@finbheara/names 0.16.0 → 0.18.0

package/README.md

@@ -180,6 +180,39 @@ index.related(personName("Mary Larson"));        // [{ name: Conley-Larson<<Mary
 tolerance the pair used, and `gateSafe` is false when any of them is one a gate must not
 admit on its own.
 
+### The given-name tables, for another language
+
+The tables behind `givenFamilyTable` and `givenFamilyList` print as JSON, for a consumer
+that is not JavaScript:
+
+```sh
+npx @finbheara/names tables given-name > given-name-tables.json   # or bunx
+```
+
+`--output json` and `--schema 1` are the defaults. The PRINTED form is the contract: a
+schema keeps its shape for as long as the package offers it, and a new shape is a new
+schema, so a command written today keeps working. The files the tables are stored in are
+not part of it. Schema 1 is
+
+```jsonc
+{
+  "$comment": [ /* how to read it */ ],
+  "table": "given-name", "schema": 1,
+  "variants":  { "CONNOR": ["CONOR"], … },        // canonical -> its other spellings
+  "families":  { "CATHERINE": ["KATE", …], … },   // the curated gate tier: root -> forms
+  "notAtGate": ["MEGAN", …],                      // forms a gate must not admit as forms
+  "extended":  { "BETHANY": { "BETH": 1 }, … }    // root -> form -> tier (1 common .. 3 rare)
+}
+```
+
+Read it as the library does, and as `$comment` says: read each name as its canonical
+spelling first; two names relate only when they share a root, a form meets EVERY root it
+sits under, and nothing is transitive (`EVIE` meets `EVE` and `EVA`, which do not meet each
+other). A shared root is `given-family` unless one name reaches it only as a `notAtGate`
+form; a name that is the root itself always passes. An `extended` pair's tier is its rarer
+member's, the root counting as 1. To avoid comparing every pair, index each name under
+every root it has and compare within a root, never on one root per name.
+
 ## Phonetic and blocking keys
 
 ```ts

package/dist/chunks/{index-ye6j60jn.js → index-17zzg7f3.js}

@@ -3,9 +3,9 @@ import {
   Lexicon2
 } from "./index-0wpvbdgh.js";
 import {
-  lexiconNameOrder,
+  lexiconNameOrder2,
   defaultPersonNameFactory2
-} from "./index-tk01chxw.js";
+} from "./index-gxw36cg3.js";
 import {
   isNameParticle2
 } from "./index-23rzmwe2.js";
@@ -73,7 +73,7 @@ function createNameCensus2(options = {}) {
       return;
     }
     persons += times;
-    if (order === "unknown" && !print.includes(",") && lexiconNameOrder(print, roles) === undefined) {
+    if (order === "unknown" && !print.includes(",") && lexiconNameOrder2(print, roles) === undefined) {
       orderUndecided.add("order undecided", print, times);
     }
     for (const run of runs(name)) {
@@ -133,5 +133,5 @@ function createNameCensus2(options = {}) {
 
 export { createNameCensus2 };
 
-//# debugId=42B6EBFD51EC241764756E2164756E21
-//# sourceMappingURL=index-ye6j60jn.js.map
+//# debugId=C23DFDD520B2D39C64756E2164756E21
+//# sourceMappingURL=index-17zzg7f3.js.map

package/dist/chunks/{index-ye6j60jn.js.map → index-17zzg7f3.js.map}

@@ -4,7 +4,7 @@
   "sourcesContent": [
     "/**\n * A census of name-shaped strings: run a stream of printed names through the\n * library and report what it could not read, and why.\n *\n * The report is the maintenance loop. An UNKNOWN TOKEN is a word no `names.txt`\n * entry knows; an UNKNOWN PAIR is two adjacent unknown words, the usual shape of\n * a school or place phrase the lexicon should carry as a `B` entry. A misread is\n * fixed by making the name known in `names.txt`, so the unknown tokens, ranked\n * by how often they print, are the candidate additions.\n *\n * Nothing here reads a file or keeps global state; a census holds only what was\n * added to it.\n */\n\nimport type { PhraseClassifier } from \"../classifier/phrase.ts\";\nimport { Lexicon } from \"../lexicon/index.ts\";\nimport { lookupKey } from \"../lexicon/keys.ts\";\nimport type { PersonName, PersonNameFactory } from \"../normalize/factory.ts\";\nimport { defaultPersonNameFactory } from \"../normalize/factory.ts\";\nimport { type NameOrder, lexiconNameOrder } from \"../normalize/name-split.ts\";\nimport { isNameParticle } from \"../particles/particle-vocabulary.ts\";\n\nexport interface NameCensusOptions {\n\t/** Parses each print. Default: `defaultPersonNameFactory()`. */\n\treadonly factory?: PersonNameFactory;\n\t/** What counts as known. Default: `Lexicon.default()`. */\n\treadonly lexicon?: Lexicon;\n\t/**\n\t * Also classify each distinct print as person / school / location and report\n\t * the low-confidence ones. Off unless given, since it loads the bloom tables.\n\t */\n\treadonly classifier?: PhraseClassifier;\n\t/** Below this `classifyNameNormalized` confidence a print is reported. Default 0.3. */\n\treadonly lowConfidence?: number;\n\t/** Examples kept per reported item. Default 5. */\n\treadonly examples?: number;\n}\n\n/** One reported item: how often it printed, and a few prints it came from. */\nexport interface CensusItem {\n\treadonly key: string;\n\treadonly count: number;\n\treadonly examples: readonly string[];\n}\n\nexport interface NameCensusReport {\n\t/** prints added, counting repeats */\n\treadonly total: number;\n\t/** distinct (print, order) pairs */\n\treadonly distinct: number;\n\t/** parsed as a person */\n\treadonly persons: number;\n\t/** read as a team designation (`Team A`, `Foireann B`) */\n\treadonly teams: number;\n\t/** empty cells */\n\treadonly empty: number;\n\t/** refused as no person's name, by rule (`refuse-digit`, `unsplittable`, …) */\n\treadonly refused: readonly CensusItem[];\n\t/** words of person names that no lexicon entry knows, most frequent first */\n\treadonly unknownTokens: readonly CensusItem[];\n\t/** adjacent unknown words, most frequent first: candidate `B` entries */\n\treadonly unknownPairs: readonly CensusItem[];\n\t/** comma-less prints of undeclared order the lexicon could not orient */\n\treadonly orderUndecided: CensusItem;\n\t/** low-confidence classifications, when a classifier was given */\n\treadonly lowConfidence: readonly CensusItem[];\n}\n\nexport interface NameCensus {\n\t/** Count one printed cell. `order` is what the source declared, if anything. */\n\tadd(print: string, order?: NameOrder): void;\n\treport(options?: { readonly limit?: number }): NameCensusReport;\n}\n\nclass Tally {\n\treadonly items = new Map<string, { count: number; examples: string[] }>();\n\tconstructor(private readonly keep: number) {}\n\tadd(key: string, example: string, times = 1): void {\n\t\tlet item = this.items.get(key);\n\t\tif (item === undefined) this.items.set(key, (item = { count: 0, examples: [] }));\n\t\titem.count += times;\n\t\tif (item.examples.length < this.keep && !item.examples.includes(example)) {\n\t\t\titem.examples.push(example);\n\t\t}\n\t}\n\ttop(limit: number): CensusItem[] {\n\t\treturn [...this.items]\n\t\t\t.map(([key, { count, examples }]) => ({ key, count, examples }))\n\t\t\t.sort((a, b) => b.count - a.count || (a.key < b.key ? -1 : a.key > b.key ? 1 : 0))\n\t\t\t.slice(0, limit);\n\t}\n}\n\n/**\n * Could this word be a missing names.txt entry? Not an initial (`M`, `M.`), not\n * a surname particle (`O`, `de`), and not empty after folding.\n */\nfunction isNameWord(word: string, key: string): boolean {\n\tif (key === \"\") return false;\n\tif (word.replace(/\\./gu, \"\").length <= 1) return false;\n\treturn !isNameParticle(word);\n}\n\n/** A census over the default factory and lexicon, or the ones given. */\nexport function createNameCensus(options: NameCensusOptions = {}): NameCensus {\n\tconst factory = options.factory ?? defaultPersonNameFactory();\n\tconst lexicon = options.lexicon ?? Lexicon.default();\n\tconst classifier = options.classifier;\n\tconst threshold = options.lowConfidence ?? 0.3;\n\tconst keep = options.examples ?? 5;\n\tconst roles = lexicon.roles();\n\n\tconst seen = new Map<string, number>();\n\tlet total = 0;\n\tlet persons = 0;\n\tlet teams = 0;\n\tlet empty = 0;\n\tconst refused = new Tally(keep);\n\tconst unknownTokens = new Tally(keep);\n\tconst unknownPairs = new Tally(keep);\n\tconst orderUndecided = new Tally(keep);\n\tconst lowConfidence = new Tally(keep);\n\n\t/**\n\t * The words of the scrubbed print IN PRINT ORDER, one run per comma-separated\n\t * field, so a pair is two words the page printed side by side.\n\t */\n\tfunction runs(name: PersonName): string[][] {\n\t\tconst text = name.verdict.kind === \"name\" ? name.verdict.scrubbed : name.nn0;\n\t\treturn text.split(\",\").map((f) => f.split(/[\\s-]+/u).filter((w) => w !== \"\"));\n\t}\n\n\tfunction measure(print: string, order: NameOrder, times: number): void {\n\t\tconst name = factory.parse(print, order);\n\t\tif (name.verdict.kind === \"empty\") {\n\t\t\tempty += times;\n\t\t\treturn;\n\t\t}\n\t\tif (name.isTeam()) {\n\t\t\tteams += times;\n\t\t\treturn;\n\t\t}\n\t\tif (!name.isPerson()) {\n\t\t\trefused.add(\n\t\t\t\tname.verdict.kind === \"refused\" ? name.verdict.refusal.rule : \"unsplittable\",\n\t\t\t\tprint,\n\t\t\t\ttimes,\n\t\t\t);\n\t\t\treturn;\n\t\t}\n\t\tpersons += times;\n\t\tif (\n\t\t\torder === \"unknown\" &&\n\t\t\t!print.includes(\",\") &&\n\t\t\tlexiconNameOrder(print, roles) === undefined\n\t\t) {\n\t\t\torderUndecided.add(\"order undecided\", print, times);\n\t\t}\n\t\tfor (const run of runs(name)) {\n\t\t\tlet previousUnknown: string | undefined;\n\t\t\tfor (const w of run) {\n\t\t\t\tconst key = lookupKey(w);\n\t\t\t\tconst unknown = isNameWord(w, key) && lexicon.getNameFlags(w) === null;\n\t\t\t\tif (unknown) {\n\t\t\t\t\tunknownTokens.add(key, print, times);\n\t\t\t\t\tif (previousUnknown !== undefined) {\n\t\t\t\t\t\tunknownPairs.add(`${previousUnknown} ${key}`, print, times);\n\t\t\t\t\t}\n\t\t\t\t}\n\t\t\t\tpreviousUnknown = unknown ? key : undefined;\n\t\t\t}\n\t\t}\n\t\tif (classifier !== undefined) {\n\t\t\tconst r = classifier.classifyNameNormalized(print);\n\t\t\tif (r.confidence < threshold) lowConfidence.add(r.category, print, times);\n\t\t}\n\t}\n\n\treturn {\n\t\tadd(print, order = \"unknown\") {\n\t\t\ttotal++;\n\t\t\tconst k = `${order}\\u0000${print}`;\n\t\t\tseen.set(k, (seen.get(k) ?? 0) + 1);\n\t\t},\n\t\treport({ limit = 50 } = {}) {\n\t\t\tpersons = teams = empty = 0;\n\t\t\tfor (const t of [refused, unknownTokens, unknownPairs, orderUndecided, lowConfidence]) {\n\t\t\t\tt.items.clear();\n\t\t\t}\n\t\t\tfor (const [k, times] of seen) {\n\t\t\t\tconst at = k.indexOf(\"\\u0000\");\n\t\t\t\tmeasure(k.slice(at + 1), k.slice(0, at) as NameOrder, times);\n\t\t\t}\n\t\t\treturn {\n\t\t\t\ttotal,\n\t\t\t\tdistinct: seen.size,\n\t\t\t\tpersons,\n\t\t\t\tteams,\n\t\t\t\tempty,\n\t\t\t\trefused: refused.top(limit),\n\t\t\t\tunknownTokens: unknownTokens.top(limit),\n\t\t\t\tunknownPairs: unknownPairs.top(limit),\n\t\t\t\torderUndecided: orderUndecided.top(1)[0] ?? {\n\t\t\t\t\tkey: \"order undecided\",\n\t\t\t\t\tcount: 0,\n\t\t\t\t\texamples: [],\n\t\t\t\t},\n\t\t\t\tlowConfidence: lowConfidence.top(limit),\n\t\t\t};\n\t\t},\n\t};\n}\n"
   ],
-  "mappings": ";;;;;;;;;;;;;AA0EA,MAAM,MAAM;AAAA,EAEkB;AAAA,EADpB,QAAQ,IAAI;AAAA,EACrB,WAAW,CAAkB,MAAc;AAAA,IAAd;AAAA;AAAA,EAC7B,GAAG,CAAC,KAAa,SAAiB,QAAQ,GAAS;AAAA,IAClD,IAAI,OAAO,KAAK,MAAM,IAAI,GAAG;AAAA,IAC7B,IAAI,SAAS;AAAA,MAAW,KAAK,MAAM,IAAI,KAAM,OAAO,EAAE,OAAO,GAAG,UAAU,CAAC,EAAE,CAAE;AAAA,IAC/E,KAAK,SAAS;AAAA,IACd,IAAI,KAAK,SAAS,SAAS,KAAK,QAAQ,CAAC,KAAK,SAAS,SAAS,OAAO,GAAG;AAAA,MACzE,KAAK,SAAS,KAAK,OAAO;AAAA,IAC3B;AAAA;AAAA,EAED,GAAG,CAAC,OAA6B;AAAA,IAChC,OAAO,CAAC,GAAG,KAAK,KAAK,EACnB,IAAI,EAAE,OAAO,OAAO,iBAAiB,EAAE,KAAK,OAAO,SAAS,EAAE,EAC9D,KAAK,CAAC,GAAG,MAAM,EAAE,QAAQ,EAAE,UAAU,EAAE,MAAM,EAAE,MAAM,KAAK,EAAE,MAAM,EAAE,MAAM,IAAI,EAAE,EAChF,MAAM,GAAG,KAAK;AAAA;AAElB;AAMA,SAAS,UAAU,CAAC,MAAc,KAAsB;AAAA,EACvD,IAAI,QAAQ;AAAA,IAAI,OAAO;AAAA,EACvB,IAAI,KAAK,QAAQ,QAAQ,EAAE,EAAE,UAAU;AAAA,IAAG,OAAO;AAAA,EACjD,OAAO,CAAC,gBAAe,IAAI;AAAA;AAIrB,SAAS,iBAAgB,CAAC,UAA6B,CAAC,GAAe;AAAA,EAC7E,MAAM,UAAU,QAAQ,WAAW,0BAAyB;AAAA,EAC5D,MAAM,UAAU,QAAQ,WAAW,SAAQ,QAAQ;AAAA,EACnD,MAAM,aAAa,QAAQ;AAAA,EAC3B,MAAM,YAAY,QAAQ,iBAAiB;AAAA,EAC3C,MAAM,OAAO,QAAQ,YAAY;AAAA,EACjC,MAAM,QAAQ,QAAQ,MAAM;AAAA,EAE5B,MAAM,OAAO,IAAI;AAAA,EACjB,IAAI,QAAQ;AAAA,EACZ,IAAI,UAAU;AAAA,EACd,IAAI,QAAQ;AAAA,EACZ,IAAI,QAAQ;AAAA,EACZ,MAAM,UAAU,IAAI,MAAM,IAAI;AAAA,EAC9B,MAAM,gBAAgB,IAAI,MAAM,IAAI;AAAA,EACpC,MAAM,eAAe,IAAI,MAAM,IAAI;AAAA,EACnC,MAAM,iBAAiB,IAAI,MAAM,IAAI;AAAA,EACrC,MAAM,gBAAgB,IAAI,MAAM,IAAI;AAAA,EAMpC,SAAS,IAAI,CAAC,MAA8B;AAAA,IAC3C,MAAM,OAAO,KAAK,QAAQ,SAAS,SAAS,KAAK,QAAQ,WAAW,KAAK;AAAA,IACzE,OAAO,KAAK,MAAM,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,MAAM,SAAS,EAAE,OAAO,CAAC,MAAM,MAAM,EAAE,CAAC;AAAA;AAAA,EAG7E,SAAS,OAAO,CAAC,OAAe,OAAkB,OAAqB;AAAA,IACtE,MAAM,OAAO,QAAQ,MAAM,OAAO,KAAK;AAAA,IACvC,IAAI,KAAK,QAAQ,SAAS,SAAS;AAAA,MAClC,SAAS;AAAA,MACT;AAAA,IACD;AAAA,IACA,IAAI,KAAK,OAAO,GAAG;AAAA,MAClB,SAAS;AAAA,MACT;AAAA,IACD;AAAA,IACA,IAAI,CAAC,KAAK,SAAS,GAAG;AAAA,MACrB,QAAQ,IACP,KAAK,QAAQ,SAAS,YAAY,KAAK,QAAQ,QAAQ,OAAO,gBAC9D,OACA,KACD;AAAA,MACA;AAAA,IACD;AAAA,IACA,WAAW;AAAA,IACX,IACC,UAAU,aACV,CAAC,MAAM,SAAS,GAAG,KACnB,iBAAiB,OAAO,KAAK,MAAM,WAClC;AAAA,MACD,eAAe,IAAI,mBAAmB,OAAO,KAAK;AAAA,IACnD;AAAA,IACA,WAAW,OAAO,KAAK,IAAI,GAAG;AAAA,MAC7B,IAAI;AAAA,MACJ,WAAW,KAAK,KAAK;AAAA,QACpB,MAAM,MAAM,WAAU,CAAC;AAAA,QACvB,MAAM,UAAU,WAAW,GAAG,GAAG,KAAK,QAAQ,aAAa,CAAC,MAAM;AAAA,QAClE,IAAI,SAAS;AAAA,UACZ,cAAc,IAAI,KAAK,OAAO,KAAK;AAAA,UACnC,IAAI,oBAAoB,WAAW;AAAA,YAClC,aAAa,IAAI,GAAG,mBAAmB,OAAO,OAAO,KAAK;AAAA,UAC3D;AAAA,QACD;AAAA,QACA,kBAAkB,UAAU,MAAM;AAAA,MACnC;AAAA,IACD;AAAA,IACA,IAAI,eAAe,WAAW;AAAA,MAC7B,MAAM,IAAI,WAAW,uBAAuB,KAAK;AAAA,MACjD,IAAI,EAAE,aAAa;AAAA,QAAW,cAAc,IAAI,EAAE,UAAU,OAAO,KAAK;AAAA,IACzE;AAAA;AAAA,EAGD,OAAO;AAAA,IACN,GAAG,CAAC,OAAO,QAAQ,WAAW;AAAA,MAC7B;AAAA,MACA,MAAM,IAAI,GAAG,YAAc;AAAA,MAC3B,KAAK,IAAI,IAAI,KAAK,IAAI,CAAC,KAAK,KAAK,CAAC;AAAA;AAAA,IAEnC,MAAM,GAAG,QAAQ,OAAO,CAAC,GAAG;AAAA,MAC3B,UAAU,QAAQ,QAAQ;AAAA,MAC1B,WAAW,KAAK,CAAC,SAAS,eAAe,cAAc,gBAAgB,aAAa,GAAG;AAAA,QACtF,EAAE,MAAM,MAAM;AAAA,MACf;AAAA,MACA,YAAY,GAAG,UAAU,MAAM;AAAA,QAC9B,MAAM,KAAK,EAAE,QAAQ,MAAQ;AAAA,QAC7B,QAAQ,EAAE,MAAM,KAAK,CAAC,GAAG,EAAE,MAAM,GAAG,EAAE,GAAgB,KAAK;AAAA,MAC5D;AAAA,MACA,OAAO;AAAA,QACN;AAAA,QACA,UAAU,KAAK;AAAA,QACf;AAAA,QACA;AAAA,QACA;AAAA,QACA,SAAS,QAAQ,IAAI,KAAK;AAAA,QAC1B,eAAe,cAAc,IAAI,KAAK;AAAA,QACtC,cAAc,aAAa,IAAI,KAAK;AAAA,QACpC,gBAAgB,eAAe,IAAI,CAAC,EAAE,MAAM;AAAA,UAC3C,KAAK;AAAA,UACL,OAAO;AAAA,UACP,UAAU,CAAC;AAAA,QACZ;AAAA,QACA,eAAe,cAAc,IAAI,KAAK;AAAA,MACvC;AAAA;AAAA,EAEF;AAAA;",
-  "debugId": "42B6EBFD51EC241764756E2164756E21",
+  "mappings": ";;;;;;;;;;;;;AA0EA,MAAM,MAAM;AAAA,EAEkB;AAAA,EADpB,QAAQ,IAAI;AAAA,EACrB,WAAW,CAAkB,MAAc;AAAA,IAAd;AAAA;AAAA,EAC7B,GAAG,CAAC,KAAa,SAAiB,QAAQ,GAAS;AAAA,IAClD,IAAI,OAAO,KAAK,MAAM,IAAI,GAAG;AAAA,IAC7B,IAAI,SAAS;AAAA,MAAW,KAAK,MAAM,IAAI,KAAM,OAAO,EAAE,OAAO,GAAG,UAAU,CAAC,EAAE,CAAE;AAAA,IAC/E,KAAK,SAAS;AAAA,IACd,IAAI,KAAK,SAAS,SAAS,KAAK,QAAQ,CAAC,KAAK,SAAS,SAAS,OAAO,GAAG;AAAA,MACzE,KAAK,SAAS,KAAK,OAAO;AAAA,IAC3B;AAAA;AAAA,EAED,GAAG,CAAC,OAA6B;AAAA,IAChC,OAAO,CAAC,GAAG,KAAK,KAAK,EACnB,IAAI,EAAE,OAAO,OAAO,iBAAiB,EAAE,KAAK,OAAO,SAAS,EAAE,EAC9D,KAAK,CAAC,GAAG,MAAM,EAAE,QAAQ,EAAE,UAAU,EAAE,MAAM,EAAE,MAAM,KAAK,EAAE,MAAM,EAAE,MAAM,IAAI,EAAE,EAChF,MAAM,GAAG,KAAK;AAAA;AAElB;AAMA,SAAS,UAAU,CAAC,MAAc,KAAsB;AAAA,EACvD,IAAI,QAAQ;AAAA,IAAI,OAAO;AAAA,EACvB,IAAI,KAAK,QAAQ,QAAQ,EAAE,EAAE,UAAU;AAAA,IAAG,OAAO;AAAA,EACjD,OAAO,CAAC,gBAAe,IAAI;AAAA;AAIrB,SAAS,iBAAgB,CAAC,UAA6B,CAAC,GAAe;AAAA,EAC7E,MAAM,UAAU,QAAQ,WAAW,0BAAyB;AAAA,EAC5D,MAAM,UAAU,QAAQ,WAAW,SAAQ,QAAQ;AAAA,EACnD,MAAM,aAAa,QAAQ;AAAA,EAC3B,MAAM,YAAY,QAAQ,iBAAiB;AAAA,EAC3C,MAAM,OAAO,QAAQ,YAAY;AAAA,EACjC,MAAM,QAAQ,QAAQ,MAAM;AAAA,EAE5B,MAAM,OAAO,IAAI;AAAA,EACjB,IAAI,QAAQ;AAAA,EACZ,IAAI,UAAU;AAAA,EACd,IAAI,QAAQ;AAAA,EACZ,IAAI,QAAQ;AAAA,EACZ,MAAM,UAAU,IAAI,MAAM,IAAI;AAAA,EAC9B,MAAM,gBAAgB,IAAI,MAAM,IAAI;AAAA,EACpC,MAAM,eAAe,IAAI,MAAM,IAAI;AAAA,EACnC,MAAM,iBAAiB,IAAI,MAAM,IAAI;AAAA,EACrC,MAAM,gBAAgB,IAAI,MAAM,IAAI;AAAA,EAMpC,SAAS,IAAI,CAAC,MAA8B;AAAA,IAC3C,MAAM,OAAO,KAAK,QAAQ,SAAS,SAAS,KAAK,QAAQ,WAAW,KAAK;AAAA,IACzE,OAAO,KAAK,MAAM,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,MAAM,SAAS,EAAE,OAAO,CAAC,MAAM,MAAM,EAAE,CAAC;AAAA;AAAA,EAG7E,SAAS,OAAO,CAAC,OAAe,OAAkB,OAAqB;AAAA,IACtE,MAAM,OAAO,QAAQ,MAAM,OAAO,KAAK;AAAA,IACvC,IAAI,KAAK,QAAQ,SAAS,SAAS;AAAA,MAClC,SAAS;AAAA,MACT;AAAA,IACD;AAAA,IACA,IAAI,KAAK,OAAO,GAAG;AAAA,MAClB,SAAS;AAAA,MACT;AAAA,IACD;AAAA,IACA,IAAI,CAAC,KAAK,SAAS,GAAG;AAAA,MACrB,QAAQ,IACP,KAAK,QAAQ,SAAS,YAAY,KAAK,QAAQ,QAAQ,OAAO,gBAC9D,OACA,KACD;AAAA,MACA;AAAA,IACD;AAAA,IACA,WAAW;AAAA,IACX,IACC,UAAU,aACV,CAAC,MAAM,SAAS,GAAG,KACnB,kBAAiB,OAAO,KAAK,MAAM,WAClC;AAAA,MACD,eAAe,IAAI,mBAAmB,OAAO,KAAK;AAAA,IACnD;AAAA,IACA,WAAW,OAAO,KAAK,IAAI,GAAG;AAAA,MAC7B,IAAI;AAAA,MACJ,WAAW,KAAK,KAAK;AAAA,QACpB,MAAM,MAAM,WAAU,CAAC;AAAA,QACvB,MAAM,UAAU,WAAW,GAAG,GAAG,KAAK,QAAQ,aAAa,CAAC,MAAM;AAAA,QAClE,IAAI,SAAS;AAAA,UACZ,cAAc,IAAI,KAAK,OAAO,KAAK;AAAA,UACnC,IAAI,oBAAoB,WAAW;AAAA,YAClC,aAAa,IAAI,GAAG,mBAAmB,OAAO,OAAO,KAAK;AAAA,UAC3D;AAAA,QACD;AAAA,QACA,kBAAkB,UAAU,MAAM;AAAA,MACnC;AAAA,IACD;AAAA,IACA,IAAI,eAAe,WAAW;AAAA,MAC7B,MAAM,IAAI,WAAW,uBAAuB,KAAK;AAAA,MACjD,IAAI,EAAE,aAAa;AAAA,QAAW,cAAc,IAAI,EAAE,UAAU,OAAO,KAAK;AAAA,IACzE;AAAA;AAAA,EAGD,OAAO;AAAA,IACN,GAAG,CAAC,OAAO,QAAQ,WAAW;AAAA,MAC7B;AAAA,MACA,MAAM,IAAI,GAAG,YAAc;AAAA,MAC3B,KAAK,IAAI,IAAI,KAAK,IAAI,CAAC,KAAK,KAAK,CAAC;AAAA;AAAA,IAEnC,MAAM,GAAG,QAAQ,OAAO,CAAC,GAAG;AAAA,MAC3B,UAAU,QAAQ,QAAQ;AAAA,MAC1B,WAAW,KAAK,CAAC,SAAS,eAAe,cAAc,gBAAgB,aAAa,GAAG;AAAA,QACtF,EAAE,MAAM,MAAM;AAAA,MACf;AAAA,MACA,YAAY,GAAG,UAAU,MAAM;AAAA,QAC9B,MAAM,KAAK,EAAE,QAAQ,MAAQ;AAAA,QAC7B,QAAQ,EAAE,MAAM,KAAK,CAAC,GAAG,EAAE,MAAM,GAAG,EAAE,GAAgB,KAAK;AAAA,MAC5D;AAAA,MACA,OAAO;AAAA,QACN;AAAA,QACA,UAAU,KAAK;AAAA,QACf;AAAA,QACA;AAAA,QACA;AAAA,QACA,SAAS,QAAQ,IAAI,KAAK;AAAA,QAC1B,eAAe,cAAc,IAAI,KAAK;AAAA,QACtC,cAAc,aAAa,IAAI,KAAK;AAAA,QACpC,gBAAgB,eAAe,IAAI,CAAC,EAAE,MAAM;AAAA,UAC3C,KAAK;AAAA,UACL,OAAO;AAAA,UACP,UAAU,CAAC;AAAA,QACZ;AAAA,QACA,eAAe,cAAc,IAAI,KAAK;AAAA,MACvC;AAAA;AAAA,EAEF;AAAA;",
+  "debugId": "C23DFDD520B2D39C64756E2164756E21",
   "names": []
 }

package/dist/chunks/index-18z8akw8.js

@@ -0,0 +1,42 @@
+import {
+  isGiven2,
+  isAmbiguous2,
+  isPlace2,
+  isTradename2
+} from "./index-0wpvbdgh.js";
+
+// src/lexicon/place.ts
+var JOINING_COMMA = /(?<=[^\s,]),(?=[^\s,])/g;
+function splitWords(s) {
+  return s.replace(JOINING_COMMA, (comma, at) => /\d/.test(s[at - 1]) && /\d/.test(s[at + 1]) ? comma : ", ").split(/\s+/).filter(Boolean);
+}
+var FIELD_END = /[,-]+$/;
+function isPlaceWordFlags(flags) {
+  return isPlace2(flags) && !isGiven2(flags) && !isAmbiguous2(flags) && !isTradename2(flags);
+}
+function isPlacePhrase(phrase, getFlags) {
+  const words = splitWords(phrase).map((w) => w.replace(FIELD_END, "")).filter((w) => w !== "");
+  if (words.length < 2)
+    return false;
+  return words.every((w) => isPlaceWordFlags(getFlags(w) ?? ""));
+}
+function isCommaPlacePhrase(phrase, getFlags) {
+  const printed = splitWords(phrase);
+  let cut = -1;
+  for (let i = 0;i < printed.length - 1; i++)
+    if (printed[i].endsWith(","))
+      cut = i;
+  if (cut < 0)
+    return false;
+  const bare = (ws) => ws.map((w) => w.replace(FIELD_END, "")).filter((w) => w !== "");
+  const words = bare(printed);
+  const after = bare(printed.slice(cut + 1));
+  if (words.length < 2 || after.length === 0)
+    return false;
+  return words.every((w) => isPlace2(getFlags(w) ?? "")) && after.every((w) => isPlaceWordFlags(getFlags(w) ?? ""));
+}
+
+export { splitWords, FIELD_END, isPlaceWordFlags, isPlacePhrase, isCommaPlacePhrase };
+
+//# debugId=56C39740150DE0D964756E2164756E21
+//# sourceMappingURL=index-18z8akw8.js.map

package/dist/chunks/{index-tqdbxd92.js.map → index-18z8akw8.js.map}

@@ -2,9 +2,9 @@
   "version": 3,
   "sources": ["../src/lexicon/place.ts"],
   "sourcesContent": [
-    "/**\n * Printed words, and which of them name a place, read off the lexicon's flags alone.\n *\n * Moved here from the classifier so that `/normalize` can ask the same question the\n * classifier asks (is this field a place?) without importing the bloom tables. The\n * classifier re-exports both functions from their old modules.\n */\n\nimport { isAmbiguous, isGiven, isPlace, isTradename } from \"./flags.ts\";\n\n/** A comma with a printed character, not a space or another comma, on each side. */\nconst JOINING_COMMA = /(?<=[^\\s,]),(?=[^\\s,])/g;\n\n/**\n * The printed words of a string, as the classifier, the segmenter and their trainers\n * all read them: split on whitespace, and at a comma that joins two words with no\n * space after it, which separates them as `, ` does (`MUNSTER,EIRE` is `MUNSTER,`\n * `EIRE`; `Quorrish,Zelvine` is `Quorrish,` `Zelvine`). The comma stays on the word before it,\n * where a reader that cares (the classifier's phrase joining) sees the field end.\n * A comma between two digits (`1,000`) is part of the number.\n */\nexport function splitWords(s: string): string[] {\n\treturn s\n\t\t.replace(JOINING_COMMA, (comma, at: number) =>\n\t\t\t/\\d/.test(s[at - 1]!) && /\\d/.test(s[at + 1]!) ? comma : \", \",\n\t\t)\n\t\t.split(/\\s+/)\n\t\t.filter(Boolean);\n}\n\n/** A printed word's trailing `,` or `-`: it ends a field, and no phrase continues through it. */\nexport const FIELD_END = /[,-]+$/;\n\n/**\n * One word's flags read as a place: a place (`G`) and not a given name (`F`, `M`, `N`),\n * an ambiguous name (`A`) or a school's name (`T`). An unknown word (`\"\"`, which\n * `classesOf` reads as `F`) is never a place.\n */\nexport function isPlaceWordFlags(flags: string): boolean {\n\treturn isPlace(flags) && !isGiven(flags) && !isAmbiguous(flags) && !isTradename(flags);\n}\n\n/**\n * A place phrase: two or more printed words (a field-ending `,` or `-` set aside), every\n * one a place (`G`) and none a given name (`F`, `M`, `N`, `A`) or a school's name (`T`).\n * `Varrow Pellinor` (`Varrow:SLG Pellinor:G`) is one; `Lirabel Pellinor` (`Lirabel:FG`, a\n * given name) and `Corrandel Pellinor` (`Corrandel:TG`, a school's name) are not. A surname that is also a place\n * does not make a person: a person's name has a given name. The runtime answers a place\n * phrase `location` without the regression, as it answers one word flagged G.\n */\nexport function isPlacePhrase(phrase: string, getFlags: (name: string) => string | null): boolean {\n\tconst words = splitWords(phrase)\n\t\t.map((w) => w.replace(FIELD_END, \"\"))\n\t\t.filter((w) => w !== \"\");\n\tif (words.length < 2) return false;\n\treturn words.every((w) => isPlaceWordFlags(getFlags(w) ?? \"\"));\n}\n"
+    "/**\n * Printed words, and which of them name a place, read off the lexicon's flags alone.\n *\n * Moved here from the classifier so that `/normalize` can ask the same question the\n * classifier asks (is this field a place?) without importing the bloom tables. The\n * classifier re-exports both functions from their old modules.\n */\n\nimport { isAmbiguous, isGiven, isPlace, isTradename } from \"./flags.ts\";\n\n/** A comma with a printed character, not a space or another comma, on each side. */\nconst JOINING_COMMA = /(?<=[^\\s,]),(?=[^\\s,])/g;\n\n/**\n * The printed words of a string, as the classifier, the segmenter and their trainers\n * all read them: split on whitespace, and at a comma that joins two words with no\n * space after it, which separates them as `, ` does (`MUNSTER,EIRE` is `MUNSTER,`\n * `EIRE`; `Quorrish,Zelvine` is `Quorrish,` `Zelvine`). The comma stays on the word before it,\n * where a reader that cares (the classifier's phrase joining) sees the field end.\n * A comma between two digits (`1,000`) is part of the number.\n */\nexport function splitWords(s: string): string[] {\n\treturn s\n\t\t.replace(JOINING_COMMA, (comma, at: number) =>\n\t\t\t/\\d/.test(s[at - 1]!) && /\\d/.test(s[at + 1]!) ? comma : \", \",\n\t\t)\n\t\t.split(/\\s+/)\n\t\t.filter(Boolean);\n}\n\n/** A printed word's trailing `,` or `-`: it ends a field, and no phrase continues through it. */\nexport const FIELD_END = /[,-]+$/;\n\n/**\n * One word's flags read as a place: a place (`G`) and not a given name (`F`, `M`, `N`),\n * an ambiguous name (`A`) or a school's name (`T`). An unknown word (`\"\"`, which\n * `classesOf` reads as `F`) is never a place.\n */\nexport function isPlaceWordFlags(flags: string): boolean {\n\treturn isPlace(flags) && !isGiven(flags) && !isAmbiguous(flags) && !isTradename(flags);\n}\n\n/**\n * A place phrase: two or more printed words (a field-ending `,` or `-` set aside), every\n * one a place (`G`) and none a given name (`F`, `M`, `N`, `A`) or a school's name (`T`).\n * `Varrow Pellinor` (`Varrow:SLG Pellinor:G`) is one; `Lirabel Pellinor` (`Lirabel:FG`, a\n * given name) and `Corrandel Pellinor` (`Corrandel:TG`, a school's name) are not. A surname that is also a place\n * does not make a person: a person's name has a given name. The runtime answers a place\n * phrase `location` without the regression, as it answers one word flagged G.\n */\nexport function isPlacePhrase(phrase: string, getFlags: (name: string) => string | null): boolean {\n\tconst words = splitWords(phrase)\n\t\t.map((w) => w.replace(FIELD_END, \"\"))\n\t\t.filter((w) => w !== \"\");\n\tif (words.length < 2) return false;\n\treturn words.every((w) => isPlaceWordFlags(getFlags(w) ?? \"\"));\n}\n\n/**\n * A place cut by a comma: two or more printed words, every one a place (`G`), at least one\n * before the last ending its field with `,`, and every word after the last such comma a\n * place and nothing else (`isPlaceWordFlags`). The words before the comma may also be a\n * given name or a school's name: `Victoria, Australia` (`Victoria:FG`, `Australia:G`) and\n * `Sydney, Australia` (`Sydney:FGT`) are ones; `Victoria Australia` (no comma) and\n * `Australia, Victoria` (a given name after the comma) are not.\n *\n * The comma is the signal only where the document prints given names first. A person is\n * printed `Surname, Given`, and under a surname-first or an undeclared order a given name\n * that is also a place may follow the comma; a school is printed `School, State`\n * (`Claddagh, Utah`) in surname-first documents. So the runtime answers this phrase\n * `location` only for a document that declares `given-first`; `isPlacePhrase` answers\n * under every order.\n */\nexport function isCommaPlacePhrase(\n\tphrase: string,\n\tgetFlags: (name: string) => string | null,\n): boolean {\n\tconst printed = splitWords(phrase);\n\tlet cut = -1;\n\tfor (let i = 0; i < printed.length - 1; i++) if (printed[i]!.endsWith(\",\")) cut = i;\n\tif (cut < 0) return false;\n\tconst bare = (ws: string[]) => ws.map((w) => w.replace(FIELD_END, \"\")).filter((w) => w !== \"\");\n\tconst words = bare(printed);\n\tconst after = bare(printed.slice(cut + 1));\n\tif (words.length < 2 || after.length === 0) return false;\n\treturn (\n\t\twords.every((w) => isPlace(getFlags(w) ?? \"\")) &&\n\t\tafter.every((w) => isPlaceWordFlags(getFlags(w) ?? \"\"))\n\t);\n}\n"
   ],
-  "mappings": ";;;;;;;;AAWA,IAAM,gBAAgB;AAUf,SAAS,UAAU,CAAC,GAAqB;AAAA,EAC/C,OAAO,EACL,QAAQ,eAAe,CAAC,OAAO,OAC/B,KAAK,KAAK,EAAE,KAAK,EAAG,KAAK,KAAK,KAAK,EAAE,KAAK,EAAG,IAAI,QAAQ,IAC1D,EACC,MAAM,KAAK,EACX,OAAO,OAAO;AAAA;AAIV,IAAM,YAAY;AAOlB,SAAS,gBAAgB,CAAC,OAAwB;AAAA,EACxD,OAAO,SAAQ,KAAK,KAAK,CAAC,SAAQ,KAAK,KAAK,CAAC,aAAY,KAAK,KAAK,CAAC,aAAY,KAAK;AAAA;AAW/E,SAAS,aAAa,CAAC,QAAgB,UAAoD;AAAA,EACjG,MAAM,QAAQ,WAAW,MAAM,EAC7B,IAAI,CAAC,MAAM,EAAE,QAAQ,WAAW,EAAE,CAAC,EACnC,OAAO,CAAC,MAAM,MAAM,EAAE;AAAA,EACxB,IAAI,MAAM,SAAS;AAAA,IAAG,OAAO;AAAA,EAC7B,OAAO,MAAM,MAAM,CAAC,MAAM,iBAAiB,SAAS,CAAC,KAAK,EAAE,CAAC;AAAA;",
-  "debugId": "A9815F8EF12D8D4964756E2164756E21",
+  "mappings": ";;;;;;;;AAWA,IAAM,gBAAgB;AAUf,SAAS,UAAU,CAAC,GAAqB;AAAA,EAC/C,OAAO,EACL,QAAQ,eAAe,CAAC,OAAO,OAC/B,KAAK,KAAK,EAAE,KAAK,EAAG,KAAK,KAAK,KAAK,EAAE,KAAK,EAAG,IAAI,QAAQ,IAC1D,EACC,MAAM,KAAK,EACX,OAAO,OAAO;AAAA;AAIV,IAAM,YAAY;AAOlB,SAAS,gBAAgB,CAAC,OAAwB;AAAA,EACxD,OAAO,SAAQ,KAAK,KAAK,CAAC,SAAQ,KAAK,KAAK,CAAC,aAAY,KAAK,KAAK,CAAC,aAAY,KAAK;AAAA;AAW/E,SAAS,aAAa,CAAC,QAAgB,UAAoD;AAAA,EACjG,MAAM,QAAQ,WAAW,MAAM,EAC7B,IAAI,CAAC,MAAM,EAAE,QAAQ,WAAW,EAAE,CAAC,EACnC,OAAO,CAAC,MAAM,MAAM,EAAE;AAAA,EACxB,IAAI,MAAM,SAAS;AAAA,IAAG,OAAO;AAAA,EAC7B,OAAO,MAAM,MAAM,CAAC,MAAM,iBAAiB,SAAS,CAAC,KAAK,EAAE,CAAC;AAAA;AAkBvD,SAAS,kBAAkB,CACjC,QACA,UACU;AAAA,EACV,MAAM,UAAU,WAAW,MAAM;AAAA,EACjC,IAAI,MAAM;AAAA,EACV,SAAS,IAAI,EAAG,IAAI,QAAQ,SAAS,GAAG;AAAA,IAAK,IAAI,QAAQ,GAAI,SAAS,GAAG;AAAA,MAAG,MAAM;AAAA,EAClF,IAAI,MAAM;AAAA,IAAG,OAAO;AAAA,EACpB,MAAM,OAAO,CAAC,OAAiB,GAAG,IAAI,CAAC,MAAM,EAAE,QAAQ,WAAW,EAAE,CAAC,EAAE,OAAO,CAAC,MAAM,MAAM,EAAE;AAAA,EAC7F,MAAM,QAAQ,KAAK,OAAO;AAAA,EAC1B,MAAM,QAAQ,KAAK,QAAQ,MAAM,MAAM,CAAC,CAAC;AAAA,EACzC,IAAI,MAAM,SAAS,KAAK,MAAM,WAAW;AAAA,IAAG,OAAO;AAAA,EACnD,OACC,MAAM,MAAM,CAAC,MAAM,SAAQ,SAAS,CAAC,KAAK,EAAE,CAAC,KAC7C,MAAM,MAAM,CAAC,MAAM,iBAAiB,SAAS,CAAC,KAAK,EAAE,CAAC;AAAA;",
+  "debugId": "56C39740150DE0D964756E2164756E21",
   "names": []
 }

package/dist/chunks/{index-ky754nsf.js → index-7j32khen.js}

@@ -6,8 +6,9 @@ import {
 import {
   splitWords,
   FIELD_END,
-  isPlacePhrase
-} from "./index-tqdbxd92.js";
+  isPlacePhrase,
+  isCommaPlacePhrase
+} from "./index-18z8akw8.js";
 import {
   TEAM_DESIGNATION_WORDS2,
   partitionTeamSuffix2,
@@ -2744,10 +2745,11 @@ function phraseClassifier2(names = Lexicon2.default(), options = {}) {
   const model = (order = "unknown") => byOrder[order] ?? classifier;
   const segmenter = createSegmenter2(namesData, options.weights?.segmenter);
   const PLACE = { person: 0.05, school: 0.05, location: 0.9 };
-  function placeOverride(phrase) {
+  function placeOverride(phrase, order = "unknown") {
     const trimmed = phrase.trim();
+    const flagsOf = (n) => names.getNameFlags(n);
     if (splitWords(trimmed).length !== 1)
-      return isPlacePhrase(trimmed, (n) => names.getNameFlags(n)) ? { ...PLACE } : null;
+      return isPlacePhrase(trimmed, flagsOf) || order === "given-first" && isCommaPlacePhrase(trimmed, flagsOf) ? { ...PLACE } : null;
     const flags = names.getNameFlags(trimmed);
     if (!flags)
       return null;
@@ -2759,7 +2761,7 @@ function phraseClassifier2(names = Lexicon2.default(), options = {}) {
     if (isGarbage(phrase))
       return { ...ZERO };
     const s = subject(phrase);
-    return placeOverride(s) ?? model(order).classifyName(s);
+    return placeOverride(s, order) ?? model(order).classifyName(s);
   }
   function category(scores) {
     const max = Math.max(scores.person, scores.school, scores.location);
@@ -2797,5 +2799,5 @@ function phraseClassifier2(names = Lexicon2.default(), options = {}) {
 
 export { CONNECTORS2, SCHOOL_KEYWORDS2, isStopword2, createSegmenter2, createClassifier2, phraseClassifier2 };
 
-//# debugId=C5025FBA3439B68D64756E2164756E21
-//# sourceMappingURL=index-ky754nsf.js.map
+//# debugId=DEAAEE822B81CADB64756E2164756E21
+//# sourceMappingURL=index-7j32khen.js.map