@kensio/pinyinjs 1.2.1 → 1.3.0

package/dist/decode/decode.d.ts

@@ -27,8 +27,14 @@ export declare function decodeSpacing(lattice: Lattice): readonly number[];
  * Stages 1 to 3 of ALGORITHM.md end to end — build every candidate reading,
  * lock the positions that cannot vary, score the rest — where the greedy
  * baseline commits to the longest match at each position and never revisits it.
+ *
+ * `before` is 汉字 standing in front of the run that the decode should see and
+ * not report: the 汉字 a number written in digits would have been written with.
+ * A run is decoded on its own without it, and a run reached that way has no idea
+ * what preceded it — 个人 alone is the word `gèrén`, while the 两 of 2个人 makes
+ * 两个 the word and leaves 人 to itself.
  */
-export declare function decodeRun(dictionary: Dictionary, run: string, rules?: readonly EdgeRule[]): readonly DecodedWord[];
+export declare function decodeRun(dictionary: Dictionary, run: string, rules?: readonly EdgeRule[], before?: string): readonly DecodedWord[];
 /**
  * Decode a Han run, keeping what each reading was chosen over.
  *
@@ -36,6 +42,10 @@ export declare function decodeRun(dictionary: Dictionary, run: string, rules?: r
  * candidates kept rather than discarded. Separate because the extra sweep is
  * only worth running for a caller that will use the answer — rendering
  * uncertain readings differently, or reporting them.
+ *
+ * `before` is the context {@link decodeRun} takes, and means the same here: it
+ * is decoded with the run and reported with neither words nor confidence of its
+ * own.
  */
-export declare function decodeRunScored(dictionary: Dictionary, run: string, rules?: readonly EdgeRule[]): readonly ScoredWord[];
+export declare function decodeRunScored(dictionary: Dictionary, run: string, rules?: readonly EdgeRule[], before?: string): readonly ScoredWord[];
 //# sourceMappingURL=decode.d.ts.map

package/dist/decode/decode.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"decode.d.ts","sourceRoot":"","sources":["../../src/decode/decode.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAE9D,OAAO,EAA2B,KAAK,OAAO,EAAE,MAAM,cAAc,CAAC;AACrE,OAAO,EAGL,KAAK,iBAAiB,EACtB,KAAK,WAAW,EAGjB,MAAM,cAAc,CAAC;AAEtB,OAAO,EAAkB,KAAK,QAAQ,EAAE,MAAM,YAAY,CAAC;AAE3D,OAAO,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AAEzD;;;;;;;GAOG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,OAAO,EAChB,UAAU,EAAE,iBAAiB,GAC5B,SAAS,WAAW,EAAE,CAmBxB;AAED;;;;;;;GAOG;AACH,wBAAgB,aAAa,CAAC,OAAO,EAAE,OAAO,GAAG,SAAS,MAAM,EAAE,CAIjE;AA2ED;;;;;;GAMG;AACH,wBAAgB,SAAS,CACvB,UAAU,EAAE,UAAU,EACtB,GAAG,EAAE,MAAM,EACX,KAAK,GAAE,SAAS,QAAQ,EAAkB,GACzC,SAAS,WAAW,EAAE,CAMxB;AAED;;;;;;;GAOG;AACH,wBAAgB,eAAe,CAC7B,UAAU,EAAE,UAAU,EACtB,GAAG,EAAE,MAAM,EACX,KAAK,GAAE,SAAS,QAAQ,EAAkB,GACzC,SAAS,UAAU,EAAE,CAavB"}
+{"version":3,"file":"decode.d.ts","sourceRoot":"","sources":["../../src/decode/decode.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAG9D,OAAO,EAAqC,KAAK,OAAO,EAAE,MAAM,cAAc,CAAC;AAC/E,OAAO,EAGL,KAAK,iBAAiB,EACtB,KAAK,WAAW,EAGjB,MAAM,cAAc,CAAC;AAEtB,OAAO,EAAkB,KAAK,QAAQ,EAAE,MAAM,YAAY,CAAC;AAE3D,OAAO,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AAEzD;;;;;;;GAOG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,OAAO,EAChB,UAAU,EAAE,iBAAiB,GAC5B,SAAS,WAAW,EAAE,CAmBxB;AAED;;;;;;;GAOG;AACH,wBAAgB,aAAa,CAAC,OAAO,EAAE,OAAO,GAAG,SAAS,MAAM,EAAE,CAIjE;AAgJD;;;;;;;;;;;;GAYG;AACH,wBAAgB,SAAS,CACvB,UAAU,EAAE,UAAU,EACtB,GAAG,EAAE,MAAM,EACX,KAAK,GAAE,SAAS,QAAQ,EAAkB,EAC1C,MAAM,SAAK,GACV,SAAS,WAAW,EAAE,CAOxB;AAED;;;;;;;;;;;GAWG;AACH,wBAAgB,eAAe,CAC7B,UAAU,EAAE,UAAU,EACtB,GAAG,EAAE,MAAM,EACX,KAAK,GAAE,SAAS,QAAQ,EAAkB,EAC1C,MAAM,SAAK,GACV,SAAS,UAAU,EAAE,CAcvB"}

package/dist/decode/decode.js

@@ -1,5 +1,6 @@
+import { toCharacters } from "../script/characters.js";
 import { scoreReadings } from "./confidence.js";
-import { buildLattice, cutPoints } from "./lattice.js";
+import { allEdges, buildLattice, cutPoints } from "./lattice.js";
 import { isSettled, projectReadings, settledUnits, unitsOf, } from "./locking.js";
 import { READING_RULES } from "./reading-rules.js";
 import { applyEdgeRules } from "./rules.js";
@@ -97,17 +98,68 @@ function wordFrom(dictionary, lattice, group) {
 function ruledLattice(dictionary, run, rules) {
     return applyEdgeRules(buildLattice(dictionary, run), dictionary, rules);
 }
+/**
+ * Whether any reading in the lattice would hold the two sides of a join
+ * together.
+ *
+ * A reading of one syllable per character can be cut anywhere; anything else —
+ * 儿化, a character with no reading — is a single claim about its whole span,
+ * so a claim spanning the join cannot be reported for the run alone. 1点儿事
+ * is that case: 一点儿 is one `yìdiǎnr` over the 一 the context supplied and
+ * the 点儿 the run did.
+ */
+function isJoinedAt(lattice, at) {
+    return allEdges(lattice).some((edge) => edge.from < at &&
+        edge.to > at &&
+        edge.reading.length !== edge.to - edge.from);
+}
+/**
+ * The lattice a run decodes over, with whatever context stands in front of it.
+ *
+ * The context is dropped where a reading would hold it to the run, leaving the
+ * run decoded on its own, which is what it would have been before there was any
+ * context to give.
+ */
+function runLattice(dictionary, run, rules, before) {
+    const alone = () => ({
+        lattice: ruledLattice(dictionary, run, rules),
+        at: 0,
+    });
+    if (before === "") {
+        return alone();
+    }
+    const lattice = ruledLattice(dictionary, before + run, rules);
+    const at = toCharacters(before).length;
+    return isJoinedAt(lattice, at) ? alone() : { lattice, at };
+}
+/**
+ * Cut a decode of the whole lattice down to the run at the end of it.
+ *
+ * Everything before `at` was context rather than run — 汉字 a caller put in
+ * front so that the decode could see them — and is dropped once it has done its
+ * work.
+ */
+function runGroups({ lattice, at }, units) {
+    return groupUnits(units.filter((unit) => unit.from >= at), decodeSpacing(lattice).filter((from) => from >= at));
+}
 /**
  * Decode a Han run with the lattice: the recommended path.
  *
  * Stages 1 to 3 of ALGORITHM.md end to end — build every candidate reading,
  * lock the positions that cannot vary, score the rest — where the greedy
  * baseline commits to the longest match at each position and never revisits it.
+ *
+ * `before` is 汉字 standing in front of the run that the decode should see and
+ * not report: the 汉字 a number written in digits would have been written with.
+ * A run is decoded on its own without it, and a run reached that way has no idea
+ * what preceded it — 个人 alone is the word `gèrén`, while the 两 of 2个人 makes
+ * 两个 the word and leaves 人 to itself.
  */
-export function decodeRun(dictionary, run, rules = READING_RULES) {
-    const lattice = ruledLattice(dictionary, run, rules);
+export function decodeRun(dictionary, run, rules = READING_RULES, before = "") {
+    const held = runLattice(dictionary, run, rules, before);
+    const { lattice } = held;
     const units = decodeReadings(lattice, projectReadings(lattice));
-    return groupUnits(units, decodeSpacing(lattice)).map((group) => wordFrom(dictionary, lattice, group));
+    return runGroups(held, units).map((group) => wordFrom(dictionary, lattice, group));
 }
 /**
  * Decode a Han run, keeping what each reading was chosen over.
@@ -116,11 +168,16 @@ export function decodeRun(dictionary, run, rules = READING_RULES) {
  * candidates kept rather than discarded. Separate because the extra sweep is
  * only worth running for a caller that will use the answer — rendering
  * uncertain readings differently, or reporting them.
+ *
+ * `before` is the context {@link decodeRun} takes, and means the same here: it
+ * is decoded with the run and reported with neither words nor confidence of its
+ * own.
  */
-export function decodeRunScored(dictionary, run, rules = READING_RULES) {
-    const lattice = ruledLattice(dictionary, run, rules);
+export function decodeRunScored(dictionary, run, rules = READING_RULES, before = "") {
+    const held = runLattice(dictionary, run, rules, before);
+    const { lattice } = held;
     const units = scoreReadings(lattice, projectReadings(lattice));
-    return groupUnits(units, decodeSpacing(lattice)).map((group) => ({
+    return runGroups(held, units).map((group) => ({
         word: wordFrom(dictionary, lattice, group),
         confidence: group.flatMap((unit) => unit.reading.map(() => ({
             isLocked: unit.isLocked,

package/dist/decode/decode.js.map

@@ -1 +1 @@
-{"version":3,"file":"decode.js","sourceRoot":"","sources":["../../src/decode/decode.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,aAAa,EAAmB,MAAM,iBAAiB,CAAC;AACjE,OAAO,EAAE,YAAY,EAAE,SAAS,EAAgB,MAAM,cAAc,CAAC;AACrE,OAAO,EACL,SAAS,EACT,eAAe,EAGf,YAAY,EACZ,OAAO,GACR,MAAM,cAAc,CAAC;AACtB,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACnD,OAAO,EAAE,cAAc,EAAiB,MAAM,YAAY,CAAC;AAC3D,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAGtE;;;;;;;GAOG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAgB,EAChB,UAA6B;IAE7B,MAAM,IAAI,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;IAChC,MAAM,KAAK,GAAkB,EAAE,CAAC;IAEhC,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;QAC3C,MAAM,EAAE,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QAC3B,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;YACrB,MAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,CAAC,EAAE,CAAC;YACpC,KAAK,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC;YAClD,SAAS;QACX,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,YAAY,CAAC,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,CAAC,EAAE,CAAC;YAChE,KAAK,CAAC,IAAI,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;QAC/B,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,aAAa,CAAC,OAAgB;IAC5C,OAAO,YAAY,CAAC,OAAO,EAAE,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,WAAW,CAAC,CAAC,GAAG,CACzE,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CACpB,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,UAAU,CACjB,KAAsB,EACtB,UAA6B;IAE7B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAChE,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,IAAI,GAAW,EAAE,CAAC;IAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC3C,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAClB,IAAI,GAAG,EAAE,CAAC;QACZ,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAClB,CAAC;IACD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpB,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACpB,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,SAAS,QAAQ,CACf,UAAsB,EACtB,OAAgB,EAChB,KAA6B;IAE7B,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACvB,MAAM,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1B,wEAAwE;IACxE,IAAI,KAAK,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;QAC9C,MAAM,IAAI,KAAK,CAAC,6CAA6C,CAAC,CAAC;IACjE,CAAC;IACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACpE,MAAM,KAAK,GAAG,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAEtC,OAAO;QACL,IAAI;QACJ,OAAO,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC;QACnD,YAAY,EAAE,KAAK,EAAE,YAAY,IAAI,KAAK;QAC1C,YAAY,EAAE,KAAK,EAAE,YAAY,IAAI,EAAE;QACvC,OAAO,EAAE,KAAK,KAAK,SAAS;KAC7B,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,YAAY,CACnB,UAAsB,EACtB,GAAW,EACX,KAA0B;IAE1B,OAAO,cAAc,CAAC,YAAY,CAAC,UAAU,EAAE,GAAG,CAAC,EAAE,UAAU,EAAE,KAAK,CAAC,CAAC;AAC1E,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,SAAS,CACvB,UAAsB,EACtB,GAAW,EACX,KAAK,GAAwB,aAAa;IAE1C,MAAM,OAAO,GAAG,YAAY,CAAC,UAAU,EAAE,GAAG,EAAE,KAAK,CAAC,CAAC;IACrD,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,eAAe,CAAC,OAAO,CAAC,CAAC,CAAC;IAChE,OAAO,UAAU,CAAC,KAAK,EAAE,aAAa,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAC7D,QAAQ,CAAC,UAAU,EAAE,OAAO,EAAE,KAAK,CAAC,CACrC,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,eAAe,CAC7B,UAAsB,EACtB,GAAW,EACX,KAAK,GAAwB,aAAa;IAE1C,MAAM,OAAO,GAAG,YAAY,CAAC,UAAU,EAAE,GAAG,EAAE,KAAK,CAAC,CAAC;IACrD,MAAM,KAAK,GAAG,aAAa,CAAC,OAAO,EAAE,eAAe,CAAC,OAAO,CAAC,CAAC,CAAC;IAE/D,OAAO,UAAU,CAAa,KAAK,EAAE,aAAa,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC3E,IAAI,EAAE,QAAQ,CAAC,UAAU,EAAE,OAAO,EAAE,KAAK,CAAC;QAC1C,UAAU,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CACjC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC;YACtB,QAAQ,EAAE,IAAI,CAAC,QAAQ;YACvB,YAAY,EAAE,IAAI,CAAC,YAAY;SAChC,CAAC,CAAC,CACJ;KACF,CAAC,CAAC,CAAC;AACN,CAAC"}
+{"version":3,"file":"decode.js","sourceRoot":"","sources":["../../src/decode/decode.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AACvD,OAAO,EAAE,aAAa,EAAmB,MAAM,iBAAiB,CAAC;AACjE,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,SAAS,EAAgB,MAAM,cAAc,CAAC;AAC/E,OAAO,EACL,SAAS,EACT,eAAe,EAGf,YAAY,EACZ,OAAO,GACR,MAAM,cAAc,CAAC;AACtB,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACnD,OAAO,EAAE,cAAc,EAAiB,MAAM,YAAY,CAAC;AAC3D,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAGtE;;;;;;;GAOG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAgB,EAChB,UAA6B;IAE7B,MAAM,IAAI,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;IAChC,MAAM,KAAK,GAAkB,EAAE,CAAC;IAEhC,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;QAC3C,MAAM,EAAE,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QAC3B,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;YACrB,MAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,CAAC,EAAE,CAAC;YACpC,KAAK,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC;YAClD,SAAS;QACX,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,YAAY,CAAC,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,CAAC,EAAE,CAAC;YAChE,KAAK,CAAC,IAAI,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;QAC/B,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,aAAa,CAAC,OAAgB;IAC5C,OAAO,YAAY,CAAC,OAAO,EAAE,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,WAAW,CAAC,CAAC,GAAG,CACzE,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CACpB,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,UAAU,CACjB,KAAsB,EACtB,UAA6B;IAE7B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAChE,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,IAAI,GAAW,EAAE,CAAC;IAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC3C,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAClB,IAAI,GAAG,EAAE,CAAC;QACZ,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAClB,CAAC;IACD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpB,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACpB,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,SAAS,QAAQ,CACf,UAAsB,EACtB,OAAgB,EAChB,KAA6B;IAE7B,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACvB,MAAM,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1B,wEAAwE;IACxE,IAAI,KAAK,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;QAC9C,MAAM,IAAI,KAAK,CAAC,6CAA6C,CAAC,CAAC;IACjE,CAAC;IACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACpE,MAAM,KAAK,GAAG,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAEtC,OAAO;QACL,IAAI;QACJ,OAAO,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC;QACnD,YAAY,EAAE,KAAK,EAAE,YAAY,IAAI,KAAK;QAC1C,YAAY,EAAE,KAAK,EAAE,YAAY,IAAI,EAAE;QACvC,OAAO,EAAE,KAAK,KAAK,SAAS;KAC7B,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,YAAY,CACnB,UAAsB,EACtB,GAAW,EACX,KAA0B;IAE1B,OAAO,cAAc,CAAC,YAAY,CAAC,UAAU,EAAE,GAAG,CAAC,EAAE,UAAU,EAAE,KAAK,CAAC,CAAC;AAC1E,CAAC;AAUD;;;;;;;;;GASG;AACH,SAAS,UAAU,CAAC,OAAgB,EAAE,EAAU;IAC9C,OAAO,QAAQ,CAAC,OAAO,CAAC,CAAC,IAAI,CAC3B,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,IAAI,GAAG,EAAE;QACd,IAAI,CAAC,EAAE,GAAG,EAAE;QACZ,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,CAC9C,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,SAAS,UAAU,CACjB,UAAsB,EACtB,GAAW,EACX,KAA0B,EAC1B,MAAc;IAEd,MAAM,KAAK,GAAG,GAAe,EAAE,CAAC,CAAC;QAC/B,OAAO,EAAE,YAAY,CAAC,UAAU,EAAE,GAAG,EAAE,KAAK,CAAC;QAC7C,EAAE,EAAE,CAAC;KACN,CAAC,CAAC;IACH,IAAI,MAAM,KAAK,EAAE,EAAE,CAAC;QAClB,OAAO,KAAK,EAAE,CAAC;IACjB,CAAC;IACD,MAAM,OAAO,GAAG,YAAY,CAAC,UAAU,EAAE,MAAM,GAAG,GAAG,EAAE,KAAK,CAAC,CAAC;IAC9D,MAAM,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IACvC,OAAO,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC;AAC7D,CAAC;AAED;;;;;;GAMG;AACH,SAAS,SAAS,CAChB,EAAE,OAAO,EAAE,EAAE,EAAc,EAC3B,KAAsB;IAEtB,OAAO,UAAU,CACf,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,IAAI,EAAE,CAAC,EACvC,aAAa,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC,CACpD,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,SAAS,CACvB,UAAsB,EACtB,GAAW,EACX,KAAK,GAAwB,aAAa,EAC1C,MAAM,GAAG,EAAE;IAEX,MAAM,IAAI,GAAG,UAAU,CAAC,UAAU,EAAE,GAAG,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC;IACxD,MAAM,EAAE,OAAO,EAAE,GAAG,IAAI,CAAC;IACzB,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,eAAe,CAAC,OAAO,CAAC,CAAC,CAAC;IAChE,OAAO,SAAS,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAC1C,QAAQ,CAAC,UAAU,EAAE,OAAO,EAAE,KAAK,CAAC,CACrC,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,eAAe,CAC7B,UAAsB,EACtB,GAAW,EACX,KAAK,GAAwB,aAAa,EAC1C,MAAM,GAAG,EAAE;IAEX,MAAM,IAAI,GAAG,UAAU,CAAC,UAAU,EAAE,GAAG,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC;IACxD,MAAM,EAAE,OAAO,EAAE,GAAG,IAAI,CAAC;IACzB,MAAM,KAAK,GAAG,aAAa,CAAC,OAAO,EAAE,eAAe,CAAC,OAAO,CAAC,CAAC,CAAC;IAE/D,OAAO,SAAS,CAAa,IAAI,EAAE,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QACxD,IAAI,EAAE,QAAQ,CAAC,UAAU,EAAE,OAAO,EAAE,KAAK,CAAC;QAC1C,UAAU,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CACjC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC;YACtB,QAAQ,EAAE,IAAI,CAAC,QAAQ;YACvB,YAAY,EAAE,IAAI,CAAC,YAAY;SAChC,CAAC,CAAC,CACJ;KACF,CAAC,CAAC,CAAC;AACN,CAAC"}

package/dist/decode/reading-rules.d.ts

@@ -49,6 +49,63 @@ export declare const MODAL_DE: EdgeRule;
  * 门铃儿 and a row of Mongol names transliterated with 儿.
  */
 export declare const ATTESTED_ERHUA: EdgeRule;
+/**
+ * 教 read as `jiāo` where it is teaching rather than a religion.
+ *
+ * The dictionary stores 教 as `jiào` with `jiāo` as an alternate, and `jiào` is
+ * right for the compounds — 教育, 教师, 宗教, 主教 — which are words and reach
+ * their reading through the word. What is left is the 教 that stands as a word
+ * of its own, and that one is the verb: 他在北京大学教了三年书 came out
+ * `jiàole sān nián shū`, and 我教英语, 谁教你法语 and 她教我如何游泳 with it.
+ * 教书 is a word and was always `jiāo shū`; nothing carried that across a 了.
+ *
+ * See {@link isTeachingAt} for what decides it. Forcing the single-character
+ * edge is not enough on its own, because a reading spanning two characters
+ * carries its own 教 into the position: 王老师教我们汉语 read `jiào` off 师教,
+ * and 来教我 off 来教. Those are pairs the dictionary holds with no part of
+ * speech and at the cost of a word nothing has ever counted — a reading rather
+ * than a word — so a *tagged* word ending in 教 is left alone, which is every
+ * one that matters: 任教, 宗教, 主教, 佛教, 传教, 执教, 请教, 家教.
+ *
+ * Measured over 88,866 lines of Tatoeba and zh.wikipedia, 181 教 decode as a
+ * word of their own and every one of them read `jiào`. This moves 158 to
+ * `jiāo` and is wrong on three: 统一教创始人, 方法教深思 and 做到了教政分离,
+ * where a nominal compound takes an object's shape, the last of them because
+ * 做到了 is a word and so the 了 in front is not a particle to look at. Of the
+ * 21 it leaves, 8 are right — 诸教中, 教外别传, 董教总, 风教, 教青局 twice, a
+ * 教室 reached through a bad split, and a 教 used as a bare noun — and 13 are
+ * misses, where jieba tags the object something this does not name.
+ */
+export declare const TEACHING_JIAO: EdgeRule;
+/**
+ * A 量词 with a number in front of it does not start the next word.
+ *
+ * 三个人 is three people and came out `sān gèrén`, three *personals*, because
+ * 个人 is a common noun and nothing weighed it against the 个 belonging to the
+ * 三 in front of it. The same swallow takes 一杯水 into 杯水, 两家俱乐部 into
+ * 家俱 and 一天天 into 天天; digits reach it too, since a run decoded after a
+ * number sees that number — see `decodeRun` — so 2个人 is the same case.
+ *
+ * What makes it decidable is that the dictionary tags the 量词: 个, 次, 天, 位,
+ * 杯 and 匹 are `q`, while the characters that only look like measure words in
+ * this position are not — 分 is `v`, 部 and 成 are `n`, 年 and 点 are `m`. So
+ * 五分钟, 三部分, 五成分, 五年级 and 三点钟 are untouched, and those are exactly
+ * the words a blanket rule would break. A word that is itself a 量词 is kept as
+ * well, since 分钟 and 点钟 are units rather than something being counted.
+ *
+ * Only the edge *starting at the 量词* is forbidden, so a number and its
+ * measure written solid survive whole where the dictionary has them: 一辈子 and
+ * 一会儿 are entries covering the number too, and the decode still reaches them.
+ *
+ * Measured over the same 88,866 lines of Tatoeba and zh.wikipedia the other
+ * rules were sized against, it forbids 561 edges across 144 shapes — 个人 203
+ * of them, then 杯水, 天一, 口气, 些小 and 家人 — and moves 53 decodes, since
+ * most of what it takes away the cost model was not going to choose anyway.
+ * Three of the 53 are wrong: 一批评, 这一名词 and 六七股灾, where the 一 and the
+ * 六七 count nothing and no tag says so. Exactly one reading changes over the
+ * corpus and it is a fix, 下了两天雨 having read 天雨 as `tiān yù`.
+ */
+export declare const COUNTED_MEASURE: EdgeRule;
 /**
  * The rules the lattice decode applies, in order.
  */

package/dist/decode/reading-rules.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"reading-rules.d.ts","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AACA,OAAO,EAEL,KAAK,QAAQ,EAId,MAAM,YAAY,CAAC;AAgCpB;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,QAAQ,EAAE,QActB,CAAC;AAEF;;;;;;;;;;;;;;;;;;;GAmBG;AACH,eAAO,MAAM,cAAc,EAAE,QAY5B,CAAC;AAEF;;GAEG;AACH,eAAO,MAAM,aAAa,EAAE,SAAS,QAAQ,EAA+B,CAAC"}
+{"version":3,"file":"reading-rules.d.ts","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAGA,OAAO,EAEL,KAAK,QAAQ,EAId,MAAM,YAAY,CAAC;AAgCpB;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,QAAQ,EAAE,QActB,CAAC;AAEF;;;;;;;;;;;;;;;;;;;GAmBG;AACH,eAAO,MAAM,cAAc,EAAE,QAY5B,CAAC;AAqDF;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,eAAO,MAAM,aAAa,EAAE,QAiB3B,CAAC;AAsCF;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,eAAe,EAAE,QAuB7B,CAAC;AAEF;;GAEG;AACH,eAAO,MAAM,aAAa,EAAE,SAAS,QAAQ,EAK5C,CAAC"}

package/dist/decode/reading-rules.js

@@ -1,3 +1,5 @@
+import { QUANTITY_CHARACTERS } from "../numerals/characters.js";
+import { toCharacters } from "../script/characters.js";
 import { tagOf, wordEndingAt, wordStartingAt, } from "./rules.js";
 /**
  * Whether a syllable is `děi`, the modal reading of 得.
@@ -100,8 +102,183 @@ export const ATTESTED_ERHUA = {
         return (reading?.at(-1)?.erhua ?? false) ? "forbid" : "keep";
     },
 };
+/**
+ * Whether a syllable is `jiāo`, the verbal reading of 教.
+ */
+function isJiao(syllable) {
+    return (syllable?.initial === "j" && syllable.final === "iao" && syllable.tone === 1);
+}
+/**
+ * What a 教 that is teaching governs: the thing or the person taught.
+ *
+ * A pronoun, a common noun, a name or one of the tags jieba gives a subject —
+ * 教我, 教英语, 教历史, 教孩子, 教玛丽. `ns` is deliberately absent, since the
+ * only 教 with a place tag after it over the corpus is 教青局, an office.
+ */
+const TAUGHT_TAGS = new Set(["r", "n", "nz", "nr", "nt", "ng"]);
+/**
+ * What follows a 教 that is teaching where no object does: aspect and the
+ * complement marker, which only a verb takes.
+ *
+ * 他教了三年书, 他教过我英语, 她教书教得很快乐. Matched as characters rather
+ * than through their tags, because what the dictionary has starting at a 得 is
+ * as likely to be 得很 as the particle itself.
+ */
+const ASPECT = new Set(["了", "过", "過", "着", "著", "得"]);
+/**
+ * The tag prefix jieba gives a particle, which nothing teaching can follow.
+ */
+const PARTICLE_TAG = "u";
+/**
+ * Whether the 教 at a position is teaching something to somebody.
+ *
+ * The object is what says so, which is the same shape the modal 得 rule takes:
+ * a 教 with a pronoun, a noun or a name after it is governing it, and a 教 with
+ * 了, 过, 着 or 得 after it is a verb whatever follows that. A particle in front
+ * of it rules it out, since a verb does not follow one.
+ */
+function isTeachingAt(context, at) {
+    if (tagOf(context, wordEndingAt(context, at)).startsWith(PARTICLE_TAG)) {
+        return false;
+    }
+    return (ASPECT.has(context.characters[at + 1] ?? "") ||
+        TAUGHT_TAGS.has(tagOf(context, wordStartingAt(context, at + 1))));
+}
+/**
+ * 教 read as `jiāo` where it is teaching rather than a religion.
+ *
+ * The dictionary stores 教 as `jiào` with `jiāo` as an alternate, and `jiào` is
+ * right for the compounds — 教育, 教师, 宗教, 主教 — which are words and reach
+ * their reading through the word. What is left is the 教 that stands as a word
+ * of its own, and that one is the verb: 他在北京大学教了三年书 came out
+ * `jiàole sān nián shū`, and 我教英语, 谁教你法语 and 她教我如何游泳 with it.
+ * 教书 is a word and was always `jiāo shū`; nothing carried that across a 了.
+ *
+ * See {@link isTeachingAt} for what decides it. Forcing the single-character
+ * edge is not enough on its own, because a reading spanning two characters
+ * carries its own 教 into the position: 王老师教我们汉语 read `jiào` off 师教,
+ * and 来教我 off 来教. Those are pairs the dictionary holds with no part of
+ * speech and at the cost of a word nothing has ever counted — a reading rather
+ * than a word — so a *tagged* word ending in 教 is left alone, which is every
+ * one that matters: 任教, 宗教, 主教, 佛教, 传教, 执教, 请教, 家教.
+ *
+ * Measured over 88,866 lines of Tatoeba and zh.wikipedia, 181 教 decode as a
+ * word of their own and every one of them read `jiào`. This moves 158 to
+ * `jiāo` and is wrong on three: 统一教创始人, 方法教深思 and 做到了教政分离,
+ * where a nominal compound takes an object's shape, the last of them because
+ * 做到了 is a word and so the 了 in front is not a particle to look at. Of the
+ * 21 it leaves, 8 are right — 诸教中, 教外别传, 董教总, 风教, 教青局 twice, a
+ * 教室 reached through a bad split, and a 教 used as a bare noun — and 13 are
+ * misses, where jieba tags the object something this does not name.
+ */
+export const TEACHING_JIAO = {
+    name: "teaching-jiao",
+    verdictFor: (context) => {
+        const { edge } = context;
+        if (edge.text === "教") {
+            return isJiao(edge.reading[0]) && isTeachingAt(context, edge.from)
+                ? "force"
+                : "keep";
+        }
+        return edge.partOfSpeech === "" &&
+            edge.text.endsWith("教") &&
+            edge.reading.length === edge.to - edge.from &&
+            !isJiao(edge.reading.at(-1)) &&
+            isTeachingAt(context, edge.to - 1)
+            ? "forbid"
+            : "keep";
+    },
+};
+/**
+ * The tag the dictionary gives a 量词.
+ */
+const MEASURE = "q";
+/**
+ * The mark that makes the number after it a position rather than a count.
+ *
+ * 第三集团军 is the Third Army Group and not three 集团军, so the 集 the
+ * dictionary tags `q` is counting nothing and 集团军 is a word like any other.
+ */
+const ORDINAL = "第";
+/**
+ * Where the number counting the character at a position starts, or undefined
+ * where nothing counts it.
+ *
+ * Two things stop a numeral in front of a 量词 from counting it. An ordinal
+ * names a position, so 第十级别 is the tenth 级别 rather than ten 级; and a
+ * numeral inside a longer word belongs to that word, so the 一 of 唯一道路 is
+ * counting no 道 and the 二 of 十二月 no 月.
+ */
+function countedFrom(context, at) {
+    const { characters } = context;
+    let from = at;
+    while (from > 0 && QUANTITY_CHARACTERS.has(characters[from - 1] ?? "")) {
+        from--;
+    }
+    if (from === at || characters[from - 1] === ORDINAL) {
+        return undefined;
+    }
+    const before = wordEndingAt(context, at);
+    const held = before === undefined ? at : at - toCharacters(before).length;
+    return held < from ? undefined : from;
+}
+/**
+ * A 量词 with a number in front of it does not start the next word.
+ *
+ * 三个人 is three people and came out `sān gèrén`, three *personals*, because
+ * 个人 is a common noun and nothing weighed it against the 个 belonging to the
+ * 三 in front of it. The same swallow takes 一杯水 into 杯水, 两家俱乐部 into
+ * 家俱 and 一天天 into 天天; digits reach it too, since a run decoded after a
+ * number sees that number — see `decodeRun` — so 2个人 is the same case.
+ *
+ * What makes it decidable is that the dictionary tags the 量词: 个, 次, 天, 位,
+ * 杯 and 匹 are `q`, while the characters that only look like measure words in
+ * this position are not — 分 is `v`, 部 and 成 are `n`, 年 and 点 are `m`. So
+ * 五分钟, 三部分, 五成分, 五年级 and 三点钟 are untouched, and those are exactly
+ * the words a blanket rule would break. A word that is itself a 量词 is kept as
+ * well, since 分钟 and 点钟 are units rather than something being counted.
+ *
+ * Only the edge *starting at the 量词* is forbidden, so a number and its
+ * measure written solid survive whole where the dictionary has them: 一辈子 and
+ * 一会儿 are entries covering the number too, and the decode still reaches them.
+ *
+ * Measured over the same 88,866 lines of Tatoeba and zh.wikipedia the other
+ * rules were sized against, it forbids 561 edges across 144 shapes — 个人 203
+ * of them, then 杯水, 天一, 口气, 些小 and 家人 — and moves 53 decodes, since
+ * most of what it takes away the cost model was not going to choose anyway.
+ * Three of the 53 are wrong: 一批评, 这一名词 and 六七股灾, where the 一 and the
+ * 六七 count nothing and no tag says so. Exactly one reading changes over the
+ * corpus and it is a fix, 下了两天雨 having read 天雨 as `tiān yù`.
+ */
+export const COUNTED_MEASURE = {
+    name: "counted-measure",
+    verdictFor: (context) => {
+        const { characters, edge } = context;
+        // A reading of a different length from its span says something about how
+        // the characters are read rather than only about where they break: 份儿 is
+        // `fènr` over two characters, and taking the edge away would take the 儿化
+        // with it.
+        if (edge.to - edge.from < 2 ||
+            edge.reading.length !== edge.to - edge.from) {
+            return "keep";
+        }
+        const counted = countedFrom(context, edge.from);
+        if (counted === undefined) {
+            return "keep";
+        }
+        return tagOf(context, characters[edge.from]) === MEASURE &&
+            edge.partOfSpeech !== MEASURE
+            ? "forbid"
+            : "keep";
+    },
+};
 /**
  * The rules the lattice decode applies, in order.
  */
-export const READING_RULES = [MODAL_DE, ATTESTED_ERHUA];
+export const READING_RULES = [
+    MODAL_DE,
+    TEACHING_JIAO,
+    ATTESTED_ERHUA,
+    COUNTED_MEASURE,
+];
 //# sourceMappingURL=reading-rules.js.map

package/dist/decode/reading-rules.js.map

@@ -1 +1 @@
-{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AACA,OAAO,EAGL,KAAK,EACL,YAAY,EACZ,cAAc,GACf,MAAM,YAAY,CAAC;AAEpB;;;;;GAKG;AACH,SAAS,KAAK,CAAC,QAA8B;IAC3C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,IAAI,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC5E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9C;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,eAAe,CAAC;AAEtC;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa;IAChC,IAAI,EAAE,UAAU;IAChB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACjD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;QAC/C,OAAO,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC7C,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACzC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,CAAC,MAAM,cAAc,GAAa;IACtC,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACzC,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,wEAAwE;QACxE,MAAM,QAAQ,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC;QACvD,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;QAC7D,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB,CAAC,QAAQ,EAAE,cAAc,CAAC,CAAC"}
+{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAEvD,OAAO,EAGL,KAAK,EACL,YAAY,EACZ,cAAc,GACf,MAAM,YAAY,CAAC;AAEpB;;;;;GAKG;AACH,SAAS,KAAK,CAAC,QAA8B;IAC3C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,IAAI,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC5E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9C;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,eAAe,CAAC;AAEtC;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa;IAChC,IAAI,EAAE,UAAU;IAChB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACjD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;QAC/C,OAAO,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC7C,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACzC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,CAAC,MAAM,cAAc,GAAa;IACtC,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACzC,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,wEAAwE;QACxE,MAAM,QAAQ,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC;QACvD,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;QAC7D,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,SAAS,MAAM,CAAC,QAA8B;IAC5C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,KAAK,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC7E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAEhE;;;;;;;GAOG;AACH,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEvD;;GAEG;AACH,MAAM,YAAY,GAAG,GAAG,CAAC;AAEzB;;;;;;;GAOG;AACH,SAAS,YAAY,CAAC,OAAoB,EAAE,EAAU;IACpD,IAAI,KAAK,CAAC,OAAO,EAAE,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QACvE,OAAO,KAAK,CAAC;IACf,CAAC;IACD,OAAO,CACL,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5C,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,cAAc,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CACjE,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,aAAa,GAAa;IACrC,IAAI,EAAE,eAAe;IACrB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,EAAE,CAAC;YACtB,OAAO,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC;gBAChE,CAAC,CAAC,OAAO;gBACT,CAAC,CAAC,MAAM,CAAC;QACb,CAAC;QACD,OAAO,IAAI,CAAC,YAAY,KAAK,EAAE;YAC7B,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI;YAC3C,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;YAC5B,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC;YAClC,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;GAKG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;;;;GAQG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAC/B,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,OAAO,IAAI,GAAG,CAAC,IAAI,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QACvE,IAAI,EAAE,CAAC;IACT,CAAC;IACD,IAAI,IAAI,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,KAAK,OAAO,EAAE,CAAC;QACpD,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IAC1E,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACxC,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAa;IACvC,IAAI,EAAE,iBAAiB;IACvB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,yEAAyE;QACzE,yEAAyE;QACzE,yEAAyE;QACzE,WAAW;QACX,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAC3C,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,OAAO,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1B,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,OAAO;YACtD,IAAI,CAAC,YAAY,KAAK,OAAO;YAC7B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB;IAChD,QAAQ;IACR,aAAa;IACb,cAAc;IACd,eAAe;CAChB,CAAC"}

package/dist/format/slug.d.ts

@@ -0,0 +1,156 @@
+import { type NumberStyle } from "../decode/convert.js";
+import type { SandhiOptions } from "../decode/sandhi.js";
+import type { Dictionary } from "../dictionary/dictionary.js";
+import type { Locale } from "../script/script.js";
+/**
+ * Whether a slug carries its tones.
+ *
+ * `numbers` writes them as digits — 中文 is `zhong1wen2` — and `none` leaves
+ * them off. Tones are on by default because dropping them collides: there are
+ * roughly 400 toneless syllables against 1300 toned ones, so a toneless slug
+ * runs 树, 书 and 输 together as `shu`. They are not a guarantee, though —
+ * 权利 and 权力 are both `quan2li4`, and only {@link SlugOptions.hash} tells
+ * those apart.
+ */
+export type SlugTones = "numbers" | "none";
+/**
+ * Whether the syllables inside one word are separated.
+ *
+ * `join` writes a word as a word — 中文 is `zhong1wen2` — which is what 正词法
+ * asks for and what a reader expects in a URL. `separate` cuts every syllable
+ * apart, `zhong1-wen2`, for a caller matching a slug syllable by syllable.
+ */
+export type SlugSyllables = "join" | "separate";
+/**
+ * How ü is written where the alphabet has no room for it.
+ *
+ * `v` is the input-method convention and keeps 绿 `lv4` apart from 路 `lu4`;
+ * `u` is prettier and merges them.
+ */
+export type SlugUmlaut = "v" | "u";
+/**
+ * How a slug is built.
+ */
+export interface SlugOptions {
+    /** Whether tones are written. Defaults to `numbers`. */
+    readonly tones?: SlugTones;
+    /**
+     * What goes between words. Defaults to `-`.
+     *
+     * Written as given rather than checked, because the separator is the one
+     * place a caller decides what kind of identifier this is: `-` for a URL, `_`
+     * for a name in code, and the empty string for a key to match against.
+     */
+    readonly separator?: string;
+    /** Whether the syllables within a word are separated. Defaults to `join`. */
+    readonly syllables?: SlugSyllables;
+    /** How ü is written. Defaults to `v`. */
+    readonly umlaut?: SlugUmlaut;
+    /**
+     * What to do with the digits in a text. Defaults to `keep`.
+     *
+     * The opposite of what a conversion defaults to, and deliberately: a slug for
+     * iPhone 15 评测 should be `iphone-15-ping2ce4` rather than
+     * `iphone-shi2wu3-ping2ce4`, because the digits are how anyone would look for
+     * it. `read` says them, for a caller that wants the text spoken.
+     */
+    readonly numbers?: NumberStyle;
+    /**
+     * A short hash of the source text on the end. Defaults to none.
+     *
+     * `true` for {@link DEFAULT_HASH_LENGTH} characters, or a length of your own,
+     * up to {@link LONGEST_HASH}. See {@link hashOf} for what is hashed and why
+     * it is the hanzi rather than the pinyin.
+     *
+     * Unconditional rather than only where two slugs collide: a function that
+     * cannot see the rest of your corpus cannot know that they did. It narrows
+     * collisions rather than closing them, and checking the slug you generated
+     * against the ones you have stored is still yours to do.
+     */
+    readonly hash?: number | boolean;
+    /**
+     * The longest the slug may be, counting any hash. Defaults to no limit.
+     *
+     * Cut at a word boundary, so a limit lands between words rather than through
+     * the middle of one. A single word longer than the limit is cut where the
+     * limit falls, since there is nowhere else to cut it, and the hash is never
+     * what gets dropped.
+     */
+    readonly maxLength?: number;
+    /**
+     * What to write where a text slugs to nothing. Defaults to the empty string.
+     *
+     * 《》！ has no letters in it, and neither does an empty text.
+     */
+    readonly fallback?: string;
+    /** Which reading standard to use. Defaults to `zh-CN`. */
+    readonly locale?: Locale;
+    /** Which tone sandhi to apply. Defaults to what a conversion applies. */
+    readonly sandhi?: SandhiOptions;
+}
+/**
+ * How many characters {@link SlugOptions.hash} writes when simply asked for.
+ *
+ * Four base-36 characters is 1.7 million values, which is far more than it
+ * looks: two texts only collide when their slugs *and* their hashes match, so
+ * the suffix is only ever telling apart the handful of texts sharing one slug,
+ * not the whole corpus.
+ */
+export declare const DEFAULT_HASH_LENGTH = 4;
+/**
+ * The longest hash there is anything to fill.
+ *
+ * {@link hashOf} is 32 bits wide, and 36⁷ is the last power of 36 under 2³²:
+ * asking for an eighth character would add a place that some values can never
+ * reach, which looks like more uniqueness while being none.
+ */
+export declare const LONGEST_HASH = 7;
+/**
+ * A 32-bit hash of a text, written in the letters and digits a slug allows.
+ *
+ * **The source text is what is hashed, not the pinyin.** Two texts needing a
+ * hash to tell them apart are, by definition, ones the pinyin already ran
+ * together: 权利 and 权力 are both `quan2li4`, and hashing that would give them
+ * the same suffix and settle nothing. Hashing the hanzi also means the suffix
+ * does not move when a later release reads a word differently, so a slug's tail
+ * is the stable half of it.
+ *
+ * FNV-1a with murmur3's finalizer, which is neither cryptographic nor trying to
+ * be: nothing here resists an attacker, and all that is being asked of it is
+ * that two texts a reader would call different land on different suffixes.
+ * Written out rather than reached for because the package has no runtime
+ * dependencies and the core imports no Node built-ins, and `crypto.subtle` —
+ * the one hash a browser ships — is asynchronous, which this is not.
+ *
+ * Normalised first, so that text typed as a decomposed sequence hashes as the
+ * same text once composed. 简体 and 繁體 are deliberately *not* folded together:
+ * a caller who wrote 头发 and 頭髮 separately meant them separately.
+ */
+export declare function hashOf(text: string, length: number): string;
+/**
+ * Convert hanzi to a URL-safe slug.
+ *
+ * 我想学中文。 is `wo3-xiang3-xue2-zhong1wen2`, and the reason to ask a
+ * dictionary for it rather than a string is everything a slugifier working
+ * after the fact cannot recover: 银行 is `yin2hang2` and not `yin2xing2`
+ * because the word was decoded rather than the characters looked up one at a
+ * time; the hyphens fall between words because 正词法 grouping put them there;
+ * and 西安 is `xi-an` rather than `xian` — a different word — because the
+ * syllable boundary was still known when the slug was cut.
+ *
+ * Everything that is not Han is folded into the same alphabet and kept, so
+ * iPhone 15 评测 is `iphone-15-ping2ce4`, and everything else — punctuation,
+ * emoji, marks of any script — becomes a boundary between words.
+ *
+ * The options generalise it past URLs: a separator of `_` writes a name in
+ * code, and one of the empty string with `tones: "none"` writes a key to match
+ * a search box against, where someone typing `zhongwen` should find 中文. See
+ * the slug page in `docs/` for those written out.
+ *
+ * **A slug is not unique and this cannot make it one.** Tones narrow the
+ * collisions, {@link SlugOptions.hash} narrows them much further, and neither
+ * is a guarantee; a slug that has to be unique still has to be checked against
+ * the ones already stored.
+ */
+export declare function slug(dictionary: Dictionary, text: string, options?: SlugOptions): string;
+//# sourceMappingURL=slug.d.ts.map

package/dist/format/slug.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"slug.d.ts","sourceRoot":"","sources":["../../src/format/slug.ts"],"names":[],"mappings":"AAAA,OAAO,EAGL,KAAK,WAAW,EACjB,MAAM,sBAAsB,CAAC;AAC9B,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACzD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAE9D,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,qBAAqB,CAAC;AAGlD;;;;;;;;;GASG;AACH,MAAM,MAAM,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;AAE3C;;;;;;GAMG;AACH,MAAM,MAAM,aAAa,GAAG,MAAM,GAAG,UAAU,CAAC;AAEhD;;;;;GAKG;AACH,MAAM,MAAM,UAAU,GAAG,GAAG,GAAG,GAAG,CAAC;AAEnC;;GAEG;AACH,MAAM,WAAW,WAAW;IAC1B,wDAAwD;IACxD,QAAQ,CAAC,KAAK,CAAC,EAAE,SAAS,CAAC;IAC3B;;;;;;OAMG;IACH,QAAQ,CAAC,SAAS,CAAC,EAAE,MAAM,CAAC;IAC5B,6EAA6E;IAC7E,QAAQ,CAAC,SAAS,CAAC,EAAE,aAAa,CAAC;IACnC,yCAAyC;IACzC,QAAQ,CAAC,MAAM,CAAC,EAAE,UAAU,CAAC;IAC7B;;;;;;;OAOG;IACH,QAAQ,CAAC,OAAO,CAAC,EAAE,WAAW,CAAC;IAC/B;;;;;;;;;;;OAWG;IACH,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC;IACjC;;;;;;;OAOG;IACH,QAAQ,CAAC,SAAS,CAAC,EAAE,MAAM,CAAC;IAC5B;;;;OAIG;IACH,QAAQ,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAC3B,0DAA0D;IAC1D,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC;IACzB,yEAAyE;IACzE,QAAQ,CAAC,MAAM,CAAC,EAAE,aAAa,CAAC;CACjC;AAED;;;;;;;GAOG;AACH,eAAO,MAAM,mBAAmB,IAAI,CAAC;AAErC;;;;;;GAMG;AACH,eAAO,MAAM,YAAY,IAAI,CAAC;AAwK9B;;;;;;;;;;;;;;;;;;;;GAoBG;AACH,wBAAgB,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,MAAM,CAgB3D;AA0BD;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AACH,wBAAgB,IAAI,CAClB,UAAU,EAAE,UAAU,EACtB,IAAI,EAAE,MAAM,EACZ,OAAO,GAAE,WAAgB,GACxB,MAAM,CAuBR"}

package/dist/format/slug.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"slug.js","sourceRoot":"","sources":["../../src/format/slug.ts"],"names":[],"mappings":"AAAA,OAAO,EAEL,qBAAqB,GAEtB,MAAM,sBAAsB,CAAC;AAG9B,OAAO,EAAE,QAAQ,EAAE,MAAM,8BAA8B,CAAC;AAExD,OAAO,EAAiB,aAAa,EAAE,MAAM,yBAAyB,CAAC;AA4FvE;;;;;;;GAOG;AACH,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,CAAC;AAErC;;;;;;GAMG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAC;AAE9B;;;;;;;;;;;GAWG;AACH,MAAM,QAAQ,GAAG,GAAG,CAAC;AAErB;;GAEG;AACH,MAAM,eAAe,GAAG,cAAc,CAAC;AAEvC;;GAEG;AACH,MAAM,eAAe,GAAG,UAAU,CAAC;AAEnC;;;;;;;GAOG;AACH,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI;SACR,WAAW,EAAE;SACb,SAAS,CAAC,KAAK,CAAC;SAChB,UAAU,CAAC,eAAe,EAAE,EAAE,CAAC;SAC/B,UAAU,CAAC,eAAe,EAAE,GAAG,EAAE,CAAC,QAAQ,CAAC,CAAC;AACjD,CAAC;AAED;;;;;;GAMG;AACH,SAAS,WAAW,CAAC,QAAgB,EAAE,MAAkB;IACvD,OAAO,QAAQ,CAAC,UAAU,CAAC,GAAG,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,CAAC;AAChD,CAAC;AAED;;;;;;;;;;;GAWG;AACH,SAAS,QAAQ,CACf,SAA8B,EAC9B,OAAoB;IAEpB,MAAM,EAAE,KAAK,GAAG,SAAS,EAAE,SAAS,EAAE,GAAG,GAAG,MAAM,EAAE,MAAM,GAAG,GAAG,EAAE,GAAG,OAAO,CAAC;IAC7E,MAAM,SAAS,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAC3C,WAAW,CACT,aAAa,CAAC,QAAQ,EAAE,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,EACjE,MAAM,CACP,CACF,CAAC;IACF,IAAI,GAAG,KAAK,UAAU,EAAE,CAAC;QACvB,OAAO,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAClC,CAAC;IACD,wEAAwE;IACxE,0EAA0E;IAC1E,OAAO,KAAK,KAAK,SAAS;QACxB,CAAC,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;QACpB,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,UAAU,CAAC,GAAG,EAAE,GAAG,EAAE,CAAC,QAAQ,CAAC,CAAC;AAC1D,CAAC;AAED;;;;;;GAMG;AACH,SAAS,UAAU,CACjB,MAAiC,EACjC,OAAoB;IAEpB,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,IAAI,GAAe,EAAE,CAAC;IAE1B,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACtB,OAAO;QACT,CAAC;QAED,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,CAAC;QACtC,IAAI,GAAG,EAAE,CAAC;IACZ,CAAC,CAAC;IAEF,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,IAAI,KAAK,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;YACjC,KAAK,EAAE,CAAC;YACR,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;YACpC,SAAS;QACX,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;IAC5B,CAAC;IACD,KAAK,EAAE,CAAC;IACR,OAAO,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;AAChC,CAAC;AAED;;;;;;GAMG;AACH,SAAS,OAAO,CAAC,KAAa;IAC5B,OAAO,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC;AAC7D,CAAC;AAED;;;;;;GAMG;AACH,SAAS,KAAK,CACZ,KAAwB,EACxB,KAAa,EACb,SAAiB,EACjB,MAAc;IAEd,MAAM,IAAI,GAAG,MAAM,KAAK,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,MAAM,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC;IAC9E,IAAI,IAAI,IAAI,CAAC,EAAE,CAAC;QACd,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,IAAI,GAAG,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC;QACtE,IAAI,IAAI,GAAG,IAAI,GAAG,IAAI,EAAE,CAAC;YACvB,MAAM;QACR,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAChB,IAAI,IAAI,IAAI,CAAC;IACf,CAAC;IACD,2EAA2E;IAC3E,0EAA0E;IAC1E,MAAM,CAAC,KAAK,CAAC,GAAG,KAAK,CAAC;IACtB,OAAO,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,KAAK,SAAS;QAC7C,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;QACxB,CAAC,CAAC,IAAI,CAAC;AACX,CAAC;AAED;;;;;;;;;;;;;;;;;;;;GAoBG;AACH,MAAM,UAAU,MAAM,CAAC,IAAY,EAAE,MAAc;IACjD,MAAM,KAAK,GAAG,IAAI,WAAW,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC;IAC9D,IAAI,IAAI,GAAG,aAAa,CAAC;IACzB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,GAAG,IAAI,EAAE,aAAa,CAAC,CAAC;IAC/C,CAAC;IACD,4EAA4E;IAC5E,2EAA2E;IAC3E,IAAI,IAAI,IAAI,KAAK,EAAE,CAAC;IACpB,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,aAAa,CAAC,CAAC;IACtC,IAAI,IAAI,IAAI,KAAK,EAAE,CAAC;IACpB,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,aAAa,CAAC,CAAC;IACtC,IAAI,IAAI,IAAI,KAAK,EAAE,CAAC;IAEpB,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;IACvE,OAAO,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,QAAQ,CAAC,YAAY,EAAE,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC;AAC9E,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CAAC,KAAmC;IACrD,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,KAAK,EAAE,CAAC;QAC3C,OAAO,CAAC,CAAC;IACX,CAAC;IACD,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;QACnB,OAAO,mBAAmB,CAAC;IAC7B,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;AAChE,CAAC;AAED;;GAEG;AACH,SAAS,OAAO,CACd,KAAwB,EACxB,IAAY,EACZ,SAAiB;IAEjB,OAAO,CAAC,GAAG,KAAK,EAAE,GAAG,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;AACpE,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AACH,MAAM,UAAU,IAAI,CAClB,UAAsB,EACtB,IAAY,EACZ,OAAO,GAAgB,EAAE;IAEzB,MAAM,EAAE,SAAS,GAAG,GAAG,EAAE,QAAQ,GAAG,EAAE,EAAE,SAAS,EAAE,GAAG,OAAO,CAAC;IAC9D,MAAM,MAAM,GAAG,qBAAqB,CAAC,UAAU,EAAE,IAAI,EAAE;QACrD,GAAG,CAAC,OAAO,CAAC,MAAM,KAAK,SAAS,IAAI,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC;QAC/D,GAAG,CAAC,OAAO,CAAC,MAAM,KAAK,SAAS,IAAI,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC;QAC/D,OAAO,EAAE,OAAO,CAAC,OAAO,IAAI,MAAM;QAClC,0EAA0E;QAC1E,yEAAyE;QACzE,2DAA2D;QAC3D,QAAQ,EAAE,MAAM;QAChB,WAAW,EAAE,MAAM;KACpB,CAAC,CAAC;IAEH,MAAM,IAAI,GAAG,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IACtC,MAAM,MAAM,GAAG,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IACpD,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;IACnD,MAAM,MAAM,GACV,SAAS,KAAK,SAAS;QACrB,CAAC,CAAC,KAAK;QACP,CAAC,CAAC,KAAK,CAAC,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,CAAC,CAAC;IAEjD,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,EAAE,MAAM,EAAE,SAAS,CAAC,CAAC;IACjD,OAAO,KAAK,KAAK,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC;AACzC,CAAC"}

package/dist/index.d.ts

@@ -35,6 +35,8 @@ export { decodeReadings, decodeRun, decodeRunScored, decodeSpacing, } from "./de
 export { convertToHtml, toHtml } from "./format/html.js";
 export { convertToWadeGiles, toTranscription, type TranscriptionOptions, type WriteWord, } from "./format/transcription.js";
 export type { HtmlOptions } from "./format/html.js";
+export { slug } from "./format/slug.js";
+export type { SlugOptions, SlugSyllables, SlugTones, SlugUmlaut, } from "./format/slug.js";
 export { decodeGreedily } from "./decode/greedy.js";
 export { allEdges, buildLattice, cutPoints, READING_CHARGE, } from "./decode/lattice.js";
 export type { Lattice, LatticeEdge } from "./decode/lattice.js";
@@ -42,13 +44,13 @@ export { isSettled, projectReadings, settledUnits, unitsOf, } from "./decode/loc
 export type { ReadingProjection, ReadingUnit } from "./decode/locking.js";
 export { applyEdgeRules, tagOf, wordEndingAt, wordStartingAt, } from "./decode/rules.js";
 export type { EdgeContext, EdgeRule, EdgeVerdict } from "./decode/rules.js";
-export { ATTESTED_ERHUA, MODAL_DE, READING_RULES, } from "./decode/reading-rules.js";
+export { ATTESTED_ERHUA, COUNTED_MEASURE, MODAL_DE, READING_RULES, TEACHING_JIAO, } from "./decode/reading-rules.js";
 export { readingCost, shortestPath, spacingCost } from "./decode/viterbi.js";
 export type { CostOf } from "./decode/viterbi.js";
 export type { DecodedWord, ScoredWord, WordSeparator } from "./decode/word.js";
 export { cardinalHanzi, LARGEST_CARDINAL } from "./numerals/cardinal.js";
 export type { CardinalOptions } from "./numerals/cardinal.js";
-export { DIGIT_CHARACTERS, numeralSyllable, UNIT_VALUES, yaoSyllable, } from "./numerals/characters.js";
+export { DIGIT_CHARACTERS, numeralSyllable, QUANTITY_CHARACTERS, UNIT_VALUES, yaoSyllable, } from "./numerals/characters.js";
 export { fractionHanzi, numeralHanzi, percentHanzi, readNumeral, readNumeralHanzi, } from "./numerals/numerals.js";
 export type { NumeralOptions, NumeralStyle } from "./numerals/numerals.js";
 export { isBopomofo, readBopomofo, writeBopomofo, writeBopomofoWord, } from "./transcription/bopomofo.js";