feat(v2): various markdown string parsing improvements/fixes (#4590)

* extract createExcerpt code in separate file + add bad test

* almost working markdown parsing refactor

* complete parseMarkdownString refactor

* fix tests

* fix blog test issue

* fix docusaurus utils imports

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

@@ -26,9 +26,10 @@ import {
   getEditUrl,
   getFolderContainingFile,
   posixPath,
+  replaceMarkdownLinks,
 } from '@docusaurus/utils';
 import {LoadContext} from '@docusaurus/types';
-import {replaceMarkdownLinks} from '@docusaurus/utils/lib/markdownLinks';
+import {assertBlogPostFrontMatter} from './blogFrontMatter';
 
 export function truncate(fileString: string, truncateMarker: RegExp): string {
   return fileString.split(truncateMarker, 1).shift()!;
@@ -140,12 +141,18 @@ export async function generateBlogPosts(
 
       const source = path.join(blogDirPath, blogSourceFile);
 
+      const {
+        frontMatter,
+        content,
+        contentTitle,
+        excerpt,
+      } = await parseMarkdownFile(source);
+      assertBlogPostFrontMatter(frontMatter);
+
       const aliasedSource = aliasedSitePath(source, siteDir);
 
       const blogFileName = path.basename(blogSourceFile);
 
-      const {frontMatter, content, excerpt} = await parseMarkdownFile(source);
-
       if (frontMatter.draft && process.env.NODE_ENV === 'production') {
         return;
       }
@@ -182,9 +189,11 @@ export async function generateBlogPosts(
         year: 'numeric',
       }).format(date);
 
+      const title = frontMatter.title ?? contentTitle ?? linkName;
+      const description = frontMatter.description ?? excerpt ?? '';
+
       const slug =
         frontMatter.slug || (match ? toUrl({date, link: linkName}) : linkName);
-      frontMatter.title = frontMatter.title || linkName;
 
       const permalink = normalizeUrl([baseUrl, routeBasePath, slug]);
 
@@ -220,16 +229,16 @@ export async function generateBlogPosts(
       }
 
       blogPosts.push({
-        id: frontMatter.slug || frontMatter.title,
+        id: frontMatter.slug ?? title,
         metadata: {
           permalink,
           editUrl: getBlogEditUrl(),
           source: aliasedSource,
-          description: frontMatter.description || excerpt,
+          title,
+          description,
           date,
           formattedDate,
-          tags: frontMatter.tags,
-          title: frontMatter.title,
+          tags: frontMatter.tags ?? [],
           readingTime: showReadingTime
             ? readingTime(content).minutes
             : undefined,