I've tried both approaches but using Paragraphs works better from a UX standpoint. I couldn't get the separate content type to work as seamlessly as intended. The user has to create one first before adding it to the Book content type.
With the Paragraphs approach, the user clicks the correct Paragraphs type (only one available) and can start entering the information.
Thanks for the help.
hepabolu → created an issue.
FYI I've created a similar tamper module: https://github.com/hepabolu/feeds_tamper_lookup_entity
This one works for me, but since this is my first Drupal module feel free to correct and improve or assimilate into your version.
FYI I have tested this:
- Drupal 10.3.10
- PHP 8.3.14
- Lando 3
- Feeds 3.0.0-rc3
- Feeds Tamper 2.0.0-beta4
- Content type Publication with field_cover = Image
- Media type Photos with field_image = Image and field_credits = Text
- Content type Model with field_photos = entity reference to media type Photos
- Feed import Publication, CSV upload, tamper format string to turn image filename into url
- Feed import Photos, CSV upload, tamper format string to turn image filename into url
- Feed import Model, CSV upload, figuring out how to set the reference to the imported Photos
Without any patch:
- the import of the image for Publication cover went fine
- the import of the Photos failed on a missing name, but the files were added anyway with the tampered name as 'filename', adding the Name field didn't help
With the patch from #105:
- I set the language from Default to English
- I explicitly set the Name field to a copy of the filename (in a separate CSV column)
Import of the media type Photos went fine, filename was correct.
I'll see if I can do more tests and report back.
Both approaches look promising. I'm going to try both and see what works best.