@storyteller-platform/align 0.1.25 → 0.1.27

package/dist/align/interpolateSentenceRanges.js

@@ -0,0 +1,101 @@
+import "../chunk-BIEQXUOY.js";
+function buildGapRanges(slots, left, right, audioFileDurations) {
+  const n = slots.length;
+  if (n === 0) return [];
+  if (left.audiofile === right.audiofile) {
+    const span = right.time - left.time;
+    return slots.map((slot, i) => ({
+      ...slot,
+      audiofile: left.audiofile,
+      start: left.time + span * i / n,
+      end: left.time + span * (i + 1) / n
+    }));
+  }
+  const leftDuration = audioFileDurations[left.audiofile] ?? left.time;
+  const leftAvail = leftDuration - left.time;
+  const rightAvail = right.time;
+  const total = leftAvail + rightAvail;
+  let n1 = total > 0 ? Math.round(n * (leftAvail / total)) : n;
+  let n2 = n - n1;
+  n1 = Math.max(0, n1);
+  n2 = n - n1;
+  const result = [];
+  if (n1 > 0) {
+    for (let i = 0; i < n1; i++) {
+      result.push({
+        // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
+        ...slots[i],
+        audiofile: left.audiofile,
+        start: left.time + leftAvail * i / n1,
+        end: left.time + leftAvail * (i + 1) / n1
+      });
+    }
+  }
+  if (n2 > 0) {
+    for (let i = 0; i < n2; i++) {
+      result.push({
+        // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
+        ...slots[n1 + i],
+        audiofile: right.audiofile,
+        start: rightAvail * i / n2,
+        end: rightAvail * (i + 1) / n2
+      });
+    }
+  }
+  return result;
+}
+function interpolateSentenceRanges(sentenceRanges, chapterSentenceCounts, audioFileDurations) {
+  if (sentenceRanges.length === 0) return [];
+  const result = [];
+  const first = sentenceRanges[0];
+  if (first.id > 0) {
+    const slots = Array.from({ length: first.id }, (_, i) => ({
+      chapterId: first.chapterId,
+      id: i
+    }));
+    const left = { time: 0, audiofile: first.audiofile };
+    const right = { time: first.start, audiofile: first.audiofile };
+    result.push(...buildGapRanges(slots, left, right, audioFileDurations));
+  }
+  result.push(first);
+  for (let idx = 1; idx < sentenceRanges.length; idx++) {
+    const prev = sentenceRanges[idx - 1];
+    const curr = sentenceRanges[idx];
+    const left = { time: prev.end, audiofile: prev.audiofile };
+    const right = { time: curr.start, audiofile: curr.audiofile };
+    const gapSlots = [];
+    if (prev.chapterId === curr.chapterId) {
+      for (let id = prev.id + 1; id < curr.id; id++) {
+        gapSlots.push({ chapterId: prev.chapterId, id });
+      }
+    } else {
+      const prevTotal = chapterSentenceCounts[prev.chapterId] ?? prev.id + 1;
+      for (let id = prev.id + 1; id < prevTotal; id++) {
+        gapSlots.push({ chapterId: prev.chapterId, id });
+      }
+      for (let id = 0; id < curr.id; id++) {
+        gapSlots.push({ chapterId: curr.chapterId, id });
+      }
+    }
+    if (gapSlots.length > 0) {
+      result.push(...buildGapRanges(gapSlots, left, right, audioFileDurations));
+    }
+    result.push(curr);
+  }
+  const last = sentenceRanges[sentenceRanges.length - 1];
+  const lastTotal = chapterSentenceCounts[last.chapterId] ?? last.id + 1;
+  if (last.id < lastTotal - 1) {
+    const slots = Array.from(
+      { length: lastTotal - 1 - last.id },
+      (_, i) => ({ chapterId: last.chapterId, id: last.id + 1 + i })
+    );
+    const fileEnd = audioFileDurations[last.audiofile] ?? last.end;
+    const left = { time: last.end, audiofile: last.audiofile };
+    const right = { time: fileEnd, audiofile: last.audiofile };
+    result.push(...buildGapRanges(slots, left, right, audioFileDurations));
+  }
+  return result;
+}
+export {
+  interpolateSentenceRanges
+};

package/dist/align/search.cjs

@@ -37,16 +37,16 @@ function buildNgramIndex(text) {
   }
   return index;
 }
+const NGRAM_SIZE = 5;
 function* ngrams(text) {
   const words = text.split("-");
-  let pos = 0;
-  for (const i of (0, import_itertools.range)(words.length - 4)) {
-    const ngram = words.slice(i, i + 5).join("-");
-    yield [ngram, pos];
-    pos += words[i].length + 1;
+  for (const i of (0, import_itertools.range)(words.length - NGRAM_SIZE - 1)) {
+    const ngram = words.slice(i, i + NGRAM_SIZE).join("-");
+    yield [ngram, i];
   }
 }
 function collectBoundaryVotes(query, document) {
+  const queryWords = query.split("-");
   const documentIndex = buildNgramIndex(document);
   let skippedNgrams = 0;
   let totalNgrams = 0;
@@ -61,7 +61,7 @@ function collectBoundaryVotes(query, document) {
     }
     for (const documentStart of documentStarts) {
       startVotes.push(documentStart - start);
-      endVotes.push(documentStart + (query.length - start));
+      endVotes.push(documentStart + (queryWords.length - start));
     }
   }
   if (skippedNgrams > totalNgrams / 2) {
@@ -97,6 +97,14 @@ function chooseBestFromBins(bins, dir) {
   }
   return dir > 0 ? (0, import_itertools.max)(best) ?? null : (0, import_itertools.min)(best) ?? null;
 }
+function getOffsetFromWordIndex(wordIndex, document) {
+  const words = document.split("-");
+  let offset = 0;
+  for (const i of (0, import_itertools.range)(Math.min(words.length, Math.max(0, wordIndex)))) {
+    offset += words[i].length + 1;
+  }
+  return offset;
+}
 function findBoundaries(query, document) {
   const boundaryVotes = collectBoundaryVotes(query, document);
   if (!boundaryVotes) return null;
@@ -111,7 +119,10 @@ function findBoundaries(query, document) {
   if (bestEnd === null) {
     return null;
   }
-  return { start: bestStart, end: bestEnd };
+  return {
+    start: getOffsetFromWordIndex(bestStart, document),
+    end: getOffsetFromWordIndex(bestEnd, document)
+  };
 }
 // Annotate the CommonJS export names for ESM import in node:
 0 && (module.exports = {

package/dist/align/search.js

@@ -12,16 +12,16 @@ function buildNgramIndex(text) {
   }
   return index;
 }
+const NGRAM_SIZE = 5;
 function* ngrams(text) {
   const words = text.split("-");
-  let pos = 0;
-  for (const i of range(words.length - 4)) {
-    const ngram = words.slice(i, i + 5).join("-");
-    yield [ngram, pos];
-    pos += words[i].length + 1;
+  for (const i of range(words.length - NGRAM_SIZE - 1)) {
+    const ngram = words.slice(i, i + NGRAM_SIZE).join("-");
+    yield [ngram, i];
   }
 }
 function collectBoundaryVotes(query, document) {
+  const queryWords = query.split("-");
   const documentIndex = buildNgramIndex(document);
   let skippedNgrams = 0;
   let totalNgrams = 0;
@@ -36,7 +36,7 @@ function collectBoundaryVotes(query, document) {
     }
     for (const documentStart of documentStarts) {
       startVotes.push(documentStart - start);
-      endVotes.push(documentStart + (query.length - start));
+      endVotes.push(documentStart + (queryWords.length - start));
     }
   }
   if (skippedNgrams > totalNgrams / 2) {
@@ -72,6 +72,14 @@ function chooseBestFromBins(bins, dir) {
   }
   return dir > 0 ? max(best) ?? null : min(best) ?? null;
 }
+function getOffsetFromWordIndex(wordIndex, document) {
+  const words = document.split("-");
+  let offset = 0;
+  for (const i of range(Math.min(words.length, Math.max(0, wordIndex)))) {
+    offset += words[i].length + 1;
+  }
+  return offset;
+}
 function findBoundaries(query, document) {
   const boundaryVotes = collectBoundaryVotes(query, document);
   if (!boundaryVotes) return null;
@@ -86,7 +94,10 @@ function findBoundaries(query, document) {
   if (bestEnd === null) {
     return null;
   }
-  return { start: bestStart, end: bestEnd };
+  return {
+    start: getOffsetFromWordIndex(bestStart, document),
+    end: getOffsetFromWordIndex(bestEnd, document)
+  };
 }
 export {
   buildNgramIndex,

package/dist/index.d.cts

@@ -7,9 +7,8 @@ import '@esfx/async-semaphore';
 import 'pino';
 import './process/AudioEncoding.cjs';
 import '@storyteller-platform/ghost-story/constants';
-import '@echogarden/text-segmentation';
 import '@storyteller-platform/epub';
-import './markup/map.cjs';
 import '@storyteller-platform/ghost-story/recognition';
 import './align/getSentenceRanges.cjs';
+import '@echogarden/text-segmentation';
 import '@storyteller-platform/transliteration';

package/dist/index.d.ts

@@ -7,9 +7,8 @@ import '@esfx/async-semaphore';
 import 'pino';
 import './process/AudioEncoding.js';
 import '@storyteller-platform/ghost-story/constants';
-import '@echogarden/text-segmentation';
 import '@storyteller-platform/epub';
-import './markup/map.js';
 import '@storyteller-platform/ghost-story/recognition';
 import './align/getSentenceRanges.js';
+import '@echogarden/text-segmentation';
 import '@storyteller-platform/transliteration';

package/dist/markup/markup.cjs

@@ -97,16 +97,11 @@ async function markup(input, output, options) {
         continue;
       }
       const chapterXml = await epub.readXhtmlItemContents(chapterId);
-      const { result: segmentation, mapping } = await (0, import_segmentation.getXhtmlSegmentation)(
-        import_epub.Epub.getXhtmlBody(chapterXml),
-        { primaryLocale }
-      );
-      const { markedUp, timing: chapterTiming } = markupChapter(
+      const { markedUp, timing: chapterTiming } = await markupChapter(
         chapterId,
         chapterXml,
         options.granularity ?? "sentence",
-        segmentation,
-        mapping
+        primaryLocale
       );
       timing.add(chapterTiming.summary());
       await epub.writeXhtmlItemContents(chapterId, markedUp);
@@ -119,7 +114,7 @@ async function markup(input, output, options) {
     __callDispose(_stack, _error, _hasError);
   }
 }
-function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping) {
+async function markupChapter(chapterId, chapterXml, granularity, locale) {
   const timing = (0, import_ghost_story.createTiming)();
   const html = import_epub.Epub.findXmlChildByName("html", chapterXml);
   if (!html) throw new Error("Invalid XHTML document: no html element");
@@ -128,8 +123,14 @@ function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping
   clearBodyElement(chapterXml);
   const taggedHtml = import_epub.Epub.findXmlChildByName("html", chapterXml);
   const taggedBody = import_epub.Epub.findXmlChildByName("body", taggedHtml["html"]);
+  const original = (0, import_parseDom.parseDom)(import_epub.Epub.getXmlChildren(body));
+  const inlined = (0, import_transform.inlineFootnotes)(original);
+  const lifted = (0, import_transform.liftText)(inlined.root);
+  const segmentation = await (0, import_segmentation.segmentChapter)(lifted.result, {
+    primaryLocale: locale
+  });
   timing.time("mark up", () => {
-    let root = (0, import_parseDom.parseDom)(import_epub.Epub.getXmlChildren(body));
+    let root = inlined.root;
     let pos = 0;
     let i = 0;
     for (const sentence of segmentation) {
@@ -140,8 +141,8 @@ function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping
           if (word.text.match(/\S/)) {
             root = (0, import_transform.addMark)(
               root,
-              mapping.invert().map(wordPos),
-              mapping.invert().map(wordPos + word.text.replace(/\n$/, "").length, -1),
+              lifted.mapping.invert().map(wordPos),
+              lifted.mapping.invert().map(wordPos + word.text.replace(/\n$/, "").length, -1),
               new import_model.Mark("span", { id: `${chapterId}-s${i}-w${j}` })
             );
             j++;
@@ -152,15 +153,21 @@ function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping
       if (sentence.text.match(/\S/)) {
         root = (0, import_transform.addMark)(
           root,
-          mapping.invert().map(pos),
-          mapping.invert().map(pos + sentence.text.replace(/\n$/, "").length, -1),
+          lifted.mapping.invert().map(pos),
+          lifted.mapping.invert().map(pos + sentence.text.replace(/\n$/, "").length, -1),
           new import_model.Mark("span", { id: `${chapterId}-s${i}` })
         );
         i++;
       }
       pos += sentence.text.replace(/\n$/, "").length;
     }
-    taggedBody["body"] = (0, import_serializeDom.serializeDom)(root);
+    const replaced = (0, import_transform.replaceFootnotes)(
+      original,
+      root,
+      inlined.footnotePairs,
+      inlined.mapping
+    );
+    taggedBody["body"] = (0, import_serializeDom.serializeDom)(replaced);
   });
   return { markedUp: chapterXml, timing };
 }

package/dist/markup/markup.d.cts

@@ -1,9 +1,7 @@
 import * as _storyteller_platform_ghost_story from '@storyteller-platform/ghost-story';
 import { TimingAggregator } from '@storyteller-platform/ghost-story';
-import { Sentence } from '@echogarden/text-segmentation';
 import { Logger } from 'pino';
 import { ParsedXml } from '@storyteller-platform/epub';
-import { Mapping } from './map.cjs';
 
 interface MarkupOptions {
     granularity?: "word" | "sentence";
@@ -12,9 +10,9 @@ interface MarkupOptions {
     logger?: Logger;
 }
 declare function markup(input: string, output: string, options: MarkupOptions): Promise<TimingAggregator>;
-declare function markupChapter(chapterId: string, chapterXml: ParsedXml, granularity: "word" | "sentence", segmentation: Sentence[], mapping: Mapping): {
+declare function markupChapter(chapterId: string, chapterXml: ParsedXml, granularity: "word" | "sentence", locale: Intl.Locale | null): Promise<{
     markedUp: ParsedXml;
     timing: _storyteller_platform_ghost_story.Timing;
-};
+}>;
 
 export { type MarkupOptions, markup, markupChapter };

package/dist/markup/markup.d.ts

@@ -1,9 +1,7 @@
 import * as _storyteller_platform_ghost_story from '@storyteller-platform/ghost-story';
 import { TimingAggregator } from '@storyteller-platform/ghost-story';
-import { Sentence } from '@echogarden/text-segmentation';
 import { Logger } from 'pino';
 import { ParsedXml } from '@storyteller-platform/epub';
-import { Mapping } from './map.js';
 
 interface MarkupOptions {
     granularity?: "word" | "sentence";
@@ -12,9 +10,9 @@ interface MarkupOptions {
     logger?: Logger;
 }
 declare function markup(input: string, output: string, options: MarkupOptions): Promise<TimingAggregator>;
-declare function markupChapter(chapterId: string, chapterXml: ParsedXml, granularity: "word" | "sentence", segmentation: Sentence[], mapping: Mapping): {
+declare function markupChapter(chapterId: string, chapterXml: ParsedXml, granularity: "word" | "sentence", locale: Intl.Locale | null): Promise<{
     markedUp: ParsedXml;
     timing: _storyteller_platform_ghost_story.Timing;
-};
+}>;
 
 export { type MarkupOptions, markup, markupChapter };

package/dist/markup/markup.js

@@ -11,9 +11,14 @@ import {
 } from "@storyteller-platform/ghost-story";
 import { Mark } from "./model.js";
 import { parseDom } from "./parseDom.js";
-import { getXhtmlSegmentation } from "./segmentation.js";
+import { segmentChapter } from "./segmentation.js";
 import { serializeDom } from "./serializeDom.js";
-import { addMark } from "./transform.js";
+import {
+  addMark,
+  inlineFootnotes,
+  liftText,
+  replaceFootnotes
+} from "./transform.js";
 async function markup(input, output, options) {
   var _stack = [];
   try {
@@ -35,16 +40,11 @@ async function markup(input, output, options) {
         continue;
       }
       const chapterXml = await epub.readXhtmlItemContents(chapterId);
-      const { result: segmentation, mapping } = await getXhtmlSegmentation(
-        Epub.getXhtmlBody(chapterXml),
-        { primaryLocale }
-      );
-      const { markedUp, timing: chapterTiming } = markupChapter(
+      const { markedUp, timing: chapterTiming } = await markupChapter(
         chapterId,
         chapterXml,
         options.granularity ?? "sentence",
-        segmentation,
-        mapping
+        primaryLocale
       );
       timing.add(chapterTiming.summary());
       await epub.writeXhtmlItemContents(chapterId, markedUp);
@@ -57,7 +57,7 @@ async function markup(input, output, options) {
     __callDispose(_stack, _error, _hasError);
   }
 }
-function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping) {
+async function markupChapter(chapterId, chapterXml, granularity, locale) {
   const timing = createTiming();
   const html = Epub.findXmlChildByName("html", chapterXml);
   if (!html) throw new Error("Invalid XHTML document: no html element");
@@ -66,8 +66,14 @@ function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping
   clearBodyElement(chapterXml);
   const taggedHtml = Epub.findXmlChildByName("html", chapterXml);
   const taggedBody = Epub.findXmlChildByName("body", taggedHtml["html"]);
+  const original = parseDom(Epub.getXmlChildren(body));
+  const inlined = inlineFootnotes(original);
+  const lifted = liftText(inlined.root);
+  const segmentation = await segmentChapter(lifted.result, {
+    primaryLocale: locale
+  });
   timing.time("mark up", () => {
-    let root = parseDom(Epub.getXmlChildren(body));
+    let root = inlined.root;
     let pos = 0;
     let i = 0;
     for (const sentence of segmentation) {
@@ -78,8 +84,8 @@ function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping
           if (word.text.match(/\S/)) {
             root = addMark(
               root,
-              mapping.invert().map(wordPos),
-              mapping.invert().map(wordPos + word.text.replace(/\n$/, "").length, -1),
+              lifted.mapping.invert().map(wordPos),
+              lifted.mapping.invert().map(wordPos + word.text.replace(/\n$/, "").length, -1),
               new Mark("span", { id: `${chapterId}-s${i}-w${j}` })
             );
             j++;
@@ -90,15 +96,21 @@ function markupChapter(chapterId, chapterXml, granularity, segmentation, mapping
       if (sentence.text.match(/\S/)) {
         root = addMark(
           root,
-          mapping.invert().map(pos),
-          mapping.invert().map(pos + sentence.text.replace(/\n$/, "").length, -1),
+          lifted.mapping.invert().map(pos),
+          lifted.mapping.invert().map(pos + sentence.text.replace(/\n$/, "").length, -1),
           new Mark("span", { id: `${chapterId}-s${i}` })
         );
         i++;
       }
       pos += sentence.text.replace(/\n$/, "").length;
     }
-    taggedBody["body"] = serializeDom(root);
+    const replaced = replaceFootnotes(
+      original,
+      root,
+      inlined.footnotePairs,
+      inlined.mapping
+    );
+    taggedBody["body"] = serializeDom(replaced);
   });
   return { markedUp: chapterXml, timing };
 }

package/dist/markup/model.cjs

@@ -18,14 +18,17 @@ var __copyProps = (to, from, except, desc) => {
 var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
 var model_exports = {};
 __export(model_exports, {
+  FootnoteNode: () => FootnoteNode,
   Mark: () => Mark,
   Node: () => Node,
+  NoterefNode: () => NoterefNode,
   Root: () => Root,
   TextNode: () => TextNode,
   descendants: () => descendants
 });
 module.exports = __toCommonJS(model_exports);
 var import_itertools = require("itertools");
+var import_resolvedPos = require("./resolvedPos.cjs");
 var import_semantics = require("./semantics.cjs");
 class Root {
   constructor(children) {
@@ -39,6 +42,12 @@ class Root {
   get textContent() {
     return this.children.reduce((acc, child) => acc + child.textContent, "");
   }
+  get nodeSize() {
+    return this.border + (this.children.reduce((acc, child) => acc + child.nodeSize, 0) || 1) + this.border;
+  }
+  get contentSize() {
+    return this.nodeSize - this.border * 2;
+  }
   split(at) {
     const children = [];
     let pos = this.border;
@@ -57,6 +66,58 @@ class Root {
   copy(opts = {}) {
     return new Root(opts.children ?? this.children);
   }
+  findIndex(pos) {
+    if (pos === 0) return { index: 0, offset: pos };
+    if (pos === this.contentSize) {
+      return { index: this.children.length, offset: pos };
+    }
+    if (pos > this.contentSize || pos < 0) {
+      throw new RangeError(`Position ${pos} outside of fragment`);
+    }
+    for (let i = 0, curPos = 0; ; i++) {
+      const cur = this.children[i];
+      const end = curPos + cur.nodeSize;
+      if (end >= pos) {
+        if (end === pos) return { index: i + 1, offset: end };
+        return { index: i, offset: curPos };
+      }
+      curPos = end;
+    }
+  }
+  replace(at, withNode) {
+    const children = [];
+    let pos = this.border;
+    for (const child of this.children) {
+      if (at === pos) {
+        children.push(withNode);
+      } else if (at > pos && at < pos + child.nodeSize) {
+        if (child instanceof TextNode) {
+          throw new Error("Tried to replace at a position within a text node");
+        }
+        children.push(child.replace(at - pos, withNode));
+      } else {
+        children.push(child);
+      }
+      pos += children.at(-1).nodeSize;
+    }
+    return this.copy({ children });
+  }
+  cut(pos) {
+    let currentPos = this.border;
+    for (const child of this.children) {
+      if (pos === currentPos) {
+        return child;
+      }
+      if (pos > currentPos && pos < currentPos + child.nodeSize) {
+        return child.cut(pos - currentPos);
+      }
+      currentPos += child.nodeSize;
+    }
+    return null;
+  }
+  resolve(pos) {
+    return import_resolvedPos.ResolvedPos.resolve(this, pos);
+  }
 }
 class Node {
   constructor(tagName, attrs = {}, children = [], marks = []) {
@@ -80,6 +141,9 @@ class Node {
   get nodeSize() {
     return this.border + (this.children.reduce((acc, child) => acc + child.nodeSize, 0) || 1) + this.border;
   }
+  get contentSize() {
+    return this.nodeSize - this.border * 2;
+  }
   get textContent() {
     return this.children.reduce((acc, child) => acc + child.textContent, "");
   }
@@ -102,14 +166,74 @@ class Node {
     }
     return this.copy({ children });
   }
+  static instance() {
+    return this;
+  }
+  static create(klass, ...args) {
+    return new klass(...args);
+  }
   copy(opts = {}) {
-    return new Node(
+    return Node.create(
+      this.constructor,
       this.tagName,
       opts.attrs ?? this.attrs,
       opts.children ?? this.children,
       opts.marks ?? this.marks
     );
   }
+  replace(at, withNode) {
+    const children = [];
+    let pos = this.border;
+    for (const child of this.children) {
+      if (at === pos) {
+        children.push(withNode);
+      } else if (at > pos && at < pos + child.nodeSize) {
+        if (child instanceof TextNode) {
+          throw new Error("Tried to replace at a position within a text node");
+        }
+        children.push(child.replace(at - pos, withNode));
+      } else {
+        children.push(child);
+      }
+      pos += children.at(-1).nodeSize;
+    }
+    return this.copy({ children });
+  }
+  cut(pos) {
+    let currentPos = this.border;
+    for (const child of this.children) {
+      if (pos === currentPos) {
+        return child;
+      }
+      if (pos > currentPos && pos < currentPos + child.nodeSize) {
+        return child.cut(pos - currentPos);
+      }
+      currentPos += child.nodeSize;
+    }
+    return null;
+  }
+  findIndex(pos) {
+    if (pos === 0) return { index: 0, offset: pos };
+    if (pos === this.contentSize) {
+      return { index: this.children.length, offset: pos };
+    }
+    if (pos > this.contentSize || pos < 0) {
+      throw new RangeError(`Position ${pos} outside of fragment`);
+    }
+    for (let i = 0, curPos = 0; ; i++) {
+      const cur = this.children[i];
+      const end = curPos + cur.nodeSize;
+      if (end >= pos) {
+        if (end === pos) return { index: i + 1, offset: end };
+        return { index: i, offset: curPos };
+      }
+      curPos = end;
+    }
+  }
+}
+class NoterefNode extends Node {
+}
+class FootnoteNode extends Node {
 }
 class Mark {
   constructor(tagName, attrs = {}) {
@@ -138,6 +262,9 @@ class TextNode {
   get nodeSize() {
     return this.text.length;
   }
+  get contentSize() {
+    return this.nodeSize;
+  }
   get textContent() {
     return this.text;
   }
@@ -152,20 +279,26 @@ class TextNode {
   copy(opts = {}) {
     return new TextNode(this.text, opts.marks ?? this.marks);
   }
+  cut(pos) {
+    return new TextNode(this.text.slice(pos));
+  }
 }
-function descendants(root, cb, pos = 0) {
-  for (const [i, child] of (0, import_itertools.enumerate)(root.children)) {
-    const descend = cb(child, pos, root, i);
+function descendants(node, cb, pos = 0) {
+  pos += node.border;
+  for (const [i, child] of (0, import_itertools.enumerate)(node.children)) {
+    const descend = cb(child, pos, node, i);
     if (descend && !child.isLeaf) {
-      descendants(child, cb, pos + child.border);
+      descendants(child, cb, pos);
     }
     pos += child.nodeSize;
   }
 }
 // Annotate the CommonJS export names for ESM import in node:
 0 && (module.exports = {
+  FootnoteNode,
   Mark,
   Node,
+  NoterefNode,
   Root,
   TextNode,
   descendants