fix(api): search vectorstore using only last message

[alekst23]

Description

Adjust the RAG implementation to search the vector store using the last user message instead of the first.

Add a needle in a haystack test to confirm problem and fix.

Apply fix for indexing already existing files to return the found file instead of 500.

[netlify]

✅ Deploy Preview for leapfrogai-docs ready!

Name	Link
🔨 Latest commit	9e31599
🔍 Latest deploy log	https://app.netlify.com/sites/leapfrogai-docs/deploys/66d8b1e3b89b060008bae5df
😎 Deploy Preview	https://deploy-preview-939--leapfrogai-docs.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.
Lighthouse	1 paths audited Performance: 43 (🟢 up 2 from production) Accessibility: 98 (no change from production) Best Practices: 100 (no change from production) SEO: 92 (no change from production) PWA: - View the detailed breakdown and full score reports

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[alekst23]

There's one issue to call out here, which is the changes to src/leapfrogai_api/backend/rag/index.py. I was encountering an error during testing where the file to be attached to an assistant was already in the vector store. In this case, the endpoint was returning error 500. I don't have an OpenAI conformance test for this yet, but it did not have the same problem during testing.