wolfy-reader 0.1.0

package/dist/core/lookup.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"lookup.js","sourceRoot":"","sources":["../../src/core/lookup.ts"],"names":[],"mappings":"AAEA;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAC,QAA4B;IACxD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAmB,CAAC;IACxC,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;QAC/B,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YAAE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,EAAE,OAAO,CAAC,CAAC;IAC3D,CAAC;IACD,OAAO,CAAC,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;AAC9B,CAAC"}

package/dist/core/open.d.ts

@@ -0,0 +1,12 @@
+import type { Book } from './book.ts';
+import type { BookFormat } from './format.ts';
+import type { BookInput } from './source.ts';
+import type { StorageAdapter } from './storage.ts';
+export interface OpenOptions {
+    /** Tried in order; the first format whose sniff claims the input decodes it. */
+    readonly formats: readonly BookFormat[];
+    readonly storage?: StorageAdapter;
+}
+/** Bytes in, Book out: runs the registered sniffers in order and decodes with the first claimant. */
+export declare function open(input: BookInput, options: OpenOptions): Promise<Book>;
+//# sourceMappingURL=open.d.ts.map

package/dist/core/open.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"open.d.ts","sourceRoot":"","sources":["../../src/core/open.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AACtC,OAAO,KAAK,EAAE,UAAU,EAAiB,MAAM,aAAa,CAAC;AAC7D,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,aAAa,CAAC;AAC7C,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,cAAc,CAAC;AAInD,MAAM,WAAW,WAAW;IAC1B,gFAAgF;IAChF,QAAQ,CAAC,OAAO,EAAE,SAAS,UAAU,EAAE,CAAC;IACxC,QAAQ,CAAC,OAAO,CAAC,EAAE,cAAc,CAAC;CACnC;AAED,qGAAqG;AACrG,wBAAsB,IAAI,CAAC,KAAK,EAAE,SAAS,EAAE,OAAO,EAAE,WAAW,GAAG,OAAO,CAAC,IAAI,CAAC,CAgBhF"}

package/dist/core/open.js

@@ -0,0 +1,23 @@
+import { BookError, CorruptContainerError, UnrecognizedFormatError } from './errors.js';
+import { toByteSource } from './source.js';
+/** Bytes in, Book out: runs the registered sniffers in order and decodes with the first claimant. */
+export async function open(input, options) {
+    const source = toByteSource(input);
+    const context = options.storage === undefined ? {} : { storage: options.storage };
+    for (const format of options.formats) {
+        if (!(await format.sniff(source)))
+            continue;
+        try {
+            return await format.decode(source, context);
+        }
+        catch (error) {
+            if (error instanceof BookError)
+                throw error;
+            throw new CorruptContainerError(`${format.name} failed to decode the input`, {
+                cause: error,
+            });
+        }
+    }
+    throw new UnrecognizedFormatError('no registered format recognizes this input');
+}
+//# sourceMappingURL=open.js.map

package/dist/core/open.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"open.js","sourceRoot":"","sources":["../../src/core/open.ts"],"names":[],"mappings":"AAIA,OAAO,EAAE,SAAS,EAAE,qBAAqB,EAAE,uBAAuB,EAAE,MAAM,aAAa,CAAC;AACxF,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAQ3C,qGAAqG;AACrG,MAAM,CAAC,KAAK,UAAU,IAAI,CAAC,KAAgB,EAAE,OAAoB;IAC/D,MAAM,MAAM,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IACnC,MAAM,OAAO,GACX,OAAO,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,CAAC;IACpE,KAAK,MAAM,MAAM,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;QACrC,IAAI,CAAC,CAAC,MAAM,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;YAAE,SAAS;QAC5C,IAAI,CAAC;YACH,OAAO,MAAM,MAAM,CAAC,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;QAC9C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,IAAI,KAAK,YAAY,SAAS;gBAAE,MAAM,KAAK,CAAC;YAC5C,MAAM,IAAI,qBAAqB,CAAC,GAAG,MAAM,CAAC,IAAI,6BAA6B,EAAE;gBAC3E,KAAK,EAAE,KAAK;aACb,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IACD,MAAM,IAAI,uBAAuB,CAAC,4CAA4C,CAAC,CAAC;AAClF,CAAC"}

package/dist/core/position.d.ts

@@ -0,0 +1,121 @@
+/**
+ * A content-addressed anchor into a section's text, Hypothesis-style. A match is
+ * found by content (the exact quote plus surrounding context), never by offset:
+ * `offset` exists only to break ties between identical quotes and is never
+ * load-bearing on its own.
+ */
+export interface TextAnchor {
+    /** The verbatim run of text the position points at. */
+    readonly exact: string;
+    /** Text immediately before `exact`, bounded length; boundary-snapped. */
+    readonly prefix: string;
+    /** Text immediately after `exact`, bounded length; boundary-snapped. */
+    readonly suffix: string;
+    /**
+     * Grapheme-cluster offset of `exact` within the section text. A tiebreak only —
+     * used to pick among duplicate quotes, never to locate a match alone.
+     */
+    readonly offset: number;
+}
+/**
+ * A reading position within a book. Its `serialized` form is an opaque string a
+ * host persists verbatim and never parses; `progress` rides alongside for host
+ * UI (scrollbars, "42%") without decoding the string. Backed by a content anchor
+ * so it survives reflow, font-size, and layout changes.
+ */
+export interface Position {
+    /** The section this position falls in. */
+    readonly sectionId: string;
+    /** Fraction through the section, 0..1. For host progress UI only. */
+    readonly progress: number;
+    /** Opaque, versioned, round-trippable interchange string. Hosts store it as-is. */
+    readonly serialized: string;
+    /** The content anchor. Resolution matches on this, not on `progress`. */
+    readonly anchor: TextAnchor;
+}
+/**
+ * Where a `Position` resolved to in a piece of target text. Offsets are grapheme
+ * indices into that text, matching `TextAnchor.offset`. A resolution *miss* is not
+ * this type — it is `undefined`, returned by `resolvePosition`.
+ */
+export interface ResolvedPosition {
+    readonly sectionId: string;
+    /** Grapheme index where the matched quote starts in the target text. */
+    readonly offset: number;
+    /** Length of the matched quote in graphemes. */
+    readonly length: number;
+}
+export interface CapturePositionOptions {
+    /**
+     * Locale for `Intl.Segmenter` word snapping. Defaults to the runtime default
+     * locale. Grapheme snapping is locale-independent.
+     */
+    readonly locale?: string;
+    /** Maximum length (in graphemes) of the captured quote. */
+    readonly quoteLength?: number;
+    /** Maximum length (in graphemes) of each context window. */
+    readonly contextLength?: number;
+}
+/**
+ * Captures a `Position` at `offset` in a section's plain text. Core is headless:
+ * the caller supplies already-extracted text and a UTF-16 code-unit offset into
+ * it; no DOM, HTML, or Range is touched here. The offset is snapped to grapheme
+ * and word boundaries with `Intl.Segmenter` so an anchor never splits a surrogate
+ * pair, combining sequence, or emoji cluster.
+ */
+export declare function capturePosition(text: string, offset: number, sectionId: string, options?: CapturePositionOptions): Position;
+/** Serializes a position to its opaque, versioned interchange string. */
+export declare function serializePosition(position: Position): string;
+/** One sentence of a section, with a `Position` anchored over its whole span. */
+export interface SentenceRange {
+    /** The sentence text, trailing whitespace trimmed. */
+    readonly text: string;
+    /**
+     * A `Position` whose anchored quote is the whole sentence, so `resolvePosition`
+     * returns its full span — a host can `goTo` it (jump) and `decorate` it
+     * (highlight the entire sentence, not just its start).
+     */
+    readonly position: Position;
+}
+/**
+ * Segments a section's plain text into sentences with `Intl.Segmenter`, each
+ * carrying a `Position` anchored over the whole sentence. This is the TTS *enabler*
+ * (PLAN M3-6): the library exposes sentence ranges and reuses `decorate` as the
+ * highlight primitive so a host can build text-to-speech on top — it speaks nothing
+ * and stores nothing. Whitespace-only segments are dropped. Headless: the caller
+ * supplies already-extracted text, exactly like `capturePosition`.
+ *
+ * `Intl.Segmenter` follows UAX #29, which ends a "sentence" at every hard line
+ * break — but the reading text is section `textContent`, whose source newlines the
+ * rendered page collapses via `white-space:normal`. Segmenting it verbatim would
+ * split one visual sentence at every wrapped source line (a heading or short line
+ * becoming a two-word "sentence"). So adjacent Intl segments are merged until the
+ * run actually ends on sentence-terminal punctuation; the anchor still spans the
+ * raw text (newlines included) so it content-matches the frame's section text,
+ * while the returned `text` collapses whitespace runs to single spaces for a clean
+ * TTS string. Abbreviations (`J. R. R.`) still split — that break is Intl's, and a
+ * period genuinely can end a sentence, so we do not second-guess it.
+ */
+export declare function segmentSentences(text: string, sectionId: string, options?: CapturePositionOptions): SentenceRange[];
+/**
+ * Parses a serialized position back into a `Position`. Round-trips losslessly
+ * with `serializePosition`. Throws `CorruptContainerError` when the string is
+ * malformed or carries an unknown version — a version this reader cannot honor is
+ * a structural failure the host catches by class, distinct from a resolution miss
+ * (which the resolver returns as a value, never an error).
+ */
+export declare function parsePosition(serialized: string): Position;
+/**
+ * Resolves a `Position` against a piece of section text — which may have changed
+ * since capture — returning where its anchor now lands, or `undefined` when the
+ * quote is gone. A miss is a value, never a throw: matching live content is an
+ * expected outcome the caller degrades on (skip the restore, drop the decoration),
+ * mirroring `Section.resolve()` returning `undefined`.
+ *
+ * The match is by content. Every occurrence of the exact quote is found, then
+ * scored by how much of the stored prefix/suffix context still surrounds it; the
+ * best-scoring occurrence wins. The stored offset breaks a tie only when context
+ * cannot — it never selects a location the content contradicts.
+ */
+export declare function resolvePosition(position: Position, text: string): ResolvedPosition | undefined;
+//# sourceMappingURL=position.d.ts.map

package/dist/core/position.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"position.d.ts","sourceRoot":"","sources":["../../src/core/position.ts"],"names":[],"mappings":"AAQA;;;;;GAKG;AACH,MAAM,WAAW,UAAU;IACzB,uDAAuD;IACvD,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,yEAAyE;IACzE,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,wEAAwE;IACxE,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB;;;OAGG;IACH,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;CACzB;AAED;;;;;GAKG;AACH,MAAM,WAAW,QAAQ;IACvB,0CAA0C;IAC1C,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,qEAAqE;IACrE,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,mFAAmF;IACnF,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAC5B,yEAAyE;IACzE,QAAQ,CAAC,MAAM,EAAE,UAAU,CAAC;CAC7B;AAED;;;;GAIG;AACH,MAAM,WAAW,gBAAgB;IAC/B,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,wEAAwE;IACxE,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,gDAAgD;IAChD,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;CACzB;AAED,MAAM,WAAW,sBAAsB;IACrC;;;OAGG;IACH,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC;IACzB,2DAA2D;IAC3D,QAAQ,CAAC,WAAW,CAAC,EAAE,MAAM,CAAC;IAC9B,4DAA4D;IAC5D,QAAQ,CAAC,aAAa,CAAC,EAAE,MAAM,CAAC;CACjC;AAmBD;;;;;;GAMG;AACH,wBAAgB,eAAe,CAC7B,IAAI,EAAE,MAAM,EACZ,MAAM,EAAE,MAAM,EACd,SAAS,EAAE,MAAM,EACjB,OAAO,GAAE,sBAA2B,GACnC,QAAQ,CAEV;AAwDD,yEAAyE;AACzE,wBAAgB,iBAAiB,CAAC,QAAQ,EAAE,QAAQ,GAAG,MAAM,CAE5D;AAED,iFAAiF;AACjF,MAAM,WAAW,aAAa;IAC5B,sDAAsD;IACtD,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB;;;;OAIG;IACH,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC;CAC7B;AAED;;;;;;;;;;;;;;;;;;GAkBG;AACH,wBAAgB,gBAAgB,CAC9B,IAAI,EAAE,MAAM,EACZ,SAAS,EAAE,MAAM,EACjB,OAAO,GAAE,sBAA2B,GACnC,aAAa,EAAE,CAwCjB;AAyBD;;;;;;GAMG;AACH,wBAAgB,aAAa,CAAC,UAAU,EAAE,MAAM,GAAG,QAAQ,CA8B1D;AAED;;;;;;;;;;;GAWG;AACH,wBAAgB,eAAe,CAC7B,QAAQ,EAAE,QAAQ,EAClB,IAAI,EAAE,MAAM,GACX,gBAAgB,GAAG,SAAS,CA8B9B"}

package/dist/core/position.js

@@ -0,0 +1,327 @@
+import { CorruptContainerError } from './errors.js';
+import { codeUnitOffsetToGraphemeIndex, countGraphemes, segmentGraphemes, } from './graphemes.js';
+const DEFAULT_QUOTE_LENGTH = 32;
+const DEFAULT_CONTEXT_LENGTH = 32;
+/** Version tag that leads every serialized position. Version-first, non-negotiable. */
+const SERIAL_PREFIX = 'wr1:';
+const SERIAL_VERSION = 1;
+/**
+ * Captures a `Position` at `offset` in a section's plain text. Core is headless:
+ * the caller supplies already-extracted text and a UTF-16 code-unit offset into
+ * it; no DOM, HTML, or Range is touched here. The offset is snapped to grapheme
+ * and word boundaries with `Intl.Segmenter` so an anchor never splits a surrogate
+ * pair, combining sequence, or emoji cluster.
+ */
+export function capturePosition(text, offset, sectionId, options = {}) {
+    return captureSegmented(segmentText(text), offset, sectionId, options);
+}
+function segmentText(text) {
+    return { text, graphemes: segmentGraphemes(text), wordStarts: null };
+}
+/** `capturePosition` over an already-segmented text; behavior is identical. */
+function captureSegmented(segmented, offset, sectionId, options) {
+    const { graphemes } = segmented;
+    const total = graphemes.length;
+    const rawIndex = codeUnitOffsetToGraphemeIndex(graphemes, offset);
+    const start = snapToWordStart(segmented, rawIndex, options.locale);
+    const quoteLength = options.quoteLength ?? DEFAULT_QUOTE_LENGTH;
+    const contextLength = options.contextLength ?? DEFAULT_CONTEXT_LENGTH;
+    const quoteEnd = Math.min(start + quoteLength, total);
+    const prefixStart = Math.max(start - contextLength, 0);
+    const suffixEnd = Math.min(quoteEnd + contextLength, total);
+    const anchor = {
+        exact: join(graphemes, start, quoteEnd),
+        prefix: join(graphemes, prefixStart, start),
+        suffix: join(graphemes, quoteEnd, suffixEnd),
+        offset: start,
+    };
+    const progress = total === 0 ? 0 : start / total;
+    return {
+        sectionId,
+        progress,
+        anchor,
+        serialized: serializeParts(sectionId, progress, anchor),
+    };
+}
+/** Serializes a position to its opaque, versioned interchange string. */
+export function serializePosition(position) {
+    return serializeParts(position.sectionId, position.progress, position.anchor);
+}
+/**
+ * Segments a section's plain text into sentences with `Intl.Segmenter`, each
+ * carrying a `Position` anchored over the whole sentence. This is the TTS *enabler*
+ * (PLAN M3-6): the library exposes sentence ranges and reuses `decorate` as the
+ * highlight primitive so a host can build text-to-speech on top — it speaks nothing
+ * and stores nothing. Whitespace-only segments are dropped. Headless: the caller
+ * supplies already-extracted text, exactly like `capturePosition`.
+ *
+ * `Intl.Segmenter` follows UAX #29, which ends a "sentence" at every hard line
+ * break — but the reading text is section `textContent`, whose source newlines the
+ * rendered page collapses via `white-space:normal`. Segmenting it verbatim would
+ * split one visual sentence at every wrapped source line (a heading or short line
+ * becoming a two-word "sentence"). So adjacent Intl segments are merged until the
+ * run actually ends on sentence-terminal punctuation; the anchor still spans the
+ * raw text (newlines included) so it content-matches the frame's section text,
+ * while the returned `text` collapses whitespace runs to single spaces for a clean
+ * TTS string. Abbreviations (`J. R. R.`) still split — that break is Intl's, and a
+ * period genuinely can end a sentence, so we do not second-guess it.
+ */
+export function segmentSentences(text, sectionId, options = {}) {
+    const out = [];
+    // Segment graphemes (and, lazily, words) once for the whole text; every
+    // sentence capture reuses the same arrays instead of re-segmenting — this is
+    // what keeps a long chapter linear instead of quadratic.
+    const segmented = segmentText(text);
+    // Accumulate a run of Intl segments and its code-unit start until the run ends on
+    // sentence-terminal punctuation, then anchor over the raw span. A run start of -1
+    // means no run is open (all leading/inter-sentence whitespace is skipped).
+    let runStart = -1;
+    let runText = '';
+    const flush = () => {
+        const trimmed = runText.replace(/\s+$/u, '');
+        if (trimmed.trim().length !== 0) {
+            const quoteLength = countGraphemes(trimmed);
+            // Anchor over the whole (raw, newline-bearing) run so decorate highlights all
+            // of it against the frame's section text.
+            const position = captureSegmented(segmented, runStart, sectionId, {
+                ...options,
+                quoteLength,
+            });
+            out.push({ text: collapseWhitespace(trimmed), position });
+        }
+        runStart = -1;
+        runText = '';
+    };
+    for (const segment of sentenceSegmenter(options.locale).segment(text)) {
+        if (segment.segment.trim().length === 0) {
+            // Whitespace-only Intl segment: part of an open run, else inter-sentence space.
+            if (runStart >= 0)
+                runText += segment.segment;
+            continue;
+        }
+        if (runStart < 0)
+            runStart = segment.index;
+        runText += segment.segment;
+        if (endsOnSentenceTerminal(runText))
+            flush();
+    }
+    if (runStart >= 0)
+        flush();
+    return out;
+}
+/** Sentence-final punctuation, incl. CJK and other-script terminators. */
+const SENTENCE_TERMINAL = /[.!?…。！？؟।]/u;
+/** Trailing closing punctuation a terminator may hide behind (quotes, brackets). */
+const TRAILING_CLOSERS = /[)\]}"'”’»›]+$/u;
+/**
+ * Whether a run, ignoring trailing whitespace and closing punctuation, ends on a
+ * sentence terminator — the seam that decides a real sentence break from a mere
+ * line wrap. `"the lazy dog."` and `"he said 'go.'"` end a sentence; `"The quick
+ * brown\n"` (a wrapped line) does not.
+ */
+function endsOnSentenceTerminal(run) {
+    const core = run.replace(/\s+$/u, '').replace(TRAILING_CLOSERS, '');
+    const chars = [...core];
+    const last = chars[chars.length - 1];
+    return last !== undefined && SENTENCE_TERMINAL.test(last);
+}
+/** Collapses every run of whitespace to a single space and trims the ends. */
+function collapseWhitespace(text) {
+    return text.replace(/\s+/gu, ' ').trim();
+}
+/**
+ * Parses a serialized position back into a `Position`. Round-trips losslessly
+ * with `serializePosition`. Throws `CorruptContainerError` when the string is
+ * malformed or carries an unknown version — a version this reader cannot honor is
+ * a structural failure the host catches by class, distinct from a resolution miss
+ * (which the resolver returns as a value, never an error).
+ */
+export function parsePosition(serialized) {
+    if (!serialized.startsWith(SERIAL_PREFIX)) {
+        throw new CorruptContainerError(`not a wolfy-reader position: missing '${SERIAL_PREFIX}' version prefix`);
+    }
+    let payload;
+    try {
+        payload = JSON.parse(serialized.slice(SERIAL_PREFIX.length));
+    }
+    catch (cause) {
+        throw new CorruptContainerError('position payload is not valid JSON', { cause });
+    }
+    if (!isSerialPayload(payload)) {
+        throw new CorruptContainerError('position payload is missing required fields');
+    }
+    if (payload.v !== SERIAL_VERSION) {
+        throw new CorruptContainerError(`unsupported position version ${payload.v}; this reader understands version ${SERIAL_VERSION}`);
+    }
+    const anchor = {
+        exact: payload.e,
+        prefix: payload.pre,
+        suffix: payload.suf,
+        offset: payload.o,
+    };
+    return { sectionId: payload.s, progress: payload.p, anchor, serialized };
+}
+/**
+ * Resolves a `Position` against a piece of section text — which may have changed
+ * since capture — returning where its anchor now lands, or `undefined` when the
+ * quote is gone. A miss is a value, never a throw: matching live content is an
+ * expected outcome the caller degrades on (skip the restore, drop the decoration),
+ * mirroring `Section.resolve()` returning `undefined`.
+ *
+ * The match is by content. Every occurrence of the exact quote is found, then
+ * scored by how much of the stored prefix/suffix context still surrounds it; the
+ * best-scoring occurrence wins. The stored offset breaks a tie only when context
+ * cannot — it never selects a location the content contradicts.
+ */
+export function resolvePosition(position, text) {
+    const { anchor, sectionId } = position;
+    const graphemes = segmentGraphemes(text);
+    if (anchor.exact === '') {
+        const offset = Math.min(anchor.offset, graphemes.length);
+        return { sectionId, offset, length: 0 };
+    }
+    const occurrences = findGraphemeOccurrences(graphemes, anchor.exact);
+    if (occurrences.length === 0)
+        return undefined;
+    const quoteLength = segmentGraphemes(anchor.exact).length;
+    // Segment the stored context once, outside the occurrence loop.
+    const prefix = segmentGraphemes(anchor.prefix).map((g) => g.segment);
+    const suffix = segmentGraphemes(anchor.suffix).map((g) => g.segment);
+    let best;
+    for (const offset of occurrences) {
+        const score = contextScore(graphemes, offset, quoteLength, prefix, suffix);
+        if (best === undefined ||
+            score > best.score ||
+            (score === best.score &&
+                Math.abs(offset - anchor.offset) < Math.abs(best.offset - anchor.offset))) {
+            best = { offset, score };
+        }
+    }
+    return { sectionId, offset: best.offset, length: quoteLength };
+}
+/** Grapheme start indices at which `quote` occurs in `graphemes`. */
+function findGraphemeOccurrences(graphemes, quote) {
+    const quoteGraphemes = segmentGraphemes(quote).map((g) => g.segment);
+    const hits = [];
+    if (quoteGraphemes.length === 0)
+        return hits;
+    const last = graphemes.length - quoteGraphemes.length;
+    for (let start = 0; start <= last; start += 1) {
+        let matched = true;
+        for (let offset = 0; offset < quoteGraphemes.length; offset += 1) {
+            if (graphemes[start + offset].segment !== quoteGraphemes[offset]) {
+                matched = false;
+                break;
+            }
+        }
+        if (matched)
+            hits.push(start);
+    }
+    return hits;
+}
+/**
+ * Count of trailing prefix graphemes + leading suffix graphemes that still match.
+ * `prefix`/`suffix` are the anchor's context pre-segmented into grapheme strings —
+ * segmented once by the caller, never per occurrence.
+ */
+function contextScore(graphemes, quoteStart, quoteLength, prefix, suffix) {
+    let score = 0;
+    for (let back = 1; back <= prefix.length; back += 1) {
+        const here = graphemes[quoteStart - back];
+        const want = prefix[prefix.length - back];
+        if (here === undefined || here.segment !== want)
+            break;
+        score += 1;
+    }
+    const afterStart = quoteStart + quoteLength;
+    for (let ahead = 0; ahead < suffix.length; ahead += 1) {
+        const here = graphemes[afterStart + ahead];
+        if (here === undefined || here.segment !== suffix[ahead])
+            break;
+        score += 1;
+    }
+    return score;
+}
+function serializeParts(sectionId, progress, anchor) {
+    const payload = {
+        v: SERIAL_VERSION,
+        s: sectionId,
+        p: progress,
+        e: anchor.exact,
+        pre: anchor.prefix,
+        suf: anchor.suffix,
+        o: anchor.offset,
+    };
+    return SERIAL_PREFIX + JSON.stringify(payload);
+}
+function isSerialPayload(value) {
+    if (typeof value !== 'object' || value === null)
+        return false;
+    const record = value;
+    return (typeof record.v === 'number' &&
+        typeof record.s === 'string' &&
+        typeof record.p === 'number' &&
+        typeof record.e === 'string' &&
+        typeof record.pre === 'string' &&
+        typeof record.suf === 'string' &&
+        typeof record.o === 'number');
+}
+const wordSegmenters = new Map();
+const sentenceSegmenters = new Map();
+function sentenceSegmenter(locale) {
+    const key = locale ?? '';
+    let segmenter = sentenceSegmenters.get(key);
+    if (segmenter === undefined) {
+        segmenter = new Intl.Segmenter(locale, { granularity: 'sentence' });
+        sentenceSegmenters.set(key, segmenter);
+    }
+    return segmenter;
+}
+function wordSegmenter(locale) {
+    const key = locale ?? '';
+    let segmenter = wordSegmenters.get(key);
+    if (segmenter === undefined) {
+        segmenter = new Intl.Segmenter(locale, { granularity: 'word' });
+        wordSegmenters.set(key, segmenter);
+    }
+    return segmenter;
+}
+function join(graphemes, start, end) {
+    let out = '';
+    for (let index = start; index < end; index += 1)
+        out += graphemes[index].segment;
+    return out;
+}
+/**
+ * Snaps a grapheme index to the start of the word segment it falls in, so a
+ * captured quote begins on a word boundary rather than mid-token. Word segments
+ * tile the text, so the containing segment is the one with the greatest start
+ * offset <= the grapheme's offset — a binary search over the word-start array,
+ * which is built once per `SegmentedText` and reused by every capture.
+ */
+function snapToWordStart(segmented, index, locale) {
+    const { graphemes } = segmented;
+    if (index <= 0 || index >= graphemes.length)
+        return index;
+    if (segmented.wordStarts === null) {
+        const starts = [];
+        for (const word of wordSegmenter(locale).segment(segmented.text))
+            starts.push(word.index);
+        segmented.wordStarts = starts;
+    }
+    const starts = segmented.wordStarts;
+    const unitOffset = graphemes[index].index;
+    // Greatest word start <= unitOffset. starts[0] is 0 and unitOffset > 0 here,
+    // so the search never underflows.
+    let low = 0;
+    let high = starts.length - 1;
+    while (low < high) {
+        const mid = (low + high + 1) >>> 1;
+        if (starts[mid] <= unitOffset)
+            low = mid;
+        else
+            high = mid - 1;
+    }
+    return codeUnitOffsetToGraphemeIndex(graphemes, starts[low]);
+}
+//# sourceMappingURL=position.js.map

package/dist/core/position.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"position.js","sourceRoot":"","sources":["../../src/core/position.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,qBAAqB,EAAE,MAAM,aAAa,CAAC;AACpD,OAAO,EACL,6BAA6B,EAC7B,cAAc,EACd,gBAAgB,GAEjB,MAAM,gBAAgB,CAAC;AAgExB,MAAM,oBAAoB,GAAG,EAAE,CAAC;AAChC,MAAM,sBAAsB,GAAG,EAAE,CAAC;AAElC,uFAAuF;AACvF,MAAM,aAAa,GAAG,MAAM,CAAC;AAC7B,MAAM,cAAc,GAAG,CAAC,CAAC;AAYzB;;;;;;GAMG;AACH,MAAM,UAAU,eAAe,CAC7B,IAAY,EACZ,MAAc,EACd,SAAiB,EACjB,OAAO,GAA2B,EAAE;IAEpC,OAAO,gBAAgB,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,CAAC,CAAC;AACzE,CAAC;AAeD,SAAS,WAAW,CAAC,IAAY;IAC/B,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,gBAAgB,CAAC,IAAI,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC;AACvE,CAAC;AAED,+EAA+E;AAC/E,SAAS,gBAAgB,CACvB,SAAwB,EACxB,MAAc,EACd,SAAiB,EACjB,OAA+B;IAE/B,MAAM,EAAE,SAAS,EAAE,GAAG,SAAS,CAAC;IAChC,MAAM,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC;IAE/B,MAAM,QAAQ,GAAG,6BAA6B,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;IAClE,MAAM,KAAK,GAAG,eAAe,CAAC,SAAS,EAAE,QAAQ,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;IAEnE,MAAM,WAAW,GAAG,OAAO,CAAC,WAAW,IAAI,oBAAoB,CAAC;IAChE,MAAM,aAAa,GAAG,OAAO,CAAC,aAAa,IAAI,sBAAsB,CAAC;IAEtE,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,WAAW,EAAE,KAAK,CAAC,CAAC;IACtD,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,aAAa,EAAE,CAAC,CAAC,CAAC;IACvD,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,QAAQ,GAAG,aAAa,EAAE,KAAK,CAAC,CAAC;IAE5D,MAAM,MAAM,GAAe;QACzB,KAAK,EAAE,IAAI,CAAC,SAAS,EAAE,KAAK,EAAE,QAAQ,CAAC;QACvC,MAAM,EAAE,IAAI,CAAC,SAAS,EAAE,WAAW,EAAE,KAAK,CAAC;QAC3C,MAAM,EAAE,IAAI,CAAC,SAAS,EAAE,QAAQ,EAAE,SAAS,CAAC;QAC5C,MAAM,EAAE,KAAK;KACd,CAAC;IAEF,MAAM,QAAQ,GAAG,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC;IAEjD,OAAO;QACL,SAAS;QACT,QAAQ;QACR,MAAM;QACN,UAAU,EAAE,cAAc,CAAC,SAAS,EAAE,QAAQ,EAAE,MAAM,CAAC;KACxD,CAAC;AACJ,CAAC;AAED,yEAAyE;AACzE,MAAM,UAAU,iBAAiB,CAAC,QAAkB;IAClD,OAAO,cAAc,CAAC,QAAQ,CAAC,SAAS,EAAE,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;AAChF,CAAC;AAcD;;;;;;;;;;;;;;;;;;GAkBG;AACH,MAAM,UAAU,gBAAgB,CAC9B,IAAY,EACZ,SAAiB,EACjB,OAAO,GAA2B,EAAE;IAEpC,MAAM,GAAG,GAAoB,EAAE,CAAC;IAChC,wEAAwE;IACxE,6EAA6E;IAC7E,yDAAyD;IACzD,MAAM,SAAS,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;IAEpC,kFAAkF;IAClF,kFAAkF;IAClF,2EAA2E;IAC3E,IAAI,QAAQ,GAAG,CAAC,CAAC,CAAC;IAClB,IAAI,OAAO,GAAG,EAAE,CAAC;IACjB,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;QAC7C,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAChC,MAAM,WAAW,GAAG,cAAc,CAAC,OAAO,CAAC,CAAC;YAC5C,8EAA8E;YAC9E,0CAA0C;YAC1C,MAAM,QAAQ,GAAG,gBAAgB,CAAC,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE;gBAChE,GAAG,OAAO;gBACV,WAAW;aACZ,CAAC,CAAC;YACH,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,kBAAkB,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,CAAC,CAAC;QAC5D,CAAC;QACD,QAAQ,GAAG,CAAC,CAAC,CAAC;QACd,OAAO,GAAG,EAAE,CAAC;IACf,CAAC,CAAC;IAEF,KAAK,MAAM,OAAO,IAAI,iBAAiB,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC;QACtE,IAAI,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACxC,gFAAgF;YAChF,IAAI,QAAQ,IAAI,CAAC;gBAAE,OAAO,IAAI,OAAO,CAAC,OAAO,CAAC;YAC9C,SAAS;QACX,CAAC;QACD,IAAI,QAAQ,GAAG,CAAC;YAAE,QAAQ,GAAG,OAAO,CAAC,KAAK,CAAC;QAC3C,OAAO,IAAI,OAAO,CAAC,OAAO,CAAC;QAC3B,IAAI,sBAAsB,CAAC,OAAO,CAAC;YAAE,KAAK,EAAE,CAAC;IAC/C,CAAC;IACD,IAAI,QAAQ,IAAI,CAAC;QAAE,KAAK,EAAE,CAAC;IAC3B,OAAO,GAAG,CAAC;AACb,CAAC;AAED,0EAA0E;AAC1E,MAAM,iBAAiB,GAAG,cAAc,CAAC;AACzC,oFAAoF;AACpF,MAAM,gBAAgB,GAAG,iBAAiB,CAAC;AAE3C;;;;;GAKG;AACH,SAAS,sBAAsB,CAAC,GAAW;IACzC,MAAM,IAAI,GAAG,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,gBAAgB,EAAE,EAAE,CAAC,CAAC;IACpE,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC;IACxB,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACrC,OAAO,IAAI,KAAK,SAAS,IAAI,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC5D,CAAC;AAED,8EAA8E;AAC9E,SAAS,kBAAkB,CAAC,IAAY;IACtC,OAAO,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AAC3C,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAC,UAAkB;IAC9C,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC,aAAa,CAAC,EAAE,CAAC;QAC1C,MAAM,IAAI,qBAAqB,CAC7B,yCAAyC,aAAa,kBAAkB,CACzE,CAAC;IACJ,CAAC;IAED,IAAI,OAAgB,CAAC;IACrB,IAAI,CAAC;QACH,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,KAAK,CAAC,aAAa,CAAC,MAAM,CAAC,CAAC,CAAC;IAC/D,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,IAAI,qBAAqB,CAAC,oCAAoC,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC;IACnF,CAAC;IAED,IAAI,CAAC,eAAe,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9B,MAAM,IAAI,qBAAqB,CAAC,6CAA6C,CAAC,CAAC;IACjF,CAAC;IACD,IAAI,OAAO,CAAC,CAAC,KAAK,cAAc,EAAE,CAAC;QACjC,MAAM,IAAI,qBAAqB,CAC7B,gCAAgC,OAAO,CAAC,CAAC,qCAAqC,cAAc,EAAE,CAC/F,CAAC;IACJ,CAAC;IAED,MAAM,MAAM,GAAe;QACzB,KAAK,EAAE,OAAO,CAAC,CAAC;QAChB,MAAM,EAAE,OAAO,CAAC,GAAG;QACnB,MAAM,EAAE,OAAO,CAAC,GAAG;QACnB,MAAM,EAAE,OAAO,CAAC,CAAC;KAClB,CAAC;IACF,OAAO,EAAE,SAAS,EAAE,OAAO,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC,EAAE,MAAM,EAAE,UAAU,EAAE,CAAC;AAC3E,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,eAAe,CAC7B,QAAkB,EAClB,IAAY;IAEZ,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,QAAQ,CAAC;IACvC,MAAM,SAAS,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAC;IAEzC,IAAI,MAAM,CAAC,KAAK,KAAK,EAAE,EAAE,CAAC;QACxB,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,MAAM,CAAC,CAAC;QACzD,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IAC1C,CAAC;IAED,MAAM,WAAW,GAAG,uBAAuB,CAAC,SAAS,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC;IACrE,IAAI,WAAW,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAE/C,MAAM,WAAW,GAAG,gBAAgB,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC;IAC1D,gEAAgE;IAChE,MAAM,MAAM,GAAG,gBAAgB,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;IACrE,MAAM,MAAM,GAAG,gBAAgB,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;IACrE,IAAI,IAAmD,CAAC;IACxD,KAAK,MAAM,MAAM,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,KAAK,GAAG,YAAY,CAAC,SAAS,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;QAC3E,IACE,IAAI,KAAK,SAAS;YAClB,KAAK,GAAG,IAAI,CAAC,KAAK;YAClB,CAAC,KAAK,KAAK,IAAI,CAAC,KAAK;gBACnB,IAAI,CAAC,GAAG,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,EAC3E,CAAC;YACD,IAAI,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;QAC3B,CAAC;IACH,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,IAAK,CAAC,MAAM,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;AAClE,CAAC;AAED,qEAAqE;AACrE,SAAS,uBAAuB,CAAC,SAAqB,EAAE,KAAa;IACnE,MAAM,cAAc,GAAG,gBAAgB,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;IACrE,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,IAAI,cAAc,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAC7C,MAAM,IAAI,GAAG,SAAS,CAAC,MAAM,GAAG,cAAc,CAAC,MAAM,CAAC;IACtD,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,IAAI,IAAI,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QAC9C,IAAI,OAAO,GAAG,IAAI,CAAC;QACnB,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,cAAc,CAAC,MAAM,EAAE,MAAM,IAAI,CAAC,EAAE,CAAC;YACjE,IAAI,SAAS,CAAC,KAAK,GAAG,MAAM,CAAE,CAAC,OAAO,KAAK,cAAc,CAAC,MAAM,CAAC,EAAE,CAAC;gBAClE,OAAO,GAAG,KAAK,CAAC;gBAChB,MAAM;YACR,CAAC;QACH,CAAC;QACD,IAAI,OAAO;YAAE,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAChC,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;;GAIG;AACH,SAAS,YAAY,CACnB,SAA8B,EAC9B,UAAkB,EAClB,WAAmB,EACnB,MAAyB,EACzB,MAAyB;IAEzB,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,IAAI,MAAM,CAAC,MAAM,EAAE,IAAI,IAAI,CAAC,EAAE,CAAC;QACpD,MAAM,IAAI,GAAG,SAAS,CAAC,UAAU,GAAG,IAAI,CAAC,CAAC;QAC1C,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,IAAI,CAAC,CAAC;QAC1C,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI;YAAE,MAAM;QACvD,KAAK,IAAI,CAAC,CAAC;IACb,CAAC;IACD,MAAM,UAAU,GAAG,UAAU,GAAG,WAAW,CAAC;IAC5C,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACtD,MAAM,IAAI,GAAG,SAAS,CAAC,UAAU,GAAG,KAAK,CAAC,CAAC;QAC3C,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,OAAO,KAAK,MAAM,CAAC,KAAK,CAAC;YAAE,MAAM;QAChE,KAAK,IAAI,CAAC,CAAC;IACb,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,cAAc,CAAC,SAAiB,EAAE,QAAgB,EAAE,MAAkB;IAC7E,MAAM,OAAO,GAAkB;QAC7B,CAAC,EAAE,cAAc;QACjB,CAAC,EAAE,SAAS;QACZ,CAAC,EAAE,QAAQ;QACX,CAAC,EAAE,MAAM,CAAC,KAAK;QACf,GAAG,EAAE,MAAM,CAAC,MAAM;QAClB,GAAG,EAAE,MAAM,CAAC,MAAM;QAClB,CAAC,EAAE,MAAM,CAAC,MAAM;KACjB,CAAC;IACF,OAAO,aAAa,GAAG,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC;AACjD,CAAC;AAED,SAAS,eAAe,CAAC,KAAc;IACrC,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IAC9D,MAAM,MAAM,GAAG,KAAgC,CAAC;IAChD,OAAO,CACL,OAAO,MAAM,CAAC,CAAC,KAAK,QAAQ;QAC5B,OAAO,MAAM,CAAC,CAAC,KAAK,QAAQ;QAC5B,OAAO,MAAM,CAAC,CAAC,KAAK,QAAQ;QAC5B,OAAO,MAAM,CAAC,CAAC,KAAK,QAAQ;QAC5B,OAAO,MAAM,CAAC,GAAG,KAAK,QAAQ;QAC9B,OAAO,MAAM,CAAC,GAAG,KAAK,QAAQ;QAC9B,OAAO,MAAM,CAAC,CAAC,KAAK,QAAQ,CAC7B,CAAC;AACJ,CAAC;AAED,MAAM,cAAc,GAAG,IAAI,GAAG,EAA0B,CAAC;AACzD,MAAM,kBAAkB,GAAG,IAAI,GAAG,EAA0B,CAAC;AAE7D,SAAS,iBAAiB,CAAC,MAA0B;IACnD,MAAM,GAAG,GAAG,MAAM,IAAI,EAAE,CAAC;IACzB,IAAI,SAAS,GAAG,kBAAkB,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAC5C,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC5B,SAAS,GAAG,IAAI,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,EAAE,WAAW,EAAE,UAAU,EAAE,CAAC,CAAC;QACpE,kBAAkB,CAAC,GAAG,CAAC,GAAG,EAAE,SAAS,CAAC,CAAC;IACzC,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,aAAa,CAAC,MAA0B;IAC/C,MAAM,GAAG,GAAG,MAAM,IAAI,EAAE,CAAC;IACzB,IAAI,SAAS,GAAG,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IACxC,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC5B,SAAS,GAAG,IAAI,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,EAAE,WAAW,EAAE,MAAM,EAAE,CAAC,CAAC;QAChE,cAAc,CAAC,GAAG,CAAC,GAAG,EAAE,SAAS,CAAC,CAAC;IACrC,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,IAAI,CAAC,SAA8B,EAAE,KAAa,EAAE,GAAW;IACtE,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,IAAI,KAAK,GAAG,KAAK,EAAE,KAAK,GAAG,GAAG,EAAE,KAAK,IAAI,CAAC;QAAE,GAAG,IAAI,SAAS,CAAC,KAAK,CAAE,CAAC,OAAO,CAAC;IAClF,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;;;;GAMG;AACH,SAAS,eAAe,CACtB,SAAwB,EACxB,KAAa,EACb,MAA0B;IAE1B,MAAM,EAAE,SAAS,EAAE,GAAG,SAAS,CAAC;IAChC,IAAI,KAAK,IAAI,CAAC,IAAI,KAAK,IAAI,SAAS,CAAC,MAAM;QAAE,OAAO,KAAK,CAAC;IAE1D,IAAI,SAAS,CAAC,UAAU,KAAK,IAAI,EAAE,CAAC;QAClC,MAAM,MAAM,GAAa,EAAE,CAAC;QAC5B,KAAK,MAAM,IAAI,IAAI,aAAa,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,SAAS,CAAC,IAAI,CAAC;YAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC1F,SAAS,CAAC,UAAU,GAAG,MAAM,CAAC;IAChC,CAAC;IACD,MAAM,MAAM,GAAG,SAAS,CAAC,UAAU,CAAC;IAEpC,MAAM,UAAU,GAAG,SAAS,CAAC,KAAK,CAAE,CAAC,KAAK,CAAC;IAC3C,6EAA6E;IAC7E,kCAAkC;IAClC,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,IAAI,IAAI,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC;IAC7B,OAAO,GAAG,GAAG,IAAI,EAAE,CAAC;QAClB,MAAM,GAAG,GAAG,CAAC,GAAG,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;QACnC,IAAI,MAAM,CAAC,GAAG,CAAE,IAAI,UAAU;YAAE,GAAG,GAAG,GAAG,CAAC;;YACrC,IAAI,GAAG,GAAG,GAAG,CAAC,CAAC;IACtB,CAAC;IACD,OAAO,6BAA6B,CAAC,SAAS,EAAE,MAAM,CAAC,GAAG,CAAE,CAAC,CAAC;AAChE,CAAC"}

package/dist/core/reading-text.d.ts

@@ -0,0 +1,77 @@
+/**
+ * The canonical "reading text" policy — the single definition of which parts of a
+ * section's markup become the text a reader actually sees in the frame.
+ *
+ * Two independent consumers must agree on it, and this module is how they stay one
+ * model rather than two:
+ *
+ * - `src/view` sanitizes and applies resources before rendering: discarded elements
+ *   vanish with their children (`sanitize.ts`), and an `<img>` that cannot be served
+ *   is replaced by its `alt` text (`resources.ts`). The frame's measured section
+ *   text — what a `Position` resolves against — is the `textContent` of that output.
+ * - `src/search` captures hit anchors headlessly, without a DOM, over the same text.
+ *   Its extractor (`src/search/extract.ts`) mirrors the pipeline using the tables
+ *   and predicates here, so a captured quote/context window exists verbatim in the
+ *   text the frame resolves against.
+ *
+ * Internal to the library: nothing here is re-exported from `src/core/index.ts`
+ * (same status as `graphemes.ts`), so it is shared implementation, not public
+ * surface. It is headless by rule — string tables and string predicates only —
+ * because `src/search` may import only `src/core` and platform primitives.
+ */
+/**
+ * HTML elements the sanitizer removes with everything inside them, because their
+ * children are not prose: raw-text elements whose content is source rather than
+ * text, form controls, and void elements with no children to keep. Everything
+ * else absent from the allowlist is unwrapped instead — the element goes, its
+ * text stays. The unwrap-vs-discard rationale lives with the allowlist
+ * (`src/view/allowlist.ts`), which re-exports these under their frozen names.
+ */
+export declare const DISCARDED_HTML_ELEMENTS: ReadonlySet<string>;
+/**
+ * SVG elements discarded whole. SMIL animation is on this list, not merely absent
+ * from the allowlist, because `<animate attributeName="href" to="javascript:...">`
+ * turns an inert element into a live one after sanitization has already run.
+ */
+export declare const DISCARDED_SVG_ELEMENTS: ReadonlySet<string>;
+export type ReferenceKind = 'empty' | 'fragment' | 'relative' | 'scheme';
+export interface ClassifiedReference {
+    readonly kind: ReferenceKind;
+    /** The reference with URL-insignificant whitespace removed. */
+    readonly value: string;
+    /** Lowercased scheme name, present only when kind is 'scheme'. */
+    readonly scheme?: string;
+}
+export declare function classifyReference(raw: string): ClassifiedReference;
+/** Collapses `.` and `..` segments, keeping leading `..` that cannot be resolved. */
+export declare function normalizeReference(reference: string): string;
+/** The bare media type: parameters trimmed, lowercased. */
+export declare function baseMediaType(mediaType: string): string;
+/**
+ * Only these media types are minted a `data:` URL by the resource layer. A
+ * reference resolving is not evidence it should be served: the seam answers for
+ * the whole manifest, so a `text/javascript` entry resolves as readily as an
+ * image — and a URL is a capability, not a convenience.
+ */
+export declare function isServableResource(mediaType: string): boolean;
+/** The synchronous face of `Section.resolve` — all the image policy needs. */
+export type ReadingResolver = (reference: string) => {
+    readonly mediaType: string;
+} | undefined;
+/**
+ * The text an `<img>` contributes to the reading text: `''` when the element
+ * survives as an image (a served resource, or a remote/`data:` reference left for
+ * the CSP to refuse), or its `alt` when the resource layer substitutes the alt
+ * text for an image it cannot serve (`applyResources` in `src/view/resources.ts`
+ * — Gutenberg drop caps are the motivating case). An empty/absent `alt` on an
+ * unservable image means decorative: the element is removed and contributes
+ * nothing, which `''` also expresses.
+ *
+ * This is the headless mirror of the frame-side decision; the one edge it cannot
+ * see is a resource whose bytes fail to *load* after resolving servable — the
+ * frame substitutes alt there too, and a search anchor over that rare case
+ * degrades to the ordinary soft miss. Change this and the `applyResources` image
+ * branch together.
+ */
+export declare function imageReadingText(src: string | null, alt: string | null, resolve: ReadingResolver | undefined): string;
+//# sourceMappingURL=reading-text.d.ts.map

package/dist/core/reading-text.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"reading-text.d.ts","sourceRoot":"","sources":["../../src/core/reading-text.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAMH;;;;;;;GAOG;AACH,eAAO,MAAM,uBAAuB,qBA0BnC,CAAC;AAEF;;;;GAIG;AACH,eAAO,MAAM,sBAAsB,qBAclC,CAAC;AAEF,MAAM,MAAM,aAAa,GAAG,OAAO,GAAG,UAAU,GAAG,UAAU,GAAG,QAAQ,CAAC;AAEzE,MAAM,WAAW,mBAAmB;IAClC,QAAQ,CAAC,IAAI,EAAE,aAAa,CAAC;IAC7B,+DAA+D;IAC/D,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,kEAAkE;IAClE,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC;CAC1B;AAyBD,wBAAgB,iBAAiB,CAAC,GAAG,EAAE,MAAM,GAAG,mBAAmB,CAOlE;AAED,qFAAqF;AACrF,wBAAgB,kBAAkB,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAa5D;AAcD,2DAA2D;AAC3D,wBAAgB,aAAa,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAEvD;AAED;;;;;GAKG;AACH,wBAAgB,kBAAkB,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAG7D;AAED,8EAA8E;AAC9E,MAAM,MAAM,eAAe,GAAG,CAAC,SAAS,EAAE,MAAM,KAAK;IAAE,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAA;CAAE,GAAG,SAAS,CAAC;AAEhG;;;;;;;;;;;;;;GAcG;AACH,wBAAgB,gBAAgB,CAC9B,GAAG,EAAE,MAAM,GAAG,IAAI,EAClB,GAAG,EAAE,MAAM,GAAG,IAAI,EAClB,OAAO,EAAE,eAAe,GAAG,SAAS,GACnC,MAAM,CAQR"}