thejayman77/upbeatBytes
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Use BBC's clean image variant (cpsprodpb) instead of the branded one

Browse Source 
BBC's og:image comes from the "branded_news" CDN path with a "BBC NEWS" logo
baked into the picture (shows as "…EWS" once the hero crops it). The identical
photo is served under "cpsprodpb" with no logo, so rewrite branded_news →
cpsprodpb. Best of both: full-resolution hero, no burned-in branding. Re-enriched
recent briefs so live images swap over. 99 tests pass.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
jay
2026-06-02 07:51:51 +00:00
parent 2145622b59
commit acbc06a9e5
2 changed files with 18 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_enrich.py
+3 -2
View File
@@ -77,8 +77,9 @@ def test_rejects_generic_share_images():
assert og_image_from_html(b'<meta property="og:image" content="https://x.com/og-default.jpg">') is None
# a real article image comes through
assert og_image_from_html(b'<meta property="og:image" content="https://x.com/real-photo.jpg">') == "https://x.com/real-photo.jpg"
# BBC's branded_news path is a REAL photo path, not a placeholder — keep it
assert og_image_from_html(b'<meta property="og:image" content="https://ichef.bbci.co.uk/news/1024/branded_news/x.jpg">') == "https://ichef.bbci.co.uk/news/1024/branded_news/x.jpg"
# BBC's branded_news path is a real photo (logo baked in) — keep it, but swap
# to the clean cpsprodpb variant so the hero isn't branded
assert og_image_from_html(b'<meta property="og:image" content="https://ichef.bbci.co.uk/news/1024/branded_news/x.jpg">') == "https://ichef.bbci.co.uk/news/1024/cpsprodpb/x.jpg"
# if the first og is a generic placeholder but a later one is real, take the real one
html = (b'<meta property="og:image" content="https://x.com/og-default.jpg">'
b'<meta property="og:image" content="https://x.com/article.jpg">')