openshift / openshift/lightspeed-rag-content
Support metadata based filtering
Nobody has claimed this yet.
- Dominant language
- HTML
- Stars
- 19
- Forks
- 38
- Avg merge
- 17h 57m
- Merged PRs (30d)
- 4
Description
Description
We would like to confirm whether Lightspeed supports metadata-based filtering during RAG document retrieval.
The documentation appears to support metadata fields such as title and url, but it is unclear whether these fields—or custom metadata fields—can be used to filter documents before the retrieved context is sent to the model.
Use case
We are building a Lightspeed “bring your own knowledge” RAG image for GitOps documentation.
The available documentation can vary significantly between GitOps versions, including:
- Configuration differences
- Feature differences
- Newly introduced features
- Deprecated or removed functionality
When a RAG image is built using only the latest documentation, the model may return inaccurate or unhelpful answers for customers who are using an older GitOps version.
We would like to attach metadata to each document or chunk, for example:
{
"version": "1.15",
"priority": 10,
"title": "GitOps installation configuration",
"url": "https://example.com/docs"
}
So that at query time, we can filter the content based on metadata such as the GitOps version.
We would like to maintain a single RAG image containing documentation for multiple GitOps versions and retrieve only the content relevant to the version used.
Without metadata based filtering, we may need to maintain separate RAG images and configurations for every supported GitOps version, which would significantly increase maintenance and deployment complexity.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue does not name a file, test, or entry point. First confirm whether Lightspeed supports filtering on built-in or custom metadata during retrieval, then document the supported behavior or define the needed feature. Done should address version-based filtering for a single RAG image.
Written by the indexing model from the issue text.
Assessment
- Domain
- data, search
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100