Some images are also referenced by plugin files, e.g. the WooCommerce placeholder image.
Hi Gal,
Most post meta scanning is already covered:
- Generic post meta scan walks all postmeta arrays/serialized data
looking for attachment ID and URL keys (handles Rank Math Schema
video thumbnails, custom plugin meta, etc.)
- Dedicated parsers for Yoast SEO, Rank Math, AIOSEO, and SEOPress
for their OpenGraph / Twitter image fields
- WooCommerce placeholder image is also detected via the options scan
Theme template files with hardcoded image paths in PHP cannot be
detected reliably, that’s a fundamental limitation of database-only
scanning and is documented under “Important Precautions” in the
readme.
If a specific Rank Math Schema image isn’t being picked up after a
full scan, please share the meta key (or a screenshot of the field)
and I’ll verify it’s matched by the parser.
Regards,
The video thumbnail meta key is rank_math_schema_VideoObject. The meta value is an array containing absolute URLs for the images.
Some files are listed (file name only) in _mfrh_history, but this is just for the Media File Renamer plugin and should probably be skipped.
The key _original_filename can probably be ignored too.
There’s a file being referenced (absolute URL) in a theme’s parts/sidebar.html file on a site that uses the Site Editor and loads all of the template files from disk. This image (youtube-channel.jpg) is loaded on many pages (like this one) in the sidebar, but is listed as unused.
On the same site, there are custom meta fields attached to products and product variations, which contain image references (absolute URLs). I really need those to NOT be deleted 😊
It’ll be good to provide some way to add meta key patterns to the scanner, then, to cater for custom fields. The plugin code should be able to handle the scan already, so this just needs the settings on the admin page.
Side note: some images can have the same file name, but a different path. I’ve found some of those and had to dig a little to realise what was going on. It may be good to list the full path in the report, and even (ideally), indicate that the file name has other matches. Sometimes, images are uploade more than once on different months, and duplicates need cleaning with confidence.
This seems to cover it for me, but once implemeted, most of the images will disappear for me, and we might discover some more.
