Using split text & replace text nodes is the quite dirty way I guess, maybe possible creating an XML data type from the html and parse it using the tree nodes, but you're more skilled than me from some previous discussions of yours regarding that (I think there is a feature request also to parse HTML into a list).

Here is a starting composition that retrieves tweets from the #NationalDoctorsDay twitter URL and display them changing every x seconds. I used the "live" tweets URL (https://twitter.com/search?q=%23NationalDoctorsDay&src=trend_click&f=live) but for some reasons it's still the trending ones that seem to be loaded. Maybe creating a wrong list using wrong splitters (perhaps live tweets don't start with "data-aria-label-part="0" or I cut the live tweets HTML code part, or it's cached somewhere somehow, to be checked).

It converts html characters to text & turn hashtag codes into hashtags, but not sure if this works if several hashtags are used in a tweet, and doesn't remove general links or picture links.
Maybe using the same technique you can also retrieve the @ of who sent the tweet.

Hope this helps.

PS why don’t they postpone the wedding ?