Vector Index & Semantic Search Page

[ElioNeto]

Vector Index & Semantic Search Page

Create an Angular page for managing vector embeddings and performing semantic search using the existing VectorIndex.

Backend Endpoints Needed

GET    /admin/vectors                     — List vector index stats (dimension, count)
POST   /admin/vectors/query               — Search by embedding vector
POST   /admin/vectors/search              — Search by text (backend generates embedding)
GET    /admin/vectors/{key}               — Get embedding for a key
DELETE /admin/vectors/{key}               — Remove a key from the index
POST   /admin/vectors/ingest              — Batch ingest vectors from CSV/JSON

UI Requirements

Vector Index Overview

• Stats: Dimension, vector count, memory usage
• Index Health: Number of vectors, any anomalies

Semantic Search

• Query Input: Text input for natural language query
• Results List: Ranked results with similarity scores
• Key Display: Show matched key + preview of associated data
• Result Count: Configurable top-K (slider 1-100)

Vector Management

• Ingest Form: Upload embeddings file (CSV with key,vector columns)
• Vector Browser: Browse stored vectors with key search
• Dimension Check: Warn on dimension mismatch

Component Structure

app/
  pages/
    vector-search/
      vector-search.component.ts
      vector-search.component.html
      vector-search.component.scss

Acceptance Criteria

• Vector index stats displayed
• Text search returns ranked results
• Results show similarity scores
• Top-K slider works
• Loading and empty states handled
• Responsive layout

Parent Epic

#290

[ElioNeto]

Resolved via feat/notes-engine-v2 branch (PR #321)