xb

Vector Quickstart (English)

English summary of xb/doc/VECTOR_QUICKSTART.md. It focuses on starting a vector project with xb + Qdrant/pgvector.

---

Setup checklist

1. Define DTOs (struct CodeVector { ... }).
2. Create embeddings (OpenAI, local model, etc.).
3. Store vectors + metadata in your database of choice.
4. Use xb to build both SQL and vector queries.

---

Example workflow

queryVec := embedder.Encode(prompt)

json, _ := xb.Of(&CodeVector{}).
    Custom(xb.NewQdrantBuilder().Build()).
    Eq("tenant_id", tenant).
    VectorSearch("embedding", queryVec, 10).
    Build().
    JsonOfSelect()

For pgvector:

sql, args, _ := xb.Of(&CodeVector{}).
    Eq("tenant_id", tenant).
    X("embedding <#> ?", queryVec). // custom operator
    Limit(10).
    Build().
    SqlOfSelect()

---

Tips

• Normalize vectors upfront.
• Store timestamps and durability metadata to support re-indexing.
• Add regression tests to ensure vector + scalar filters compose correctly.

---

Related docs

• doc/en/VECTOR_GUIDE.md
• doc/en/QDRANT_GUIDE.md

This site is open source. Improve this page.