@storyteller-platform/align 0.1.8 → 0.1.10

package/dist/markup/segmentation.js

@@ -1,66 +1,19 @@
 import "../chunk-BIEQXUOY.js";
 import {
-  WordSequence,
   segmentText
 } from "@echogarden/text-segmentation";
-import { Epub } from "@storyteller-platform/epub";
-import { BLOCKS } from "./semantics.js";
+import { parseDom } from "./parseDom.js";
+import { liftText } from "./transform.js";
 async function getXhtmlSegmentation(xml, options) {
-  const result = {
-    words: new WordSequence(),
-    sentences: [],
-    segmentSentenceRanges: []
-  };
-  let stagedText = "";
-  for (const child of xml) {
-    if (Epub.isXmlTextNode(child)) {
-      stagedText += child["#text"];
-      continue;
-    }
-    const childName = Epub.getXmlElementName(child);
-    if (!BLOCKS.includes(childName)) {
-      stagedText += Epub.getXhtmlTextContent(Epub.getXmlChildren(child));
-      continue;
-    }
-    mergeSegmentations(
-      result,
-      await segmentText(collapseWhitespace(stagedText), {
-        ...options.primaryLocale && {
-          language: options.primaryLocale.language
-        },
-        enableEastAsianPostprocessing: true
-      })
-    );
-    stagedText = "";
-    mergeSegmentations(
-      result,
-      await getXhtmlSegmentation(Epub.getXmlChildren(child), options)
-    );
-  }
-  mergeSegmentations(
-    result,
-    await segmentText(collapseWhitespace(stagedText), {
-      ...options.primaryLocale && {
-        language: options.primaryLocale.language
-      },
-      enableEastAsianPostprocessing: true
-    })
-  );
-  return result;
-}
-function collapseWhitespace(text) {
-  return text.replace(/^\s*/, "").replace(/\s*$/, "").replaceAll(/\s+/g, " ");
-}
-function mergeSegmentations(first, second) {
-  for (const wordEntry of second.words.entries) {
-    first.words.addWord(
-      wordEntry.text,
-      wordEntry.startOffset,
-      wordEntry.isPunctuation
-    );
-  }
-  first.sentences.push(...second.sentences);
-  first.segmentSentenceRanges.push(...second.segmentSentenceRanges);
+  const root = parseDom(xml);
+  const { result: text, mapping } = liftText(root);
+  const result = await segmentText(text, {
+    ...options.primaryLocale && {
+      language: options.primaryLocale.language
+    },
+    enableEastAsianPostprocessing: true
+  });
+  return { result: result.sentences, mapping };
 }
 export {
   getXhtmlSegmentation

package/dist/markup/serializeDom.cjs

@@ -0,0 +1,87 @@
+"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 serializeDom_exports = {};
+__export(serializeDom_exports, {
+  serializeDom: () => serializeDom,
+  serializeDomNode: () => serializeDomNode
+});
+module.exports = __toCommonJS(serializeDom_exports);
+var import_epub = require("@storyteller-platform/epub");
+var import_model = require("./model.cjs");
+function serializeDom(doc) {
+  return doc.children.map((child) => serializeDomNode(child));
+}
+function serializeDomNode(node) {
+  if (node instanceof import_model.TextNode) {
+    return import_epub.Epub.createXmlTextNode(node.text);
+  }
+  return import_epub.Epub.createXmlElement(
+    node.tagName,
+    node.attrs,
+    serializeDomNodes(node.children)
+  );
+}
+function serializeDomNodes(nodes) {
+  const partitioned = nodes.reduce((acc, child) => {
+    const lastPartition = acc.at(-1);
+    if (!lastPartition) {
+      return [[child]];
+    }
+    const lastChild = lastPartition.at(-1);
+    if (!lastChild) {
+      return [...acc.slice(0, acc.length), [child]];
+    }
+    const childFirstMark = child.marks[0];
+    const lastChildFirstMark = lastChild.marks[0];
+    if (childFirstMark === lastChildFirstMark || (childFirstMark == null ? void 0 : childFirstMark.eq(lastChildFirstMark))) {
+      return [
+        ...acc.slice(0, acc.length - 1),
+        [...lastPartition.slice(0, lastPartition.length), child]
+      ];
+    }
+    return [...acc, [child]];
+  }, []);
+  const xmlChildren = [];
+  for (const partition of partitioned) {
+    xmlChildren.push(...serializePartition(partition));
+  }
+  return xmlChildren;
+}
+function serializePartition(nodes) {
+  const firstChild = nodes[0];
+  if (!firstChild) return [];
+  const firstMark = firstChild.marks[0];
+  if (!firstMark) {
+    return nodes.map((child) => serializeDomNode(child));
+  }
+  return [
+    import_epub.Epub.createXmlElement(
+      firstMark.tagName,
+      firstMark.attrs,
+      serializeDomNodes(
+        nodes.map((node) => node.copy({ marks: node.marks.slice(1) }))
+      )
+    )
+  ];
+}
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  serializeDom,
+  serializeDomNode
+});

package/dist/markup/serializeDom.d.cts

@@ -0,0 +1,7 @@
+import { ParsedXml, XmlNode } from '@storyteller-platform/epub';
+import { Root, Node, TextNode } from './model.cjs';
+
+declare function serializeDom(doc: Root): ParsedXml;
+declare function serializeDomNode(node: Node | TextNode): XmlNode;
+
+export { serializeDom, serializeDomNode };

package/dist/markup/serializeDom.d.ts

@@ -0,0 +1,7 @@
+import { ParsedXml, XmlNode } from '@storyteller-platform/epub';
+import { Root, Node, TextNode } from './model.js';
+
+declare function serializeDom(doc: Root): ParsedXml;
+declare function serializeDomNode(node: Node | TextNode): XmlNode;
+
+export { serializeDom, serializeDomNode };

package/dist/markup/serializeDom.js

@@ -0,0 +1,63 @@
+import "../chunk-BIEQXUOY.js";
+import { Epub } from "@storyteller-platform/epub";
+import { TextNode } from "./model.js";
+function serializeDom(doc) {
+  return doc.children.map((child) => serializeDomNode(child));
+}
+function serializeDomNode(node) {
+  if (node instanceof TextNode) {
+    return Epub.createXmlTextNode(node.text);
+  }
+  return Epub.createXmlElement(
+    node.tagName,
+    node.attrs,
+    serializeDomNodes(node.children)
+  );
+}
+function serializeDomNodes(nodes) {
+  const partitioned = nodes.reduce((acc, child) => {
+    const lastPartition = acc.at(-1);
+    if (!lastPartition) {
+      return [[child]];
+    }
+    const lastChild = lastPartition.at(-1);
+    if (!lastChild) {
+      return [...acc.slice(0, acc.length), [child]];
+    }
+    const childFirstMark = child.marks[0];
+    const lastChildFirstMark = lastChild.marks[0];
+    if (childFirstMark === lastChildFirstMark || (childFirstMark == null ? void 0 : childFirstMark.eq(lastChildFirstMark))) {
+      return [
+        ...acc.slice(0, acc.length - 1),
+        [...lastPartition.slice(0, lastPartition.length), child]
+      ];
+    }
+    return [...acc, [child]];
+  }, []);
+  const xmlChildren = [];
+  for (const partition of partitioned) {
+    xmlChildren.push(...serializePartition(partition));
+  }
+  return xmlChildren;
+}
+function serializePartition(nodes) {
+  const firstChild = nodes[0];
+  if (!firstChild) return [];
+  const firstMark = firstChild.marks[0];
+  if (!firstMark) {
+    return nodes.map((child) => serializeDomNode(child));
+  }
+  return [
+    Epub.createXmlElement(
+      firstMark.tagName,
+      firstMark.attrs,
+      serializeDomNodes(
+        nodes.map((node) => node.copy({ marks: node.marks.slice(1) }))
+      )
+    )
+  ];
+}
+export {
+  serializeDom,
+  serializeDomNode
+};

package/dist/markup/transform.cjs

@@ -0,0 +1,92 @@
+"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 transform_exports = {};
+__export(transform_exports, {
+  addMark: () => addMark,
+  liftText: () => liftText
+});
+module.exports = __toCommonJS(transform_exports);
+var import_map = require("./map.cjs");
+var import_model = require("./model.cjs");
+function addMark(root, from, to, mark) {
+  const result = root.split(from).split(to);
+  let pos = 0;
+  const children = [];
+  for (const child of result.children) {
+    children.push(addMarkToNode(child, pos, from, to, mark));
+    pos += child.nodeSize;
+  }
+  return result.copy({ children });
+}
+function addMarkToNode(node, pos, from, to, mark) {
+  if (from >= pos + node.nodeSize || to <= pos) {
+    return node;
+  }
+  if (node.isLeaf) {
+    return node.copy({ marks: [mark, ...node.marks] });
+  }
+  let childPos = node.border;
+  const children = [];
+  for (const child of node.children) {
+    children.push(addMarkToNode(child, pos + childPos, from, to, mark));
+    childPos += child.nodeSize;
+  }
+  return node.copy({ children });
+}
+function liftText(root) {
+  const mapping = new import_map.Mapping();
+  let text = "";
+  let textLength = 0;
+  let lastTextEnd = 0;
+  (0, import_model.descendants)(root, (node, pos, parent, index) => {
+    if (node.isBlock) {
+      return !!node.textContent.match(/\S/);
+    }
+    if (!(node instanceof import_model.TextNode)) return true;
+    if (mapping.map(pos) - mapping.map(lastTextEnd)) {
+      mapping.appendMap(
+        new import_map.StepMap([
+          mapping.map(lastTextEnd),
+          mapping.map(pos) - mapping.map(lastTextEnd),
+          0
+        ])
+      );
+    }
+    lastTextEnd = pos + node.nodeSize;
+    let result = node.text.replaceAll(/\n/g, " ");
+    const hasBlockSiblings = parent.children.some((child) => child.isBlock);
+    if (hasBlockSiblings && !result.match(/\S/)) {
+      mapping.appendMap(new import_map.StepMap([textLength, result.length, 0]));
+      result = "";
+    }
+    if (parent.isBlock && index === parent.children.length - 1 && !(text + result).endsWith("\n")) {
+      result += "\n";
+      textLength--;
+    }
+    text += result;
+    textLength += result.length;
+    return true;
+  });
+  return { result: text, mapping };
+}
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  addMark,
+  liftText
+});

package/dist/markup/transform.d.cts

@@ -0,0 +1,11 @@
+import { Mapping } from './map.cjs';
+import { Root, Mark } from './model.cjs';
+import '@storyteller-platform/epub';
+
+declare function addMark(root: Root, from: number, to: number, mark: Mark): Root;
+declare function liftText(root: Root): {
+    result: string;
+    mapping: Mapping;
+};
+
+export { addMark, liftText };

package/dist/markup/transform.d.ts

@@ -0,0 +1,11 @@
+import { Mapping } from './map.js';
+import { Root, Mark } from './model.js';
+import '@storyteller-platform/epub';
+
+declare function addMark(root: Root, from: number, to: number, mark: Mark): Root;
+declare function liftText(root: Root): {
+    result: string;
+    mapping: Mapping;
+};
+
+export { addMark, liftText };

package/dist/markup/transform.js

@@ -0,0 +1,71 @@
+import "../chunk-BIEQXUOY.js";
+import { Mapping, StepMap } from "./map.js";
+import {
+  TextNode,
+  descendants
+} from "./model.js";
+function addMark(root, from, to, mark) {
+  const result = root.split(from).split(to);
+  let pos = 0;
+  const children = [];
+  for (const child of result.children) {
+    children.push(addMarkToNode(child, pos, from, to, mark));
+    pos += child.nodeSize;
+  }
+  return result.copy({ children });
+}
+function addMarkToNode(node, pos, from, to, mark) {
+  if (from >= pos + node.nodeSize || to <= pos) {
+    return node;
+  }
+  if (node.isLeaf) {
+    return node.copy({ marks: [mark, ...node.marks] });
+  }
+  let childPos = node.border;
+  const children = [];
+  for (const child of node.children) {
+    children.push(addMarkToNode(child, pos + childPos, from, to, mark));
+    childPos += child.nodeSize;
+  }
+  return node.copy({ children });
+}
+function liftText(root) {
+  const mapping = new Mapping();
+  let text = "";
+  let textLength = 0;
+  let lastTextEnd = 0;
+  descendants(root, (node, pos, parent, index) => {
+    if (node.isBlock) {
+      return !!node.textContent.match(/\S/);
+    }
+    if (!(node instanceof TextNode)) return true;
+    if (mapping.map(pos) - mapping.map(lastTextEnd)) {
+      mapping.appendMap(
+        new StepMap([
+          mapping.map(lastTextEnd),
+          mapping.map(pos) - mapping.map(lastTextEnd),
+          0
+        ])
+      );
+    }
+    lastTextEnd = pos + node.nodeSize;
+    let result = node.text.replaceAll(/\n/g, " ");
+    const hasBlockSiblings = parent.children.some((child) => child.isBlock);
+    if (hasBlockSiblings && !result.match(/\S/)) {
+      mapping.appendMap(new StepMap([textLength, result.length, 0]));
+      result = "";
+    }
+    if (parent.isBlock && index === parent.children.length - 1 && !(text + result).endsWith("\n")) {
+      result += "\n";
+      textLength--;
+    }
+    text += result;
+    textLength += result.length;
+    return true;
+  });
+  return { result: text, mapping };
+}
+export {
+  addMark,
+  liftText
+};

package/dist/types/node-gyp-build.d.cjs

@@ -0,0 +1 @@
+"use strict";

package/dist/types/node-gyp-build.d.d.cts

@@ -0,0 +1,3 @@
+declare module "node-gyp-build" {
+  export default function requireBinding(dirname: string): unknown
+}

package/dist/types/node-gyp-build.d.d.ts

@@ -0,0 +1,3 @@
+declare module "node-gyp-build" {
+  export default function requireBinding(dirname: string): unknown
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@storyteller-platform/align",
-  "version": "0.1.8",
+  "version": "0.1.10",
   "description": "A library and CLI for automatically aligning audiobooks and EPUBs to produce Media Overlays",
   "author": "Shane Friedman",
   "license": "MIT",
@@ -46,6 +46,7 @@
   "scripts": {
     "compile": "NODE_OPTIONS=--experimental-import-meta-resolve ./scripts/bundle.js && node --build-sea sea-config.json",
     "build": "tsup",
+    "install": "node-gyp-build",
     "prepack": "yarn build",
     "test": "yarn tsx --test"
   },
@@ -56,17 +57,20 @@
     "@esfx/async-semaphore": "^1.0.0",
     "@optique/core": "^0.10.7",
     "@optique/run": "^0.10.7",
-    "@storyteller-platform/audiobook": "^0.3.8",
-    "@storyteller-platform/epub": "^0.4.7",
-    "@storyteller-platform/ghost-story": "^0.1.4",
+    "@storyteller-platform/audiobook": "^0.3.9",
+    "@storyteller-platform/epub": "^0.4.8",
+    "@storyteller-platform/ghost-story": "^0.1.5",
     "@storyteller-platform/transliteration": "^3.1.0",
     "chalk": "^5.4.1",
     "cli-progress": "^3.12.0",
     "esbuild": "^0.27.3",
+    "itertools": "^2.6.0",
     "locale-currency": "^1.0.0",
     "memoize": "^10.2.0",
+    "node-gyp-build": "^4.8.4",
     "pino": "^10.3.1",
     "pino-pretty": "^13.1.3",
+    "runes2": "^1.1.4",
     "to-words": "^5.3.0",
     "zod": "^3.24.0"
   },
@@ -77,6 +81,9 @@
     "@types/cli-progress": "^3",
     "@types/node": "^24.0.0",
     "eslint": "^8.0.0",
+    "node-addon-api": "^8.3.1",
+    "node-gyp": "^11.2.0",
+    "prebuildify": "^6.0.1",
     "tsup": "^8.5.0",
     "tsx": "^4.19.2",
     "typescript": "~5.8.3"