A simple IDB vector store.
<script type="module">
import { VectorDB } from "idb-vector";
const db = new VectorDB({
vectorPath: "embedding"
});
await db.insert({ embedding: [1, 2, 3], "text": "ASDASINDASDASZd" });
await db.insert({ embedding: [2, 3, 4], "text": "GTFSDGRG" });
await db.insert({ embedding: [73, -213, 3], "text": "hYTRTERFR" });
// Query returns a list ordered by the entries closest to the vector (cosine similarity)
console.log(await db.query([1, 2, 3], { limit: 20 }));
</script>