@bendyline/squisq 2.11.5 → 2.11.7

package/dist/{chunk-AIB5WEK2.js → chunk-3YNFMFNK.js}

@@ -26,15 +26,6 @@ import {
   DEFAULT_THEME,
   FRONTMATTER_CUSTOM_TEMPLATES_KEY,
   FRONTMATTER_CUSTOM_THEMES_KEY,
-  oklchDarken,
-  pickContrastingText,
-  relativeLuminance,
-  resolveFontFamily,
-  validateCustomTemplateDefinition,
-  validateTheme,
-  withAlpha
-} from "./chunk-RXYKU6EX.js";
-import {
   VIEWPORT_PRESETS,
   cloneData,
   createTemplateContext,
@@ -42,8 +33,15 @@ import {
   getViewportOrientation,
   isPersistentLayerTemplate,
   isTemplateBlock,
-  scaledFontSize2 as scaledFontSize
-} from "./chunk-BAOV476U.js";
+  oklchDarken,
+  pickContrastingText,
+  relativeLuminance,
+  resolveFontFamily,
+  scaledFontSize2 as scaledFontSize,
+  validateCustomTemplateDefinition,
+  validateTheme,
+  withAlpha
+} from "./chunk-Y4D6TDZT.js";
 import {
   KNOWN_BLOCK_META_KEYS,
   TABLE_FED_TEMPLATES,
@@ -67,10 +65,7 @@ import {
 } from "./chunk-4VOD55SX.js";
 import {
   estimateReadingTime
-} from "./chunk-XOFT65EX.js";
-import {
-  estimateSpokenWordCount
-} from "./chunk-67L6WRJN.js";
+} from "./chunk-GHOPHGOK.js";
 import {
   SeededRandom,
   hashString
@@ -3184,7 +3179,7 @@ function imageWithCaption(input, context) {
     return layers;
   }
   if (caption) {
-    const captionBgY = captionPosition === "top" ? "7%" : captionPosition === "center" ? "42%" : "68%";
+    const captionBgY = captionPosition === "top" ? "8%" : captionPosition === "center" ? "43%" : "69%";
     const captionY = captionPosition === "top" ? "13%" : captionPosition === "center" ? "48%" : "74%";
     layers.push({
       type: "shape",
@@ -3197,7 +3192,7 @@ function imageWithCaption(input, context) {
         x: 0,
         y: captionBgY,
         width: "100%",
-        height: "16%"
+        height: "10%"
       }
     });
     layers.push({
@@ -3869,8 +3864,8 @@ function photoGrid(input, context) {
   if (caption) {
     const captionFontSize = themedFontSize(layout.stackColumns ? 24 : 26, context, false);
     const captionY = layout.stackColumns ? "78%" : "74%";
-    const captionBgY = layout.stackColumns ? "72%" : "68%";
-    const captionHeight = layout.stackColumns ? "13%" : "16%";
+    const captionBgY = layout.stackColumns ? "73%" : "70%";
+    const captionHeight = layout.stackColumns ? "10%" : "8%";
     const bg = theme.colors.background;
     layers.push({
       type: "shape",
@@ -4457,7 +4452,7 @@ function videoWithCaption(input, context) {
   ];
   if (caption) {
     const bg = theme.colors.background;
-    const captionBgY = captionPosition === "top" ? "7%" : captionPosition === "center" ? "42%" : "68%";
+    const captionBgY = captionPosition === "top" ? "8%" : captionPosition === "center" ? "43%" : "69%";
     const captionY = captionPosition === "top" ? "13%" : captionPosition === "center" ? "48%" : "74%";
     layers.push({
       type: "shape",
@@ -4470,7 +4465,7 @@ function videoWithCaption(input, context) {
         x: 0,
         y: captionBgY,
         width: "100%",
-        height: "16%"
+        height: "10%"
       }
     });
     layers.push({
@@ -10499,6 +10494,7 @@ function expandDocBlocks(blocks, options = {}) {
     blocksBySegment.get(segmentIndex).push({ block, originalIndex: index });
   });
   const expandedBlocks = new Array(blocks.length);
+  const splitParts = /* @__PURE__ */ new Map();
   for (const [segmentIndex, segmentBlocks] of blocksBySegment) {
     const audioSegment = audioSegments[segmentIndex];
     if (!audioSegment) {
@@ -10660,6 +10656,7 @@ function expandDocBlocks(blocks, options = {}) {
           const partDuration = block.duration / numParts;
           if (partDuration >= MIN_TRANSITION_GAP) {
             block.duration = partDuration;
+            const parts = [];
             for (let p = 1; p < numParts; p++) {
               const splitBlock = {
                 id: `${block.id}-split-${p}`,
@@ -10673,13 +10670,23 @@ function expandDocBlocks(blocks, options = {}) {
                 transition: { type: "dissolve", duration: 1 },
                 template: block.template
               };
-              expandedBlocks.push(splitBlock);
+              parts.push(splitBlock);
             }
+            splitParts.set(block, parts);
           }
         }
       }
     }
   }
+  if (splitParts.size > 0) {
+    const sequenced = [];
+    for (const block of expandedBlocks) {
+      sequenced.push(block);
+      const parts = block ? splitParts.get(block) : void 0;
+      if (parts) sequenced.push(...parts);
+    }
+    expandedBlocks.splice(0, expandedBlocks.length, ...sequenced);
+  }
   return mergeShortBlocks ? expandedBlocks.filter((block) => block && block.duration > 0) : expandedBlocks.filter(Boolean);
 }
 function getAvailableTemplates() {
@@ -11960,320 +11967,6 @@ function buildStartBlock(markdownDoc, rootBlocks, fileName) {
   return config;
 }
 
-// src/narration/syllables.ts
-var BOUNDARY_PUNCT = /^[.,!?;:'"()[\]{}]+|[.,!?;:'"()[\]{}]+$/g;
-var SYLLABLES_PER_SPOKEN_WORD = 1.4;
-function estimateSyllables(token) {
-  const stripped = token.replace(BOUNDARY_PUNCT, "");
-  if (!stripped) return 1;
-  if (/^[-+]?[\d.,]+$/.test(stripped)) {
-    const num = parseFloat(stripped.replace(/,/g, ""));
-    if (!Number.isNaN(num) && Number.isFinite(num)) {
-      return Math.max(1, Math.round(estimateSpokenWordCount(stripped) * SYLLABLES_PER_SPOKEN_WORD));
-    }
-  }
-  if (/^[A-Z]{2,5}$/.test(stripped) && !/[AEIOUY]/.test(stripped)) {
-    return stripped.length;
-  }
-  const parts = stripped.toLowerCase().split(/[^a-z']+/).filter((p) => p.length > 0);
-  if (parts.length === 0) return 1;
-  let total = 0;
-  for (const part of parts) {
-    total += countAlphaSyllables(part.replace(/'/g, ""));
-  }
-  return Math.max(1, total);
-}
-function countAlphaSyllables(word) {
-  if (!word) return 0;
-  const groups = word.match(/[aeiouy]+/g);
-  let count = groups ? groups.length : 0;
-  if (count === 0) return 1;
-  if (count > 1 && /[^aeiouy]e$/.test(word)) {
-    count -= 1;
-  }
-  if (count > 1 && /[^aeiouy]ed$/.test(word) && !/[td]ed$/.test(word)) {
-    count -= 1;
-  }
-  if (count > 1 && /[^aeiouy]es$/.test(word) && !/(?:[sxz]|[cs]h)es$/.test(word)) {
-    count -= 1;
-  }
-  if (/[^aeiouy]le$/.test(word)) {
-    count += 1;
-  }
-  return Math.max(1, count);
-}
-
-// src/narration/script.ts
-function blockSpokenText(block, includeTitles) {
-  const title = includeTitles && block.title ? block.title.trim() : "";
-  const body = getBlockBodyText2(block);
-  if (title && body) return `${title}
-${body}`;
-  return title || body;
-}
-function buildNarrationScript(doc, options) {
-  const includeTitles = options?.includeTitles ?? true;
-  const separator = options?.blockSeparator ?? "\n\n";
-  const pieces = [];
-  for (const block of flattenRenderableBlocks(doc.blocks)) {
-    const text = blockSpokenText(block, includeTitles);
-    if (!text) continue;
-    pieces.push({
-      blockId: block.id,
-      heading: block.title || void 0,
-      text
-    });
-  }
-  let sourceText = "";
-  const blocks = [];
-  for (let i = 0; i < pieces.length; i++) {
-    if (i > 0) sourceText += separator;
-    const charStart = sourceText.length;
-    sourceText += pieces[i].text;
-    blocks.push({
-      blockId: pieces[i].blockId,
-      ...pieces[i].heading !== void 0 ? { heading: pieces[i].heading } : {},
-      tokenStart: 0,
-      tokenEnd: 0,
-      charStart,
-      charEnd: sourceText.length
-    });
-  }
-  const tokens = [];
-  const tokenRe = /\S+/g;
-  let blockIndex = 0;
-  let match;
-  while ((match = tokenRe.exec(sourceText)) !== null) {
-    const charOffset = match.index;
-    const charEnd = charOffset + match[0].length;
-    while (blockIndex < blocks.length - 1 && charOffset >= blocks[blockIndex].charEnd) {
-      blockIndex++;
-    }
-    const spoken = hasSpokenContent(match[0]);
-    tokens.push({
-      text: match[0],
-      charOffset,
-      charEnd,
-      blockId: blocks[blockIndex]?.blockId ?? "",
-      blockIndex,
-      syllables: spoken ? estimateSyllables(match[0]) : 0,
-      spokenWordEquiv: spoken ? estimateSpokenWordCount(stripBoundaryPunct(match[0])) : 0,
-      pauseAfter: 0,
-      spoken
-    });
-  }
-  for (const range of blocks) {
-    range.tokenStart = tokens.length;
-    range.tokenEnd = 0;
-  }
-  for (let i = 0; i < tokens.length; i++) {
-    const range = blocks[tokens[i].blockIndex];
-    if (range) {
-      if (i < range.tokenStart) range.tokenStart = i;
-      range.tokenEnd = i + 1;
-    }
-  }
-  for (const range of blocks) {
-    if (range.tokenStart > range.tokenEnd) {
-      range.tokenStart = range.tokenEnd;
-    }
-  }
-  for (let i = 0; i < tokens.length; i++) {
-    const token = tokens[i];
-    const next = tokens[i + 1];
-    if (!next || next.blockIndex !== token.blockIndex) {
-      token.pauseAfter = 3;
-    } else if (sourceText.slice(token.charEnd, next.charOffset).includes("\n")) {
-      token.pauseAfter = 2;
-    } else if (endsWithClausePunctuation(token.text)) {
-      token.pauseAfter = 1;
-    }
-  }
-  const cumulativeSyllables = new Array(tokens.length + 1);
-  cumulativeSyllables[0] = 0;
-  for (let i = 0; i < tokens.length; i++) {
-    cumulativeSyllables[i + 1] = cumulativeSyllables[i] + tokens[i].syllables;
-  }
-  return {
-    sourceText,
-    tokens,
-    blocks,
-    totalSyllables: cumulativeSyllables[tokens.length],
-    cumulativeSyllables
-  };
-}
-var TRAILING_CLOSERS = /["')\]}»”’]+$/;
-function hasSpokenContent(token) {
-  return /[\p{L}\p{N}]/u.test(token);
-}
-function stripBoundaryPunct(token) {
-  return token.replace(/^[.,!?;:'"()[\]{}]+|[.,!?;:'"()[\]{}]+$/g, "");
-}
-function endsWithClausePunctuation(token) {
-  const bare = token.replace(TRAILING_CLOSERS, "");
-  const last = bare.charAt(bare.length - 1);
-  if (!last) return false;
-  if (last === ",") {
-    return true;
-  }
-  return last === "." || last === ";" || last === ":" || last === "!" || last === "?";
-}
-function clamp2(value, min, max) {
-  return value < min ? min : value > max ? max : value;
-}
-function expectedSyllablesAt(script, wordPos) {
-  const n = script.tokens.length;
-  if (n === 0) return 0;
-  const pos = clamp2(wordPos, 0, n);
-  const idx = Math.floor(pos);
-  if (idx >= n) return script.totalSyllables;
-  return script.cumulativeSyllables[idx] + (pos - idx) * script.tokens[idx].syllables;
-}
-function wordPosAtExpectedSyllables(script, syllables) {
-  const n = script.tokens.length;
-  if (n === 0) return 0;
-  const target = clamp2(syllables, 0, script.totalSyllables);
-  let lo = 0;
-  let hi = n;
-  while (lo < hi) {
-    const mid = lo + hi + 1 >> 1;
-    if (script.cumulativeSyllables[mid] <= target) lo = mid;
-    else hi = mid - 1;
-  }
-  if (lo >= n) return n;
-  const within = target - script.cumulativeSyllables[lo];
-  const sylls = script.tokens[lo].syllables;
-  return lo + (sylls > 0 ? within / sylls : 0);
-}
-function wordIndexAtChar(script, charOffset) {
-  const tokens = script.tokens;
-  if (tokens.length === 0) return 0;
-  let lo = 0;
-  let hi = tokens.length - 1;
-  while (lo < hi) {
-    const mid = lo + hi >> 1;
-    if (tokens[mid].charEnd <= charOffset) lo = mid + 1;
-    else hi = mid;
-  }
-  return lo;
-}
-function wordIndexAtTime(words, tSec) {
-  if (words.length === 0) return 0;
-  let lo = 0;
-  let hi = words.length - 1;
-  while (lo < hi) {
-    const mid = lo + hi + 1 >> 1;
-    if (words[mid].tSec <= tSec) lo = mid;
-    else hi = mid - 1;
-  }
-  return words[lo].tSec <= tSec ? lo : 0;
-}
-
-// src/narration/sidecar.ts
-function buildNarrationTimingJson(script, alignment, durationSec, options) {
-  const bookmarks = alignment.words.map((word) => {
-    const token = script.tokens[word.tokenIndex];
-    return {
-      id: `word-${word.tokenIndex}`,
-      time: word.tSec,
-      charOffset: token?.charOffset ?? 0,
-      textFragment: token?.text ?? ""
-    };
-  });
-  const blocks = alignment.blocks.map((range) => ({
-    ...range.blockId ? { blockId: range.blockId } : {},
-    ...range.heading !== void 0 ? { heading: range.heading } : {},
-    blockIndex: range.blockIndex,
-    charStart: range.charStart,
-    charEnd: range.charEnd,
-    startSec: range.startSec,
-    endSec: range.endSec
-  }));
-  return {
-    version: 3,
-    sourceText: script.sourceText,
-    duration: Number.isFinite(durationSec) && durationSec >= 0 ? durationSec : 0,
-    bookmarks,
-    blocks,
-    // Write only what `parseNarrationTimingJson` will read back, so the codec
-    // is a round-trip pair. A non-finite value would serialize to `null`
-    // (JSON has no NaN) and be dropped on parse anyway — omitting it here
-    // keeps the emitted sidecar honest rather than carrying a dead field.
-    ...finiteNumber(options?.cameraOffsetSec) ? { cameraOffsetSec: options.cameraOffsetSec } : {},
-    generator: {
-      name: "squisq-teleprompter",
-      method: "dsp-align",
-      ...options?.baseWpm !== void 0 ? { baseWpm: options.baseWpm } : {}
-    }
-  };
-}
-function finiteNonNegative(value) {
-  return typeof value === "number" && Number.isFinite(value) && value >= 0;
-}
-function finiteNumber(value) {
-  return typeof value === "number" && Number.isFinite(value);
-}
-function parseNarrationTimingJson(data) {
-  try {
-    const text = typeof data === "string" ? data : new TextDecoder().decode(data instanceof Uint8Array ? data : new Uint8Array(data));
-    const parsed = JSON.parse(text);
-    if (typeof parsed !== "object" || parsed === null) return null;
-    const raw = parsed;
-    if (typeof raw.sourceText !== "string" || !finiteNonNegative(raw.duration)) return null;
-    const bookmarks = Array.isArray(raw.bookmarks) ? raw.bookmarks.flatMap((b) => {
-      if (typeof b !== "object" || b === null) return [];
-      const bookmark = b;
-      if (!finiteNonNegative(bookmark.time) || !finiteNonNegative(bookmark.charOffset)) {
-        return [];
-      }
-      return [
-        {
-          id: typeof bookmark.id === "string" ? bookmark.id : `bookmark-${bookmark.charOffset}`,
-          time: bookmark.time,
-          charOffset: bookmark.charOffset,
-          ...typeof bookmark.textFragment === "string" ? { textFragment: bookmark.textFragment } : {}
-        }
-      ];
-    }) : [];
-    let blocks = Array.isArray(raw.blocks) ? raw.blocks.flatMap((b) => {
-      if (typeof b !== "object" || b === null) return [];
-      const block = b;
-      if (!finiteNonNegative(block.startSec) || !finiteNonNegative(block.endSec) || !finiteNonNegative(block.charStart) || !finiteNonNegative(block.charEnd)) {
-        return [];
-      }
-      return [
-        {
-          ...typeof block.blockId === "string" ? { blockId: block.blockId } : {},
-          ...typeof block.heading === "string" ? { heading: block.heading } : {},
-          blockIndex: finiteNonNegative(block.blockIndex) ? block.blockIndex : 0,
-          charStart: block.charStart,
-          charEnd: block.charEnd,
-          startSec: block.startSec,
-          endSec: Math.max(block.startSec, block.endSec)
-        }
-      ];
-    }) : [];
-    blocks = blocks.sort((a, b) => a.startSec - b.startSec);
-    let cursor = 0;
-    for (const block of blocks) {
-      if (block.startSec < cursor) block.startSec = cursor;
-      if (block.endSec < block.startSec) block.endSec = block.startSec;
-      cursor = block.startSec;
-    }
-    return {
-      version: 3,
-      sourceText: raw.sourceText,
-      duration: raw.duration,
-      bookmarks,
-      blocks,
-      ...finiteNumber(raw.cameraOffsetSec) ? { cameraOffsetSec: raw.cameraOffsetSec } : {},
-      ...typeof raw.generator === "object" && raw.generator !== null ? { generator: raw.generator } : {}
-    };
-  } catch {
-    return null;
-  }
-}
-
 export {
   expandPersistentLayers,
   getPersistentLayersFromTheme,
@@ -12399,13 +12092,5 @@ export {
   flattenRenderableBlocks,
   countBlocks,
   getBlockDepth,
-  getBlockBodyText2 as getBlockBodyText,
-  estimateSyllables,
-  buildNarrationScript,
-  expectedSyllablesAt,
-  wordPosAtExpectedSyllables,
-  wordIndexAtChar,
-  wordIndexAtTime,
-  buildNarrationTimingJson,
-  parseNarrationTimingJson
+  getBlockBodyText2 as getBlockBodyText
 };

package/dist/chunk-44BJ42GT.js

@@ -0,0 +1,165 @@
+import {
+  estimateTimeFromText
+} from "./chunk-GHOPHGOK.js";
+
+// src/timing/captions.ts
+var DEFAULT_VERSION = 1;
+var DEFAULT_TARGET_WORDS = 10;
+var DEFAULT_MIN_DURATION = 1.5;
+var DEFAULT_MAX_DURATION = 6;
+var DEFAULT_WORDS_PER_SECOND = 2.5;
+function phrasesFromTiming(source, segmentIndex, targetWords) {
+  const { segment, timing } = source;
+  if (!timing || timing.bookmarks.length === 0) return [];
+  const segmentEnd = segment.startTime + segment.duration;
+  const phrases = [];
+  let words = [];
+  let phraseText = "";
+  for (let index = 0; index < timing.bookmarks.length; index += 1) {
+    const bookmark = timing.bookmarks[index];
+    const text = bookmark.textFragment?.trim() ?? "";
+    if (!text) continue;
+    const next = timing.bookmarks[index + 1];
+    const startTime = Math.min(
+      segmentEnd,
+      Math.max(segment.startTime, segment.startTime + bookmark.time)
+    );
+    const endTime = Math.min(
+      segmentEnd,
+      Math.max(startTime, segment.startTime + (next?.time ?? timing.duration))
+    );
+    const word = {
+      text,
+      startTime,
+      endTime
+    };
+    words.push(word);
+    phraseText += `${phraseText ? " " : ""}${text}`;
+    const flush = /[.!?]$/.test(text) && words.length >= 4 || /[,;:]$/.test(text) && words.length >= targetWords - 2 || words.length >= targetWords + 3 || index === timing.bookmarks.length - 1;
+    if (flush) {
+      phrases.push({
+        text: phraseText,
+        startTime: words[0].startTime,
+        endTime: words[words.length - 1].endTime,
+        audioSegment: segmentIndex,
+        words
+      });
+      words = [];
+      phraseText = "";
+    }
+  }
+  return phrases;
+}
+function phrasesFromText(source, segmentIndex, options) {
+  const { segment } = source;
+  const sourceText = source.text?.trim();
+  if (!sourceText) return [];
+  const prefix = source.spokenPrefix ?? "";
+  const prefixWords = prefix.split(/\s+/).filter(Boolean).length;
+  const prefixDuration = Math.min(
+    segment.duration,
+    prefixWords / options.wordsPerSecond + (prefix.match(/\n\n/g)?.length ?? 0) * 0.5
+  );
+  const contentDuration = Math.max(0, segment.duration - prefixDuration);
+  const phrases = [];
+  const prefixText = prefix.replace(/\n+/g, " ").trim();
+  if (prefixText && prefixDuration > 0.5) {
+    phrases.push({
+      text: prefixText,
+      startTime: segment.startTime,
+      endTime: Math.min(segment.startTime + segment.duration, segment.startTime + prefixDuration),
+      audioSegment: segmentIndex
+    });
+  }
+  const sentences = sourceText.replace(/\n+/g, " ").split(/(?<=[.!?])\s+/).map((sentence) => sentence.trim()).filter(Boolean);
+  let phraseText = "";
+  let phraseWordCount = 0;
+  let phraseStartOffset = 0;
+  let currentOffset = 0;
+  const flush = () => {
+    if (!phraseText) return;
+    const startTime = segment.startTime + prefixDuration + estimateTimeFromText(sourceText, phraseStartOffset, contentDuration);
+    const estimatedEnd = segment.startTime + prefixDuration + estimateTimeFromText(sourceText, currentOffset, contentDuration);
+    phrases.push({
+      text: phraseText,
+      startTime,
+      endTime: Math.min(
+        segment.startTime + segment.duration,
+        Math.min(
+          Math.max(startTime + options.minPhraseDuration, estimatedEnd),
+          startTime + options.maxPhraseDuration
+        )
+      ),
+      audioSegment: segmentIndex
+    });
+    phraseText = "";
+    phraseWordCount = 0;
+    phraseStartOffset = currentOffset;
+  };
+  for (const sentence of sentences) {
+    const sentenceWords = sentence.split(/\s+/).filter(Boolean).length;
+    if (phraseWordCount > 0 && phraseWordCount + sentenceWords > options.targetWordsPerPhrase * 1.5) {
+      flush();
+    }
+    phraseText += `${phraseText ? " " : ""}${sentence}`;
+    phraseWordCount += sentenceWords;
+    currentOffset += sentence.length + 1;
+    if (phraseWordCount >= options.targetWordsPerPhrase) flush();
+  }
+  flush();
+  return phrases;
+}
+function buildCaptionTrack(sources, options = {}) {
+  const targetWordsPerPhrase = options.targetWordsPerPhrase ?? DEFAULT_TARGET_WORDS;
+  const textOptions = {
+    targetWordsPerPhrase,
+    minPhraseDuration: options.minPhraseDuration ?? DEFAULT_MIN_DURATION,
+    maxPhraseDuration: options.maxPhraseDuration ?? DEFAULT_MAX_DURATION,
+    wordsPerSecond: options.wordsPerSecond ?? DEFAULT_WORDS_PER_SECOND
+  };
+  if (!Number.isInteger(targetWordsPerPhrase) || targetWordsPerPhrase < 1) {
+    throw new Error("targetWordsPerPhrase must be a positive integer");
+  }
+  if (!Number.isFinite(textOptions.wordsPerSecond) || textOptions.wordsPerSecond <= 0) {
+    throw new Error("wordsPerSecond must be a positive finite number");
+  }
+  if (!Number.isFinite(textOptions.minPhraseDuration) || !Number.isFinite(textOptions.maxPhraseDuration) || textOptions.minPhraseDuration < 0 || textOptions.maxPhraseDuration < textOptions.minPhraseDuration) {
+    throw new Error(
+      "caption phrase durations must be finite with max greater than or equal to min"
+    );
+  }
+  const phrases = sources.flatMap((source, segmentIndex) => {
+    const timed = phrasesFromTiming(source, segmentIndex, targetWordsPerPhrase);
+    return timed.length > 0 ? timed : phrasesFromText(source, segmentIndex, textOptions);
+  });
+  if (options.closePhraseGaps ?? true) {
+    for (let index = 0; index < phrases.length - 1; index += 1) {
+      if (phrases[index].audioSegment === phrases[index + 1].audioSegment) {
+        phrases[index].endTime = Math.max(phrases[index].startTime, phrases[index + 1].startTime);
+      }
+    }
+    for (let segmentIndex = 0; segmentIndex < sources.length; segmentIndex += 1) {
+      const segment = sources[segmentIndex].segment;
+      let last;
+      for (let index = phrases.length - 1; index >= 0; index -= 1) {
+        if (phrases[index].audioSegment === segmentIndex) {
+          last = phrases[index];
+          break;
+        }
+      }
+      if (last) {
+        const segmentEnd = segment.startTime + segment.duration;
+        last.endTime = Math.min(segmentEnd, Math.max(last.endTime, segmentEnd - 0.05));
+      }
+    }
+  }
+  return {
+    phrases,
+    ...options.generatedAt ? { generatedAt: options.generatedAt } : {},
+    version: options.version ?? DEFAULT_VERSION
+  };
+}
+
+export {
+  buildCaptionTrack
+};