Can I vibe code GroupBy?

Build an open-source, high-performance hybrid ecommerce search microservice using Node.js (TypeScript), Typesense, and PostgreSQL.

1. ARCHITECTURE & DATA MODEL
- Connect to PostgreSQL for catalog storage (products, variants, categories, price, stock, metadata).
- Use Typesense as the primary search index.
- Define a Product document structure in Typesense containing ID, title, description, category, brand, price, dynamic attributes (key-value strings), vector embeddings (array of floats), and calculated popularity score (float).

2. CORE FUNCTIONALITY
- Catalog Ingestion Sync: Create a webhooks listener endpoint `/api/v1/sync` that accepts SKU updates, formats fields into JSON, generates dense vector embeddings using OpenAI's `text-embedding-3-small` for title + description, and upserts them into Typesense.
- Hybrid Search Endpoint: Build `/api/v1/search` accepting query text, filter parameters (category, price range, dynamic attributes), sort rules, page, and limit.
- Combine BM25 keyword matching with vector semantic search using Typesense multi-search hybrid scoring.
- Merchandising Engine: Implement a override rules processor. Before returning results, execute logic that applies:
  a. Query redirects (e.g., query 'shipping' redirects to custom URL).
  b. Pinning (force specific Product IDs to positions 1..N for defined search queries).
  c. Burying/Boosting (apply multipliers to numerical scores based on attribute matches like brand='Nike').
- Dynamic Faceting: Automatically generate aggregation facets for category, brand, price ranges, and custom attributes based on matching candidate sets.

3. ANALYTICS & RE-RANKING FEED
- Expose an endpoint `/api/v1/telemetry/event` to track `search_click` and `conversion` events with query string and product ID.
- Aggregate daily click-through rates (CTR) and update the `popularity_score` field in Typesense to boost conversion-heavy SKUs natively.

4. FAILURE MODES & LATENCY REQUIREMENTS
- Total search query processing time must execute under 40ms.
- If vector embedding API call fails or times out during query time, fallback seamlessly to pure keyword BM25 search.
- Handle out-of-stock products by auto-applying a score penalty instead of omitting them entirely, unless explicitly requested via filter.

5. OUT OF SCOPE
- Complex enterprise authentication SAML/SSO.
- Multi-tenant multi-store visual admin dashboards.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs