microlinkhq / metascraper / 32051197496

95%

push

github

fix(author): fall back to twitter:creator (#874)

* fix(author): skip UI chrome, keep real bylines

Print/share buttons and Elsevier rights links were winning
author. Reject that chrome, read twitter:creator, and collect

Co-authored-by: Cursor <cursoragent@cursor.com>
#author-group names when citation_author is missing.

* fix(citation): collect given-name/surname pairs

Drop the Elsevier #author-group / xocs selectors. JATS
.given-name + .surname is what the markup actually shares.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(author): drop UI-chrome denylist

Site-specific Print/rights strings do not belong in isAuthor.
twitter:creator and given-name/surname already cover the cases.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(author): keep only standard author markup

Drop given-name/surname HTML scraping. Sites that omit
citation_author / schema.org / twitter:creator are not
patched with publisher CSS.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(author): drop toAuthorHandle

twitter: creator is always a handle; default author() already strips @
Co-authored-by: Cursor <cursoragent@cursor.com>

* test(author): update snapshots for twitter:creator

Co-authored-by: Cursor <cursoragent@cursor.com>

* test: omit flaky logo from silicon-angle snapshot

Co-authored-by: Cursor <cursoragent@cursor.com>

* test: assert silicon-angle logo is a URL

Co-authored-by: Cursor <cursoragent@cursor.com>

* test: validate destructured logos with url()

Co-authored-by: Cursor <cursoragent@cursor.com>

* test: allow null logo on jewish-business-news

logo-favicon only accepts ico/png; this fixture's icons are jpg

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(logo): read img inside itemprop=logo

Pin jewish-business-news to the HTML logo instead of a
network-dependent favicon fallback.

Co-authored-by: Cursor <cursoragent@cursor.com>

* test: pin HTML-only logos, drop favicon network

logo-favicon reachability is not determinis... (continued)

720 of 773 branches covered (93.14%)

Branch coverage included in aggregate %.

2446 of 2596 relevant lines covered (94.22%)

47.88 hits per line