@helloao/tools 0.0.5 → 0.0.6-alpha

package/dist/cjs/generation/dataset.cjs

@@ -0,0 +1,158 @@
+"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 dataset_exports = {};
+__export(dataset_exports, {
+  generateDataset: () => generateDataset
+});
+module.exports = __toCommonJS(dataset_exports);
+var import_usfm_parser = require("../parser/usfm-parser.js");
+var import_usx_parser = require("../parser/usx-parser.js");
+var import_book_order = require("./book-order.js");
+var import_lodash = require("lodash");
+var import_audio = require("./audio.js");
+var import_codex_parser = require("../parser/codex-parser.js");
+function generateDataset(files, parser = new globalThis.DOMParser()) {
+  let output = {
+    translations: []
+  };
+  let usfmParser = new import_usfm_parser.UsfmParser();
+  let usxParser = new import_usx_parser.USXParser(parser);
+  let codexParser = new import_codex_parser.CodexParser();
+  let parsedTranslations = /* @__PURE__ */ new Map();
+  const unknownLanguages = /* @__PURE__ */ new Set();
+  for (let file of files) {
+    try {
+      let parser2;
+      if (file.fileType === "usfm") {
+        parser2 = usfmParser;
+      } else if (file.fileType === "usx") {
+        parser2 = usxParser;
+      } else if (file.fileType === "json") {
+        parser2 = codexParser;
+      } else {
+        console.warn(
+          "[generate] File does not have a valid type!",
+          file.name
+        );
+        continue;
+      }
+      const parsed = parser2.parse(file.content);
+      const id = parsed.id;
+      if (!id) {
+        console.warn(
+          "[generate] File does not have a valid book ID!",
+          file.name,
+          id
+        );
+        continue;
+      }
+      const order = import_book_order.bookOrderMap.get(id);
+      if (typeof order !== "number") {
+        console.warn("[generate] Book does not have an order!", id);
+        continue;
+      }
+      const bookMap = import_book_order.bookIdMap.get(file.metadata.translation.language);
+      if (!bookMap) {
+        if (!unknownLanguages.has(file.metadata.translation.language)) {
+          console.warn(
+            "[generate] File does not have a known language!",
+            file.name,
+            file.metadata.translation.language
+          );
+          unknownLanguages.add(file.metadata.translation.language);
+        }
+      }
+      const bookName = bookMap?.get(id);
+      if (!!bookMap && !bookName) {
+        console.warn(
+          "[generate] Book name not found for ID!",
+          file.name,
+          id
+        );
+      }
+      let translation = parsedTranslations.get(
+        file.metadata.translation.id
+      );
+      if (!translation) {
+        translation = {
+          ...(0, import_lodash.omit)(file.metadata.translation, "direction"),
+          textDirection: file.metadata.translation.direction,
+          books: []
+        };
+        output.translations.push(translation);
+        parsedTranslations.set(
+          file.metadata.translation.id,
+          translation
+        );
+      }
+      const name = parsed.header ?? bookName?.commonName ?? parsed.title;
+      if (!name) {
+        console.warn(
+          "[generate] Book does not have a name!",
+          file.name,
+          id
+        );
+        continue;
+      }
+      const commonName = bookName?.commonName ?? parsed.header ?? parsed.title ?? id;
+      const book = {
+        id,
+        name,
+        commonName,
+        title: parsed.title ?? null,
+        order,
+        chapters: []
+      };
+      for (let content of parsed.content) {
+        if (content.type === "chapter") {
+          book.chapters.push({
+            chapter: {
+              number: content.number,
+              content: content.content,
+              footnotes: content.footnotes
+            },
+            thisChapterAudioLinks: (0, import_audio.getAudioUrlsForChapter)(
+              translation.id,
+              id,
+              content.number
+            )
+          });
+        }
+      }
+      book.chapters = (0, import_lodash.sortBy)(book.chapters, (c) => c.chapter.number);
+      const index = (0, import_lodash.sortedIndexBy)(
+        translation.books,
+        book,
+        (b) => b.order
+      );
+      translation.books.splice(index, 0, book);
+    } catch (err) {
+      console.error(
+        `[generate] Error occurred while parsing ${file.name}`,
+        err
+      );
+    }
+  }
+  return output;
+}
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  generateDataset
+});
+//# sourceMappingURL=dataset.cjs.map

package/dist/cjs/generation/dataset.cjs.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../../../generation/dataset.ts"],
+  "sourcesContent": ["import { UsfmParser } from '../parser/usfm-parser.js';\nimport { USXParser } from '../parser/usx-parser.js';\nimport {\n    InputFile,\n    Translation,\n    TranslationBook,\n    TranslationBookChapter,\n} from './common-types.js';\nimport { bookIdMap, bookOrderMap } from './book-order.js';\nimport { omit, sortBy, sortedIndexBy } from 'lodash';\nimport { getAudioUrlsForChapter } from './audio.js';\nimport { CodexParser } from '../parser/codex-parser.js';\n\n/**\n * Defines an interface that contains generated dataset info.\n */\nexport interface DatasetOutput {\n    /**\n     * The list of translations that are available in the dataset.\n     */\n    translations: DatasetTranslation[];\n}\n\n/**\n * Defines a translation that is used in the dataset.\n */\nexport interface DatasetTranslation extends Translation {\n    /**\n     * The list of books that are available for the translation.\n     */\n    books: DatasetTranslationBook[];\n}\n\n/**\n * Defines a translation book that is used in the dataset.\n */\nexport interface DatasetTranslationBook extends TranslationBook {\n    /**\n     * The list of chapters that are available for the book.\n     */\n    chapters: TranslationBookChapter[];\n}\n\n/**\n * Generates a list of output files from the given list of input files.\n * @param file The list of files.\n */\nexport function generateDataset(\n    files: InputFile[],\n    parser: DOMParser = new globalThis.DOMParser()\n): DatasetOutput {\n    let output: DatasetOutput = {\n        translations: [],\n    };\n\n    let usfmParser = new UsfmParser();\n    let usxParser = new USXParser(parser);\n    let codexParser = new CodexParser();\n\n    let parsedTranslations = new Map<string, DatasetTranslation>();\n\n    const unknownLanguages = new Set<string>();\n    for (let file of files) {\n        try {\n            let parser: UsfmParser | USXParser | CodexParser;\n            if (file.fileType === 'usfm') {\n                parser = usfmParser;\n            } else if (file.fileType === 'usx') {\n                parser = usxParser;\n            } else if (file.fileType === 'json') {\n                parser = codexParser;\n            } else {\n                console.warn(\n                    '[generate] File does not have a valid type!',\n                    file.name\n                );\n                continue;\n            }\n\n            const parsed = parser.parse(file.content);\n            const id = parsed.id;\n\n            if (!id) {\n                console.warn(\n                    '[generate] File does not have a valid book ID!',\n                    file.name,\n                    id\n                );\n                continue;\n            }\n\n            const order = bookOrderMap.get(id);\n\n            if (typeof order !== 'number') {\n                console.warn('[generate] Book does not have an order!', id);\n                continue;\n            }\n\n            const bookMap = bookIdMap.get(file.metadata.translation.language);\n\n            if (!bookMap) {\n                if (!unknownLanguages.has(file.metadata.translation.language)) {\n                    console.warn(\n                        '[generate] File does not have a known language!',\n                        file.name,\n                        file.metadata.translation.language\n                    );\n                    unknownLanguages.add(file.metadata.translation.language);\n                }\n            }\n\n            const bookName = bookMap?.get(id);\n\n            if (!!bookMap && !bookName) {\n                console.warn(\n                    '[generate] Book name not found for ID!',\n                    file.name,\n                    id\n                );\n            }\n\n            let translation = parsedTranslations.get(\n                file.metadata.translation.id\n            );\n\n            if (!translation) {\n                translation = {\n                    ...omit(file.metadata.translation, 'direction'),\n                    textDirection: file.metadata.translation.direction,\n                    books: [],\n                };\n                output.translations.push(translation);\n                parsedTranslations.set(\n                    file.metadata.translation.id,\n                    translation\n                );\n            }\n\n            const name = parsed.header ?? bookName?.commonName ?? parsed.title;\n\n            if (!name) {\n                console.warn(\n                    '[generate] Book does not have a name!',\n                    file.name,\n                    id\n                );\n                continue;\n            }\n\n            const commonName =\n                bookName?.commonName ?? parsed.header ?? parsed.title ?? id;\n\n            const book: DatasetTranslationBook = {\n                id: id,\n                name,\n                commonName,\n                title: parsed.title ?? null,\n                order,\n                chapters: [],\n            };\n\n            for (let content of parsed.content) {\n                if (content.type === 'chapter') {\n                    book.chapters.push({\n                        chapter: {\n                            number: content.number,\n                            content: content.content,\n                            footnotes: content.footnotes,\n                        },\n                        thisChapterAudioLinks: getAudioUrlsForChapter(\n                            translation.id,\n                            id,\n                            content.number\n                        ),\n                    });\n                }\n            }\n\n            book.chapters = sortBy(book.chapters, (c) => c.chapter.number);\n\n            const index = sortedIndexBy(\n                translation.books,\n                book,\n                (b) => b.order\n            );\n            translation.books.splice(index, 0, book);\n        } catch (err) {\n            console.error(\n                `[generate] Error occurred while parsing ${file.name}`,\n                err\n            );\n        }\n    }\n\n    return output;\n}\n"],
+  "mappings": ";;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,yBAA2B;AAC3B,wBAA0B;AAO1B,wBAAwC;AACxC,oBAA4C;AAC5C,mBAAuC;AACvC,0BAA4B;AAoCrB,SAAS,gBACZ,OACA,SAAoB,IAAI,WAAW,UAAU,GAChC;AACb,MAAI,SAAwB;AAAA,IACxB,cAAc,CAAC;AAAA,EACnB;AAEA,MAAI,aAAa,IAAI,8BAAW;AAChC,MAAI,YAAY,IAAI,4BAAU,MAAM;AACpC,MAAI,cAAc,IAAI,gCAAY;AAElC,MAAI,qBAAqB,oBAAI,IAAgC;AAE7D,QAAM,mBAAmB,oBAAI,IAAY;AACzC,WAAS,QAAQ,OAAO;AACpB,QAAI;AACA,UAAIA;AACJ,UAAI,KAAK,aAAa,QAAQ;AAC1B,QAAAA,UAAS;AAAA,MACb,WAAW,KAAK,aAAa,OAAO;AAChC,QAAAA,UAAS;AAAA,MACb,WAAW,KAAK,aAAa,QAAQ;AACjC,QAAAA,UAAS;AAAA,MACb,OAAO;AACH,gBAAQ;AAAA,UACJ;AAAA,UACA,KAAK;AAAA,QACT;AACA;AAAA,MACJ;AAEA,YAAM,SAASA,QAAO,MAAM,KAAK,OAAO;AACxC,YAAM,KAAK,OAAO;AAElB,UAAI,CAAC,IAAI;AACL,gBAAQ;AAAA,UACJ;AAAA,UACA,KAAK;AAAA,UACL;AAAA,QACJ;AACA;AAAA,MACJ;AAEA,YAAM,QAAQ,+BAAa,IAAI,EAAE;AAEjC,UAAI,OAAO,UAAU,UAAU;AAC3B,gBAAQ,KAAK,2CAA2C,EAAE;AAC1D;AAAA,MACJ;AAEA,YAAM,UAAU,4BAAU,IAAI,KAAK,SAAS,YAAY,QAAQ;AAEhE,UAAI,CAAC,SAAS;AACV,YAAI,CAAC,iBAAiB,IAAI,KAAK,SAAS,YAAY,QAAQ,GAAG;AAC3D,kBAAQ;AAAA,YACJ;AAAA,YACA,KAAK;AAAA,YACL,KAAK,SAAS,YAAY;AAAA,UAC9B;AACA,2BAAiB,IAAI,KAAK,SAAS,YAAY,QAAQ;AAAA,QAC3D;AAAA,MACJ;AAEA,YAAM,WAAW,SAAS,IAAI,EAAE;AAEhC,UAAI,CAAC,CAAC,WAAW,CAAC,UAAU;AACxB,gBAAQ;AAAA,UACJ;AAAA,UACA,KAAK;AAAA,UACL;AAAA,QACJ;AAAA,MACJ;AAEA,UAAI,cAAc,mBAAmB;AAAA,QACjC,KAAK,SAAS,YAAY;AAAA,MAC9B;AAEA,UAAI,CAAC,aAAa;AACd,sBAAc;AAAA,UACV,OAAG,oBAAK,KAAK,SAAS,aAAa,WAAW;AAAA,UAC9C,eAAe,KAAK,SAAS,YAAY;AAAA,UACzC,OAAO,CAAC;AAAA,QACZ;AACA,eAAO,aAAa,KAAK,WAAW;AACpC,2BAAmB;AAAA,UACf,KAAK,SAAS,YAAY;AAAA,UAC1B;AAAA,QACJ;AAAA,MACJ;AAEA,YAAM,OAAO,OAAO,UAAU,UAAU,cAAc,OAAO;AAE7D,UAAI,CAAC,MAAM;AACP,gBAAQ;AAAA,UACJ;AAAA,UACA,KAAK;AAAA,UACL;AAAA,QACJ;AACA;AAAA,MACJ;AAEA,YAAM,aACF,UAAU,cAAc,OAAO,UAAU,OAAO,SAAS;AAE7D,YAAM,OAA+B;AAAA,QACjC;AAAA,QACA;AAAA,QACA;AAAA,QACA,OAAO,OAAO,SAAS;AAAA,QACvB;AAAA,QACA,UAAU,CAAC;AAAA,MACf;AAEA,eAAS,WAAW,OAAO,SAAS;AAChC,YAAI,QAAQ,SAAS,WAAW;AAC5B,eAAK,SAAS,KAAK;AAAA,YACf,SAAS;AAAA,cACL,QAAQ,QAAQ;AAAA,cAChB,SAAS,QAAQ;AAAA,cACjB,WAAW,QAAQ;AAAA,YACvB;AAAA,YACA,2BAAuB;AAAA,cACnB,YAAY;AAAA,cACZ;AAAA,cACA,QAAQ;AAAA,YACZ;AAAA,UACJ,CAAC;AAAA,QACL;AAAA,MACJ;AAEA,WAAK,eAAW,sBAAO,KAAK,UAAU,CAAC,MAAM,EAAE,QAAQ,MAAM;AAE7D,YAAM,YAAQ;AAAA,QACV,YAAY;AAAA,QACZ;AAAA,QACA,CAAC,MAAM,EAAE;AAAA,MACb;AACA,kBAAY,MAAM,OAAO,OAAO,GAAG,IAAI;AAAA,IAC3C,SAAS,KAAK;AACV,cAAQ;AAAA,QACJ,2CAA2C,KAAK,IAAI;AAAA,QACpD;AAAA,MACJ;AAAA,IACJ;AAAA,EACJ;AAEA,SAAO;AACX;",
+  "names": ["parser"]
+}

package/dist/cjs/generation/index.cjs

@@ -0,0 +1,51 @@
+"use strict";
+var __create = Object.create;
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __getProtoOf = Object.getPrototypeOf;
+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 __reExport = (target, mod, secondTarget) => (__copyProps(target, mod, "default"), secondTarget && __copyProps(secondTarget, mod, "default"));
+var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
+  // If the importer is in node compatibility mode or this is not an ESM
+  // file that has been converted to a CommonJS file using a Babel-
+  // compatible transform (i.e. "__esModule" has not been set), then set
+  // "default" to the CommonJS "module.exports" for node compatibility.
+  isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
+  mod
+));
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var generation_exports = {};
+__export(generation_exports, {
+  api: () => api,
+  audio: () => audio,
+  bookOrder: () => bookOrder,
+  dataset: () => dataset
+});
+module.exports = __toCommonJS(generation_exports);
+var api = __toESM(require("./api.js"));
+var audio = __toESM(require("./audio.js"));
+var dataset = __toESM(require("./dataset.js"));
+var bookOrder = __toESM(require("./book-order.js"));
+__reExport(generation_exports, require("./common-types.js"), module.exports);
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  api,
+  audio,
+  bookOrder,
+  dataset,
+  ...require("./common-types.js")
+});
+//# sourceMappingURL=index.cjs.map

package/dist/cjs/generation/index.cjs.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../../../generation/index.ts"],
+  "sourcesContent": ["import * as api from './api.js';\nimport * as audio from './audio.js';\nimport * as dataset from './dataset.js';\nimport * as bookOrder from './book-order.js';\nexport * from './common-types.js';\n\nexport { api, audio, dataset, bookOrder };\n"],
+  "mappings": ";;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,UAAqB;AACrB,YAAuB;AACvB,cAAyB;AACzB,gBAA2B;AAC3B,+BAAc,8BAJd;",
+  "names": []
+}

package/dist/cjs/index.cjs

@@ -0,0 +1,45 @@
+"use strict";
+var __create = Object.create;
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __getProtoOf = Object.getPrototypeOf;
+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 __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
+  // If the importer is in node compatibility mode or this is not an ESM
+  // file that has been converted to a CommonJS file using a Babel-
+  // compatible transform (i.e. "__esModule" has not been set), then set
+  // "default" to the CommonJS "module.exports" for node compatibility.
+  isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
+  mod
+));
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var helloao_tools_exports = {};
+__export(helloao_tools_exports, {
+  generation: () => generation,
+  parser: () => parser,
+  utils: () => utils
+});
+module.exports = __toCommonJS(helloao_tools_exports);
+var utils = __toESM(require("./utils.js"));
+var generation = __toESM(require("./generation/index.js"));
+var parser = __toESM(require("./parser/index.js"));
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  generation,
+  parser,
+  utils
+});
+//# sourceMappingURL=index.cjs.map

package/dist/cjs/index.cjs.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../../index.ts"],
+  "sourcesContent": ["import * as utils from './utils.js';\nimport * as generation from './generation/index.js';\nimport * as parser from './parser/index.js';\n\nexport { utils, generation, parser };\n"],
+  "mappings": ";;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,YAAuB;AACvB,iBAA4B;AAC5B,aAAwB;",
+  "names": []
+}

package/dist/cjs/parser/codex-parser.cjs

@@ -0,0 +1,176 @@
+"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 codex_parser_exports = {};
+__export(codex_parser_exports, {
+  CodexParser: () => CodexParser,
+  chapterHeadingSchema: () => chapterHeadingSchema,
+  codexSchema: () => codexSchema,
+  metadataSchema: () => metadataSchema
+});
+module.exports = __toCommonJS(codex_parser_exports);
+var import_zod = require("zod");
+var import_utils = require("../utils.js");
+const chapterHeadingSchema = import_zod.z.object({
+  type: import_zod.z.literal("chapter-heading"),
+  data: import_zod.z.object({
+    chapter: import_zod.z.coerce.number().int()
+  })
+});
+const metadataSchema = import_zod.z.discriminatedUnion("type", [
+  chapterHeadingSchema
+]);
+const codexSchema = import_zod.z.object({
+  cells: import_zod.z.array(
+    import_zod.z.object({
+      // kind: z.number().int(),
+      language: import_zod.z.string(),
+      value: import_zod.z.string(),
+      metadata: import_zod.z.object({
+        type: import_zod.z.string()
+      }).passthrough().nullable().optional()
+    })
+  )
+});
+class CodexParser {
+  // TODO:
+  // /**
+  //  * Whether to preserve markdown in the parsed content.
+  //  */
+  // preserveMarkdown: boolean = true;
+  _noteCounter = 0;
+  /**
+   * Parses the specified USX content.
+   *
+   * @param usx The USX content to parse.
+   * @returns The parse tree that was generated.
+   */
+  parse(codex) {
+    const json = JSON.parse(codex);
+    const data = codexSchema.parse(json);
+    let root = {
+      type: "root",
+      content: []
+    };
+    let chapters = /* @__PURE__ */ new Map();
+    let lastChapter = null;
+    function addChapterContent(chapterNumber, content) {
+      let chapter = chapters.get(chapterNumber);
+      if (!chapter) {
+        chapter = {
+          type: "chapter",
+          number: chapterNumber,
+          content: [],
+          footnotes: []
+        };
+        lastChapter = chapter;
+        chapters.set(chapterNumber, chapter);
+      }
+      chapter.content.push(...content);
+    }
+    function addReference(ref, content) {
+      addChapterContent(ref.chapter, [
+        {
+          type: "verse",
+          number: ref.verse,
+          content
+        }
+      ]);
+    }
+    for (let cell of data.cells) {
+      if (cell.language === "scripture") {
+        const lines = cell.value.split("\n");
+        const references = lines.map(
+          (l) => (0, import_utils.parseVerseReference)(l) ?? l
+        );
+        let currentRef = null;
+        let content = [];
+        for (let i = 0; i < references.length; i++) {
+          const ref = references[i];
+          const nextRef = references[i + 1];
+          if (typeof ref === "string") {
+            if (ref === "") {
+              if (typeof nextRef === "object") {
+                if (currentRef && currentRef.chapter === nextRef.chapter) {
+                  addChapterContent(currentRef.chapter, [
+                    {
+                      type: "line_break"
+                    }
+                  ]);
+                } else {
+                }
+              } else {
+                content.push({
+                  lineBreak: true
+                });
+              }
+            } else if (ref) {
+              content.push(ref);
+            }
+          } else {
+            if (!root.id) {
+              root.id = ref.book;
+            }
+            if (currentRef) {
+              currentRef = null;
+              content = [];
+            }
+            currentRef = ref;
+            if (ref.content) {
+              content.push(ref.content);
+            }
+            addReference(ref, content);
+          }
+        }
+      } else if (cell.language === "markdown") {
+        if (cell.metadata) {
+          if (cell.metadata.type === "chapter-heading") {
+            const metadata = metadataSchema.parse(cell.metadata);
+            const chapter = metadata.data.chapter;
+            addChapterContent(chapter, [
+              {
+                type: "heading",
+                content: [cell.value]
+              }
+            ]);
+          }
+        } else {
+          if (lastChapter) {
+            lastChapter.footnotes.push({
+              noteId: this._noteCounter++,
+              text: cell.value,
+              caller: null
+            });
+          }
+        }
+      }
+    }
+    for (let chapter of chapters.values()) {
+      root.content.push(chapter);
+    }
+    return root;
+  }
+}
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  CodexParser,
+  chapterHeadingSchema,
+  codexSchema,
+  metadataSchema
+});
+//# sourceMappingURL=codex-parser.cjs.map

package/dist/cjs/parser/codex-parser.cjs.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../../../parser/codex-parser.ts"],
+  "sourcesContent": ["import { z } from 'zod';\nimport {\n    Chapter,\n    ChapterContent,\n    ParseTree,\n    Verse,\n    VerseContent,\n} from './types.js';\nimport { parseVerseReference, VerseRef } from '../utils.js';\n\nexport const chapterHeadingSchema = z.object({\n    type: z.literal('chapter-heading'),\n    data: z.object({\n        chapter: z.coerce.number().int(),\n    }),\n});\n\nexport const metadataSchema = z.discriminatedUnion('type', [\n    chapterHeadingSchema,\n]);\n\nexport const codexSchema = z.object({\n    cells: z.array(\n        z.object({\n            // kind: z.number().int(),\n            language: z.string(),\n            value: z.string(),\n            metadata: z\n                .object({\n                    type: z.string(),\n                })\n                .passthrough()\n                .nullable()\n                .optional(),\n        })\n    ),\n});\n\n/**\n * Defines a parser for codex files.\n */\nexport class CodexParser {\n    // TODO:\n    // /**\n    //  * Whether to preserve markdown in the parsed content.\n    //  */\n    // preserveMarkdown: boolean = true;\n\n    private _noteCounter: number = 0;\n\n    /**\n     * Parses the specified USX content.\n     *\n     * @param usx The USX content to parse.\n     * @returns The parse tree that was generated.\n     */\n    public parse(codex: string): ParseTree {\n        const json = JSON.parse(codex);\n        const data = codexSchema.parse(json);\n\n        let root: ParseTree = {\n            type: 'root',\n            content: [],\n        };\n\n        let chapters: Map<number, Chapter> = new Map();\n        let lastChapter: Chapter | null = null;\n\n        function addChapterContent(\n            chapterNumber: number,\n            content: ChapterContent[]\n        ) {\n            let chapter = chapters.get(chapterNumber);\n            if (!chapter) {\n                chapter = {\n                    type: 'chapter',\n                    number: chapterNumber,\n                    content: [],\n                    footnotes: [],\n                };\n                lastChapter = chapter;\n                chapters.set(chapterNumber, chapter);\n            }\n\n            chapter.content.push(...content);\n        }\n\n        function addReference(ref: VerseRef, content: Verse['content']) {\n            addChapterContent(ref.chapter, [\n                {\n                    type: 'verse',\n                    number: ref.verse,\n                    content: content,\n                },\n            ]);\n        }\n\n        for (let cell of data.cells) {\n            if (cell.language === 'scripture') {\n                const lines = cell.value.split('\\n');\n                const references = lines.map(\n                    (l) => parseVerseReference(l) ?? l\n                );\n                let currentRef: VerseRef | null = null;\n                let content: Verse['content'] = [];\n                for (let i = 0; i < references.length; i++) {\n                    const ref = references[i];\n                    const nextRef = references[i + 1];\n\n                    if (typeof ref === 'string') {\n                        // continuation on existing verse\n                        // or line break between verses\n\n                        if (ref === '') {\n                            // line break\n\n                            if (typeof nextRef === 'object') {\n                                // between verses\n\n                                if (\n                                    currentRef &&\n                                    currentRef.chapter === nextRef.chapter\n                                ) {\n                                    // in same chapter\n                                    addChapterContent(currentRef.chapter, [\n                                        {\n                                            type: 'line_break',\n                                        },\n                                    ]);\n                                } else {\n                                    // do nothing since line breaks between chapters are ignored\n                                }\n                            } else {\n                                // inside verse\n                                content.push({\n                                    lineBreak: true,\n                                });\n                            }\n                        } else if (ref) {\n                            content.push(ref);\n                        }\n                    } else {\n                        if (!root.id) {\n                            root.id = ref.book;\n                        }\n\n                        // new verse\n                        if (currentRef) {\n                            // finish previous verse\n                            currentRef = null;\n                            content = [];\n                        }\n\n                        currentRef = ref;\n\n                        if (ref.content) {\n                            content.push(ref.content);\n                        }\n\n                        addReference(ref, content);\n                    }\n                }\n            } else if (cell.language === 'markdown') {\n                if (cell.metadata) {\n                    if (cell.metadata.type === 'chapter-heading') {\n                        const metadata = metadataSchema.parse(cell.metadata);\n\n                        const chapter = metadata.data.chapter;\n                        addChapterContent(chapter, [\n                            {\n                                type: 'heading',\n                                content: [cell.value],\n                            },\n                        ]);\n                    }\n                } else {\n                    if (lastChapter) {\n                        (lastChapter as Chapter).footnotes.push({\n                            noteId: this._noteCounter++,\n                            text: cell.value,\n                            caller: null,\n                        });\n                    }\n                }\n            }\n        }\n\n        for (let chapter of chapters.values()) {\n            root.content.push(chapter);\n        }\n\n        return root;\n    }\n}\n"],
+  "mappings": ";;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,iBAAkB;AAQlB,mBAA8C;AAEvC,MAAM,uBAAuB,aAAE,OAAO;AAAA,EACzC,MAAM,aAAE,QAAQ,iBAAiB;AAAA,EACjC,MAAM,aAAE,OAAO;AAAA,IACX,SAAS,aAAE,OAAO,OAAO,EAAE,IAAI;AAAA,EACnC,CAAC;AACL,CAAC;AAEM,MAAM,iBAAiB,aAAE,mBAAmB,QAAQ;AAAA,EACvD;AACJ,CAAC;AAEM,MAAM,cAAc,aAAE,OAAO;AAAA,EAChC,OAAO,aAAE;AAAA,IACL,aAAE,OAAO;AAAA;AAAA,MAEL,UAAU,aAAE,OAAO;AAAA,MACnB,OAAO,aAAE,OAAO;AAAA,MAChB,UAAU,aACL,OAAO;AAAA,QACJ,MAAM,aAAE,OAAO;AAAA,MACnB,CAAC,EACA,YAAY,EACZ,SAAS,EACT,SAAS;AAAA,IAClB,CAAC;AAAA,EACL;AACJ,CAAC;AAKM,MAAM,YAAY;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOb,eAAuB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQxB,MAAM,OAA0B;AACnC,UAAM,OAAO,KAAK,MAAM,KAAK;AAC7B,UAAM,OAAO,YAAY,MAAM,IAAI;AAEnC,QAAI,OAAkB;AAAA,MAClB,MAAM;AAAA,MACN,SAAS,CAAC;AAAA,IACd;AAEA,QAAI,WAAiC,oBAAI,IAAI;AAC7C,QAAI,cAA8B;AAElC,aAAS,kBACL,eACA,SACF;AACE,UAAI,UAAU,SAAS,IAAI,aAAa;AACxC,UAAI,CAAC,SAAS;AACV,kBAAU;AAAA,UACN,MAAM;AAAA,UACN,QAAQ;AAAA,UACR,SAAS,CAAC;AAAA,UACV,WAAW,CAAC;AAAA,QAChB;AACA,sBAAc;AACd,iBAAS,IAAI,eAAe,OAAO;AAAA,MACvC;AAEA,cAAQ,QAAQ,KAAK,GAAG,OAAO;AAAA,IACnC;AAEA,aAAS,aAAa,KAAe,SAA2B;AAC5D,wBAAkB,IAAI,SAAS;AAAA,QAC3B;AAAA,UACI,MAAM;AAAA,UACN,QAAQ,IAAI;AAAA,UACZ;AAAA,QACJ;AAAA,MACJ,CAAC;AAAA,IACL;AAEA,aAAS,QAAQ,KAAK,OAAO;AACzB,UAAI,KAAK,aAAa,aAAa;AAC/B,cAAM,QAAQ,KAAK,MAAM,MAAM,IAAI;AACnC,cAAM,aAAa,MAAM;AAAA,UACrB,CAAC,UAAM,kCAAoB,CAAC,KAAK;AAAA,QACrC;AACA,YAAI,aAA8B;AAClC,YAAI,UAA4B,CAAC;AACjC,iBAAS,IAAI,GAAG,IAAI,WAAW,QAAQ,KAAK;AACxC,gBAAM,MAAM,WAAW,CAAC;AACxB,gBAAM,UAAU,WAAW,IAAI,CAAC;AAEhC,cAAI,OAAO,QAAQ,UAAU;AAIzB,gBAAI,QAAQ,IAAI;AAGZ,kBAAI,OAAO,YAAY,UAAU;AAG7B,oBACI,cACA,WAAW,YAAY,QAAQ,SACjC;AAEE,oCAAkB,WAAW,SAAS;AAAA,oBAClC;AAAA,sBACI,MAAM;AAAA,oBACV;AAAA,kBACJ,CAAC;AAAA,gBACL,OAAO;AAAA,gBAEP;AAAA,cACJ,OAAO;AAEH,wBAAQ,KAAK;AAAA,kBACT,WAAW;AAAA,gBACf,CAAC;AAAA,cACL;AAAA,YACJ,WAAW,KAAK;AACZ,sBAAQ,KAAK,GAAG;AAAA,YACpB;AAAA,UACJ,OAAO;AACH,gBAAI,CAAC,KAAK,IAAI;AACV,mBAAK,KAAK,IAAI;AAAA,YAClB;AAGA,gBAAI,YAAY;AAEZ,2BAAa;AACb,wBAAU,CAAC;AAAA,YACf;AAEA,yBAAa;AAEb,gBAAI,IAAI,SAAS;AACb,sBAAQ,KAAK,IAAI,OAAO;AAAA,YAC5B;AAEA,yBAAa,KAAK,OAAO;AAAA,UAC7B;AAAA,QACJ;AAAA,MACJ,WAAW,KAAK,aAAa,YAAY;AACrC,YAAI,KAAK,UAAU;AACf,cAAI,KAAK,SAAS,SAAS,mBAAmB;AAC1C,kBAAM,WAAW,eAAe,MAAM,KAAK,QAAQ;AAEnD,kBAAM,UAAU,SAAS,KAAK;AAC9B,8BAAkB,SAAS;AAAA,cACvB;AAAA,gBACI,MAAM;AAAA,gBACN,SAAS,CAAC,KAAK,KAAK;AAAA,cACxB;AAAA,YACJ,CAAC;AAAA,UACL;AAAA,QACJ,OAAO;AACH,cAAI,aAAa;AACb,YAAC,YAAwB,UAAU,KAAK;AAAA,cACpC,QAAQ,KAAK;AAAA,cACb,MAAM,KAAK;AAAA,cACX,QAAQ;AAAA,YACZ,CAAC;AAAA,UACL;AAAA,QACJ;AAAA,MACJ;AAAA,IACJ;AAEA,aAAS,WAAW,SAAS,OAAO,GAAG;AACnC,WAAK,QAAQ,KAAK,OAAO;AAAA,IAC7B;AAEA,WAAO;AAAA,EACX;AACJ;",
+  "names": []
+}

package/dist/cjs/parser/index.cjs

@@ -0,0 +1,48 @@
+"use strict";
+var __create = Object.create;
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __getProtoOf = Object.getPrototypeOf;
+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 __reExport = (target, mod, secondTarget) => (__copyProps(target, mod, "default"), secondTarget && __copyProps(secondTarget, mod, "default"));
+var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
+  // If the importer is in node compatibility mode or this is not an ESM
+  // file that has been converted to a CommonJS file using a Babel-
+  // compatible transform (i.e. "__esModule" has not been set), then set
+  // "default" to the CommonJS "module.exports" for node compatibility.
+  isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
+  mod
+));
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var parser_exports = {};
+__export(parser_exports, {
+  iterators: () => iterators
+});
+module.exports = __toCommonJS(parser_exports);
+__reExport(parser_exports, require("./types.js"), module.exports);
+__reExport(parser_exports, require("./usfm-parser.js"), module.exports);
+__reExport(parser_exports, require("./usx-parser.js"), module.exports);
+__reExport(parser_exports, require("./codex-parser.js"), module.exports);
+var iterators = __toESM(require("./iterators.js"));
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  iterators,
+  ...require("./types.js"),
+  ...require("./usfm-parser.js"),
+  ...require("./usx-parser.js"),
+  ...require("./codex-parser.js")
+});
+//# sourceMappingURL=index.cjs.map

package/dist/cjs/parser/index.cjs.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../../../parser/index.ts"],
+  "sourcesContent": ["export * from './types.js';\nexport * from './usfm-parser.js';\nexport * from './usx-parser.js';\nexport * from './codex-parser.js';\nimport * as iterators from './iterators.js';\n\nexport { iterators };\n"],
+  "mappings": ";;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,2BAAc,uBAAd;AACA,2BAAc,6BADd;AAEA,2BAAc,4BAFd;AAGA,2BAAc,8BAHd;AAIA,gBAA2B;",
+  "names": []
+}