Split out HTML processing from markdown conversion for API generation (…

…#614)

This splits out all of the HTML helper functions from `htmlToMd.ts` into
`processHtml.ts`. `processHmtl.ts` is where we have all of our Cheerio
code that deals with HMTL -> HTML, whereas `htmlToMd.ts` goes from HTML
-> MD.

This PR avoids making refactors to the source code to keep this easier
to review. A follow up PR will add tests.

scripts/lib/api/HtmlToMdResult.ts

@@ -12,10 +12,15 @@
 
 import { Metadata } from "./Metadata";
 
+export type Image = {
+  src: string;
+  dest: string;
+};
+
 export type HtmlToMdResult = {
   markdown: string;
   meta: Metadata;
-  images: Array<{ src: string; dest: string }>;
+  images: Image[];
   isReleaseNotes: boolean;
 };