Fix for article rich metadata not being populated

[KonradIT]

• sharedHttpClient has no User-Agent header, fixed it, as well as fix for content-type. Use chrome user agent instead, as well as use Accept: text/html header to get websites to send over the og:xxxxx tags
• Use Bluesky's existing infra to populate rich metadata: https://cardyb.bsky.app/v1/extract?url=$website returns a JSON formatted response with the data, this is what bluesky-social uses. Failover to existing logic.

Here is the app populating rich metadata previews after my fix is applied:

[0xZhangKe]

Looks like you’ve contributed a lot of changes, really appreciate it. I’ve been pretty busy lately, but I’ll make some time to review and merge them.