Skip to main content

search

Searches for vectors similar to a given vector or content string.

search(
    data: string | number[],
    options?: {
      topK?: number
      measurement?: Measurement
      filters?: Record<string, any>
      includeVector?: boolean
      includeMetadata?: boolean
    }
): Promise<SearchResponse[]>

Reference​

import { myVectorStore } from "#elements";
export default async function () {
  const results = await myVectorStore.search("content");
  console.log(JSON.stringify(results, null, 2));
}

Parameters​

  • data: The vector or content string to search for.
  • options: Optional configuration parameters, including:
    • topK: (optional) The number of results to return, defaults to 10 and capped at 100.
    • measurement: (optional) The measurement method for semantic similarity, defaults to euclidean.
    • includeVector: (optional) Whether to include the vector in the result, defaults to false.
    • includeMetadata: (optional) Whether to include the metadata in the result, defaults to true.
    • filters: (optional) Filters to apply to the search.

Returns​

Promise of search results, each with an ID, a distance to the given content, and optional vector and metadata.

  • id: The ID of the result.
  • distance: The distance to the given content.
  • vector: (Optional) The vector of the result.
  • metadata: (Optional) The metadata of the result.

Caveats​

  • The returned data is sorted by distance in ascending order. The smaller the distance, the more similar the vector.

Examples​

In the HTTP Element, search data in the VectorStore Element that matches the key-value pairs included in filters:

import * as Koa from "koa";
import { myVectorStore } from "#elements";

export default async function (
  request: Koa.Request,
  response: Koa.Response,
  ctx: Koa.Context
) {
  const result = await myVectorStore.search("what is babel?", {
    topK: 3,
    filters: {
      source: "https://docs.babel.cloud/docs/overview",
    },
  });
  return result;
}