fix(title): pick best <title> when a page has multiple (Medium fix)

[rebelchris]

Problem

Recently-scraped Medium articles are being stored with the title Medium.

Medium serves two <title> tags on article pages — the real one (… | by Author | … | Medium) and a bare <title>Medium</title>. In the non-JS HTML our scraper (post-scraper-one-ai / yggdrasil via this newspaper fork) receives, the bare tag can appear first, and get_title() blindly used title_element[0] → Medium.

Fix

Instead of always taking the first <title>, choose the best candidate among all <title> tags:

1. prefer the one whose filtered text contains og:title (the descriptive article title), else
2. the longest candidate.

The existing |/- splitter then trims the | by … | Medium suffix as before. Single-<title> pages are unaffected.

Verification

Ran against the real Medium article HTML (both tag orderings) + a normal single-title page:

as-fetched (real first)    -> 'High performance has a human requirement'
medium-first (worst case)  -> 'High performance has a human requirement'
single-title sanity        -> 'My Great Post'

Notes / follow-up

This handles the reachable-article case. Genuinely walled/404 Medium pages still only contain <title>Medium</title> (no real title tag), so a defensive guard in post-scraper-one-ai to reject a final title of Medium/Just a moment... is still worth adding separately.