@storyteller-platform/align 0.1.6 → 0.1.7

package/dist/align/align.cjs

@@ -89,10 +89,12 @@ var import_ffmpeg = require("../common/ffmpeg.cjs");
 var import_segmentation = require("../markup/segmentation.cjs");
 var import_fuzzy = require("./fuzzy.cjs");
 var import_getSentenceRanges = require("./getSentenceRanges.cjs");
+var import_slugify = require("./slugify.cjs");
 const OFFSET_SEARCH_WINDOW_SIZE = 5e3;
 async function align(input, output, transcriptionsDir, audiobookDir, options) {
   var _stack = [];
   try {
+    await (0, import_promises.mkdir)((0, import_posix.dirname)(output), { recursive: true });
     await (0, import_promises.copyFile)(input, output);
     const audiobookFiles = await (0, import_promises.readdir)(audiobookDir).then(
       (filenames) => filenames.filter((f) => (0, import_audiobook.isAudioFile)(f)).map((f) => (0, import_node_path.join)(audiobookDir, f))
@@ -118,6 +120,7 @@ async function align(input, output, transcriptionsDir, audiobookDir, options) {
       options.logger
     );
     const timing = await aligner.alignBook(options.onProgress);
+    await epub.saveAndClose();
     if (options.reportsPath) {
       await (0, import_promises.mkdir)((0, import_node_path.dirname)(options.reportsPath), { recursive: true });
       await (0, import_promises.writeFile)(
@@ -152,7 +155,7 @@ class Aligner {
   report = {
     chapters: []
   };
-  findBestOffset(epubSentences, transcriptionText, lastMatchOffset) {
+  findBestOffset(epubSentences, transcriptionText, lastMatchOffset, mapping) {
     let i = 0;
     while (i < transcriptionText.length) {
       let startSentence = 0;
@@ -164,11 +167,13 @@ class Aligner {
       let startSeen = null;
       let endSeen = null;
       for (const aligned of this.alignedChapters) {
-        if (startSeen !== null && endSeen === aligned.startOffset) {
-          endSeen = aligned.endOffset;
+        const alignedStart = mapping.map(aligned.startOffset, -1);
+        const alignedEnd = mapping.map(aligned.endOffset, -1);
+        if (startSeen !== null && endSeen === alignedStart) {
+          endSeen = alignedEnd;
         } else {
-          startSeen = aligned.startOffset;
-          endSeen = aligned.endOffset;
+          startSeen = alignedStart;
+          endSeen = alignedEnd;
         }
         if (startIndex >= startSeen && startIndex < endSeen) {
           startIndex = endSeen;
@@ -183,7 +188,7 @@ class Aligner {
           endIndex
         );
         while (startSentence < epubSentences.length) {
-          const queryString = epubSentences.slice(startSentence, startSentence + 6).join(" ");
+          const queryString = epubSentences.slice(startSentence, startSentence + 6).join("-");
           const firstMatch = (0, import_fuzzy.findNearestMatch)(
             queryString.toLowerCase(),
             transcriptionTextSlice.toLowerCase(),
@@ -309,7 +314,7 @@ class Aligner {
       }, [])
     });
   }
-  async alignChapter(startSentence, chapterId, transcriptionOffset, lastSentenceRange) {
+  async alignChapter(startSentence, chapterId, transcriptionOffset, locale, lastSentenceRange) {
     const timing = (0, import_ghost_story.createTiming)();
     timing.start("read contents");
     const manifest = await this.epub.getManifest();
@@ -329,6 +334,7 @@ class Aligner {
       this.transcription,
       chapterSentences,
       transcriptionOffset,
+      locale,
       lastSentenceRange
     );
     timing.end("align sentences");
@@ -369,53 +375,67 @@ class Aligner {
     };
   }
   async alignBook(onProgress) {
-    var _a, _b, _c, _d, _e, _f;
-    this.timing.setMetadata(
-      "language",
-      ((_a = this.languageOverride ?? await this.epub.getLanguage()) == null ? void 0 : _a.language) ?? "unknown"
-    );
+    var _a, _b, _c, _d, _e, _f, _g;
+    const locale = this.languageOverride ?? await this.epub.getLanguage() ?? new Intl.Locale("en-US");
+    this.timing.setMetadata("language", locale.toString());
     this.timing.setMetadata("granularity", this.granularity);
     const spine = await this.epub.getSpineItems();
-    const transcriptionText = this.transcription.transcript;
+    const manifest = await this.epub.getManifest();
+    const { result: transcriptionText, mapping } = await (0, import_slugify.slugify)(
+      this.transcription.transcript,
+      locale
+    );
     let lastTranscriptionOffset = 0;
     let lastSentenceRange = null;
     for (let index = 0; index < spine.length; index++) {
       onProgress == null ? void 0 : onProgress(index / spine.length);
       const spineItem = spine[index];
-      (_b = this.logger) == null ? void 0 : _b.info(
+      (_a = this.logger) == null ? void 0 : _a.info(
         `Aligning epub item #${index} : ${(0, import_posix.basename)(spineItem.href)}`
       );
       const chapterId = spineItem.id;
+      if ((_c = (_b = manifest[chapterId]) == null ? void 0 : _b.properties) == null ? void 0 : _c.includes("nav")) {
+        continue;
+      }
       const chapterSentences = await this.getChapterSentences(chapterId);
+      const slugifiedChapterSentences = [];
+      for (const chapterSentence of chapterSentences) {
+        slugifiedChapterSentences.push(
+          (await (0, import_slugify.slugify)(chapterSentence, locale)).result
+        );
+      }
       if (chapterSentences.length === 0) {
-        (_c = this.logger) == null ? void 0 : _c.info(`Chapter #${index} has no text; skipping`);
+        (_d = this.logger) == null ? void 0 : _d.info(`Chapter #${index} has no text; skipping`);
         continue;
       }
       if (chapterSentences.length < 2 && // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
       chapterSentences[0].split(" ").length < 4) {
-        (_d = this.logger) == null ? void 0 : _d.info(
+        (_e = this.logger) == null ? void 0 : _e.info(
           `Chapter #${index} is fewer than four words; skipping`
         );
         continue;
       }
-      const { startSentence, transcriptionOffset } = this.findBestOffset(
-        chapterSentences,
+      const { startSentence, transcriptionOffset: slugifiedOffset } = this.findBestOffset(
+        slugifiedChapterSentences,
         transcriptionText,
-        lastTranscriptionOffset
+        mapping.map(lastTranscriptionOffset, -1),
+        mapping
       );
+      const transcriptionOffset = slugifiedOffset && mapping.invert().map(slugifiedOffset, -1);
       if (transcriptionOffset === null) {
-        (_e = this.logger) == null ? void 0 : _e.info(
+        (_f = this.logger) == null ? void 0 : _f.info(
           `Couldn't find matching transcription for chapter #${index}`
         );
         continue;
       }
-      (_f = this.logger) == null ? void 0 : _f.info(
+      (_g = this.logger) == null ? void 0 : _g.info(
         `Chapter #${index} best matches transcription at offset ${transcriptionOffset}, starting at sentence ${startSentence}`
       );
       const result = await this.alignChapter(
         startSentence,
         chapterId,
         transcriptionOffset,
+        locale,
         lastSentenceRange
       );
       lastSentenceRange = result.lastSentenceRange;

package/dist/align/align.js

@@ -23,10 +23,12 @@ import {
   getSentenceRanges,
   interpolateSentenceRanges
 } from "./getSentenceRanges.js";
+import { slugify } from "./slugify.js";
 const OFFSET_SEARCH_WINDOW_SIZE = 5e3;
 async function align(input, output, transcriptionsDir, audiobookDir, options) {
   var _stack = [];
   try {
+    await mkdir(dirname(output), { recursive: true });
     await copyFile(input, output);
     const audiobookFiles = await readdir(audiobookDir).then(
       (filenames) => filenames.filter((f) => isAudioFile(f)).map((f) => autoJoin(audiobookDir, f))
@@ -52,6 +54,7 @@ async function align(input, output, transcriptionsDir, audiobookDir, options) {
       options.logger
     );
     const timing = await aligner.alignBook(options.onProgress);
+    await epub.saveAndClose();
     if (options.reportsPath) {
       await mkdir(autoDirname(options.reportsPath), { recursive: true });
       await writeFile(
@@ -86,7 +89,7 @@ class Aligner {
   report = {
     chapters: []
   };
-  findBestOffset(epubSentences, transcriptionText, lastMatchOffset) {
+  findBestOffset(epubSentences, transcriptionText, lastMatchOffset, mapping) {
     let i = 0;
     while (i < transcriptionText.length) {
       let startSentence = 0;
@@ -98,11 +101,13 @@ class Aligner {
       let startSeen = null;
       let endSeen = null;
       for (const aligned of this.alignedChapters) {
-        if (startSeen !== null && endSeen === aligned.startOffset) {
-          endSeen = aligned.endOffset;
+        const alignedStart = mapping.map(aligned.startOffset, -1);
+        const alignedEnd = mapping.map(aligned.endOffset, -1);
+        if (startSeen !== null && endSeen === alignedStart) {
+          endSeen = alignedEnd;
         } else {
-          startSeen = aligned.startOffset;
-          endSeen = aligned.endOffset;
+          startSeen = alignedStart;
+          endSeen = alignedEnd;
         }
         if (startIndex >= startSeen && startIndex < endSeen) {
           startIndex = endSeen;
@@ -117,7 +122,7 @@ class Aligner {
           endIndex
         );
         while (startSentence < epubSentences.length) {
-          const queryString = epubSentences.slice(startSentence, startSentence + 6).join(" ");
+          const queryString = epubSentences.slice(startSentence, startSentence + 6).join("-");
           const firstMatch = findNearestMatch(
             queryString.toLowerCase(),
             transcriptionTextSlice.toLowerCase(),
@@ -243,7 +248,7 @@ class Aligner {
       }, [])
     });
   }
-  async alignChapter(startSentence, chapterId, transcriptionOffset, lastSentenceRange) {
+  async alignChapter(startSentence, chapterId, transcriptionOffset, locale, lastSentenceRange) {
     const timing = createTiming();
     timing.start("read contents");
     const manifest = await this.epub.getManifest();
@@ -263,6 +268,7 @@ class Aligner {
       this.transcription,
       chapterSentences,
       transcriptionOffset,
+      locale,
       lastSentenceRange
     );
     timing.end("align sentences");
@@ -303,53 +309,67 @@ class Aligner {
     };
   }
   async alignBook(onProgress) {
-    var _a, _b, _c, _d, _e, _f;
-    this.timing.setMetadata(
-      "language",
-      ((_a = this.languageOverride ?? await this.epub.getLanguage()) == null ? void 0 : _a.language) ?? "unknown"
-    );
+    var _a, _b, _c, _d, _e, _f, _g;
+    const locale = this.languageOverride ?? await this.epub.getLanguage() ?? new Intl.Locale("en-US");
+    this.timing.setMetadata("language", locale.toString());
     this.timing.setMetadata("granularity", this.granularity);
     const spine = await this.epub.getSpineItems();
-    const transcriptionText = this.transcription.transcript;
+    const manifest = await this.epub.getManifest();
+    const { result: transcriptionText, mapping } = await slugify(
+      this.transcription.transcript,
+      locale
+    );
     let lastTranscriptionOffset = 0;
     let lastSentenceRange = null;
     for (let index = 0; index < spine.length; index++) {
       onProgress == null ? void 0 : onProgress(index / spine.length);
       const spineItem = spine[index];
-      (_b = this.logger) == null ? void 0 : _b.info(
+      (_a = this.logger) == null ? void 0 : _a.info(
         `Aligning epub item #${index} : ${basename(spineItem.href)}`
       );
       const chapterId = spineItem.id;
+      if ((_c = (_b = manifest[chapterId]) == null ? void 0 : _b.properties) == null ? void 0 : _c.includes("nav")) {
+        continue;
+      }
       const chapterSentences = await this.getChapterSentences(chapterId);
+      const slugifiedChapterSentences = [];
+      for (const chapterSentence of chapterSentences) {
+        slugifiedChapterSentences.push(
+          (await slugify(chapterSentence, locale)).result
+        );
+      }
       if (chapterSentences.length === 0) {
-        (_c = this.logger) == null ? void 0 : _c.info(`Chapter #${index} has no text; skipping`);
+        (_d = this.logger) == null ? void 0 : _d.info(`Chapter #${index} has no text; skipping`);
         continue;
       }
       if (chapterSentences.length < 2 && // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
       chapterSentences[0].split(" ").length < 4) {
-        (_d = this.logger) == null ? void 0 : _d.info(
+        (_e = this.logger) == null ? void 0 : _e.info(
           `Chapter #${index} is fewer than four words; skipping`
         );
         continue;
       }
-      const { startSentence, transcriptionOffset } = this.findBestOffset(
-        chapterSentences,
+      const { startSentence, transcriptionOffset: slugifiedOffset } = this.findBestOffset(
+        slugifiedChapterSentences,
         transcriptionText,
-        lastTranscriptionOffset
+        mapping.map(lastTranscriptionOffset, -1),
+        mapping
       );
+      const transcriptionOffset = slugifiedOffset && mapping.invert().map(slugifiedOffset, -1);
       if (transcriptionOffset === null) {
-        (_e = this.logger) == null ? void 0 : _e.info(
+        (_f = this.logger) == null ? void 0 : _f.info(
           `Couldn't find matching transcription for chapter #${index}`
         );
         continue;
       }
-      (_f = this.logger) == null ? void 0 : _f.info(
+      (_g = this.logger) == null ? void 0 : _g.info(
         `Chapter #${index} best matches transcription at offset ${transcriptionOffset}, starting at sentence ${startSentence}`
       );
       const result = await this.alignChapter(
         startSentence,
         chapterId,
         transcriptionOffset,
+        locale,
         lastSentenceRange
       );
       lastSentenceRange = result.lastSentenceRange;

package/dist/align/fuzzy.cjs

@@ -108,7 +108,7 @@ function expand(subsequence, sequence, maxDist) {
 function* levenshteinNgram(subsequence, sequence, maxDist) {
   const subsequenceLength = subsequence.length;
   const sequenceLength = sequence.length;
-  const ngramLength = Math.round(subsequenceLength / (maxDist + 1));
+  const ngramLength = Math.floor(subsequenceLength / (maxDist + 1));
   if (ngramLength === 0) {
     throw new Error("The subsequence length must be greater than maxDist");
   }

package/dist/align/fuzzy.js

@@ -86,7 +86,7 @@ function expand(subsequence, sequence, maxDist) {
 function* levenshteinNgram(subsequence, sequence, maxDist) {
   const subsequenceLength = subsequence.length;
   const sequenceLength = sequence.length;
-  const ngramLength = Math.round(subsequenceLength / (maxDist + 1));
+  const ngramLength = Math.floor(subsequenceLength / (maxDist + 1));
   if (ngramLength === 0) {
     throw new Error("The subsequence length must be greater than maxDist");
   }

package/dist/align/getSentenceRanges.cjs

@@ -28,6 +28,7 @@ module.exports = __toCommonJS(getSentenceRanges_exports);
 var import_text_segmentation = require("@echogarden/text-segmentation");
 var import_ffmpeg = require("../common/ffmpeg.cjs");
 var import_fuzzy = require("./fuzzy.cjs");
+var import_slugify = require("./slugify.cjs");
 async function getSentencesWithOffsets(text) {
   const sentences = await (0, import_text_segmentation.segmentText)(text).then(
     (r) => r.sentences.map((s) => s.text)
@@ -75,7 +76,7 @@ function getWindowIndexFromOffset(window, offset) {
 function collapseWhitespace(input) {
   return input.replaceAll(/\s+/g, " ");
 }
-async function getSentenceRanges(startSentence, transcription, sentences, chapterOffset, lastSentenceRange) {
+async function getSentenceRanges(startSentence, transcription, sentences, chapterOffset, locale, lastSentenceRange) {
   const sentenceRanges = [];
   const fullTranscriptionText = transcription.transcript;
   const transcriptionText = fullTranscriptionText.slice(chapterOffset);
@@ -83,13 +84,15 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
     transcriptionText
   ).then((s) => s.map((sentence) => sentence.toLowerCase()));
   let startSentenceEntry = startSentence;
-  const sentenceEntries = sentences.map((sentence, index) => [index, sentence]).filter(([index, sentence]) => {
-    if (sentence.replaceAll(/[.-_()[\],/?!@#$%^^&*`~;:='"<>+ˌˈ]/g, "").length <= 3) {
-      if (index < startSentence) startSentenceEntry--;
-      return false;
+  const sentenceEntries = [];
+  for (let i = 0; i < sentences.length; i++) {
+    const sentence = (await (0, import_slugify.slugify)(sentences[i], locale)).result;
+    if (sentence.length <= 3) {
+      if (i < startSentence) startSentenceEntry--;
+      continue;
     }
-    return true;
-  });
+    sentenceEntries.push([i, sentence]);
+  }
   let transcriptionWindowIndex = 0;
   let transcriptionWindowOffset = 0;
   let lastGoodTranscriptionWindow = 0;
@@ -102,7 +105,11 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
       transcriptionWindowIndex,
       transcriptionWindowIndex + 10
     );
-    const transcriptionWindow = transcriptionWindowList.join("").slice(transcriptionWindowOffset);
+    const { result: transcriptionWindow, mapping } = await (0, import_slugify.slugify)(
+      transcriptionWindowList.join("-").slice(transcriptionWindowOffset),
+      locale
+    );
+    const inverted = mapping.invert();
     const query = collapseWhitespace(sentence.trim()).toLowerCase();
     const firstMatch = (0, import_fuzzy.findNearestMatch)(
       query,
@@ -125,8 +132,13 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
       continue;
     }
     const transcriptionOffset = transcriptionSentences.slice(0, transcriptionWindowIndex).join("").length;
+    const matchStart = inverted.map(firstMatch.index, 1);
+    const matchEnd = inverted.map(
+      firstMatch.index + firstMatch.match.length,
+      -1
+    );
     const startResult = findStartTimestamp(
-      firstMatch.index + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
+      matchStart + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
       transcription
     );
     if (!startResult) {
@@ -136,7 +148,7 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
     let start = startResult.start;
     const audiofile = startResult.audiofile;
     const end = findEndTimestamp(
-      firstMatch.index + firstMatch.match.length + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
+      matchEnd + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
       transcription
     ) ?? startResult.end;
     if (sentenceRanges.length > 0) {
@@ -177,10 +189,10 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
       audiofile
     });
     notFound = 0;
-    lastMatchEnd = firstMatch.index + firstMatch.match.length + transcriptionOffset + transcriptionWindowOffset + chapterOffset;
+    lastMatchEnd = matchEnd + transcriptionOffset + transcriptionWindowOffset + chapterOffset;
     const windowIndexResult = getWindowIndexFromOffset(
       transcriptionWindowList,
-      firstMatch.index + firstMatch.match.length + transcriptionWindowOffset
+      matchEnd + transcriptionWindowOffset
     );
     transcriptionWindowIndex += windowIndexResult.index;
     transcriptionWindowOffset = windowIndexResult.offset;

package/dist/align/getSentenceRanges.d.cts

@@ -14,7 +14,7 @@ type SentenceRange = {
     audiofile: string;
 };
 declare function findEndTimestamp(matchEndIndex: number, transcription: StorytellerTranscription): number | null;
-declare function getSentenceRanges(startSentence: number, transcription: StorytellerTranscription, sentences: string[], chapterOffset: number, lastSentenceRange: SentenceRange | null): Promise<{
+declare function getSentenceRanges(startSentence: number, transcription: StorytellerTranscription, sentences: string[], chapterOffset: number, locale: Intl.Locale, lastSentenceRange: SentenceRange | null): Promise<{
     sentenceRanges: SentenceRange[];
     transcriptionOffset: number;
 }>;

package/dist/align/getSentenceRanges.d.ts

@@ -14,7 +14,7 @@ type SentenceRange = {
     audiofile: string;
 };
 declare function findEndTimestamp(matchEndIndex: number, transcription: StorytellerTranscription): number | null;
-declare function getSentenceRanges(startSentence: number, transcription: StorytellerTranscription, sentences: string[], chapterOffset: number, lastSentenceRange: SentenceRange | null): Promise<{
+declare function getSentenceRanges(startSentence: number, transcription: StorytellerTranscription, sentences: string[], chapterOffset: number, locale: Intl.Locale, lastSentenceRange: SentenceRange | null): Promise<{
     sentenceRanges: SentenceRange[];
     transcriptionOffset: number;
 }>;

package/dist/align/getSentenceRanges.js

@@ -2,6 +2,7 @@ import "../chunk-BIEQXUOY.js";
 import { segmentText } from "@echogarden/text-segmentation";
 import { getTrackDuration } from "../common/ffmpeg.js";
 import { findNearestMatch } from "./fuzzy.js";
+import { slugify } from "./slugify.js";
 async function getSentencesWithOffsets(text) {
   const sentences = await segmentText(text).then(
     (r) => r.sentences.map((s) => s.text)
@@ -49,7 +50,7 @@ function getWindowIndexFromOffset(window, offset) {
 function collapseWhitespace(input) {
   return input.replaceAll(/\s+/g, " ");
 }
-async function getSentenceRanges(startSentence, transcription, sentences, chapterOffset, lastSentenceRange) {
+async function getSentenceRanges(startSentence, transcription, sentences, chapterOffset, locale, lastSentenceRange) {
   const sentenceRanges = [];
   const fullTranscriptionText = transcription.transcript;
   const transcriptionText = fullTranscriptionText.slice(chapterOffset);
@@ -57,13 +58,15 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
     transcriptionText
   ).then((s) => s.map((sentence) => sentence.toLowerCase()));
   let startSentenceEntry = startSentence;
-  const sentenceEntries = sentences.map((sentence, index) => [index, sentence]).filter(([index, sentence]) => {
-    if (sentence.replaceAll(/[.-_()[\],/?!@#$%^^&*`~;:='"<>+ˌˈ]/g, "").length <= 3) {
-      if (index < startSentence) startSentenceEntry--;
-      return false;
+  const sentenceEntries = [];
+  for (let i = 0; i < sentences.length; i++) {
+    const sentence = (await slugify(sentences[i], locale)).result;
+    if (sentence.length <= 3) {
+      if (i < startSentence) startSentenceEntry--;
+      continue;
     }
-    return true;
-  });
+    sentenceEntries.push([i, sentence]);
+  }
   let transcriptionWindowIndex = 0;
   let transcriptionWindowOffset = 0;
   let lastGoodTranscriptionWindow = 0;
@@ -76,7 +79,11 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
       transcriptionWindowIndex,
       transcriptionWindowIndex + 10
     );
-    const transcriptionWindow = transcriptionWindowList.join("").slice(transcriptionWindowOffset);
+    const { result: transcriptionWindow, mapping } = await slugify(
+      transcriptionWindowList.join("-").slice(transcriptionWindowOffset),
+      locale
+    );
+    const inverted = mapping.invert();
     const query = collapseWhitespace(sentence.trim()).toLowerCase();
     const firstMatch = findNearestMatch(
       query,
@@ -99,8 +106,13 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
       continue;
     }
     const transcriptionOffset = transcriptionSentences.slice(0, transcriptionWindowIndex).join("").length;
+    const matchStart = inverted.map(firstMatch.index, 1);
+    const matchEnd = inverted.map(
+      firstMatch.index + firstMatch.match.length,
+      -1
+    );
     const startResult = findStartTimestamp(
-      firstMatch.index + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
+      matchStart + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
       transcription
     );
     if (!startResult) {
@@ -110,7 +122,7 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
     let start = startResult.start;
     const audiofile = startResult.audiofile;
     const end = findEndTimestamp(
-      firstMatch.index + firstMatch.match.length + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
+      matchEnd + transcriptionOffset + transcriptionWindowOffset + chapterOffset,
       transcription
     ) ?? startResult.end;
     if (sentenceRanges.length > 0) {
@@ -151,10 +163,10 @@ async function getSentenceRanges(startSentence, transcription, sentences, chapte
       audiofile
     });
     notFound = 0;
-    lastMatchEnd = firstMatch.index + firstMatch.match.length + transcriptionOffset + transcriptionWindowOffset + chapterOffset;
+    lastMatchEnd = matchEnd + transcriptionOffset + transcriptionWindowOffset + chapterOffset;
     const windowIndexResult = getWindowIndexFromOffset(
       transcriptionWindowList,
-      firstMatch.index + firstMatch.match.length + transcriptionWindowOffset
+      matchEnd + transcriptionWindowOffset
     );
     transcriptionWindowIndex += windowIndexResult.index;
     transcriptionWindowOffset = windowIndexResult.offset;

package/dist/align/slugify.cjs

@@ -0,0 +1,125 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var slugify_exports = {};
+__export(slugify_exports, {
+  slugify: () => slugify
+});
+module.exports = __toCommonJS(slugify_exports);
+var import_locale_currency = require("locale-currency");
+var import_to_words = require("to-words");
+var import_transliteration = require("@storyteller-platform/transliteration");
+const replacerMap = /* @__PURE__ */ new WeakMap();
+function createReplacers(locale) {
+  const maximizedLocale = locale.maximize();
+  const demoNumber = 123456.789;
+  const currencyFormat = new Intl.NumberFormat(locale, {
+    style: "currency",
+    // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
+    currency: (0, import_locale_currency.getCurrency)(locale.maximize().region)
+  });
+  const currencyParts = currencyFormat.formatToParts(demoNumber);
+  const currencySymbols = currencyParts.reduce(
+    (acc, part, index) => {
+      if (part.type === "group") {
+        return {
+          ...acc,
+          group: part.value
+        };
+      }
+      if (part.type === "decimal") {
+        return {
+          ...acc,
+          decimal: part.value
+        };
+      }
+      if (part.type === "currency") {
+        return {
+          ...acc,
+          currency: part.value,
+          currencyLeading: index === 0
+        };
+      }
+      return acc;
+    },
+    { group: "", decimal: "", currency: "", currencyLeading: true }
+  );
+  const numeralRegexPart = `(\\p{Number}[\\p{Number}${currencySymbols.group}]*(?:[${currencySymbols.decimal}]\\p{Number}*)?)`;
+  const currencyRegex = currencySymbols.currencyLeading ? new RegExp(`[${currencySymbols.currency}]\\s?${numeralRegexPart}`, "gu") : new RegExp(`${numeralRegexPart}\\s?[${currencySymbols.currency}]`, "gu");
+  function currencyReplacer(match) {
+    const numeralMatch = match[1];
+    if (!numeralMatch) return match[0];
+    const normalizedNumeral = numeralMatch.replaceAll(new RegExp(`\\${currencySymbols.group}`, "g"), "").replace(new RegExp(`\\${currencySymbols.decimal}`), ".");
+    const number = parseFloat(normalizedNumeral);
+    return (0, import_to_words.toWords)(number, {
+      localeCode: `${maximizedLocale.language}-${maximizedLocale.region}`,
+      currency: true,
+      doNotAddOnly: true
+    });
+  }
+  const numberFormat = new Intl.NumberFormat(locale);
+  const numberParts = numberFormat.formatToParts(demoNumber);
+  const numberSymbols = numberParts.reduce(
+    (acc, part) => {
+      if (part.type === "group") {
+        return {
+          ...acc,
+          group: part.value
+        };
+      }
+      if (part.type === "decimal") {
+        return {
+          ...acc,
+          decimal: part.value
+        };
+      }
+      return acc;
+    },
+    { group: "", decimal: "" }
+  );
+  const numberRegex = new RegExp(
+    `(\\p{Number}[\\p{Number}${numberSymbols.group}]*(?:[${numberSymbols.decimal}]\\p{Number}*)?)`,
+    "gu"
+  );
+  function numberReplacer(match) {
+    const numeralMatch = match[1];
+    if (!numeralMatch) return match[0];
+    const normalizedNumeral = numeralMatch.replaceAll(new RegExp(`\\${numberSymbols.group}`, "g"), "").replace(new RegExp(`\\${numberSymbols.decimal}`), ".");
+    const number = parseFloat(normalizedNumeral);
+    return (0, import_to_words.toWords)(number, {
+      localeCode: `${maximizedLocale.language}-${maximizedLocale.region}`
+    });
+  }
+  return [
+    [currencyRegex, currencyReplacer],
+    [numberRegex, numberReplacer]
+  ];
+}
+async function slugify(text, locale) {
+  const replacers = replacerMap.get(locale) ?? createReplacers(locale);
+  replacerMap.set(locale, replacers);
+  const { result, mapping } = await (0, import_transliteration.slugify)(text, {
+    allowedChars: "a-zA-Z0-9",
+    replace: replacers
+  });
+  return { result, mapping };
+}
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  slugify
+});

package/dist/align/slugify.d.cts

@@ -0,0 +1,8 @@
+import * as _storyteller_platform_transliteration from '@storyteller-platform/transliteration';
+
+declare function slugify(text: string, locale: Intl.Locale): Promise<{
+    result: string;
+    mapping: _storyteller_platform_transliteration.Mapping;
+}>;
+
+export { slugify };