feature: improve website search functionality
This is a Feature Request
The current Kubeflow website uses a Google Custom Search Engine implementation, which has limitations in terms of customization, relevance, and user experience. While this provides basic functionality, it has several limitations:
- Users are redirected away from the Kubeflow site to view search results.
- The relevancy and filtering of results are limited.
- There is no in-site integration for highlighting, faceting, or advanced query capabilities.
What would you like to be added
To enhance the search experience for users, I propose migrating to a modern search solution like Algolia or Lunr.
Proposed options:
-
Algolia DocSearch: It a free, hosted search solution designed specifically for technical documentation and open-source projects. It crawls the site, indexes content, and delivers a fast, typo-tolerant search experience with minimal setup.
- Sites with this implemented: https://docs.cilium.io/en/stable/, https://docs.litmuschaos.io/
-
Lunr: It is a lightweight, open-source, client-side search library that enables offline-capable, self-hosted search functionality. It’s ideal for static websites and projects prioritizing privacy, control, and zero external dependencies.
- Sites with this implemented: https://kubectl.docs.kubernetes.io/
Labels
/area website
Comments
Love this idea? Give it a 👍.
more options: https://gohugo.io/tools/search/
Sounds good.. we can also implement a custom retrieval based searching algo. Sounds complex but its simple in real-life. Considering the now not so massive but elaborate docs, we can actually do this.
I am not sure we have agreed on a specific implementation @SanthoshToorpu so I am not sure what you are assigning yourself to.
cc @kubeflow/kubeflow-steering-committee for discussion on if google search is sufficient, and if not so what alternative we will adopt.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
This issue has been automatically closed because it has not had recent activity. Please comment "/reopen" to reopen it.