import {
  GlideClient,
  GlideClusterClient,
  GlideJson,
  GlideFt,
  GlideBf,
} from "@glidemq/speedkey";

// Standalone
const client = await GlideClient.createClient({
  addresses: [{ host: "localhost", port: 6379 }],
});

// Cluster
const cluster = await GlideClusterClient.createClient({
  addresses: [{ host: "127.0.0.1", port: 7000 }],
});

await GlideJson.set(client, "user:1", "$", '{"name":"Alice","age":30}');
const user = await GlideJson.get(client, "user:1", { path: "$" });

// Batch set
await GlideJson.mset(client, [
  { key: "user:2", path: "$", value: '{"name":"Bob"}' },
  { key: "user:3", path: "$", value: '{"name":"Carol"}' },
]);

// Create index with vector field (required by valkey-search)
await GlideFt.create(client, "docs-idx", [
  { type: "TAG", name: "category" },
  { type: "NUMERIC", name: "score" },
  {
    type: "VECTOR", name: "embedding",
    attributes: {
      algorithm: "HNSW",
      dimensions: 384,
      distanceMetric: "COSINE",
      type: "FLOAT32",
    },
  },
], { dataType: "HASH", prefixes: ["doc:"] });

// Store document with embedding
const embedding = new Float32Array(384); // from your embedding model
await client.hset("doc:1", {
  category: "ml",
  score: "95",
  embedding: Buffer.from(embedding.buffer),
});

// KNN search with pre-filter
const [total, results] = await GlideFt.search(
  client, "docs-idx",
  "@category:{ml}=>[KNN 5 @embedding $VEC AS dist]",
  { params: [{ key: "VEC", value: Buffer.from(embedding.buffer) }] },
);

await GlideBf.reserve(client, "emails", 0.001, 100000);
await GlideBf.add(client, "emails", "alice@example.com");

const exists = await GlideBf.exists(client, "emails", "alice@example.com"); // true
const missing = await GlideBf.exists(client, "emails", "unknown@x.com");   // false

const info = await GlideBf.info(client, "emails");
// { capacity: 100000, size: ..., numberOfFilters: 1, numberOfItems: 1, expansionRate: 2 }