feat(utils): JSDoc for all APIs (#6980)

* feat(utils): JSDoc for all APIs

* fix tests

packages/docusaurus-plugin-content-blog/src/types.ts

@@ -5,10 +5,7 @@
  * LICENSE file in the root directory of this source tree.
  */
 
-import type {
-  BrokenMarkdownLink,
-  ContentPaths,
-} from '@docusaurus/utils/lib/markdownLinks';
+import type {BrokenMarkdownLink, ContentPaths} from '@docusaurus/utils';
 import type {BlogPostMetadata} from '@docusaurus/plugin-content-blog';
 import type {Metadata as BlogPaginatedMetadata} from '@theme/BlogListPage';
 

packages/docusaurus-plugin-content-docs/src/markdown/__tests__/linkify.test.ts

@@ -13,7 +13,7 @@ import type {
   DocsMarkdownOption,
   SourceToPermalink,
   VersionMetadata,
-  BrokenMarkdownLink,
+  DocBrokenMarkdownLink,
 } from '../../types';
 import {VERSIONED_DOCS_DIR, CURRENT_VERSION_NAME} from '../../constants';
 
@@ -156,22 +156,22 @@ describe('linkify', () => {
       filePath: doc5,
       link: 'docNotExist1.md',
       contentPaths: versionCurrent,
-    } as BrokenMarkdownLink);
+    } as DocBrokenMarkdownLink);
     expect(onBrokenMarkdownLink).toHaveBeenNthCalledWith(2, {
       filePath: doc5,
       link: './docNotExist2.mdx',
       contentPaths: versionCurrent,
-    } as BrokenMarkdownLink);
+    } as DocBrokenMarkdownLink);
     expect(onBrokenMarkdownLink).toHaveBeenNthCalledWith(3, {
       filePath: doc5,
       link: '../docNotExist3.mdx',
       contentPaths: versionCurrent,
-    } as BrokenMarkdownLink);
+    } as DocBrokenMarkdownLink);
     expect(onBrokenMarkdownLink).toHaveBeenNthCalledWith(4, {
       filePath: doc5,
       link: './subdir/docNotExist4.md',
       contentPaths: versionCurrent,
-    } as BrokenMarkdownLink);
+    } as DocBrokenMarkdownLink);
   });
 
   it('transforms absolute links in versioned docs', async () => {

packages/docusaurus-plugin-content-docs/src/sidebars/utils.ts

@@ -20,7 +20,7 @@ import type {
 } from './types';
 
 import _ from 'lodash';
-import {getElementsAround, toMessageRelativeFilePath} from '@docusaurus/utils';
+import {toMessageRelativeFilePath} from '@docusaurus/utils';
 import type {DocMetadataBase, DocNavLink} from '../types';
 
 export function isCategoriesShorthand(
@@ -225,11 +225,11 @@ export function createSidebarsUtils(sidebars: Sidebars): SidebarsUtils {
       return {sidebarName, next: undefined, previous: undefined};
     }
 
-    const {previous, next} = getElementsAround(
-      navigationItems,
-      currentItemIndex,
-    );
-    return {sidebarName, previous, next};
+    return {
+      sidebarName,
+      previous: navigationItems[currentItemIndex - 1],
+      next: navigationItems[currentItemIndex + 1],
+    };
   }
 
   function getCategoryGeneratedIndexList(): SidebarItemCategoryWithGeneratedIndex[] {
@@ -268,11 +268,11 @@ export function createSidebarsUtils(sidebars: Sidebars): SidebarsUtils {
     const currentItemIndex = navigationItems.findIndex(
       isCurrentCategoryGeneratedIndexItem,
     );
-    const {previous, next} = getElementsAround(
-      navigationItems,
-      currentItemIndex,
-    );
-    return {sidebarName, previous, next};
+    return {
+      sidebarName,
+      previous: navigationItems[currentItemIndex - 1],
+      next: navigationItems[currentItemIndex + 1],
+    };
   }
 
   function checkSidebarsDocIds(validDocIds: string[], sidebarFilePath: string) {

packages/docusaurus-plugin-content-docs/src/types.ts

@@ -8,11 +8,12 @@
 /// <reference types="@docusaurus/module-type-aliases" />
 
 import type {Sidebars} from './sidebars/types';
-import type {Tag, FrontMatterTag} from '@docusaurus/utils';
 import type {
-  BrokenMarkdownLink as IBrokenMarkdownLink,
+  Tag,
+  FrontMatterTag,
+  BrokenMarkdownLink,
   ContentPaths,
-} from '@docusaurus/utils/lib/markdownLinks';
+} from '@docusaurus/utils';
 import type {VersionBanner} from '@docusaurus/plugin-content-docs';
 
 export type DocFile = {
@@ -133,11 +134,11 @@ export type LoadedContent = {
   loadedVersions: LoadedVersion[];
 };
 
-export type BrokenMarkdownLink = IBrokenMarkdownLink<VersionMetadata>;
+export type DocBrokenMarkdownLink = BrokenMarkdownLink<VersionMetadata>;
 
 export type DocsMarkdownOption = {
   versionsMetadata: VersionMetadata[];
   siteDir: string;
   sourceToPermalink: SourceToPermalink;
-  onBrokenMarkdownLink: (brokenMarkdownLink: BrokenMarkdownLink) => void;
+  onBrokenMarkdownLink: (brokenMarkdownLink: DocBrokenMarkdownLink) => void;
 };

packages/docusaurus-theme-classic/src/theme-classic.d.ts

@@ -898,7 +898,8 @@ declare module '@theme/TagsListByLetter' {
 }
 
 declare module '@theme/TagsListInline' {
-  export type Tag = Readonly<{label: string; permalink: string}>;
+  import type {Tag} from '@docusaurus/utils';
+
   export interface Props {
     readonly tags: readonly Tag[];
   }

packages/docusaurus-utils-validation/package.json

@@ -20,6 +20,7 @@
   "dependencies": {
     "@docusaurus/logger": "2.0.0-beta.17",
     "@docusaurus/utils": "2.0.0-beta.17",
+    "js-yaml": "^4.1.0",
     "joi": "^17.6.0",
     "tslib": "^2.3.1"
   },

packages/docusaurus-utils-validation/src/JoiFrontMatter.ts

@@ -18,11 +18,13 @@ const JoiFrontMatterString: Joi.Extension = {
     return {value};
   },
 };
+
 /**
- * Enhance the default Joi.string() type so that it can convert number to
+ * Enhance the default `Joi.string()` type so that it can convert number to
  * strings. If user use front matter "tag: 2021", we shouldn't need to ask her
  * to write "tag: '2021'". Also yaml tries to convert patterns like "2019-01-01"
  * to dates automatically.
+ *
  * @see https://github.com/facebook/docusaurus/issues/4642
  * @see https://github.com/sideway/joi/issues/1442#issuecomment-823997884
  */

packages/docusaurus-utils-validation/src/validationUtils.ts

@@ -7,8 +7,10 @@
 
 import type Joi from './Joi';
 import logger from '@docusaurus/logger';
+import Yaml from 'js-yaml';
 import {PluginIdSchema} from './validationSchemas';
 
+/** Print warnings returned from Joi validation. */
 export function printWarning(warning?: Joi.ValidationError): void {
   if (warning) {
     const warningMessages = warning.details
@@ -18,9 +20,14 @@ export function printWarning(warning?: Joi.ValidationError): void {
   }
 }
 
+/**
+ * The callback that should be used to validate plugin options. Handles plugin
+ * IDs on a generic level: no matter what the schema declares, this callback
+ * would require a string ID or default to "default".
+ */
 export function normalizePluginOptions<T extends {id?: string}>(
   schema: Joi.ObjectSchema<T>,
-  // This allows us to automatically normalize undefined to {id: 'default'}
+  // This allows us to automatically normalize undefined to { id: "default" }
   options: Partial<T> = {},
 ): T {
   // All plugins can be provided an "id" option (multi-instance support)
@@ -41,6 +48,10 @@ export function normalizePluginOptions<T extends {id?: string}>(
   return value;
 }
 
+/**
+ * The callback that should be used to validate theme config. No matter what the
+ * schema declares, this callback would allow unknown attributes.
+ */
 export function normalizeThemeConfig<T>(
   schema: Joi.ObjectSchema<T>,
   themeConfig: Partial<T>,
@@ -62,6 +73,9 @@ export function normalizeThemeConfig<T>(
   return value;
 }
 
+/**
+ * Validate front matter with better error message
+ */
 export function validateFrontMatter<T>(
   frontMatter: Record<string, unknown>,
   schema: Joi.ObjectSchema<T>,
@@ -75,13 +89,13 @@ export function validateFrontMatter<T>(
   printWarning(warning);
 
   if (error) {
-    const frontMatterString = JSON.stringify(frontMatter, null, 2);
     const errorDetails = error.details;
     const invalidFields = errorDetails.map(({path}) => path).join(', ');
 
     logger.error`The following front matter:
-${logger.yellow(frontMatterString)}
-contains invalid values for field(s): ${logger.yellow(invalidFields)}.
+---
+${Yaml.dump(frontMatter)}---
+contains invalid values for field(s): code=${invalidFields}.
 ${errorDetails.map(({message}) => message)}
 `;
     throw error;

packages/docusaurus-utils/src/__tests__/globUtils.test.ts

@@ -103,7 +103,7 @@ describe('createAbsoluteFilePathMatcher', () => {
     expect(() =>
       matcher('/bad/path/myDoc.md'),
     ).toThrowErrorMatchingInlineSnapshot(
-      `"createAbsoluteFilePathMatcher unexpected error, absoluteFilePath=/bad/path/myDoc.md was not contained in any of the root folders [\\"/_root/docs\\",\\"/root/_docs/\\",\\"/__test__/website/src\\"]"`,
+      `"createAbsoluteFilePathMatcher unexpected error, absoluteFilePath=/bad/path/myDoc.md was not contained in any of the root folders: /_root/docs, /root/_docs/, /__test__/website/src"`,
     );
   });
 });

packages/docusaurus-utils/src/__tests__/jsUtils.test.ts

@@ -9,7 +9,6 @@ import {jest} from '@jest/globals';
 import {
   removeSuffix,
   removePrefix,
-  getElementsAround,
   mapAsyncSequential,
   findAsyncSequential,
   reportMessage,
@@ -38,40 +37,6 @@ describe('removePrefix', () => {
   });
 });
 
-describe('getElementsAround', () => {
-  it('returns elements around', () => {
-    expect(getElementsAround(['a', 'b', 'c', 'd'], 0)).toEqual({
-      previous: undefined,
-      next: 'b',
-    });
-    expect(getElementsAround(['a', 'b', 'c', 'd'], 1)).toEqual({
-      previous: 'a',
-      next: 'c',
-    });
-    expect(getElementsAround(['a', 'b', 'c', 'd'], 2)).toEqual({
-      previous: 'b',
-      next: 'd',
-    });
-    expect(getElementsAround(['a', 'b', 'c', 'd'], 3)).toEqual({
-      previous: 'c',
-      next: undefined,
-    });
-  });
-
-  it('throws if bad index is provided', () => {
-    expect(() =>
-      getElementsAround(['a', 'b', 'c', 'd'], -1),
-    ).toThrowErrorMatchingInlineSnapshot(
-      `"Valid \\"aroundIndex\\" for array (of size 4) are between 0 and 3, but you provided -1."`,
-    );
-    expect(() =>
-      getElementsAround(['a', 'b', 'c', 'd'], 4),
-    ).toThrowErrorMatchingInlineSnapshot(
-      `"Valid \\"aroundIndex\\" for array (of size 4) are between 0 and 3, but you provided 4."`,
-    );
-  });
-});
-
 describe('mapAsyncSequential', () => {
   function sleep(timeout: number): Promise<void> {
     return new Promise((resolve) => {

packages/docusaurus-utils/src/__tests__/markdownUtils.test.ts

@@ -113,9 +113,13 @@ describe('createExcerpt', () => {
           import Component from '@site/src/components/Component'
           import './styles.css';
 
-          export function ItemCol(props) { return <Item {...props} className={'col col--6 margin-bottom--lg'}/> }
+          export function ItemCol(props) {
+            return <Item {...props} className={'col col--6 margin-bottom--lg'}/>
+          }
 
-          export function ItemCol(props) { return <Item {...props} className={'col col--6 margin-bottom--lg'}/> };
+          export function ItemCol(props) {
+            return <Item {...props} className={'col col--6 margin-bottom--lg'}/>
+          };
 
           Lorem **ipsum** dolor sit \`amet\`[^1], consectetur _adipiscing_ elit. [**Vestibulum**](https://wiktionary.org/wiki/vestibulum) ex urna[^note], ~~molestie~~ et sagittis ut, varius ac justo :wink:.
 
@@ -146,6 +150,18 @@ describe('createExcerpt', () => {
         `),
     ).toBe('Lorem ipsum dolor sit amet, consectetur adipiscing elit.');
   });
+
+  it('creates excerpt after multi-line imports', () => {
+    expect(
+      createExcerpt(dedent`
+          import React, {
+            type ReactNode,
+          } from 'react';
+
+          Lorem \`ipsum\` dolor sit amet, consectetur \`adipiscing elit\`.
+        `),
+    ).toBe('Lorem ipsum dolor sit amet, consectetur adipiscing elit.');
+  });
 });
 
 describe('parseMarkdownContentTitle', () => {

packages/docusaurus-utils/src/__tests__/tags.test.ts

@@ -5,62 +5,60 @@
  * LICENSE file in the root directory of this source tree.
  */
 
-import {
-  normalizeFrontMatterTag,
-  normalizeFrontMatterTags,
-  groupTaggedItems,
-  type Tag,
-} from '../tags';
-
-describe('normalizeFrontMatterTag', () => {
-  type Input = Parameters<typeof normalizeFrontMatterTag>[1];
-  type Output = ReturnType<typeof normalizeFrontMatterTag>;
+import {normalizeFrontMatterTags, groupTaggedItems, type Tag} from '../tags';
 
+describe('normalizeFrontMatterTags', () => {
   it('normalizes simple string tag', () => {
     const tagsPath = '/all/tags';
-    const input: Input = 'tag';
-    const expectedOutput: Output = {
+    const input = 'tag';
+    const expectedOutput = {
       label: 'tag',
       permalink: `${tagsPath}/tag`,
     };
-    expect(normalizeFrontMatterTag(tagsPath, input)).toEqual(expectedOutput);
+    expect(normalizeFrontMatterTags(tagsPath, [input])).toEqual([
+      expectedOutput,
+    ]);
   });
 
   it('normalizes complex string tag', () => {
     const tagsPath = '/all/tags';
-    const input: Input = 'some more Complex_tag';
-    const expectedOutput: Output = {
+    const input = 'some more Complex_tag';
+    const expectedOutput = {
       label: 'some more Complex_tag',
       permalink: `${tagsPath}/some-more-complex-tag`,
     };
-    expect(normalizeFrontMatterTag(tagsPath, input)).toEqual(expectedOutput);
+    expect(normalizeFrontMatterTags(tagsPath, [input])).toEqual([
+      expectedOutput,
+    ]);
   });
 
   it('normalizes simple object tag', () => {
     const tagsPath = '/all/tags';
-    const input: Input = {label: 'tag', permalink: 'tagPermalink'};
-    const expectedOutput: Output = {
+    const input = {label: 'tag', permalink: 'tagPermalink'};
+    const expectedOutput = {
       label: 'tag',
       permalink: `${tagsPath}/tagPermalink`,
     };
-    expect(normalizeFrontMatterTag(tagsPath, input)).toEqual(expectedOutput);
+    expect(normalizeFrontMatterTags(tagsPath, [input])).toEqual([
+      expectedOutput,
+    ]);
   });
 
   it('normalizes complex string tag with object tag', () => {
     const tagsPath = '/all/tags';
-    const input: Input = {
+    const input = {
       label: 'tag complex Label',
       permalink: '/MoreComplex/Permalink',
     };
-    const expectedOutput: Output = {
+    const expectedOutput = {
       label: 'tag complex Label',
       permalink: `${tagsPath}/MoreComplex/Permalink`,
     };
-    expect(normalizeFrontMatterTag(tagsPath, input)).toEqual(expectedOutput);
+    expect(normalizeFrontMatterTags(tagsPath, [input])).toEqual([
+      expectedOutput,
+    ]);
   });
-});
 
-describe('normalizeFrontMatterTags', () => {
   type Input = Parameters<typeof normalizeFrontMatterTags>[1];
   type Output = ReturnType<typeof normalizeFrontMatterTags>;
 

packages/docusaurus-utils/src/constants.ts

@@ -5,34 +5,86 @@
  * LICENSE file in the root directory of this source tree.
  */
 
+/** Node major version, directly read from env. */
 export const NODE_MAJOR_VERSION = parseInt(
   process.versions.node.split('.')[0]!,
   10,
 );
+/** Node minor version, directly read from env. */
 export const NODE_MINOR_VERSION = parseInt(
   process.versions.node.split('.')[1]!,
   10,
 );
 
-// Can be overridden with cli option --out-dir
+/**
+ * Can be overridden with cli option `--out-dir`. Code should generally use
+ * `context.outDir` instead (which is always absolute and localized).
+ */
 export const DEFAULT_BUILD_DIR_NAME = 'build';
 
-// Can be overridden with cli option --config
+/**
+ * Can be overridden with cli option `--config`. Code should generally use
+ * `context.siteConfigPath` instead (which is always absolute).
+ */
 export const DEFAULT_CONFIG_FILE_NAME = 'docusaurus.config.js';
 
+/** Can be absolute or relative to site directory. */
 export const BABEL_CONFIG_FILE_NAME =
-  process.env.DOCUSAURUS_BABEL_CONFIG_FILE_NAME || 'babel.config.js';
+  process.env.DOCUSAURUS_BABEL_CONFIG_FILE_NAME ?? 'babel.config.js';
 
+/**
+ * Can be absolute or relative to site directory. Code should generally use
+ * `context.generatedFilesDir` instead (which is always absolute).
+ */
 export const GENERATED_FILES_DIR_NAME =
-  process.env.DOCUSAURUS_GENERATED_FILES_DIR_NAME || '.docusaurus';
+  process.env.DOCUSAURUS_GENERATED_FILES_DIR_NAME ?? '.docusaurus';
 
+/**
+ * We would assume all of the site's JS code lives in here and not outside.
+ * Relative to the site directory.
+ */
 export const SRC_DIR_NAME = 'src';
-export const STATIC_DIR_NAME = 'static';
-export const OUTPUT_STATIC_ASSETS_DIR_NAME = 'assets'; // files handled by webpack, hashed (can be cached aggressively)
+
+/**
+ * Can be overridden with `config.staticDirectories`. Code should use
+ * `context.siteConfig.staticDirectories` instead (which is always absolute).
+ */
+export const DEFAULT_STATIC_DIR_NAME = 'static';
+
+/**
+ * Files here are handled by webpack, hashed (can be cached aggressively).
+ * Relative to the build output folder.
+ */
+export const OUTPUT_STATIC_ASSETS_DIR_NAME = 'assets';
+
+/**
+ * Components in this directory will receive the `@theme` alias and be able to
+ * shadow default theme components.
+ */
 export const THEME_PATH = `${SRC_DIR_NAME}/theme`;
+
+/**
+ * All translation-related data live here, relative to site directory. Content
+ * will be namespaced by locale.
+ */
+export const I18N_DIR_NAME = 'i18n';
+
+/**
+ * Translations for React code.
+ */
+export const CODE_TRANSLATIONS_FILE_NAME = 'code.json';
+
+/** Dev server opens on this port by default. */
 export const DEFAULT_PORT = 3000;
+
+/** Default plugin ID. */
 export const DEFAULT_PLUGIN_ID = 'default';
 
-// Temporary fix for https://github.com/facebook/docusaurus/issues/5493
+/**
+ * Allow overriding the limit after which the url loader will no longer inline
+ * assets.
+ *
+ * @see https://github.com/facebook/docusaurus/issues/5493
+ */
 export const WEBPACK_URL_LOADER_LIMIT =
   process.env.WEBPACK_URL_LOADER_LIMIT ?? 10000;

packages/docusaurus-utils/src/dataFileUtils.ts

@@ -13,15 +13,25 @@ import type {ContentPaths} from './markdownLinks';
 import logger from '@docusaurus/logger';
 
 type DataFileParams = {
+  /** Path to the potential data file, relative to `contentPaths` */
   filePath: string;
+  /**
+   * Includes the base path and localized path, both of which are eligible for
+   * sourcing data files. Both paths should be absolute.
+   */
   contentPaths: ContentPaths;
 };
 
+/**
+ * Looks for a data file in the potential content paths; loads a localized data
+ * file in priority.
+ *
+ * @returns An absolute path to the data file, or `undefined` if there isn't one.
+ */
 export async function getDataFilePath({
   filePath,
   contentPaths,
 }: DataFileParams): Promise<string | undefined> {
-  // Loads a localized data file in priority
   const contentPath = await findFolderContainingFile(
     getContentPathList(contentPaths),
     filePath,
@@ -33,11 +43,17 @@ export async function getDataFilePath({
 }
 
 /**
- * Looks up for a data file in the content paths, returns the normalized object.
- * Throws when validation fails; returns undefined when file not found
+ * Looks up for a data file in the content paths, returns the object validated
+ * and normalized according to the `validate` callback.
+ *
+ * @returns `undefined` when file not found
+ * @throws Throws when validation fails, displaying a helpful context message.
  */
 export async function getDataFileData<T>(
-  params: DataFileParams & {fileType: string},
+  params: DataFileParams & {
+    /** Used for the "The X file looks invalid" message. */
+    fileType: string;
+  },
   validate: (content: unknown) => T,
 ): Promise<T | undefined> {
   const filePath = await getDataFilePath(params);
@@ -54,12 +70,21 @@ export async function getDataFileData<T>(
   }
 }
 
-// Order matters: we look in priority in localized folder
+/**
+ * Takes the `contentPaths` data structure and returns an ordered path list
+ * indicating their priorities. For all data, we look in the localized folder
+ * in priority.
+ */
 export function getContentPathList(contentPaths: ContentPaths): string[] {
   return [contentPaths.contentPathLocalized, contentPaths.contentPath];
 }
 
-// return the first folder path in which the file exists in
+/**
+ * @param folderPaths a list of absolute paths.
+ * @param relativeFilePath file path relative to each `folderPaths`.
+ * @returns the first folder path in which the file exists, or `undefined` if
+ * none is found.
+ */
 export async function findFolderContainingFile(
   folderPaths: string[],
   relativeFilePath: string,
@@ -69,6 +94,16 @@ export async function findFolderContainingFile(
   );
 }
 
+/**
+ * Fail-fast alternative to `findFolderContainingFile`.
+ *
+ * @param folderPaths a list of absolute paths.
+ * @param relativeFilePath file path relative to each `folderPaths`.
+ * @returns the first folder path in which the file exists.
+ * @throws Throws if no file can be found. You should use this method only when
+ * you actually know the file exists (e.g. when the `relativeFilePath` is read
+ * with a glob and you are just trying to localize it)
+ */
 export async function getFolderContainingFile(
   folderPaths: string[],
   relativeFilePath: string,
@@ -77,12 +112,10 @@ export async function getFolderContainingFile(
     folderPaths,
     relativeFilePath,
   );
-  // should never happen, as the source was read from the FS anyway...
   if (!maybeFolderPath) {
     throw new Error(
-      `File "${relativeFilePath}" does not exist in any of these folders:\n- ${folderPaths.join(
-        '\n- ',
-      )}`,
+      `File "${relativeFilePath}" does not exist in any of these folders:
+- ${folderPaths.join('\n- ')}`,
     );
   }
   return maybeFolderPath;

packages/docusaurus-utils/src/emitUtils.ts

@@ -13,6 +13,16 @@ import {findAsyncSequential} from './jsUtils';
 
 const fileHash = new Map<string, string>();
 
+/**
+ * Outputs a file to the generated files directory. Only writes files if content
+ * differs from cache (for hot reload performance).
+ *
+ * @param generatedFilesDir Absolute path.
+ * @param file Path relative to `generatedFilesDir`.
+ * @param content String content to write.
+ * @param skipCache If `true` (defaults as `true` for production), file is
+ * force-rewritten, skipping cache.
+ */
 export async function generate(
   generatedFilesDir: string,
   file: string,
@@ -23,14 +33,21 @@ export async function generate(
 
   if (skipCache) {
     await fs.outputFile(filepath, content);
+    // Cache still needs to be reset, otherwise, writing "A", "B", and "A" where
+    // "B" skips cache will cause the last "A" not be able to overwrite as the
+    // first "A" remains in cache. But if the file never existed in cache, no
+    // need to register it.
+    if (fileHash.get(filepath)) {
+      fileHash.set(filepath, createHash('md5').update(content).digest('hex'));
+    }
     return;
   }
 
   let lastHash = fileHash.get(filepath);
 
-  // If file already exists but its not in runtime cache yet,
-  // we try to calculate the content hash and then compare
-  // This is to avoid unnecessary overwriting and we can reuse old file.
+  // If file already exists but it's not in runtime cache yet, we try to
+  // calculate the content hash and then compare. This is to avoid unnecessary
+  // overwriting and we can reuse old file.
   if (!lastHash && (await fs.pathExists(filepath))) {
     const lastContent = await fs.readFile(filepath, 'utf8');
     lastHash = createHash('md5').update(lastContent).digest('hex');
@@ -45,7 +62,7 @@ export async function generate(
   }
 }
 
-const chunkNameCache = new Map();
+const chunkNameCache = new Map<string, string>();
 
 /**
  * Generate unique chunk name given a module path.
@@ -56,7 +73,7 @@ export function genChunkName(
   preferredName?: string,
   shortId: boolean = process.env.NODE_ENV === 'production',
 ): string {
-  let chunkName: string | undefined = chunkNameCache.get(modulePath);
+  let chunkName = chunkNameCache.get(modulePath);
   if (!chunkName) {
     if (shortId) {
       chunkName = simpleHash(modulePath, 8);
@@ -82,6 +99,8 @@ export function genChunkName(
  * @returns This returns a buffer, which you have to decode string yourself if
  * needed. (Not always necessary since the output isn't for human consumption
  * anyways, and most HTML manipulation libs accept buffers)
+ * @throws Throws when the HTML file is not found at any of the potential paths.
+ * This should never happen as it would lead to a 404.
  */
 export async function readOutputHTMLFile(
   permalink: string,

packages/docusaurus-utils/src/gitUtils.ts

@@ -8,23 +8,67 @@
 import path from 'path';
 import shell from 'shelljs';
 
+/** Custom error thrown when git is not found in `PATH`. */
 export class GitNotFoundError extends Error {}
 
+/** Custom error thrown when the current file is not tracked by git. */
 export class FileNotTrackedError extends Error {}
 
+/**
+ * Fetches the git history of a file and returns a relevant commit date.
+ * It gets the commit date instead of author date so that amended commits
+ * can have their dates updated.
+ *
+ * @throws {GitNotFoundError} If git is not found in `PATH`.
+ * @throws {FileNotTrackedError} If the current file is not tracked by git.
+ * @throws Also throws when `git log` exited with non-zero, or when it outputs
+ * unexpected text.
+ */
 export function getFileCommitDate(
+  /** Absolute path to the file. */
   file: string,
-  args: {age?: 'oldest' | 'newest'; includeAuthor?: false},
+  args: {
+    /**
+     * `"oldest"` is the commit that added the file, following renames;
+     * `"newest"` is the last commit that edited the file.
+     */
+    age?: 'oldest' | 'newest';
+    /** Use `includeAuthor: true` to get the author information as well. */
+    includeAuthor?: false;
+  },
 ): {
+  /** Relevant commit date. */
   date: Date;
+  /** Timestamp in **seconds**, as returned from git. */
   timestamp: number;
 };
+/**
+ * Fetches the git history of a file and returns a relevant commit date.
+ * It gets the commit date instead of author date so that amended commits
+ * can have their dates updated.
+ *
+ * @throws {GitNotFoundError} If git is not found in `PATH`.
+ * @throws {FileNotTrackedError} If the current file is not tracked by git.
+ * @throws Also throws when `git log` exited with non-zero, or when it outputs
+ * unexpected text.
+ */
 export function getFileCommitDate(
+  /** Absolute path to the file. */
   file: string,
-  args: {age?: 'oldest' | 'newest'; includeAuthor: true},
+  args: {
+    /**
+     * `"oldest"` is the commit that added the file, following renames;
+     * `"newest"` is the last commit that edited the file.
+     */
+    age?: 'oldest' | 'newest';
+    includeAuthor: true;
+  },
 ): {
+  /** Relevant commit date. */
   date: Date;
+  /** Timestamp in **seconds**, as returned from git. */
   timestamp: number;
+  /** The author's name, as returned from git. */
   author: string;
 };
 export function getFileCommitDate(
@@ -53,8 +97,6 @@ export function getFileCommitDate(
     );
   }
 
-  // Commit time and author name; not using author time so that amended commits
-  // can have their dates updated
   let formatArg = '--format=%ct';
   if (includeAuthor) {
     formatArg += ',%an';

packages/docusaurus-utils/src/globUtils.ts

@@ -10,24 +10,31 @@
 import Micromatch from 'micromatch'; // Note: Micromatch is used by Globby
 import path from 'path';
 
+/** A re-export of the globby instance. */
 export {default as Globby} from 'globby';
 
-// The default patterns we ignore when globbing
-// using _ prefix for exclusion by convention
+/**
+ * The default glob patterns we ignore when sourcing content.
+ * - Ignore files and folders starting with `_` recursively
+ * - Ignore tests
+ */
 export const GlobExcludeDefault = [
-  // Ignore files starting with _
   '**/_*.{js,jsx,ts,tsx,md,mdx}',
-
-  // Ignore folders starting with _ (including folder content)
   '**/_*/**',
-
-  // Ignore tests
   '**/*.test.{js,jsx,ts,tsx}',
   '**/__tests__/**',
 ];
 
 type Matcher = (str: string) => boolean;
 
+/**
+ * A very thin wrapper around `Micromatch.makeRe`.
+ *
+ * @see {@link createAbsoluteFilePathMatcher}
+ * @param patterns A list of glob patterns.
+ * @returns A matcher handle that tells if a file path is matched by any of the
+ * patterns.
+ */
 export function createMatcher(patterns: string[]): Matcher {
   const regexp = new RegExp(
     patterns.map((pattern) => Micromatch.makeRe(pattern).source).join('|'),
@@ -35,10 +42,19 @@ export function createMatcher(patterns: string[]): Matcher {
   return (str) => regexp.test(str);
 }
 
-// We use match patterns like '**/_*/**',
-// This function permits to help to:
-// Match /user/sebastien/website/docs/_partials/xyz.md
-// Ignore /user/_sebastien/website/docs/partials/xyz.md
+/**
+ * We use match patterns like `"** /_* /**"` (ignore the spaces), where `"_*"`
+ * should only be matched within a subfolder. This function would:
+ * - Match `/user/sebastien/website/docs/_partials/xyz.md`
+ * - Ignore `/user/_sebastien/website/docs/partials/xyz.md`
+ *
+ * @param patterns A list of glob patterns.
+ * @param rootFolders A list of root folders to resolve the glob from.
+ * @returns A matcher handle that tells if a file path is matched by any of the
+ * patterns, resolved from the first root folder that contains the path.
+ * @throws Throws when the returned matcher receives a path that doesn't belong
+ * to any of the `rootFolders`.
+ */
 export function createAbsoluteFilePathMatcher(
   patterns: string[],
   rootFolders: string[],
@@ -51,8 +67,8 @@ export function createAbsoluteFilePathMatcher(
     );
     if (!rootFolder) {
       throw new Error(
-        `createAbsoluteFilePathMatcher unexpected error, absoluteFilePath=${absoluteFilePath} was not contained in any of the root folders ${JSON.stringify(
-          rootFolders,
+        `createAbsoluteFilePathMatcher unexpected error, absoluteFilePath=${absoluteFilePath} was not contained in any of the root folders: ${rootFolders.join(
+          ', ',
         )}`,
       );
     }

packages/docusaurus-utils/src/hashUtils.ts

@@ -9,20 +9,21 @@ import {createHash} from 'crypto';
 import _ from 'lodash';
 import {shortName, isNameTooLong} from './pathUtils';
 
+/** Thin wrapper around `crypto.createHash("md5")`. */
 export function md5Hash(str: string): string {
   return createHash('md5').update(str).digest('hex');
 }
 
+/** Creates an MD5 hash and truncates it to the given length. */
 export function simpleHash(str: string, length: number): string {
-  return md5Hash(str).substr(0, length);
+  return md5Hash(str).substring(0, length);
 }
 
 // Based on https://github.com/gatsbyjs/gatsby/pull/21518/files
 /**
- * Given an input string, convert to kebab-case and append a hash.
- * Avoid str collision.
- * Also removes part of the string if its larger than the allowed
- * filename per OS. Avoids ERRNAMETOOLONG error.
+ * Given an input string, convert to kebab-case and append a hash, avoiding name
+ * collision. Also removes part of the string if its larger than the allowed
+ * filename per OS, avoiding `ERRNAMETOOLONG` error.
  */
 export function docuHash(str: string): string {
   if (str === '/') {

packages/docusaurus-utils/src/i18nUtils.ts

@@ -8,16 +8,21 @@
 import path from 'path';
 import _ from 'lodash';
 import type {TranslationFileContent, TranslationFile} from '@docusaurus/types';
-import {DEFAULT_PLUGIN_ID} from './constants';
+import {DEFAULT_PLUGIN_ID, I18N_DIR_NAME} from './constants';
 
+/**
+ * Takes a list of translation file contents, and shallow-merges them into one.
+ */
 export function mergeTranslations(
   contents: TranslationFileContent[],
 ): TranslationFileContent {
   return contents.reduce((acc, content) => ({...acc, ...content}), {});
 }
 
-// Useful to update all the messages of a translation file
-// Used in tests to simulate translations
+/**
+ * Useful to update all the messages of a translation file. Used in tests to
+ * simulate translations.
+ */
 export function updateTranslationFileMessages(
   translationFile: TranslationFile,
   updateMessage: (message: string) => string,
@@ -31,6 +36,10 @@ export function updateTranslationFileMessages(
   };
 }
 
+/**
+ * Takes everything needed and constructs a plugin i18n path. Plugins should
+ * expect everything it needs for translations to be found under this path.
+ */
 export function getPluginI18nPath({
   siteDir,
   locale,
@@ -46,7 +55,7 @@ export function getPluginI18nPath({
 }): string {
   return path.join(
     siteDir,
-    'i18n',
+    I18N_DIR_NAME,
     // namespace first by locale: convenient to work in a single folder for a
     // translator
     locale,

packages/docusaurus-utils/src/index.ts

@@ -13,9 +13,11 @@ export {
   BABEL_CONFIG_FILE_NAME,
   GENERATED_FILES_DIR_NAME,
   SRC_DIR_NAME,
-  STATIC_DIR_NAME,
+  DEFAULT_STATIC_DIR_NAME,
   OUTPUT_STATIC_ASSETS_DIR_NAME,
   THEME_PATH,
+  I18N_DIR_NAME,
+  CODE_TRANSLATIONS_FILE_NAME,
   DEFAULT_PORT,
   DEFAULT_PLUGIN_ID,
   WEBPACK_URL_LOADER_LIMIT,
@@ -34,7 +36,6 @@ export {
 export {
   removeSuffix,
   removePrefix,
-  getElementsAround,
   mapAsyncSequential,
   findAsyncSequential,
   reportMessage,
@@ -56,8 +57,6 @@ export {
 export {
   type Tag,
   type FrontMatterTag,
-  type TaggedItemGroup,
-  normalizeFrontMatterTag,
   normalizeFrontMatterTags,
   groupTaggedItems,
 } from './tags';
@@ -73,8 +72,6 @@ export {
 export {
   type ContentPaths,
   type BrokenMarkdownLink,
-  type ReplaceMarkdownLinksParams,
-  type ReplaceMarkdownLinksReturn,
   replaceMarkdownLinks,
 } from './markdownLinks';
 export {type SluggerOptions, type Slugger, createSlugger} from './slugger';

packages/docusaurus-utils/src/jsUtils.ts

@@ -8,36 +8,27 @@
 import type {ReportingSeverity} from '@docusaurus/types';
 import logger from '@docusaurus/logger';
 
+/** Removes a given string suffix from `str`. */
 export function removeSuffix(str: string, suffix: string): string {
   if (suffix === '') {
-    return str; // always returns "" otherwise!
+    // str.slice(0, 0) is ""
+    return str;
   }
   return str.endsWith(suffix) ? str.slice(0, -suffix.length) : str;
 }
 
+/** Removes a given string prefix from `str`. */
 export function removePrefix(str: string, prefix: string): string {
   return str.startsWith(prefix) ? str.slice(prefix.length) : str;
 }
 
-export function getElementsAround<T>(
-  array: T[],
-  aroundIndex: number,
-): {
-  next: T | undefined;
-  previous: T | undefined;
-} {
-  const min = 0;
-  const max = array.length - 1;
-  if (aroundIndex < min || aroundIndex > max) {
-    throw new Error(
-      `Valid "aroundIndex" for array (of size ${array.length}) are between ${min} and ${max}, but you provided ${aroundIndex}.`,
-    );
-  }
-  const previous = aroundIndex === min ? undefined : array[aroundIndex - 1];
-  const next = aroundIndex === max ? undefined : array[aroundIndex + 1];
-  return {previous, next};
-}
-
+/**
+ * `Array#map` for async operations where order matters.
+ * @param array The array to traverse.
+ * @param action An async action to be performed on every array item. Will be
+ * awaited before working on the next.
+ * @returns The list of results returned from every `action(item)`
+ */
 export async function mapAsyncSequential<T, R>(
   array: T[],
   action: (t: T) => Promise<R>,
@@ -50,6 +41,14 @@ export async function mapAsyncSequential<T, R>(
   return results;
 }
 
+/**
+ * `Array#find` for async operations where order matters.
+ * @param array The array to traverse.
+ * @param predicate An async predicate to be called on every array item. Should
+ * return a boolean indicating whether the currently element should be returned.
+ * @returns The function immediately returns the first item on which `predicate`
+ * returns `true`, or `undefined` if none matches the predicate.
+ */
 export async function findAsyncSequential<T>(
   array: T[],
   predicate: (t: T) => Promise<boolean>,
@@ -62,6 +61,21 @@ export async function findAsyncSequential<T>(
   return undefined;
 }
 
+/**
+ * Takes a message and reports it according to the severity that the user wants.
+ *
+ * - `ignore`: completely no-op
+ * - `log`: uses the `INFO` log level
+ * - `warn`: uses the `WARN` log level
+ * - `error`: uses the `ERROR` log level
+ * - `throw`: aborts the process, throws the error.
+ *
+ * Since the logger doesn't have logging level filters yet, these severities
+ * mostly just differ by their colors.
+ *
+ * @throws In addition to throwing when `reportingSeverity === "throw"`, this
+ * function also throws if `reportingSeverity` is not one of the above.
+ */
 export function reportMessage(
   message: string,
   reportingSeverity: ReportingSeverity,

packages/docusaurus-utils/src/markdownLinks.ts

@@ -6,41 +6,79 @@
  */
 
 import path from 'path';
+import {getContentPathList} from './dataFileUtils';
 import {aliasedSitePath} from './pathUtils';
 
+/**
+ * Content plugins have a base path and a localized path to source content from.
+ * We will look into the localized path in priority.
+ */
 export type ContentPaths = {
+  /**
+   * The absolute path to the base content directory, like `"<siteDir>/docs"`.
+   */
   contentPath: string;
+  /**
+   * The absolute path to the localized content directory, like
+   * `"<siteDir>/i18n/zh-Hans/plugin-content-docs"`.
+   */
   contentPathLocalized: string;
 };
 
+/** Data structure representing each broken Markdown link to be reported. */
 export type BrokenMarkdownLink<T extends ContentPaths> = {
+  /** Absolute path to the file containing this link. */
   filePath: string;
+  /**
+   * This is generic because it may contain extra metadata like version name,
+   * which the reporter can provide for context.
+   */
   contentPaths: T;
+  /**
+   * The content of the link, like `"./brokenFile.md"`
+   */
   link: string;
 };
 
-export type ReplaceMarkdownLinksParams<T extends ContentPaths> = {
-  siteDir: string;
-  fileString: string;
-  filePath: string;
-  contentPaths: T;
-  sourceToPermalink: Record<string, string>;
-};
-
-export type ReplaceMarkdownLinksReturn<T extends ContentPaths> = {
-  newContent: string;
-  brokenMarkdownLinks: BrokenMarkdownLink<T>[];
-};
-
+/**
+ * Takes a Markdown file and replaces relative file references with their URL
+ * counterparts, e.g. `[link](./intro.md)` => `[link](/docs/intro)`, preserving
+ * everything else.
+ *
+ * This method uses best effort to find a matching file. The file reference can
+ * be relative to the directory of the current file (most likely) or any of the
+ * content paths (so `/tutorials/intro.md` can be resolved as
+ * `<siteDir>/docs/tutorials/intro.md`). Links that contain the `http(s):` or
+ * `@site/` prefix will always be ignored.
+ */
 export function replaceMarkdownLinks<T extends ContentPaths>({
   siteDir,
   fileString,
   filePath,
   contentPaths,
   sourceToPermalink,
-}: ReplaceMarkdownLinksParams<T>): ReplaceMarkdownLinksReturn<T> {
-  const {contentPath, contentPathLocalized} = contentPaths;
-
+}: {
+  /** Absolute path to the site directory, used to resolve aliased paths. */
+  siteDir: string;
+  /** The Markdown file content to be processed. */
+  fileString: string;
+  /** Absolute path to the current file containing `fileString`. */
+  filePath: string;
+  /** The content paths which the file reference may live in. */
+  contentPaths: T;
+  /**
+   * A map from source paths to their URLs. Source paths are `@site` aliased.
+   */
+  sourceToPermalink: Record<string, string>;
+}): {
+  /**
+   * The content with all Markdown file references replaced with their URLs.
+   * Unresolved links are left as-is.
+   */
+  newContent: string;
+  /** The list of broken links,  */
+  brokenMarkdownLinks: BrokenMarkdownLink<T>[];
+} {
   const brokenMarkdownLinks: BrokenMarkdownLink<T>[] = [];
 
   // Replace internal markdown linking (except in fenced blocks).
@@ -64,9 +102,8 @@ export function replaceMarkdownLinks<T extends ContentPaths>({
 
     let modifiedLine = line;
     // Replace inline-style links or reference-style links e.g:
-    // This is [Document 1](doc1.md) -> we replace this doc1.md with correct
-    // ink
-    // [doc1]: doc1.md -> we replace this doc1.md with correct link
+    // This is [Document 1](doc1.md)
+    // [doc1]: doc1.md
     const mdRegex =
       /(?:\]\(|\]:\s*)(?!https?:\/\/|@site\/)(?<filename>[^'")\]\s>]+\.mdx?)/g;
     let mdMatch = mdRegex.exec(modifiedLine);
@@ -75,10 +112,9 @@ export function replaceMarkdownLinks<T extends ContentPaths>({
       const mdLink = mdMatch.groups!.filename!;
 
       const sourcesToTry = [
-        path.resolve(path.dirname(filePath), decodeURIComponent(mdLink)),
-        `${contentPathLocalized}/${decodeURIComponent(mdLink)}`,
-        `${contentPath}/${decodeURIComponent(mdLink)}`,
-      ];
+        path.dirname(filePath),
+        ...getContentPathList(contentPaths),
+      ].map((p) => path.join(p, decodeURIComponent(mdLink)));
 
       const aliasedSourceMatch = sourcesToTry
         .map((source) => aliasedSitePath(source, siteDir))

packages/docusaurus-utils/src/markdownUtils.ts

@@ -7,12 +7,25 @@
 
 import logger from '@docusaurus/logger';
 import matter from 'gray-matter';
-import {createSlugger, type Slugger} from './slugger';
+import {createSlugger, type Slugger, type SluggerOptions} from './slugger';
 
-// Input: ## Some heading {#some-heading}
-// Output: {text: "## Some heading", id: "some-heading"}
+// Some utilities for parsing Markdown content. These things are only used on
+// server-side when we infer metadata like `title` and `description` from the
+// content. Most parsing is still done in MDX through the mdx-loader.
+
+/**
+ * Parses custom ID from a heading. The ID must be composed of letters,
+ * underscores, and dashes only.
+ *
+ * @param heading e.g. `## Some heading {#some-heading}` where the last
+ * character must be `}` for the ID to be recognized
+ */
 export function parseMarkdownHeadingId(heading: string): {
+  /**
+   * The heading content sans the ID part, right-trimmed. e.g. `## Some heading`
+   */
   text: string;
+  /** The heading ID. e.g. `some-heading` */
   id?: string;
 } {
   const customHeadingIdRegex = /\s*\{#(?<id>[\w-]+)\}$/;
@@ -26,26 +39,40 @@ export function parseMarkdownHeadingId(heading: string): {
   return {text: heading, id: undefined};
 }
 
-// Hacky way of stripping out import statements from the excerpt
 // TODO: Find a better way to do so, possibly by compiling the Markdown content,
 // stripping out HTML tags and obtaining the first line.
+/**
+ * Creates an excerpt of a Markdown file. This function will:
+ *
+ * - Ignore h1 headings (setext or atx)
+ * - Ignore import/export
+ * - Ignore code blocks
+ *
+ * And for the first contentful line, it will strip away most Markdown
+ * syntax, including HTML tags, emphasis, links (keeping the text), etc.
+ */
 export function createExcerpt(fileString: string): string | undefined {
   const fileLines = fileString
-    .trimLeft()
+    .trimStart()
     // Remove Markdown alternate title
     .replace(/^[^\n]*\n[=]+/g, '')
     .split('\n');
   let inCode = false;
+  let inImport = false;
   let lastCodeFence = '';
 
   for (const fileLine of fileLines) {
+    if (fileLine === '' && inImport) {
+      inImport = false;
+    }
     // Skip empty line.
     if (!fileLine.trim()) {
       continue;
     }
 
     // Skip import/export declaration.
-    if (/^(?:import|export)\s.*/.test(fileLine)) {
+    if ((/^(?:import|export)\s.*/.test(fileLine) || inImport) && !inCode) {
+      inImport = true;
       continue;
     }
 
@@ -102,8 +129,22 @@ export function createExcerpt(fileString: string): string | undefined {
   return undefined;
 }
 
+/**
+ * Takes a raw Markdown file content, and parses the front matter using
+ * gray-matter. Worth noting that gray-matter accepts TOML and other markup
+ * languages as well.
+ *
+ * @throws Throws when gray-matter throws. e.g.:
+ * ```md
+ * ---
+ * foo: : bar
+ * ---
+ * ```
+ */
 export function parseFrontMatter(markdownFileContent: string): {
+  /** Front matter as parsed by gray-matter. */
   frontMatter: Record<string, unknown>;
+  /** The remaining content, trimmed. */
   content: string;
 } {
   const {data, content} = matter(markdownFileContent);
@@ -113,11 +154,6 @@ export function parseFrontMatter(markdownFileContent: string): {
   };
 }
 
-/**
- * Try to convert markdown heading to text. Does not need to be perfect, it is
- * only used as a fallback when frontMatter.title is not provided. For now, we
- * just unwrap possible inline code blocks (# `config.js`)
- */
 function toTextContentTitle(contentTitle: string): string {
   if (contentTitle.startsWith('`') && contentTitle.endsWith('`')) {
     return contentTitle.substring(1, contentTitle.length - 1);
@@ -125,10 +161,36 @@ function toTextContentTitle(contentTitle: string): string {
   return contentTitle;
 }
 
+type ParseMarkdownContentTitleOptions = {
+  /**
+   * If `true`, the matching title will be removed from the returned content.
+   * We can promise that at least one empty line will be left between the
+   * content before and after, but you shouldn't make too much assumption
+   * about what's left.
+   */
+  removeContentTitle?: boolean;
+};
+
+/**
+ * Takes the raw Markdown content, without front matter, and tries to find an h1
+ * title (setext or atx) to be used as metadata.
+ *
+ * It only searches until the first contentful paragraph, ignoring import/export
+ * declarations.
+ *
+ * It will try to convert markdown to reasonable text, but won't be best effort,
+ * since it's only used as a fallback when `frontMatter.title` is not provided.
+ * For now, we just unwrap inline code (``# `config.js` `` => `config.js`).
+ */
 export function parseMarkdownContentTitle(
   contentUntrimmed: string,
-  options?: {removeContentTitle?: boolean},
-): {content: string; contentTitle: string | undefined} {
+  options?: ParseMarkdownContentTitleOptions,
+): {
+  /** The content, optionally without the content title. */
+  content: string;
+  /** The title, trimmed and without the `#`. */
+  contentTitle: string | undefined;
+} {
   const removeContentTitleOption = options?.removeContentTitle ?? false;
 
   const content = contentUntrimmed.trim();
@@ -171,17 +233,28 @@ export function parseMarkdownContentTitle(
   };
 }
 
-type ParsedMarkdown = {
-  frontMatter: Record<string, unknown>;
-  content: string;
-  contentTitle: string | undefined;
-  excerpt: string | undefined;
-};
-
+/**
+ * Makes a full-round parse.
+ *
+ * @throws Throws when `parseFrontMatter` throws, usually because of invalid
+ * syntax.
+ */
 export function parseMarkdownString(
   markdownFileContent: string,
-  options?: {removeContentTitle?: boolean},
-): ParsedMarkdown {
+  options?: ParseMarkdownContentTitleOptions,
+): {
+  /** @see {@link parseFrontMatter} */
+  frontMatter: Record<string, unknown>;
+  /** @see {@link parseMarkdownContentTitle} */
+  contentTitle: string | undefined;
+  /** @see {@link createExcerpt} */
+  excerpt: string | undefined;
+  /**
+   * Content without front matter and (optionally) without title, depending on
+   * the `removeContentTitle` option.
+   */
+  content: string;
+} {
   try {
     const {frontMatter, content: contentWithoutFrontMatter} =
       parseFrontMatter(markdownFileContent);
@@ -229,11 +302,16 @@ function addHeadingId(
   return `${headingHashes}${headingText} {#${slug}}`;
 }
 
-export type WriteHeadingIDOptions = {
-  maintainCase?: boolean;
+export type WriteHeadingIDOptions = SluggerOptions & {
+  /** Overwrite existing heading IDs. */
   overwrite?: boolean;
 };
 
+/**
+ * Takes Markdown content, returns new content with heading IDs written.
+ * Respects existing IDs (unless `overwrite=true`) and never generates colliding
+ * IDs (through the slugger).
+ */
 export function writeMarkdownHeadingId(
   content: string,
   options: WriteHeadingIDOptions = {maintainCase: false, overwrite: false},

packages/docusaurus-utils/src/pathUtils.ts

@@ -24,7 +24,7 @@ export const isNameTooLong = (str: string): boolean =>
     ? str.length + SPACE_FOR_APPENDING > MAX_PATH_SEGMENT_CHARS // MacOS (APFS) and Windows (NTFS) filename length limit (255 chars)
     : Buffer.from(str).length + SPACE_FOR_APPENDING > MAX_PATH_SEGMENT_BYTES; // Other (255 bytes)
 
-export const shortName = (str: string): string => {
+export function shortName(str: string): string {
   if (isMacOs() || isWindows()) {
     const overflowingChars = str.length - MAX_PATH_SEGMENT_CHARS;
     return str.slice(
@@ -41,7 +41,7 @@ export const shortName = (str: string): string => {
       Buffer.byteLength(strBuffer) - overflowingBytes - SPACE_FOR_APPENDING - 1,
     )
     .toString();
-};
+}
 
 /**
  * Convert Windows backslash paths to posix style paths.

packages/docusaurus-utils/src/slugger.ts

@@ -10,12 +10,24 @@ import GithubSlugger from 'github-slugger';
 // We create our own abstraction on top of the lib:
 // - unify usage everywhere in the codebase
 // - ability to add extra options
-export type SluggerOptions = {maintainCase?: boolean};
+export type SluggerOptions = {
+  /** Keep the headings' casing, otherwise make all lowercase. */
+  maintainCase?: boolean;
+};
 
 export type Slugger = {
+  /**
+   * Takes a Markdown heading like "Josh Cena" and sluggifies it according to
+   * GitHub semantics (in this case `josh-cena`). Stateful, because if you try
+   * to sluggify "Josh Cena" again it would return `josh-cena-1`.
+   */
   slug: (value: string, options?: SluggerOptions) => string;
 };
 
+/**
+ * A thin wrapper around github-slugger. This is a factory function that returns
+ * a stateful Slugger object.
+ */
 export function createSlugger(): Slugger {
   const githubSlugger = new GithubSlugger();
   return {

packages/docusaurus-utils/src/tags.ts

@@ -10,12 +10,13 @@ import {normalizeUrl} from './urlUtils';
 
 export type Tag = {
   label: string;
+  /** Permalink to this tag's page, without the `/tags/` base path. */
   permalink: string;
 };
 
 export type FrontMatterTag = string | Tag;
 
-export function normalizeFrontMatterTag(
+function normalizeFrontMatterTag(
   tagsPath: string,
   frontMatterTag: FrontMatterTag,
 ): Tag {
@@ -45,8 +46,19 @@ export function normalizeFrontMatterTag(
   };
 }
 
+/**
+ * Takes tag objects as they are defined in front matter, and normalizes each
+ * into a standard tag object. The permalink is created by appending the
+ * sluggified label to `tagsPath`. Front matter tags already containing
+ * permalinks would still have `tagsPath` prepended.
+ *
+ * The result will always be unique by permalinks. The behavior with colliding
+ * permalinks is undetermined.
+ */
 export function normalizeFrontMatterTags(
+  /** Base path to append the tag permalinks to. */
   tagsPath: string,
+  /** Can be `undefined`, so that we can directly pipe in `frontMatter.tags`. */
   frontMatterTags: FrontMatterTag[] | undefined = [],
 ): Tag[] {
   const tags = frontMatterTags.map((tag) =>
@@ -56,29 +68,34 @@ export function normalizeFrontMatterTags(
   return _.uniqBy(tags, (tag) => tag.permalink);
 }
 
-export type TaggedItemGroup<Item> = {
+type TaggedItemGroup<Item> = {
   tag: Tag;
   items: Item[];
 };
 
 /**
- * Permits to group docs/blogPosts by tag (provided by front matter)
- * Note: groups are indexed by permalink, because routes must be unique in the
- * end. Labels may vary on 2 md files but they are normalized. Docs with
- * label='some label' and label='some-label' should end-up in the same
- * group/page in the end. We can't create 2 routes /some-label because one would
- * override the other
+ * Permits to group docs/blog posts by tag (provided by front matter).
+ *
+ * @returns a map from tag permalink to the items and other relevant tag data.
+ * The record is indexed by permalink, because routes must be unique in the end.
+ * Labels may vary on 2 MD files but they are normalized. Docs with
+ * label='some label' and label='some-label' should end up in the same page.
  */
 export function groupTaggedItems<Item>(
   items: readonly Item[],
+  /**
+   * A callback telling me how to get the tags list of the current item. Usually
+   * simply getting it from some metadata of the current item.
+   */
   getItemTags: (item: Item) => readonly Tag[],
-): Record<string, TaggedItemGroup<Item>> {
-  const result: Record<string, TaggedItemGroup<Item>> = {};
+): {[permalink: string]: TaggedItemGroup<Item>} {
+  const result: {[permalink: string]: TaggedItemGroup<Item>} = {};
 
-  function handleItemTag(item: Item, tag: Tag) {
+  items.forEach((item) => {
+    getItemTags(item).forEach((tag) => {
       // Init missing tag groups
-    // TODO: it's not really clear what should be the behavior if 2 items have
-    // the same tag but the permalink is different for each
+      // TODO: it's not really clear what should be the behavior if 2 tags have
+      // the same permalink but the label is different for each
       // For now, the first tag found wins
       result[tag.permalink] ??= {
         tag,
@@ -87,11 +104,6 @@ export function groupTaggedItems<Item>(
 
       // Add item to group
       result[tag.permalink]!.items.push(item);
-  }
-
-  items.forEach((item) => {
-    getItemTags(item).forEach((tag) => {
-      handleItemTag(item, tag);
     });
   });
 

packages/docusaurus-utils/src/urlUtils.ts

@@ -8,6 +8,18 @@
 import {removeSuffix} from './jsUtils';
 import resolvePathnameUnsafe from 'resolve-pathname';
 
+/**
+ * Much like `path.join`, but much better. Takes an array of URL segments, and
+ * joins them into a reasonable URL.
+ *
+ * - `["file:", "/home", "/user/", "website"]` => `file:///home/user/website`
+ * - `["file://", "home", "/user/", "website"]` => `file://home/user/website` (relative!)
+ * - Remove trailing slash before parameters or hash.
+ * - Replace `?` in query parameters with `&`.
+ * - Dedupe forward slashes in the entire path, avoiding protocol slashes.
+ *
+ * @throws {TypeError} If any of the URL segment is not a string, this throws.
+ */
 export function normalizeUrl(rawUrls: string[]): string {
   const urls = [...rawUrls];
   const resultArray = [];
@@ -75,8 +87,8 @@ export function normalizeUrl(rawUrls: string[]): string {
   }
 
   let str = resultArray.join('/');
-  // Each input component is now separated by a single slash
-  // except the possible first plain protocol part.
+  // Each input component is now separated by a single slash except the possible
+  // first plain protocol part.
 
   // Remove trailing slash before parameters or hash.
   str = str.replace(/\/(?<search>\?|&|#[^!])/g, '$1');
@@ -94,6 +106,11 @@ export function normalizeUrl(rawUrls: string[]): string {
   return str;
 }
 
+/**
+ * Takes a file's path, relative to its content folder, and computes its edit
+ * URL. If `editUrl` is `undefined`, this returns `undefined`, as is the case
+ * when the user doesn't want an edit URL in her config.
+ */
 export function getEditUrl(
   fileRelativePath: string,
   editUrl?: string,
@@ -105,8 +122,8 @@ export function getEditUrl(
 }
 
 /**
- * Convert filepath to url path.
- * Example: 'index.md' -> '/', 'foo/bar.js' -> '/foo/bar',
+ * Converts file path to a reasonable URL path, e.g. `'index.md'` -> `'/'`,
+ * `'foo/bar.js'` -> `'/foo/bar'`
  */
 export function fileToPath(file: string): string {
   const indexRE = /(?<dirname>^|.*\/)index\.(?:mdx?|jsx?|tsx?)$/i;
@@ -118,6 +135,13 @@ export function fileToPath(file: string): string {
   return `/${file.replace(extRE, '').replace(/\\/g, '/')}`;
 }
 
+/**
+ * Similar to `encodeURI`, but uses `encodeURIComponent` and assumes there's no
+ * query.
+ *
+ * `encodeURI("/question?/answer")` => `"/question?/answer#section"`;
+ * `encodePath("/question?/answer#section")` => `"/question%3F/answer%23foo"`
+ */
 export function encodePath(userPath: string): string {
   return userPath
     .split('/')
@@ -125,6 +149,10 @@ export function encodePath(userPath: string): string {
     .join('/');
 }
 
+/**
+ * Whether `str` is a valid pathname. It must be absolute, and not contain
+ * special characters.
+ */
 export function isValidPathname(str: string): boolean {
   if (!str.startsWith('/')) {
     return false;
@@ -138,22 +166,31 @@ export function isValidPathname(str: string): boolean {
   }
 }
 
-// resolve pathname and fail fast if resolution fails
+/**
+ * Resolve pathnames and fail-fast if resolution fails. Uses standard URL
+ * semantics (provided by `resolve-pathname` which is used internally by React
+ * router)
+ */
 export function resolvePathname(to: string, from?: string): string {
   return resolvePathnameUnsafe(to, from);
 }
+/** Appends a leading slash to `str`, if one doesn't exist. */
 export function addLeadingSlash(str: string): string {
   return str.startsWith('/') ? str : `/${str}`;
 }
 
 // TODO deduplicate: also present in @docusaurus/utils-common
+/** Appends a trailing slash to `str`, if one doesn't exist. */
 export function addTrailingSlash(str: string): string {
   return str.endsWith('/') ? str : `${str}/`;
 }
+
+/** Removes the trailing slash from `str`. */
 export function removeTrailingSlash(str: string): string {
   return removeSuffix(str, '/');
 }
 
+/** Constructs an SSH URL that can be used to push to GitHub. */
 export function buildSshUrl(
   githubHost: string,
   organizationName: string,
@@ -166,6 +203,7 @@ export function buildSshUrl(
   return `git@${githubHost}:${organizationName}/${projectName}.git`;
 }
 
+/** Constructs an HTTP URL that can be used to push to GitHub. */
 export function buildHttpsUrl(
   gitCredentials: string,
   githubHost: string,
@@ -179,6 +217,11 @@ export function buildHttpsUrl(
   return `https://${gitCredentials}@${githubHost}/${organizationName}/${projectName}.git`;
 }
 
+/**
+ * Whether the current URL is an SSH protocol. In addition to looking for
+ * `ssh:`, it will also allow protocol-less URLs like
+ * `git@github.com:facebook/docusaurus.git`.
+ */
 export function hasSSHProtocol(sourceRepoUrl: string): boolean {
   try {
     if (new URL(sourceRepoUrl).protocol === 'ssh:') {
@@ -187,6 +230,6 @@ export function hasSSHProtocol(sourceRepoUrl: string): boolean {
     return false;
   } catch {
     // Fails when there isn't a protocol
-    return /^(?:[\w-]+@)?[\w.-]+:[\w./-]+/.test(sourceRepoUrl); // git@github.com:facebook/docusaurus.git
+    return /^(?:[\w-]+@)?[\w.-]+:[\w./-]+/.test(sourceRepoUrl);
   }
 }

packages/docusaurus-utils/src/webpackUtils.ts

@@ -31,7 +31,11 @@ type FileLoaderUtils = {
   };
 };
 
-// Inspired by https://github.com/gatsbyjs/gatsby/blob/8e6e021014da310b9cc7d02e58c9b3efe938c665/packages/gatsby/src/utils/webpack-utils.ts#L447
+/**
+ * Returns unified loader configurations to be used for various file types.
+ *
+ * Inspired by https://github.com/gatsbyjs/gatsby/blob/8e6e021014da310b9cc7d02e58c9b3efe938c665/packages/gatsby/src/utils/webpack-utils.ts#L447
+ */
 export function getFileLoaderUtils(): FileLoaderUtils {
   // files/images < urlLoaderLimit will be inlined as base64 strings directly in
   // the html
@@ -39,7 +43,11 @@ export function getFileLoaderUtils(): FileLoaderUtils {
 
   // defines the path/pattern of the assets handled by webpack
   const fileLoaderFileName = (folder: AssetFolder) =>
-    `${OUTPUT_STATIC_ASSETS_DIR_NAME}/${folder}/[name]-[contenthash].[ext]`;
+    path.posix.join(
+      OUTPUT_STATIC_ASSETS_DIR_NAME,
+      folder,
+      '[name]-[contenthash].[ext]',
+    );
 
   const loaders: FileLoaderUtils['loaders'] = {
     file: (options: {folder: AssetFolder}) => ({

packages/docusaurus/src/server/configValidation.ts

@@ -6,7 +6,10 @@
  */
 
 import type {DocusaurusConfig, I18nConfig} from '@docusaurus/types';
-import {DEFAULT_CONFIG_FILE_NAME, STATIC_DIR_NAME} from '@docusaurus/utils';
+import {
+  DEFAULT_CONFIG_FILE_NAME,
+  DEFAULT_STATIC_DIR_NAME,
+} from '@docusaurus/utils';
 import {Joi, URISchema, printWarning} from '@docusaurus/utils-validation';
 
 const DEFAULT_I18N_LOCALE = 'en';
@@ -53,7 +56,7 @@ export const DEFAULT_CONFIG: Pick<
   noIndex: false,
   tagline: '',
   baseUrlIssueBanner: true,
-  staticDirectories: [STATIC_DIR_NAME],
+  staticDirectories: [DEFAULT_STATIC_DIR_NAME],
 };
 
 function createPluginSchema(theme: boolean) {

packages/docusaurus/src/server/translations/translations.ts

@@ -14,7 +14,12 @@ import type {
   TranslationMessage,
   InitializedPlugin,
 } from '@docusaurus/types';
-import {getPluginI18nPath, toMessageRelativeFilePath} from '@docusaurus/utils';
+import {
+  getPluginI18nPath,
+  toMessageRelativeFilePath,
+  I18N_DIR_NAME,
+  CODE_TRANSLATIONS_FILE_NAME,
+} from '@docusaurus/utils';
 import {Joi} from '@docusaurus/utils-validation';
 import logger from '@docusaurus/logger';
 
@@ -140,7 +145,7 @@ Maybe you should remove them? ${unknownKeys}`;
 
 // should we make this configurable?
 function getTranslationsDirPath(context: TranslationContext): string {
-  return path.resolve(path.join(context.siteDir, `i18n`));
+  return path.resolve(path.join(context.siteDir, I18N_DIR_NAME));
 }
 export function getTranslationsLocaleDirPath(
   context: TranslationContext,
@@ -149,7 +154,10 @@ export function getTranslationsLocaleDirPath(
 }
 
 function getCodeTranslationsFilePath(context: TranslationContext): string {
-  return path.join(getTranslationsLocaleDirPath(context), 'code.json');
+  return path.join(
+    getTranslationsLocaleDirPath(context),
+    CODE_TRANSLATIONS_FILE_NAME,
+  );
 }
 
 export async function readCodeTranslationFileContent(

project-words.txt

@@ -256,9 +256,13 @@ sebastienlorber
 sensical
 serializers
 setaf
+setext
 sida
 simen
 slorber
+sluggified
+sluggifies
+sluggify
 spâce
 stackblitz
 stackblitzrc