The front search lucene results sometimes aren't relevant.
Searching an unusual word:
- my first result is a page which only contains the word juste once in a content attachment (score of 0.32)
- the next result is a page which contains the word twice in the main content (score of 0.16)
The lucene query:
+(title:myword subtitle:myword subject:myword...) +(content-types:article content-types:faq type:resource...)
If I explain the results in Luke, the 0.32 score comes from the page having 2 article contents and a FAQ content, and not from the word search.
We should weigh the "content types" query part with 0 weight, the results would be more relevant:
+(title:myword subtitle:myword subject:myword...) +(content-types:article content-types:faq type:resource...)^0