[beebutler]

Summary

This change automatically populates the Publisher field using the source website's hostname when og:site_name metadata is missing or empty.

Existing publisher metadata is preserved, and the field remains editable.

Related to #3132.

Changes

• Preserve the website's existing og:site_name when available.
• Fall back to the source hostname when publisher metadata is missing or blank.
• Handle non-HTTP(S) URLs safely.
• Add four regression tests covering these scenarios.

Testing

• npm run lint — passed.
• QUnit — 695 assertions passed, 0 failed.
• Verified the fallback in Chrome using controlled test pages.
• Generated EPUBs and confirmed the Publisher metadata in Calibre.
• Successfully retested after restarting Chrome.

Additional context

I originally encountered this issue while experimenting with a locally adapted parser for an independent archive running AO3/OTW-based software during the ongoing AO3 outage.

I deliberately haven't requested official support for that archive. Its volunteer administrator has been dealing with harassment and server-load concerns, and I don't want to bring additional attention or traffic to the site without their approval.

That experimentation led me to investigate how WebToEpub handles Publisher metadata when og:site_name is unavailable.

This pull request addresses the general metadata issue and does not include the custom parser.

Thank you for considering this contribution.

[dteviot]

@beebutler
Thank you for your contribution.
You missed adding yourself to the credits, but I've taken care of that.