase 3f entry.

docs/design/decisions/salutation-titles.md

@@ -7,7 +7,7 @@ Closes #606 (#662 later added owura and bapak, its bullet below, and #847 cllr).
 - A word that addresses or ranks a person and is written in FRONT of the name ships as a title unless it is borne as an ordinary name in the leading position, the position test of vocabulary-collisions' C-i, TITLES having no ambiguous subset to mark a collision in, so the only answers are ship and do not ship — 2026-10-04 #606.
 - One exception to C-i, for TITLES alone: a word borne as a name only RARELY as the first word of one ships on the precedent of `graf` (`herra`, `batoni`, `hrabia`, `kreivi` and the rest of the graf class, each with its counts on its title reading's `note`), and no word in the particle or suffix sets, and no later title, takes it without its own counts — 2026-10-04 #606.
 - The cost falls on a name that BEGINS with such a word: `Greve Anna` reads title 'Greve', family 'Anna' and `Herra Wijaya` title 'Herra', family 'Wijaya'; a declared `FAMILY_FIRST` does not help (`Hrabia Anna` reads title 'Hrabia' under it); the family comma keeps the word as the family where no name word stands behind it (`Hrabia, Anna` reads given 'Anna', family 'Hrabia', and so does `Hrabia Jr., Anna`), while a name word behind it makes it the title by H6 (`Hrabia Nowak, Anna` reads title 'Hrabia', family 'Nowak'); a name that ends in the word keeps it (`Gladys Knight`, `Knight, Gladys` read family 'Knight'), unless a period follows it (`Gladys Knight.` reads title 'Knight.', family 'Gladys', by H5, accepted rather than prevented) — 2026-10-04 #606.
-- Words judged a COMMON leading-position name stay out (`thiru`, `puan`, `pan`, `rodina`, and `ông` and its pair `bà`), and words with no evidence either way are not shipped (`gróf`, `cik`, `paron`, `tikin`, `janab`), each an `excluded` entry in its language's file (`uv run python tools/vocab/query.py --excluded title --issue 606`) — 2026-10-04 #606.
+- Words judged a COMMON leading-position name stay out (`thiru`, `puan`, `pan`, `rodina`, and `ông` and its pair `bà`), and words with no evidence either way are not shipped (`gróf`, `cik`, `paron`, `tikin`, `janab`), each an `excluded` entry in its language's file (`uv run python tools/vocab/query.py --excluded title --issue 606`) — 2026-10-04 #606. `pan` alone has since become a `salutation` reading written `default = false` in cs.toml and pl.toml, which a Czech or Polish choice turns on and the default does not read — 2026-10-30 #753, phase 3f.
 - Accents are not folded, so `paní` and `härra` match only their accented spellings (`Paní Nováková` reads title 'Paní', `Pani Anna` reads given 'Pani'); folding the lexicon would turn both into collisions, and whoever proposes it re-argues the entry — 2026-10-04 #606.
 - `gravos` was dropped, not excluded, since nothing showed it is the Greek word for Count; `κόμης`, a Greek surname written first in official lists, is neither added nor excluded, and adding it is a decision still to make — 2026-10-04 #606.
 - `knt` (Knight, beside `kt`) is a suffix acronym with a `Knt` case mask: `Mary Knt` reads suffix 'Knt' — 2026-10-04 #606.
@@ -28,3 +28,4 @@ Elsewhere: `justice` and `marshal`, excluded here as common leading-position nam
 - **2026-10-17 #662 — owura and bapak join TITLES; malam and bay are excluded.** Akan and Indonesian Mr., found in the CLDR corpus (`Owura Bertram…`, `Bapak Dwi Putro bin Adinata MP` read the salutation as the given name). Counts from `tools/name_usage.py` on 2026-10-17: owura and bapak each under 100 US Census bearers and none found on Wikidata (a zero proves nothing, per the tool). The two exclusions are `excluded` entries in ha.toml and tr.toml, with their counts (`uv run python tools/vocab/query.py --issue 662`). Measurement: two corpus names move, both in corpus_cldr.jsonl and both toward CLDR's fields (`Owura Bertram Wilberforce Henry Robert Wooster MP` and `Bapak Dwi Putro bin Adinata MP`, salutation from given to title); recompute with `uv run python tools/differential/snapshot.py --since <the commit before this entry>`. The release-log reading was checked against the 1.4.0, 2.0.0 and 2.4.2 wheels.
 - **2026-10-22 #840 — the Italian abbreviated salutations join TITLES with case masks; sig and ing are excluded.** #838 made case repair write each full-stop chunk of a word as a word, so an unlisted `SIG.RA` repaired to 'Sig.Ra' where every release gave 'Sig.ra', and the remedy is a listed casing. Thirteen forms whose letters after the full stop are the word's own ending: `sig.ra` and `sig.na` (salutations, spellings of `signora` and `signorina`), `dott.ssa` and `prof.ssa` (academic), and the letter-heading styles `gent.mo`, `gent.ma`, `gent.le`, `ill.mo`, `ill.ma`, `ch.mo`, `ch.ma`, `preg.mo`, `preg.ma`, each with a `cased` mask. Listing them also settles the parse, which had read them by accident: `Sig.ra` and `Prof.ssa` were titles through a piece their joined word matched, and `Dott.ssa Anna Bianchi` read given 'Dott.ssa', middle 'Anna' on 1.4.0 and 2.5.0 (wheels, measured 2026-10-22); it reads title 'Dott.ssa', given 'Anna' now, under all three orders and every comma shape. None of the thirteen spells a name, each being an Italian abbreviation, so C-i's position test finds no bearer to weigh (a word with an interior full stop can be one in general: `J.R. Smith` reads given 'J.R.', rules/H2; the nearest shadow found is joined initials spelling `ch.ma`, `CH.MA ROSSI` now reading title 'CH.MA', with no bearer found); the committed corpora hold none (the snapshot moved only rules/R4's own `SIG.RA` row), and the class rests on standard Italian usage rather than a count. The bare `sig` (US SSA 50+ male, Wikidata 11+ given) and `ing` (nam_dict band 5 in Sweden, Wikidata 10+ given, US Census 1,814 family) are excluded with their counts (tools/name_usage.py, 2026-10-22); `Sig.` and `Ing.` already read as titles by H2's shape. A cost the `ing` exclusion does not decide: ClinicalTrials.gov officials write a bare `Ing` AFTER a comma, as a credential (52 of 536,981 write it, 43 after a comma, the same tool run), and 9 of the 28 distinct names holding it put it in the given field (`Viera Sevcikova, Ing` reads given 'Ing', family 'Viera Sevcikova'), where the dotted `Mario Rossi, Ing.` reads title 'Ing.'. That is a suffix question, a credential reading of `ing` beside the comma, and a title exclusion neither causes nor settles it. A trailing `Anna Bianchi Dott.ssa` keeps 'Dott.ssa' as the family, since H5 reads a trailing title only when it ends in a full stop, as `John Smith Sir` keeps 'Sir'; after a comma (`Maria Rossi, Sig.ra`) it is the title.
 - **2026-10-22 #847 — `cllr` joins TITLES as `councillor` abbreviated.** British style writes it without the period, so `Cllr John Smith` read given 'Cllr', middle 'John' where `Cllr. John Smith` read title by shape. `tools/name_usage.py Cllr`: US Census surnames under 100, SSA under 5 births a year, not in nam_dict (a selection, so a miss proves little), and 0 on Wikidata as given or family (a lower bound: a zero proves nothing, per the tool), so no bearer was found as a name, and a five-letter vowelless abbreviation is not a plausible given name. It opens 508 of 15,214 names in `ukec-individuals` (cache data date 2026-10-23T03:45:09Z, same corpus sha as the first fetch of the issue), all bare; with `cllr` removed from `Lexicon.default()` the same 508 `as_dict()` readings differ and no others, 505 of them also moving the old middle into given. `Duke` and `Earl` stay out as borne names, by the caution above: they open 8 and 7 of those names (`name_usage.py` counts 8 and 15 names that write either anywhere). The accepted cost is `John Smith Cllr.`, whose period-marked trailing `Cllr.` was family and is now a title (rules/H5).
+- **2026-10-30 #753, phase 3f — `pan` is a salutation off by default, for a Czech or Polish choice to turn on.** The 2026-10-04 exclusions in cs.toml and pl.toml become `salutation` readings written `default = false`, the collision with the Chinese surname Pan, which leads, kept in their notes (US Census 2010 12,664 bearers). The default does not read the word, so no parse moves; under `locales.CS` `Pan Jan Novák` reads title 'Pan' and `Pan Wei` title 'Pan', family 'Wei', the cost of choosing the language. decisions/vocabulary-model's phase 3f entry holds the measurement.