I have node with a field for paragraphs. These paragraphs can be published / unpublished so editors can add content without directly publishing that part.
The Search API has been configured with a Solr backend. The index is configured to use nodes as a datasource (paragraphs are not used as a datasource).
The Highlight processor is enabled to always highlight returned field data, highlight partial matches and create excerpt. I've built a custom processor that excludes unpublished paragraphs from the index so that the contents can't be found in Solr.
However, the excerpt that is generated includes the unpublished content. The problem occurs when you search for content in a published paragraph that preceeds an unpublished paragraph.
- Create a paragraph type with a title / body field.
- Create a node type with a field that contains paragraphs of said type.
- The paragraphs can be unpublished in the node-edit form.
- Set up search API to index nodes, add the paragraph fields to the index.
- Highlighting enabled, always highlight returned data, partial matches, create excerpt with a length of 256.
- Create a node with a published paragraph (A) and an unpublished paragraph (B).
- Search for the content of paragraph A, preferably a piece of text at the end of the content .
- The result is that in the excerpt the contents of paragraph B is visible.
The "extractItemValues" from the FieldsHelper triggers $item->getOriginalObject(). This doesn't respect the published status of the entity. Maybe it should? I'll leave that up to you :-)
Active
1.0
General code
Thanks for reporting this problem.
Does the patch in
✨
Allow excluding unpublished references from index
Active
help? If you have the “Entity status” processor enabled the code in that issue’s MR might already resolve this for you.