@kensio/pinyinjs 1.7.7 → 1.7.9

package/dist/format/html.d.ts

@@ -1,5 +1,14 @@
+/**
+ * A conversion as markup, and the options that decide what the markup carries.
+ *
+ * One element per syllable, which is what lets a page colour a tone or mark a
+ * guess. What goes inside one of those elements is `syllable-span.ts` and the
+ * ruby that puts the reading over the hanzi is `annotate.ts`.
+ */
 export { convertToAnnotatedHtml, convertToHtml, toAnnotatedHtml, } from "./annotate.js";
 import type { ConvertedPiece, ConvertOptions } from "../decode/convert.js";
+import type { TranscriptionSystem } from "../transcription/systems.js";
+export { writePiece } from "./syllable-span.js";
 /**
  * How a conversion is marked up.
  */
@@ -15,18 +24,32 @@ export interface HtmlOptions extends ConvertOptions {
      * Whether readings the decode was guessing at are marked. Defaults to true.
      *
      * `py-uncertain`, plus a `data-alternatives` attribute listing what was
-     * rejected. See {@link isUncertain} for what qualifies and how often it is
-     * wrong.
+     * rejected. See `isUncertain` for what qualifies and how often it is wrong.
      */
     readonly markUncertain?: boolean;
     /**
      * Whether each syllable declares what language it is in. Defaults to true.
      *
      * `lang="zh-Latn-CN-pinyin"`, or `zh-Latn-TW-pinyin` where the conversion
-     * reads `zh-TW`. See {@link languageTag} for why that matters, and what to
-     * do instead of turning it off.
+     * reads `zh-TW`. See `languageTag` for why that matters, and what to do
+     * instead of turning it off.
      */
     readonly lang?: boolean;
+    /**
+     * A system to write the reading in instead of pinyin.
+     *
+     * `BOPOMOFO`, `WADE_GILES`, `YALE`, `GWOYEU` or `IPA`, or a
+     * {@link TranscriptionSystem} of the caller's own. Left out, the reading is
+     * the pinyin the conversion wrote.
+     *
+     * The word segmentation is the conversion's either way and only the join
+     * changes, which is the division `toTranscription` makes and for the same
+     * reason. What a word is belongs to the language, and how its syllables are
+     * run together belongs to the system. Everything else a syllable carries is
+     * untouched, so tone classes, `py-uncertain` and `data-alternatives` mean
+     * what they always did.
+     */
+    readonly transcription?: TranscriptionSystem;
 }
 /**
  * Escape text for markup.
@@ -39,9 +62,21 @@ export interface HtmlOptions extends ConvertOptions {
  */
 export declare function escape(text: string): string;
 /**
- * Mark up one piece of a conversion.
+ * Mark up a run of pieces, with the system's join between two syllables.
+ *
+ * Pinyin needs none. Its syllables are run together and the marks that divide
+ * them are pieces of their own. Every other system decides the join for itself,
+ * and the join is what separates the five at this level, so it is written
+ * between the spans rather than inside one. `<span>Pei³</span>-<span>ching¹
+ * </span>` is one word in Wade-Giles and leaves each syllable its own element
+ * to carry a tone class on.
+ *
+ * A piece that wrote nothing does not break the run, which is what carries the
+ * join over pinyin's own hyphen: 干干净净 is four syllables of one word, and
+ * the mark between the second and the third is dropped rather than counted as a
+ * gap.
  */
-export declare function writePiece(piece: ConvertedPiece, options: HtmlOptions): string;
+export declare function writePieces(pieces: readonly ConvertedPiece[], options: HtmlOptions): string;
 /**
  * Mark up a converted text, one element per syllable.
  *

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

@@ -1 +1 @@
-{"version":3,"file":"html.d.ts","sourceRoot":"","sources":["../../src/format/html.ts"],"names":[],"mappings":"AACA,OAAO,EACL,sBAAsB,EACtB,aAAa,EACb,eAAe,GAChB,MAAM,eAAe,CAAC;AAEvB,OAAO,KAAK,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAI3E;;GAEG;AACH,MAAM,WAAW,WAAY,SAAQ,cAAc;IACjD;;;;;OAKG;IACH,QAAQ,CAAC,WAAW,CAAC,EAAE,OAAO,CAAC;IAC/B;;;;;;OAMG;IACH,QAAQ,CAAC,aAAa,CAAC,EAAE,OAAO,CAAC;IACjC;;;;;;OAMG;IACH,QAAQ,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC;CACzB;AAwDD;;;;;;;;GAQG;AACH,wBAAgB,MAAM,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAK3C;AAwBD;;GAEG;AACH,wBAAgB,UAAU,CACxB,KAAK,EAAE,cAAc,EACrB,OAAO,EAAE,WAAW,GACnB,MAAM,CA4BR;AAED;;;;;GAKG;AACH,wBAAgB,MAAM,CACpB,MAAM,EAAE,SAAS,cAAc,EAAE,EACjC,OAAO,GAAE,WAAgB,GACxB,MAAM,CAER"}
+{"version":3,"file":"html.d.ts","sourceRoot":"","sources":["../../src/format/html.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,EACL,sBAAsB,EACtB,aAAa,EACb,eAAe,GAChB,MAAM,eAAe,CAAC;AACvB,OAAO,KAAK,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAC3E,OAAO,KAAK,EAAE,mBAAmB,EAAE,MAAM,6BAA6B,CAAC;AAGvE,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD;;GAEG;AACH,MAAM,WAAW,WAAY,SAAQ,cAAc;IACjD;;;;;OAKG;IACH,QAAQ,CAAC,WAAW,CAAC,EAAE,OAAO,CAAC;IAC/B;;;;;OAKG;IACH,QAAQ,CAAC,aAAa,CAAC,EAAE,OAAO,CAAC;IACjC;;;;;;OAMG;IACH,QAAQ,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC;IACxB;;;;;;;;;;;;;OAaG;IACH,QAAQ,CAAC,aAAa,CAAC,EAAE,mBAAmB,CAAC;CAC9C;AAYD;;;;;;;;GAQG;AACH,wBAAgB,MAAM,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAK3C;AAED;;;;;;;;;;;;;;GAcG;AACH,wBAAgB,WAAW,CACzB,MAAM,EAAE,SAAS,cAAc,EAAE,EACjC,OAAO,EAAE,WAAW,GACnB,MAAM,CAgBR;AAED;;;;;GAKG;AACH,wBAAgB,MAAM,CACpB,MAAM,EAAE,SAAS,cAAc,EAAE,EACjC,OAAO,GAAE,WAAgB,GACxB,MAAM,CAER"}

package/dist/format/html.js

@@ -1,49 +1,13 @@
-import { isUncertain } from "../decode/confidence.js";
-export { convertToAnnotatedHtml, convertToHtml, toAnnotatedHtml, } from "./annotate.js";
-import { READING_CHARGE } from "../decode/lattice.js";
-import { DEFAULT_LOCALE } from "../script/script.js";
-import { writeSyllable } from "../syllable/syllable.js";
-/**
- * The class every syllable carries, whatever else it carries.
- */
-const SYLLABLE_CLASS = "py-syllable";
-const UNCERTAIN_CLASS = "py-uncertain";
 /**
- * What a syllable declares itself to be, in the reading standard it was read in.
- *
- * A syllable element holds Mandarin written in the Latin alphabet, which is
- * neither the `zh` of the surrounding page nor the `en` of a page that quotes
- * it — and nothing about `yín` on its own says so. The tag is what a screen
- * reader consults before deciding how to pronounce it, and what a browser
- * consults for hyphenation and font selection; without one, `xíng` is read as
- * whatever the page around it claims to be, which is how pinyin ends up
- * spoken as English.
- *
- * The subtags are all registered and mean exactly this: `Latn` for the script,
- * the region for the reading standard, and the `pinyin` variant, whose prefix
- * in the IANA registry is `zh-Latn`. The region is the one distinction the
- * conversion itself makes — 垃圾 is `lājī` under `zh-CN` and `lèsè` under
- * `zh-TW` — so it is read off {@link ConvertOptions.locale} rather than
- * guessed at.
+ * A conversion as markup, and the options that decide what the markup carries.
  *
- * Tone notation does not enter into it. `hang2` is pinyin spelt with a tone
- * number, not another romanisation, so it takes the same tag as `háng`.
- *
- * The cost is the tag repeated on every syllable, which is the price of
- * wrapping nothing around the whole conversion. A caller who would rather
- * declare it once can set `lang: false` and put the same tag on a wrapper of
- * their own, which is inherited by everything inside it.
+ * One element per syllable, which is what lets a page colour a tone or mark a
+ * guess. What goes inside one of those elements is `syllable-span.ts` and the
+ * ruby that puts the reading over the hanzi is `annotate.ts`.
  */
-function languageTag(locale = DEFAULT_LOCALE) {
-    switch (locale) {
-        case "zh-CN": {
-            return "zh-Latn-CN-pinyin";
-        }
-        case "zh-TW": {
-            return "zh-Latn-TW-pinyin";
-        }
-    }
-}
+export { convertToAnnotatedHtml, convertToHtml, toAnnotatedHtml, } from "./annotate.js";
+import { writePiece } from "./syllable-span.js";
+export { writePiece } from "./syllable-span.js";
 /**
  * How each character that means something in markup is written as text.
  */
@@ -69,49 +33,36 @@ export function escape(text) {
     });
 }
 /**
- * The readings a syllable was chosen over, as the reader would write them.
+ * Mark up a run of pieces, with the system's join between two syllables.
  *
- * Only the ones that made it uncertain: a rival dearer than
- * {@link READING_CHARGE} could only be taken by breaking a dictionary word
- * apart, so it was never really in the running and listing it would suggest
- * otherwise.
- */
-function alternativesOf(confidence, options) {
-    const written = confidence.alternatives
-        .filter((alternative) => alternative.cost < READING_CHARGE)
-        .map((alternative) => alternative.reading
-        .map((syllable) => writeSyllable(syllable, options.notation))
-        .join(""));
-    return [...new Set(written)];
-}
-/**
- * Mark up one piece of a conversion.
+ * Pinyin needs none. Its syllables are run together and the marks that divide
+ * them are pieces of their own. Every other system decides the join for itself,
+ * and the join is what separates the five at this level, so it is written
+ * between the spans rather than inside one. `<span>Pei³</span>-<span>ching¹
+ * </span>` is one word in Wade-Giles and leaves each syllable its own element
+ * to carry a tone class on.
+ *
+ * A piece that wrote nothing does not break the run, which is what carries the
+ * join over pinyin's own hyphen: 干干净净 is four syllables of one word, and
+ * the mark between the second and the third is dropped rather than counted as a
+ * gap.
  */
-export function writePiece(piece, options) {
-    const { toneClasses = true, markUncertain = true, lang = true } = options;
-    if (piece.syllable === undefined) {
-        return escape(piece.text);
+export function writePieces(pieces, options) {
+    const { transcription } = options;
+    if (transcription === undefined) {
+        return pieces.map((piece) => writePiece(piece, options)).join("");
     }
-    const classes = [SYLLABLE_CLASS];
-    if (toneClasses && piece.syllable.tone !== undefined) {
-        classes.push(`py-tone-${String(piece.syllable.tone)}`);
-    }
-    const alternatives = markUncertain &&
-        piece.confidence !== undefined &&
-        isUncertain(piece.confidence)
-        ? alternativesOf(piece.confidence, options)
-        : [];
-    if (alternatives.length > 0) {
-        classes.push(UNCERTAIN_CLASS);
-    }
-    const attributes = [`class="${classes.join(" ")}"`];
-    if (lang) {
-        attributes.push(`lang="${languageTag(options.locale)}"`);
-    }
-    if (alternatives.length > 0) {
-        attributes.push(`data-alternatives="${escape(alternatives.join(" "))}"`);
+    const written = [];
+    let isJoined = false;
+    for (const piece of pieces) {
+        const markup = writePiece(piece, options);
+        if (piece.syllable !== undefined && isJoined) {
+            written.push(transcription.separator);
+        }
+        written.push(markup);
+        isJoined = piece.syllable !== undefined || (isJoined && markup === "");
     }
-    return `<span ${attributes.join(" ")}>${escape(piece.text)}</span>`;
+    return written.join("");
 }
 /**
  * Mark up a converted text, one element per syllable.
@@ -120,6 +71,6 @@ export function writePiece(piece, options) {
  * once and render more than one way.
  */
 export function toHtml(pieces, options = {}) {
-    return pieces.map((piece) => writePiece(piece, options)).join("");
+    return writePieces(pieces, options);
 }
 //# sourceMappingURL=html.js.map

package/dist/format/html.js.map

@@ -1 +1 @@
-{"version":3,"file":"html.js","sourceRoot":"","sources":["../../src/format/html.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAA0B,MAAM,yBAAyB,CAAC;AAC9E,OAAO,EACL,sBAAsB,EACtB,aAAa,EACb,eAAe,GAChB,MAAM,eAAe,CAAC;AACvB,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAEtD,OAAO,EAAE,cAAc,EAAe,MAAM,qBAAqB,CAAC;AAClE,OAAO,EAAE,aAAa,EAAE,MAAM,yBAAyB,CAAC;AA+BxD;;GAEG;AACH,MAAM,cAAc,GAAG,aAAa,CAAC;AAErC,MAAM,eAAe,GAAG,cAAc,CAAC;AAEvC;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AACH,SAAS,WAAW,CAAC,MAAM,GAAW,cAAc;IAClD,QAAQ,MAAM,EAAE,CAAC;QACf,KAAK,OAAO,EAAE,CAAC;YACb,OAAO,mBAAmB,CAAC;QAC7B,CAAC;QACD,KAAK,OAAO,EAAE,CAAC;YACb,OAAO,mBAAmB,CAAC;QAC7B,CAAC;IACH,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAiB;IACtC,CAAC,GAAG,EAAE,OAAO,CAAC;IACd,CAAC,GAAG,EAAE,MAAM,CAAC;IACb,CAAC,GAAG,EAAE,MAAM,CAAC;IACb,CAAC,GAAG,EAAE,QAAQ,CAAC;CAChB,CAAC,CAAC;AAEH;;;;;;;;GAQG;AACH,MAAM,UAAU,MAAM,CAAC,IAAY;IACjC,OAAO,IAAI,CAAC,UAAU,CAAC,UAAU,EAAE,CAAC,SAAS,EAAE,EAAE;QAC/C,mEAAmE;QACnE,OAAO,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,SAAS,CAAC;IAC7C,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,cAAc,CACrB,UAA6B,EAC7B,OAAoB;IAEpB,MAAM,OAAO,GAAG,UAAU,CAAC,YAAY;SACpC,MAAM,CAAC,CAAC,WAAW,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,GAAG,cAAc,CAAC;SAC1D,GAAG,CAAC,CAAC,WAAW,EAAE,EAAE,CACnB,WAAW,CAAC,OAAO;SAChB,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC;SAC5D,IAAI,CAAC,EAAE,CAAC,CACZ,CAAC;IACJ,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC;AAC/B,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CACxB,KAAqB,EACrB,OAAoB;IAEpB,MAAM,EAAE,WAAW,GAAG,IAAI,EAAE,aAAa,GAAG,IAAI,EAAE,IAAI,GAAG,IAAI,EAAE,GAAG,OAAO,CAAC;IAC1E,IAAI,KAAK,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;QACjC,OAAO,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC5B,CAAC;IAED,MAAM,OAAO,GAAG,CAAC,cAAc,CAAC,CAAC;IACjC,IAAI,WAAW,IAAI,KAAK,CAAC,QAAQ,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QACrD,OAAO,CAAC,IAAI,CAAC,WAAW,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACzD,CAAC;IACD,MAAM,YAAY,GAChB,aAAa;QACb,KAAK,CAAC,UAAU,KAAK,SAAS;QAC9B,WAAW,CAAC,KAAK,CAAC,UAAU,CAAC;QAC3B,CAAC,CAAC,cAAc,CAAC,KAAK,CAAC,UAAU,EAAE,OAAO,CAAC;QAC3C,CAAC,CAAC,EAAE,CAAC;IACT,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,OAAO,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;IAChC,CAAC;IAED,MAAM,UAAU,GAAG,CAAC,UAAU,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IACpD,IAAI,IAAI,EAAE,CAAC;QACT,UAAU,CAAC,IAAI,CAAC,SAAS,WAAW,CAAC,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;IAC3D,CAAC;IACD,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,UAAU,CAAC,IAAI,CAAC,sBAAsB,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;IAC3E,CAAC;IACD,OAAO,SAAS,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC;AACtE,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,MAAM,CACpB,MAAiC,EACjC,OAAO,GAAgB,EAAE;IAEzB,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACpE,CAAC"}
+{"version":3,"file":"html.js","sourceRoot":"","sources":["../../src/format/html.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,EACL,sBAAsB,EACtB,aAAa,EACb,eAAe,GAChB,MAAM,eAAe,CAAC;AAGvB,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AA6ChD;;GAEG;AACH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAiB;IACtC,CAAC,GAAG,EAAE,OAAO,CAAC;IACd,CAAC,GAAG,EAAE,MAAM,CAAC;IACb,CAAC,GAAG,EAAE,MAAM,CAAC;IACb,CAAC,GAAG,EAAE,QAAQ,CAAC;CAChB,CAAC,CAAC;AAEH;;;;;;;;GAQG;AACH,MAAM,UAAU,MAAM,CAAC,IAAY;IACjC,OAAO,IAAI,CAAC,UAAU,CAAC,UAAU,EAAE,CAAC,SAAS,EAAE,EAAE;QAC/C,mEAAmE;QACnE,OAAO,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,SAAS,CAAC;IAC7C,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;;;;;;;;;;;;;GAcG;AACH,MAAM,UAAU,WAAW,CACzB,MAAiC,EACjC,OAAoB;IAEpB,MAAM,EAAE,aAAa,EAAE,GAAG,OAAO,CAAC;IAClC,IAAI,aAAa,KAAK,SAAS,EAAE,CAAC;QAChC,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACpE,CAAC;IACD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;QAC1C,IAAI,KAAK,CAAC,QAAQ,KAAK,SAAS,IAAI,QAAQ,EAAE,CAAC;YAC7C,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC;QACxC,CAAC;QACD,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrB,QAAQ,GAAG,KAAK,CAAC,QAAQ,KAAK,SAAS,IAAI,CAAC,QAAQ,IAAI,MAAM,KAAK,EAAE,CAAC,CAAC;IACzE,CAAC;IACD,OAAO,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC1B,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,MAAM,CACpB,MAAiC,EACjC,OAAO,GAAgB,EAAE;IAEzB,OAAO,WAAW,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;AACtC,CAAC"}

package/dist/format/language-tag.d.ts

@@ -0,0 +1,57 @@
+/**
+ * What a written reading declares itself to be, in BCP 47.
+ *
+ * One place that knows the subtags and the order they go in, since the tag is
+ * the same question whether the reading is pinyin or one of the five systems
+ * that can stand in for it.
+ */
+import { type Locale } from "../script/script.js";
+import type { TranscriptionSystem } from "../transcription/systems.js";
+/**
+ * The BCP 47 region subtag for a reading standard.
+ */
+export declare function regionOf(locale: Locale): string;
+/**
+ * What a syllable declares itself to be, in the system it was written in.
+ *
+ * A syllable element holds Mandarin written in some script that is not the `zh`
+ * of the surrounding page and not the `en` of a page that quotes it, and
+ * nothing about `yín` on its own says so. The tag is what a screen reader
+ * consults before deciding how to pronounce it, and what a browser consults for
+ * hyphenation and font selection. Without one, `xíng` is read as whatever the
+ * page around it claims to be, which is how pinyin ends up spoken as English.
+ *
+ * The subtags are registered and mean exactly this. The script is `Latn` for a
+ * romanisation and `Bopo` for bopomofo, the region is the reading standard, and
+ * the variant names the system where the IANA registry has one. It has `pinyin`
+ * and `wadegile`, both with the prefix `zh-Latn`, and `fonipa` for the IPA.
+ *
+ * | System | `zh-CN` | `zh-TW` |
+ * | --- | --- | --- |
+ * | pinyin | `zh-Latn-CN-pinyin` | `zh-Latn-TW-pinyin` |
+ * | bopomofo | `zh-Bopo-CN` | `zh-Bopo-TW` |
+ * | Wade-Giles | `zh-Latn-CN-wadegile` | `zh-Latn-TW-wadegile` |
+ * | Yale | `zh-Latn-CN` | `zh-Latn-TW` |
+ * | Gwoyeu Romatzyh | `zh-Latn-CN` | `zh-Latn-TW` |
+ * | IPA | `zh-Latn-CN-fonipa` | `zh-Latn-TW-fonipa` |
+ *
+ * **The region is carried by every one of them**, because the distinction it
+ * marks is in the reading rather than in the spelling. 垃圾 is `lājī` under
+ * `zh-CN` and `lèsè` under `zh-TW`, and it is two different words in bopomofo
+ * for the same reason, so the region is read off {@link ConvertOptions.locale}
+ * rather than guessed at.
+ *
+ * Yale and Gwoyeu Romatzyh name themselves nowhere, the registry having no
+ * variant for either, so both go out saying only that they are Mandarin in the
+ * Latin alphabet. See {@link TranscriptionSystem.variant}.
+ *
+ * Tone notation does not enter into it. `hang2` is pinyin spelt with a tone
+ * number, not another romanisation, so it takes the same tag as `háng`.
+ *
+ * The cost is the tag repeated on every syllable, which is the price of
+ * wrapping nothing around the whole conversion. A caller who would rather
+ * declare it once can set `lang: false` and put the same tag on a wrapper of
+ * their own, which is inherited by everything inside it.
+ */
+export declare function languageTag(locale?: Locale, system?: TranscriptionSystem): string;
+//# sourceMappingURL=language-tag.d.ts.map

package/dist/format/language-tag.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-tag.d.ts","sourceRoot":"","sources":["../../src/format/language-tag.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,EAAkB,KAAK,MAAM,EAAE,MAAM,qBAAqB,CAAC;AAClE,OAAO,KAAK,EAAE,mBAAmB,EAAE,MAAM,6BAA6B,CAAC;AASvE;;GAEG;AACH,wBAAgB,QAAQ,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAS/C;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAyCG;AACH,wBAAgB,WAAW,CACzB,MAAM,GAAE,MAAuB,EAC/B,MAAM,CAAC,EAAE,mBAAmB,GAC3B,MAAM,CAKR"}

package/dist/format/language-tag.js

@@ -0,0 +1,76 @@
+/**
+ * What a written reading declares itself to be, in BCP 47.
+ *
+ * One place that knows the subtags and the order they go in, since the tag is
+ * the same question whether the reading is pinyin or one of the five systems
+ * that can stand in for it.
+ */
+import { DEFAULT_LOCALE } from "../script/script.js";
+/**
+ * The pinyin variant's script and its own subtag, which a conversion writes by
+ * default.
+ */
+const PINYIN_SCRIPT = "Latn";
+const PINYIN_VARIANT = "pinyin";
+/**
+ * The BCP 47 region subtag for a reading standard.
+ */
+export function regionOf(locale) {
+    switch (locale) {
+        case "zh-CN": {
+            return "CN";
+        }
+        case "zh-TW": {
+            return "TW";
+        }
+    }
+}
+/**
+ * What a syllable declares itself to be, in the system it was written in.
+ *
+ * A syllable element holds Mandarin written in some script that is not the `zh`
+ * of the surrounding page and not the `en` of a page that quotes it, and
+ * nothing about `yín` on its own says so. The tag is what a screen reader
+ * consults before deciding how to pronounce it, and what a browser consults for
+ * hyphenation and font selection. Without one, `xíng` is read as whatever the
+ * page around it claims to be, which is how pinyin ends up spoken as English.
+ *
+ * The subtags are registered and mean exactly this. The script is `Latn` for a
+ * romanisation and `Bopo` for bopomofo, the region is the reading standard, and
+ * the variant names the system where the IANA registry has one. It has `pinyin`
+ * and `wadegile`, both with the prefix `zh-Latn`, and `fonipa` for the IPA.
+ *
+ * | System | `zh-CN` | `zh-TW` |
+ * | --- | --- | --- |
+ * | pinyin | `zh-Latn-CN-pinyin` | `zh-Latn-TW-pinyin` |
+ * | bopomofo | `zh-Bopo-CN` | `zh-Bopo-TW` |
+ * | Wade-Giles | `zh-Latn-CN-wadegile` | `zh-Latn-TW-wadegile` |
+ * | Yale | `zh-Latn-CN` | `zh-Latn-TW` |
+ * | Gwoyeu Romatzyh | `zh-Latn-CN` | `zh-Latn-TW` |
+ * | IPA | `zh-Latn-CN-fonipa` | `zh-Latn-TW-fonipa` |
+ *
+ * **The region is carried by every one of them**, because the distinction it
+ * marks is in the reading rather than in the spelling. 垃圾 is `lājī` under
+ * `zh-CN` and `lèsè` under `zh-TW`, and it is two different words in bopomofo
+ * for the same reason, so the region is read off {@link ConvertOptions.locale}
+ * rather than guessed at.
+ *
+ * Yale and Gwoyeu Romatzyh name themselves nowhere, the registry having no
+ * variant for either, so both go out saying only that they are Mandarin in the
+ * Latin alphabet. See {@link TranscriptionSystem.variant}.
+ *
+ * Tone notation does not enter into it. `hang2` is pinyin spelt with a tone
+ * number, not another romanisation, so it takes the same tag as `háng`.
+ *
+ * The cost is the tag repeated on every syllable, which is the price of
+ * wrapping nothing around the whole conversion. A caller who would rather
+ * declare it once can set `lang: false` and put the same tag on a wrapper of
+ * their own, which is inherited by everything inside it.
+ */
+export function languageTag(locale = DEFAULT_LOCALE, system) {
+    const script = system?.script ?? PINYIN_SCRIPT;
+    const variant = system === undefined ? PINYIN_VARIANT : system.variant;
+    const tag = `zh-${script}-${regionOf(locale)}`;
+    return variant === undefined ? tag : `${tag}-${variant}`;
+}
+//# sourceMappingURL=language-tag.js.map

package/dist/format/language-tag.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-tag.js","sourceRoot":"","sources":["../../src/format/language-tag.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,EAAE,cAAc,EAAe,MAAM,qBAAqB,CAAC;AAGlE;;;GAGG;AACH,MAAM,aAAa,GAAG,MAAM,CAAC;AAC7B,MAAM,cAAc,GAAG,QAAQ,CAAC;AAEhC;;GAEG;AACH,MAAM,UAAU,QAAQ,CAAC,MAAc;IACrC,QAAQ,MAAM,EAAE,CAAC;QACf,KAAK,OAAO,EAAE,CAAC;YACb,OAAO,IAAI,CAAC;QACd,CAAC;QACD,KAAK,OAAO,EAAE,CAAC;YACb,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAyCG;AACH,MAAM,UAAU,WAAW,CACzB,MAAM,GAAW,cAAc,EAC/B,MAA4B;IAE5B,MAAM,MAAM,GAAG,MAAM,EAAE,MAAM,IAAI,aAAa,CAAC;IAC/C,MAAM,OAAO,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC;IACvE,MAAM,GAAG,GAAG,MAAM,MAAM,IAAI,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;IAC/C,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,GAAG,IAAI,OAAO,EAAE,CAAC;AAC3D,CAAC"}

package/dist/format/syllable-span.d.ts

@@ -0,0 +1,10 @@
+import type { ConvertedPiece } from "../decode/convert.js";
+import { type HtmlOptions } from "./html.js";
+/**
+ * Mark up one piece of a conversion.
+ *
+ * A piece that is pinyin's own mark writes nothing where a transcription is
+ * asked for, since {@link writePieces} puts the system's own join in its place.
+ */
+export declare function writePiece(piece: ConvertedPiece, options: HtmlOptions): string;
+//# sourceMappingURL=syllable-span.d.ts.map

package/dist/format/syllable-span.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"syllable-span.d.ts","sourceRoot":"","sources":["../../src/format/syllable-span.ts"],"names":[],"mappings":"AAQA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAG3D,OAAO,EAAU,KAAK,WAAW,EAAE,MAAM,WAAW,CAAC;AAuDrD;;;;;GAKG;AACH,wBAAgB,UAAU,CACxB,KAAK,EAAE,cAAc,EACrB,OAAO,EAAE,WAAW,GACnB,MAAM,CAiCR"}

package/dist/format/syllable-span.js

@@ -0,0 +1,87 @@
+/**
+ * One syllable of a conversion, as the element that carries it.
+ *
+ * The classes, the language tag and the attribute naming what the decode
+ * rejected all hang off a single syllable, so they are decided in one place and
+ * `html.ts` is left with the run.
+ */
+import { isUncertain } from "../decode/confidence.js";
+import { READING_CHARGE } from "../decode/lattice.js";
+import { writeSyllable } from "../syllable/syllable.js";
+import { escape } from "./html.js";
+import { languageTag } from "./language-tag.js";
+import { capitalised, isCapitalised, isPinyinMark } from "./transcribed.js";
+/**
+ * The class every syllable carries, whatever else it carries.
+ */
+const SYLLABLE_CLASS = "py-syllable";
+const UNCERTAIN_CLASS = "py-uncertain";
+/**
+ * The readings a syllable was chosen over, as the reader would write them.
+ *
+ * Only the ones that made it uncertain: a rival dearer than
+ * {@link READING_CHARGE} could only be taken by breaking a dictionary word
+ * apart, so it was never really in the running and listing it would suggest
+ * otherwise.
+ */
+function alternativesOf(confidence, options) {
+    const written = confidence.alternatives
+        .filter((alternative) => alternative.cost < READING_CHARGE)
+        .map((alternative) => alternative.reading
+        .map((syllable) => writeSyllable(syllable, options.notation))
+        .join(""));
+    return [...new Set(written)];
+}
+/**
+ * How one syllable is written, in pinyin or in the system that replaces it.
+ *
+ * The capital comes from the pinyin the conversion settled rather than from the
+ * syllable, for the reason {@link isCapitalised} gives, and only the systems
+ * that write capitals at all take it.
+ */
+function readingOf(piece, syllable, options) {
+    const { transcription } = options;
+    if (transcription === undefined) {
+        return piece.text;
+    }
+    const written = transcription.write(syllable, options.notation !== "none");
+    return transcription.capitals && isCapitalised(piece.text)
+        ? capitalised(written)
+        : written;
+}
+/**
+ * Mark up one piece of a conversion.
+ *
+ * A piece that is pinyin's own mark writes nothing where a transcription is
+ * asked for, since {@link writePieces} puts the system's own join in its place.
+ */
+export function writePiece(piece, options) {
+    const { toneClasses = true, markUncertain = true, lang = true } = options;
+    if (piece.syllable === undefined) {
+        return options.transcription !== undefined && isPinyinMark(piece)
+            ? ""
+            : escape(piece.text);
+    }
+    const classes = [SYLLABLE_CLASS];
+    if (toneClasses && piece.syllable.tone !== undefined) {
+        classes.push(`py-tone-${String(piece.syllable.tone)}`);
+    }
+    const alternatives = markUncertain &&
+        piece.confidence !== undefined &&
+        isUncertain(piece.confidence)
+        ? alternativesOf(piece.confidence, options)
+        : [];
+    if (alternatives.length > 0) {
+        classes.push(UNCERTAIN_CLASS);
+    }
+    const attributes = [`class="${classes.join(" ")}"`];
+    if (lang) {
+        attributes.push(`lang="${languageTag(options.locale, options.transcription)}"`);
+    }
+    if (alternatives.length > 0) {
+        attributes.push(`data-alternatives="${escape(alternatives.join(" "))}"`);
+    }
+    const reading = readingOf(piece, piece.syllable, options);
+    return `<span ${attributes.join(" ")}>${escape(reading)}</span>`;
+}
+//# sourceMappingURL=syllable-span.js.map

package/dist/format/syllable-span.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"syllable-span.js","sourceRoot":"","sources":["../../src/format/syllable-span.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,EAAE,WAAW,EAA0B,MAAM,yBAAyB,CAAC;AAE9E,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAiB,aAAa,EAAE,MAAM,yBAAyB,CAAC;AACvE,OAAO,EAAE,MAAM,EAAoB,MAAM,WAAW,CAAC;AACrD,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,EAAE,WAAW,EAAE,aAAa,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAE5E;;GAEG;AACH,MAAM,cAAc,GAAG,aAAa,CAAC;AAErC,MAAM,eAAe,GAAG,cAAc,CAAC;AAEvC;;;;;;;GAOG;AACH,SAAS,cAAc,CACrB,UAA6B,EAC7B,OAAoB;IAEpB,MAAM,OAAO,GAAG,UAAU,CAAC,YAAY;SACpC,MAAM,CAAC,CAAC,WAAW,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,GAAG,cAAc,CAAC;SAC1D,GAAG,CAAC,CAAC,WAAW,EAAE,EAAE,CACnB,WAAW,CAAC,OAAO;SAChB,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC;SAC5D,IAAI,CAAC,EAAE,CAAC,CACZ,CAAC;IACJ,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC;AAC/B,CAAC;AAED;;;;;;GAMG;AACH,SAAS,SAAS,CAChB,KAAqB,EACrB,QAAkB,EAClB,OAAoB;IAEpB,MAAM,EAAE,aAAa,EAAE,GAAG,OAAO,CAAC;IAClC,IAAI,aAAa,KAAK,SAAS,EAAE,CAAC;QAChC,OAAO,KAAK,CAAC,IAAI,CAAC;IACpB,CAAC;IACD,MAAM,OAAO,GAAG,aAAa,CAAC,KAAK,CAAC,QAAQ,EAAE,OAAO,CAAC,QAAQ,KAAK,MAAM,CAAC,CAAC;IAC3E,OAAO,aAAa,CAAC,QAAQ,IAAI,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC;QACxD,CAAC,CAAC,WAAW,CAAC,OAAO,CAAC;QACtB,CAAC,CAAC,OAAO,CAAC;AACd,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,UAAU,CACxB,KAAqB,EACrB,OAAoB;IAEpB,MAAM,EAAE,WAAW,GAAG,IAAI,EAAE,aAAa,GAAG,IAAI,EAAE,IAAI,GAAG,IAAI,EAAE,GAAG,OAAO,CAAC;IAC1E,IAAI,KAAK,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;QACjC,OAAO,OAAO,CAAC,aAAa,KAAK,SAAS,IAAI,YAAY,CAAC,KAAK,CAAC;YAC/D,CAAC,CAAC,EAAE;YACJ,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACzB,CAAC;IAED,MAAM,OAAO,GAAG,CAAC,cAAc,CAAC,CAAC;IACjC,IAAI,WAAW,IAAI,KAAK,CAAC,QAAQ,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QACrD,OAAO,CAAC,IAAI,CAAC,WAAW,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACzD,CAAC;IACD,MAAM,YAAY,GAChB,aAAa;QACb,KAAK,CAAC,UAAU,KAAK,SAAS;QAC9B,WAAW,CAAC,KAAK,CAAC,UAAU,CAAC;QAC3B,CAAC,CAAC,cAAc,CAAC,KAAK,CAAC,UAAU,EAAE,OAAO,CAAC;QAC3C,CAAC,CAAC,EAAE,CAAC;IACT,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,OAAO,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;IAChC,CAAC;IAED,MAAM,UAAU,GAAG,CAAC,UAAU,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IACpD,IAAI,IAAI,EAAE,CAAC;QACT,UAAU,CAAC,IAAI,CACb,SAAS,WAAW,CAAC,OAAO,CAAC,MAAM,EAAE,OAAO,CAAC,aAAa,CAAC,GAAG,CAC/D,CAAC;IACJ,CAAC;IACD,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,UAAU,CAAC,IAAI,CAAC,sBAAsB,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;IAC3E,CAAC;IACD,MAAM,OAAO,GAAG,SAAS,CAAC,KAAK,EAAE,KAAK,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAC1D,OAAO,SAAS,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC;AACnE,CAAC"}

package/dist/format/transcribed.d.ts

@@ -0,0 +1,36 @@
+/**
+ * What the two transcription paths agree about.
+ *
+ * `toTranscription` writes a string and `toHtml` writes markup, and both have
+ * to put the capitals back and both have to decide what becomes of pinyin's own
+ * orthography. Keeping the answers here is what stops the two from drifting.
+ */
+import type { ConvertedPiece } from "../decode/convert.js";
+/**
+ * Whether a piece was written with a capital.
+ *
+ * Read off the text the conversion produced rather than recomputed, because
+ * capitalisation is settled by `convertPieces` against the whole sentence (a
+ * proper noun, or the first word after a full stop) and none of that survives
+ * in a bare syllable.
+ */
+export declare function isCapitalised(text: string): boolean;
+/**
+ * Put a capital back on a transcribed word.
+ */
+export declare function capitalised(text: string): string;
+/**
+ * Whether a piece is a mark pinyin writes and another system does not.
+ *
+ * The hyphen of `gāngān-jìngjìng` is the whole of it. GB/T 16159 puts it inside
+ * one orthographic word, and every other system joins the syllables of a word
+ * its own way. Wade-Giles hyphenates all of them, bopomofo spaces them, Yale
+ * and Gwoyeu Romatzyh and the IPA run them together, so carrying pinyin's mark
+ * across writes `ㄍㄢ ㄍㄢ-ㄐㄧㄥˋ ㄐㄧㄥˋ`, a hyphen in a script that has none.
+ *
+ * A piece naming no source is pinyin's own, which is the distinction
+ * `sourcePiece` exists to draw. The space between two words is pinyin's too and
+ * is kept, since every system puts one there.
+ */
+export declare function isPinyinMark(piece: ConvertedPiece): boolean;
+//# sourceMappingURL=transcribed.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"transcribed.d.ts","sourceRoot":"","sources":["../../src/format/transcribed.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAE3D;;;;;;;GAOG;AACH,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAGnD;AAED;;GAEG;AACH,wBAAgB,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAGhD;AAED;;;;;;;;;;;;GAYG;AACH,wBAAgB,YAAY,CAAC,KAAK,EAAE,cAAc,GAAG,OAAO,CAM3D"}

package/dist/format/transcribed.js

@@ -0,0 +1,38 @@
+/**
+ * Whether a piece was written with a capital.
+ *
+ * Read off the text the conversion produced rather than recomputed, because
+ * capitalisation is settled by `convertPieces` against the whole sentence (a
+ * proper noun, or the first word after a full stop) and none of that survives
+ * in a bare syllable.
+ */
+export function isCapitalised(text) {
+    const first = /\p{L}/u.exec(text)?.[0];
+    return first !== undefined && first !== first.toLowerCase();
+}
+/**
+ * Put a capital back on a transcribed word.
+ */
+export function capitalised(text) {
+    // Not global: only the first letter of the word takes the capital.
+    return text.replace(/\p{L}/u, (letter) => letter.toUpperCase());
+}
+/**
+ * Whether a piece is a mark pinyin writes and another system does not.
+ *
+ * The hyphen of `gāngān-jìngjìng` is the whole of it. GB/T 16159 puts it inside
+ * one orthographic word, and every other system joins the syllables of a word
+ * its own way. Wade-Giles hyphenates all of them, bopomofo spaces them, Yale
+ * and Gwoyeu Romatzyh and the IPA run them together, so carrying pinyin's mark
+ * across writes `ㄍㄢ ㄍㄢ-ㄐㄧㄥˋ ㄐㄧㄥˋ`, a hyphen in a script that has none.
+ *
+ * A piece naming no source is pinyin's own, which is the distinction
+ * `sourcePiece` exists to draw. The space between two words is pinyin's too and
+ * is kept, since every system puts one there.
+ */
+export function isPinyinMark(piece) {
+    return (piece.syllable === undefined &&
+        piece.source === undefined &&
+        piece.text.trim() !== "");
+}
+//# sourceMappingURL=transcribed.js.map

package/dist/format/transcribed.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"transcribed.js","sourceRoot":"","sources":["../../src/format/transcribed.ts"],"names":[],"mappings":"AASA;;;;;;;GAOG;AACH,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACvC,OAAO,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,KAAK,CAAC,WAAW,EAAE,CAAC;AAC9D,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,mEAAmE;IACnE,OAAO,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC,CAAC;AAClE,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,YAAY,CAAC,KAAqB;IAChD,OAAO,CACL,KAAK,CAAC,QAAQ,KAAK,SAAS;QAC5B,KAAK,CAAC,MAAM,KAAK,SAAS;QAC1B,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,CACzB,CAAC;AACJ,CAAC"}

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

@@ -1 +1 @@
-{"version":3,"file":"transcription.d.ts","sourceRoot":"","sources":["../../src/format/transcription.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,KAAK,cAAc,EACnB,KAAK,cAAc,EAEpB,MAAM,sBAAsB,CAAC;AAC9B,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAE9D,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAC;AAExD;;;;;;;;GAQG;AACH,MAAM,MAAM,SAAS,GAAG,CAAC,SAAS,EAAE,SAAS,QAAQ,EAAE,KAAK,MAAM,CAAC;AAEnE;;;;;;;;;;;;;;GAcG;AACH,MAAM,WAAW,oBAAoB;IACnC,QAAQ,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;CAC7B;AAuBD;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AACH,wBAAgB,eAAe,CAC7B,MAAM,EAAE,SAAS,cAAc,EAAE,EACjC,SAAS,EAAE,SAAS,EACpB,OAAO,GAAE,oBAAyB,GACjC,MAAM,CA6BR;AAED;;;;;;;;;;GAUG;AACH,wBAAgB,kBAAkB,CAChC,UAAU,EAAE,UAAU,EACtB,IAAI,EAAE,MAAM,EACZ,OAAO,GAAE,cAAmB,GAC3B,MAAM,CASR"}
+{"version":3,"file":"transcription.d.ts","sourceRoot":"","sources":["../../src/format/transcription.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,KAAK,cAAc,EACnB,KAAK,cAAc,EAEpB,MAAM,sBAAsB,CAAC;AAC9B,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAE9D,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAC;AAGxD;;;;;;;;GAQG;AACH,MAAM,MAAM,SAAS,GAAG,CAAC,SAAS,EAAE,SAAS,QAAQ,EAAE,KAAK,MAAM,CAAC;AAEnE;;;;;;;;;;;;;;GAcG;AACH,MAAM,WAAW,oBAAoB;IACnC,QAAQ,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;CAC7B;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AACH,wBAAgB,eAAe,CAC7B,MAAM,EAAE,SAAS,cAAc,EAAE,EACjC,SAAS,EAAE,SAAS,EACpB,OAAO,GAAE,oBAAyB,GACjC,MAAM,CAmCR;AAED;;;;;;;;;;GAUG;AACH,wBAAgB,kBAAkB,CAChC,UAAU,EAAE,UAAU,EACtB,IAAI,EAAE,MAAM,EACZ,OAAO,GAAE,cAAmB,GAC3B,MAAM,CASR"}

package/dist/format/transcription.js

@@ -1,24 +1,6 @@
 import { convertPieces, } from "../decode/convert.js";
 import { writeWadeGilesWord } from "../transcription/wade-giles.js";
-/**
- * Whether a piece was written with a capital.
- *
- * Read off the text the conversion produced rather than recomputed, because
- * capitalisation is settled by {@link convertPieces} against the whole sentence
- * — a proper noun, or the first word after a full stop — and none of that
- * survives in a bare {@link Syllable}.
- */
-function isCapitalised(text) {
-    const first = /\p{L}/u.exec(text)?.[0];
-    return first !== undefined && first !== first.toLowerCase();
-}
-/**
- * Put a capital back on a transcribed word.
- */
-function capitalised(text) {
-    // Not global: only the first letter of the word takes the capital.
-    return text.replace(/\p{L}/u, (letter) => letter.toUpperCase());
-}
+import { capitalised, isCapitalised, isPinyinMark } from "./transcribed.js";
 /**
  * Write a conversion in another transcription system, word by word.
  *
@@ -64,6 +46,12 @@ export function toTranscription(pieces, writeWord, options = {}) {
     };
     for (const piece of pieces) {
         if (piece.syllable === undefined) {
+            // A mark pinyin writes and this system does not is dropped rather than
+            // ending the word, so the system's own join covers it. See
+            // {@link isPinyinMark}.
+            if (isPinyinMark(piece)) {
+                continue;
+            }
             flush();
             written.push(piece.text);
             continue;

package/dist/format/transcription.js.map

@@ -1 +1 @@
-{"version":3,"file":"transcription.js","sourceRoot":"","sources":["../../src/format/transcription.ts"],"names":[],"mappings":"AAAA,OAAO,EAGL,aAAa,GACd,MAAM,sBAAsB,CAAC;AAE9B,OAAO,EAAE,kBAAkB,EAAE,MAAM,gCAAgC,CAAC;AAiCpE;;;;;;;GAOG;AACH,SAAS,aAAa,CAAC,IAAY;IACjC,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACvC,OAAO,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,KAAK,CAAC,WAAW,EAAE,CAAC;AAC9D,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,IAAY;IAC/B,mEAAmE;IACnE,OAAO,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC,CAAC;AAClE,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AACH,MAAM,UAAU,eAAe,CAC7B,MAAiC,EACjC,SAAoB,EACpB,OAAO,GAAyB,EAAE;IAElC,MAAM,EAAE,QAAQ,GAAG,IAAI,EAAE,GAAG,OAAO,CAAC;IACpC,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,IAAI,GAAe,EAAE,CAAC;IAC1B,IAAI,UAAU,GAAG,KAAK,CAAC;IAEvB,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACtB,OAAO;QACT,CAAC;QACD,MAAM,GAAG,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;QAC5B,OAAO,CAAC,IAAI,CAAC,UAAU,IAAI,QAAQ,CAAC,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;QAC9D,IAAI,GAAG,EAAE,CAAC;QACV,UAAU,GAAG,KAAK,CAAC;IACrB,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,KAAK,CAAC,IAAI,CAAC,CAAC;YACzB,SAAS;QACX,CAAC;QACD,wEAAwE;QACxE,yDAAyD;QACzD,UAAU,KAAK,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC9D,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;IAC5B,CAAC;IACD,KAAK,EAAE,CAAC;IACR,OAAO,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC1B,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,UAAU,kBAAkB,CAChC,UAAsB,EACtB,IAAY,EACZ,OAAO,GAAmB,EAAE;IAE5B,OAAO,eAAe,CACpB,aAAa,CAAC,UAAU,EAAE,IAAI,EAAE,OAAO,CAAC,EACxC,CAAC,SAAS,EAAE,EAAE,CACZ,kBAAkB,CAChB,SAAS,EACT,OAAO,CAAC,QAAQ,KAAK,MAAM,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CACrD,CACJ,CAAC;AACJ,CAAC"}
+{"version":3,"file":"transcription.js","sourceRoot":"","sources":["../../src/format/transcription.ts"],"names":[],"mappings":"AAAA,OAAO,EAGL,aAAa,GACd,MAAM,sBAAsB,CAAC;AAE9B,OAAO,EAAE,kBAAkB,EAAE,MAAM,gCAAgC,CAAC;AAEpE,OAAO,EAAE,WAAW,EAAE,aAAa,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAgC5E;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AACH,MAAM,UAAU,eAAe,CAC7B,MAAiC,EACjC,SAAoB,EACpB,OAAO,GAAyB,EAAE;IAElC,MAAM,EAAE,QAAQ,GAAG,IAAI,EAAE,GAAG,OAAO,CAAC;IACpC,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,IAAI,GAAe,EAAE,CAAC;IAC1B,IAAI,UAAU,GAAG,KAAK,CAAC;IAEvB,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACtB,OAAO;QACT,CAAC;QACD,MAAM,GAAG,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;QAC5B,OAAO,CAAC,IAAI,CAAC,UAAU,IAAI,QAAQ,CAAC,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;QAC9D,IAAI,GAAG,EAAE,CAAC;QACV,UAAU,GAAG,KAAK,CAAC;IACrB,CAAC,CAAC;IAEF,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,IAAI,KAAK,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;YACjC,uEAAuE;YACvE,2DAA2D;YAC3D,wBAAwB;YACxB,IAAI,YAAY,CAAC,KAAK,CAAC,EAAE,CAAC;gBACxB,SAAS;YACX,CAAC;YACD,KAAK,EAAE,CAAC;YACR,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YACzB,SAAS;QACX,CAAC;QACD,wEAAwE;QACxE,yDAAyD;QACzD,UAAU,KAAK,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC9D,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;IAC5B,CAAC;IACD,KAAK,EAAE,CAAC;IACR,OAAO,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC1B,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,UAAU,kBAAkB,CAChC,UAAsB,EACtB,IAAY,EACZ,OAAO,GAAmB,EAAE;IAE5B,OAAO,eAAe,CACpB,aAAa,CAAC,UAAU,EAAE,IAAI,EAAE,OAAO,CAAC,EACxC,CAAC,SAAS,EAAE,EAAE,CACZ,kBAAkB,CAChB,SAAS,EACT,OAAO,CAAC,QAAQ,KAAK,MAAM,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CACrD,CACJ,CAAC;AACJ,CAAC"}