feat(www): ai-search #4106
feat(www): ai-search #4106
Conversation
… hybrid search, fix types, add tests
…-practices (w/ lang support)
- Fix hardcoded index name in setup-embedder.mjs to use MEILISEARCH_PROJECT_NAME - Enables proper embedder registration for semantic search - Resolves vector search failure that caused RAG to fall back to keyword search
- Add check-embedder-status.mjs to validate embedder configuration and test vector search - Add debug-search.mjs to inspect index content and test different query types - These tools help troubleshoot RAG pipeline issues and verify semantic search functionality
|
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| 'Content-Type': 'application/json', | ||
| 'api-key': env.AZURE_KEY, | ||
| }, | ||
| body: JSON.stringify({ input: text }), |
Check warning
Code scanning / CodeQL
File data in outbound network request Medium
There was a problem hiding this comment.
If this server only functions as an API, you can build that directly into react router.
In loaders, you can run server code, and return responses from it.
You can also make sure that code only runs on the server by using server modules, see: https://reactrouter.com/api/framework-conventions/server-modules
You can then, in the loader for the API route, use data() to send what you need, see: https://reactrouter.com/api/utils/data
It would be nice to not have to spin up a separate Azure deployment.
If you need any env vars, simply add them to the Dockerfile and azure action for www
|
Preview deployments for this pull request: storybook - themebuilder - www - |
…ensitive information Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
…ensitive information Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
Resolves #3618