What causes the content-name-unstemmed weight in Confluence search results?

[Rick van Twillert (TMC)]

Hi,

We are researching what causes some pages to rank higher than others in the Confluence search and noticed that we have a page which doesnt have the following weight in its sum when using the explain.action:

13.393457 = (MATCH) weight(content-name-unstemmed:door in 216674) [BM25LSimilarity], result of:

If it would have this weight, the page would rank on top. So the question is, what is content-name-unstemmed and how can we "add" it to a page.

When looking at what influence the score according to the Atlassian KB article, the page that doesnt have the content-name-unstemmed weight, should rank on top because of it's Document Age, Title, Content and Popularity.

Thanks in advance!

[Stephen Sifers]

Hello Rick,

I’ve read through your very descript needs for added content-name-unstemmed into your pages for advance your search functionality and weight. I wasn’t able to find anything which explicitly shows how to use this however, I did find a developer doc which does reference content-name-unstemmed. You can find this documentation at LuceneContentNameSearcher. Within here there is some explanation about this field.

I hope this reference proves helpful and assists with improving your search results.

Regards,
Stephen Sifers

[Jon P]

This is old I know, but I was looking for this info too and recently found this page which refers to: content-name-unstemmed

https://developer.atlassian.com/server/confluence/extractor2-module-tutorial/

which may help. It appears it is normally a variation of the page title:

resultBuilder.add(new TextFieldDescriptor("content-name-unstemmed", page.getTitle(), Store.YES); 

possibly a non-encoded version or similar