Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Change search result ranking #3194

Open
kineticsquid opened this issue Dec 4, 2024 · 5 comments
Open

Change search result ranking #3194

kineticsquid opened this issue Dec 4, 2024 · 5 comments

Comments

@kineticsquid
Copy link
Contributor

I recently received a complaint about an extension that was misleading. It was a republication of a popular extension, sight changes to package.json, but otheriwse no changes to name, displayname or any of the links. And because it was new, it was showing up first in the search results, ahead of the much more popular extension. I'm recommending we change our search ranking to more heavily, or completely, weight on download count.

@filiptronicek @spoenemann @amvanbaren Thoughts?

@filiptronicek
Copy link
Member

@amvanbaren did some changes to sorting last year in eclipse/openvsx#672, but IIRC that was just for sorting by "Rating". I'm not sure what we sort by now by default, but download count should definitely be one of the main factors we take a look at (besides probably string distance).

@amvanbaren
Copy link
Contributor

By default it sorts by relevance, which is a combination of multiple weighting factors: https://github.com/eclipse/openvsx/blob/072bf3e7fa38eef217ebb30d9748f5556440c665/server/src/main/java/org/eclipse/openvsx/search/RelevanceService.java#L39-L48

This is how the relevance rating is calculated: https://github.com/eclipse/openvsx/blob/072bf3e7fa38eef217ebb30d9748f5556440c665/server/src/main/java/org/eclipse/openvsx/search/RelevanceService.java#L73-L109

The timestamp factor is the leading relevance factor for the open-vsx.org deployment:

search:
relevance:
rating: 0.2
downloads: 1.0
timestamp: 3.0

@kineticsquid
Copy link
Contributor Author

IMO downloads should be most heavily weighted. Not sure of the values though.

@filiptronicek
Copy link
Member

I'd propose that we switch the open-vsx.org deployment to the following:

 rating: 0.5
 downloads: 2.0
 timestamp: 1.2

This is as arbitrary as it gets, but it feels like we should make ratings and downloads the primary metrics and only have the timestamp secondary, since besides indicating that an extension is more "up-to-date", it does not signal much.

@kineticsquid
Copy link
Contributor Author

Here's a PR with those changes: #3206

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants