@docusaurus/plugin-content-blog 2.0.0-beta.8e9b829d9 → 2.0.0-beta.9

package/src/blogUtils.ts

@@ -9,14 +9,14 @@ import fs from 'fs-extra';
 import chalk from 'chalk';
 import path from 'path';
 import readingTime from 'reading-time';
-import {Feed} from 'feed';
 import {keyBy, mapValues} from 'lodash';
 import {
   PluginOptions,
   BlogPost,
-  DateLink,
   BlogContentPaths,
   BlogMarkdownLoaderOptions,
+  BlogTags,
+  ReadingTimeFunction,
 } from './types';
 import {
   parseMarkdownFile,
@@ -27,9 +27,12 @@ import {
   posixPath,
   replaceMarkdownLinks,
   Globby,
+  normalizeFrontMatterTags,
+  groupTaggedItems,
 } from '@docusaurus/utils';
 import {LoadContext} from '@docusaurus/types';
 import {validateBlogPostFrontMatter} from './blogFrontMatter';
+import {AuthorsMap, getAuthorsMap, getBlogPostAuthors} from './authors';
 
 export function truncate(fileString: string, truncateMarker: RegExp): string {
   return fileString.split(truncateMarker, 1).shift()!;
@@ -44,15 +47,46 @@ export function getSourceToPermalink(
   );
 }
 
-// YYYY-MM-DD-{name}.mdx?
-// Prefer named capture, but older Node versions do not support it.
-const DATE_FILENAME_PATTERN = /^(\d{4}-\d{1,2}-\d{1,2})-?(.*?).mdx?$/;
+export function getBlogTags(blogPosts: BlogPost[]): BlogTags {
+  const groups = groupTaggedItems(
+    blogPosts,
+    (blogPost) => blogPost.metadata.tags,
+  );
+  return mapValues(groups, (group) => {
+    return {
+      name: group.tag.label,
+      items: group.items.map((item) => item.id),
+      permalink: group.tag.permalink,
+    };
+  });
+}
 
-function toUrl({date, link}: DateLink) {
-  return `${date
-    .toISOString()
-    .substring(0, '2019-01-01'.length)
-    .replace(/-/g, '/')}/${link}`;
+const DATE_FILENAME_REGEX =
+  /^(?<date>\d{4}[-/]\d{1,2}[-/]\d{1,2})[-/]?(?<text>.*?)(\/index)?.mdx?$/;
+
+type ParsedBlogFileName = {
+  date: Date | undefined;
+  text: string;
+  slug: string;
+};
+
+export function parseBlogFileName(
+  blogSourceRelative: string,
+): ParsedBlogFileName {
+  const dateFilenameMatch = blogSourceRelative.match(DATE_FILENAME_REGEX);
+  if (dateFilenameMatch) {
+    const dateString = dateFilenameMatch.groups!.date!;
+    const text = dateFilenameMatch.groups!.text!;
+    // Always treat dates as UTC by adding the `Z`
+    const date = new Date(`${dateString}Z`);
+    const slugDate = dateString.replace(/-/g, '/');
+    const slug = `/${slugDate}/${text}`;
+    return {date, text, slug};
+  } else {
+    const text = blogSourceRelative.replace(/(\/index)?\.mdx?$/, '');
+    const slug = `/${text}`;
+    return {date: undefined, text, slug};
+  }
 }
 
 function formatBlogPostDate(locale: string, date: Date): string {
@@ -68,206 +102,193 @@ function formatBlogPostDate(locale: string, date: Date): string {
   }
 }
 
-export async function generateBlogFeed(
-  contentPaths: BlogContentPaths,
-  context: LoadContext,
-  options: PluginOptions,
-): Promise<Feed | null> {
-  if (!options.feedOptions) {
-    throw new Error(
-      'Invalid options: "feedOptions" is not expected to be null.',
-    );
-  }
-  const {siteConfig} = context;
-  const blogPosts = await generateBlogPosts(contentPaths, context, options);
-  if (!blogPosts.length) {
-    return null;
-  }
-
-  const {feedOptions, routeBasePath} = options;
-  const {url: siteUrl, baseUrl, title, favicon} = siteConfig;
-  const blogBaseUrl = normalizeUrl([siteUrl, baseUrl, routeBasePath]);
-
-  const updated =
-    (blogPosts[0] && blogPosts[0].metadata.date) ||
-    new Date('2015-10-25T16:29:00.000-07:00');
-
-  const feed = new Feed({
-    id: blogBaseUrl,
-    title: feedOptions.title || `${title} Blog`,
-    updated,
-    language: feedOptions.language,
-    link: blogBaseUrl,
-    description: feedOptions.description || `${siteConfig.title} Blog`,
-    favicon: favicon ? normalizeUrl([siteUrl, baseUrl, favicon]) : undefined,
-    copyright: feedOptions.copyright,
-  });
-
-  blogPosts.forEach((post) => {
-    const {
-      id,
-      metadata: {title: metadataTitle, permalink, date, description},
-    } = post;
-    feed.addItem({
-      title: metadataTitle,
-      id,
-      link: normalizeUrl([siteUrl, permalink]),
-      date,
-      description,
-    });
+async function parseBlogPostMarkdownFile(blogSourceAbsolute: string) {
+  const result = await parseMarkdownFile(blogSourceAbsolute, {
+    removeContentTitle: true,
   });
-
-  return feed;
+  return {
+    ...result,
+    frontMatter: validateBlogPostFrontMatter(result.frontMatter),
+  };
 }
 
-export async function generateBlogPosts(
+const defaultReadingTime: ReadingTimeFunction = ({content, options}) => {
+  return readingTime(content, options).minutes;
+};
+
+async function processBlogSourceFile(
+  blogSourceRelative: string,
   contentPaths: BlogContentPaths,
-  {siteConfig, siteDir, i18n}: LoadContext,
+  context: LoadContext,
   options: PluginOptions,
-): Promise<BlogPost[]> {
+  authorsMap?: AuthorsMap,
+): Promise<BlogPost | undefined> {
+  const {
+    siteConfig: {baseUrl},
+    siteDir,
+    i18n,
+  } = context;
   const {
-    include,
-    exclude,
     routeBasePath,
+    tagsBasePath: tagsRouteBasePath,
     truncateMarker,
     showReadingTime,
     editUrl,
   } = options;
 
-  if (!fs.existsSync(contentPaths.contentPath)) {
-    return [];
-  }
+  // Lookup in localized folder in priority
+  const blogDirPath = await getFolderContainingFile(
+    getContentPathList(contentPaths),
+    blogSourceRelative,
+  );
 
-  const {baseUrl = ''} = siteConfig;
-  const blogSourceFiles = await Globby(include, {
-    cwd: contentPaths.contentPath,
-    ignore: exclude,
-  });
+  const blogSourceAbsolute = path.join(blogDirPath, blogSourceRelative);
 
-  const blogPosts: BlogPost[] = [];
+  const {frontMatter, content, contentTitle, excerpt} =
+    await parseBlogPostMarkdownFile(blogSourceAbsolute);
 
-  async function processBlogSourceFile(blogSourceFile: string) {
-    // Lookup in localized folder in priority
-    const blogDirPath = await getFolderContainingFile(
-      getContentPathList(contentPaths),
-      blogSourceFile,
+  const aliasedSource = aliasedSitePath(blogSourceAbsolute, siteDir);
+
+  if (frontMatter.draft && process.env.NODE_ENV === 'production') {
+    return undefined;
+  }
+
+  if (frontMatter.id) {
+    console.warn(
+      chalk.yellow(
+        `"id" header option is deprecated in ${blogSourceRelative} file. Please use "slug" option instead.`,
+      ),
     );
+  }
 
-    const source = path.join(blogDirPath, blogSourceFile);
+  const parsedBlogFileName = parseBlogFileName(blogSourceRelative);
 
-    const {
-      frontMatter: unsafeFrontMatter,
-      content,
-      contentTitle,
-      excerpt,
-    } = await parseMarkdownFile(source, {removeContentTitle: true});
-    const frontMatter = validateBlogPostFrontMatter(unsafeFrontMatter);
+  async function getDate(): Promise<Date> {
+    // Prefer user-defined date.
+    if (frontMatter.date) {
+      return new Date(frontMatter.date);
+    } else if (parsedBlogFileName.date) {
+      return parsedBlogFileName.date;
+    }
+    // Fallback to file create time
+    return (await fs.stat(blogSourceAbsolute)).birthtime;
+  }
 
-    const aliasedSource = aliasedSitePath(source, siteDir);
+  const date = await getDate();
+  const formattedDate = formatBlogPostDate(i18n.currentLocale, date);
 
-    const blogFileName = path.basename(blogSourceFile);
+  const title = frontMatter.title ?? contentTitle ?? parsedBlogFileName.text;
+  const description = frontMatter.description ?? excerpt ?? '';
 
-    if (frontMatter.draft && process.env.NODE_ENV === 'production') {
-      return;
-    }
+  const slug = frontMatter.slug || parsedBlogFileName.slug;
 
-    if (frontMatter.id) {
-      console.warn(
-        chalk.yellow(
-          `"id" header option is deprecated in ${blogFileName} file. Please use "slug" option instead.`,
-        ),
-      );
-    }
+  const permalink = normalizeUrl([baseUrl, routeBasePath, slug]);
 
-    let date: Date | undefined;
-    // Extract date and title from filename.
-    const dateFilenameMatch = blogFileName.match(DATE_FILENAME_PATTERN);
-    let linkName = blogFileName.replace(/\.mdx?$/, '');
+  function getBlogEditUrl() {
+    const blogPathRelative = path.relative(
+      blogDirPath,
+      path.resolve(blogSourceAbsolute),
+    );
 
-    if (dateFilenameMatch) {
-      const [, dateString, name] = dateFilenameMatch;
-      // Always treat dates as UTC by adding the `Z`
-      date = new Date(`${dateString}Z`);
-      linkName = name;
+    if (typeof editUrl === 'function') {
+      return editUrl({
+        blogDirPath: posixPath(path.relative(siteDir, blogDirPath)),
+        blogPath: posixPath(blogPathRelative),
+        permalink,
+        locale: i18n.currentLocale,
+      });
+    } else if (typeof editUrl === 'string') {
+      const isLocalized = blogDirPath === contentPaths.contentPathLocalized;
+      const fileContentPath =
+        isLocalized && options.editLocalizedFiles
+          ? contentPaths.contentPathLocalized
+          : contentPaths.contentPath;
+
+      const contentPathEditUrl = normalizeUrl([
+        editUrl,
+        posixPath(path.relative(siteDir, fileContentPath)),
+      ]);
+
+      return getEditUrl(blogPathRelative, contentPathEditUrl);
     }
+    return undefined;
+  }
 
-    // Prefer user-defined date.
-    if (frontMatter.date) {
-      date = new Date(frontMatter.date);
-    }
+  const tagsBasePath = normalizeUrl([
+    baseUrl,
+    routeBasePath,
+    tagsRouteBasePath,
+  ]);
+  const authors = getBlogPostAuthors({authorsMap, frontMatter});
+
+  return {
+    id: frontMatter.slug ?? title,
+    metadata: {
+      permalink,
+      editUrl: getBlogEditUrl(),
+      source: aliasedSource,
+      title,
+      description,
+      date,
+      formattedDate,
+      tags: normalizeFrontMatterTags(tagsBasePath, frontMatter.tags),
+      readingTime: showReadingTime
+        ? options.readingTime({
+            content,
+            frontMatter,
+            defaultReadingTime,
+          })
+        : undefined,
+      truncated: truncateMarker?.test(content) || false,
+      authors,
+    },
+    content,
+  };
+}
 
-    // Use file create time for blog.
-    date = date ?? (await fs.stat(source)).birthtime;
-    const formattedDate = formatBlogPostDate(i18n.currentLocale, date);
-
-    const title = frontMatter.title ?? contentTitle ?? linkName;
-    const description = frontMatter.description ?? excerpt ?? '';
-
-    const slug =
-      frontMatter.slug ||
-      (dateFilenameMatch ? toUrl({date, link: linkName}) : linkName);
-
-    const permalink = normalizeUrl([baseUrl, routeBasePath, slug]);
-
-    function getBlogEditUrl() {
-      const blogPathRelative = path.relative(blogDirPath, path.resolve(source));
-
-      if (typeof editUrl === 'function') {
-        return editUrl({
-          blogDirPath: posixPath(path.relative(siteDir, blogDirPath)),
-          blogPath: posixPath(blogPathRelative),
-          permalink,
-          locale: i18n.currentLocale,
-        });
-      } else if (typeof editUrl === 'string') {
-        const isLocalized = blogDirPath === contentPaths.contentPathLocalized;
-        const fileContentPath =
-          isLocalized && options.editLocalizedFiles
-            ? contentPaths.contentPathLocalized
-            : contentPaths.contentPath;
-
-        const contentPathEditUrl = normalizeUrl([
-          editUrl,
-          posixPath(path.relative(siteDir, fileContentPath)),
-        ]);
-
-        return getEditUrl(blogPathRelative, contentPathEditUrl);
-      } else {
-        return undefined;
-      }
-    }
+export async function generateBlogPosts(
+  contentPaths: BlogContentPaths,
+  context: LoadContext,
+  options: PluginOptions,
+): Promise<BlogPost[]> {
+  const {include, exclude} = options;
 
-    blogPosts.push({
-      id: frontMatter.slug ?? title,
-      metadata: {
-        permalink,
-        editUrl: getBlogEditUrl(),
-        source: aliasedSource,
-        title,
-        description,
-        date,
-        formattedDate,
-        tags: frontMatter.tags ?? [],
-        readingTime: showReadingTime ? readingTime(content).minutes : undefined,
-        truncated: truncateMarker?.test(content) || false,
-      },
-    });
+  if (!fs.existsSync(contentPaths.contentPath)) {
+    return [];
   }
 
-  await Promise.all(
-    blogSourceFiles.map(async (blogSourceFile: string) => {
-      try {
-        return await processBlogSourceFile(blogSourceFile);
-      } catch (e) {
-        console.error(
-          chalk.red(
-            `Processing of blog source file failed for path "${blogSourceFile}"`,
-          ),
-        );
-        throw e;
-      }
-    }),
-  );
+  const blogSourceFiles = await Globby(include, {
+    cwd: contentPaths.contentPath,
+    ignore: exclude,
+  });
+
+  const authorsMap = await getAuthorsMap({
+    contentPaths,
+    authorsMapPath: options.authorsMapPath,
+  });
+
+  const blogPosts = (
+    await Promise.all(
+      blogSourceFiles.map(async (blogSourceFile: string) => {
+        try {
+          return await processBlogSourceFile(
+            blogSourceFile,
+            contentPaths,
+            context,
+            options,
+            authorsMap,
+          );
+        } catch (e) {
+          console.error(
+            chalk.red(
+              `Processing of blog source file failed for path "${blogSourceFile}"`,
+            ),
+          );
+          throw e;
+        }
+      }),
+    )
+  ).filter(Boolean) as BlogPost[];
 
   blogPosts.sort(
     (a, b) => b.metadata.date.getTime() - a.metadata.date.getTime(),

package/src/feed.ts

@@ -0,0 +1,129 @@
+/**
+ * Copyright (c) Facebook, Inc. and its affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+import {Feed, Author as FeedAuthor} from 'feed';
+import {PluginOptions, Author, BlogPost, FeedType} from './types';
+import {normalizeUrl, mdxToHtml} from '@docusaurus/utils';
+import {DocusaurusConfig} from '@docusaurus/types';
+import path from 'path';
+import fs from 'fs-extra';
+
+// TODO this is temporary until we handle mdxToHtml better
+// It's hard to convert reliably JSX/require calls  to an html feed content
+// See https://github.com/facebook/docusaurus/issues/5664
+function mdxToFeedContent(mdxContent: string): string | undefined {
+  try {
+    return mdxToHtml(mdxContent);
+  } catch (e) {
+    // TODO will we need a plugin option to configure how to handle such an error
+    // Swallow the error on purpose for now, until we understand better the problem space
+    return undefined;
+  }
+}
+
+export async function generateBlogFeed({
+  blogPosts,
+  options,
+  siteConfig,
+}: {
+  blogPosts: BlogPost[];
+  options: PluginOptions;
+  siteConfig: DocusaurusConfig;
+}): Promise<Feed | null> {
+  if (!blogPosts.length) {
+    return null;
+  }
+
+  const {feedOptions, routeBasePath} = options;
+  const {url: siteUrl, baseUrl, title, favicon} = siteConfig;
+  const blogBaseUrl = normalizeUrl([siteUrl, baseUrl, routeBasePath]);
+
+  const updated =
+    (blogPosts[0] && blogPosts[0].metadata.date) ||
+    new Date('2015-10-25T16:29:00.000-07:00'); // weird legacy magic date
+
+  const feed = new Feed({
+    id: blogBaseUrl,
+    title: feedOptions.title || `${title} Blog`,
+    updated,
+    language: feedOptions.language,
+    link: blogBaseUrl,
+    description: feedOptions.description || `${siteConfig.title} Blog`,
+    favicon: favicon ? normalizeUrl([siteUrl, baseUrl, favicon]) : undefined,
+    copyright: feedOptions.copyright,
+  });
+
+  function toFeedAuthor(author: Author): FeedAuthor {
+    // TODO ask author emails?
+    // RSS feed requires email to render authors
+    return {name: author.name, link: author.url};
+  }
+
+  blogPosts.forEach((post) => {
+    const {
+      id,
+      metadata: {title: metadataTitle, permalink, date, description, authors},
+    } = post;
+    feed.addItem({
+      title: metadataTitle,
+      id,
+      link: normalizeUrl([siteUrl, permalink]),
+      date,
+      description,
+      content: mdxToFeedContent(post.content),
+      author: authors.map(toFeedAuthor),
+    });
+  });
+
+  return feed;
+}
+
+async function createBlogFeedFile({
+  feed,
+  feedType,
+  filePath,
+}: {
+  feed: Feed;
+  feedType: FeedType;
+  filePath: string;
+}) {
+  const feedContent = feedType === 'rss' ? feed.rss2() : feed.atom1();
+  try {
+    await fs.outputFile(filePath, feedContent);
+  } catch (err) {
+    throw new Error(`Generating ${feedType} feed failed: ${err}.`);
+  }
+}
+
+export async function createBlogFeedFiles({
+  blogPosts,
+  options,
+  siteConfig,
+  outDir,
+}: {
+  blogPosts: BlogPost[];
+  options: PluginOptions;
+  siteConfig: DocusaurusConfig;
+  outDir: string;
+}): Promise<void> {
+  const feed = await generateBlogFeed({blogPosts, options, siteConfig});
+
+  const feedTypes = options.feedOptions.type;
+  if (!feed || !feedTypes) {
+    return;
+  }
+
+  await Promise.all(
+    feedTypes.map(async function (feedType) {
+      await createBlogFeedFile({
+        feed,
+        feedType,
+        filePath: path.join(outDir, options.routeBasePath, `${feedType}.xml`),
+      });
+    }),
+  );
+}