Search a vector store

Search a vector store for relevant chunks based on a query and file attributes filter.

Request

POST

https://ai.api.cloud.yandex.net/v1/vector_stores/{vector_store_id}/search

Path parameters

Name

Description

vector_store_id

Type: string

The ID of the vector store to search.

Example: ``

Body

application/json
{
          "query": "example",
          "rewrite_query": false,
          "max_num_results": 10,
          "filters": {
            "type": "eq",
            "key": "example",
            "value": "example"
          },
          "ranking_options": {
            "ranker": "auto",
            "score_threshold": 0
          }
        }

Name

Description

query
One of 2 types

  • Type: string

    Example: example

  • Type: string[]

    Example
    [
          "example"
        ]

A query string for a search

Example: example

filters
One of 2 types

  • Comparison Filter

    Type: ComparisonFilter

    A filter used to compare a specified attribute key to a given value using a defined comparison operation.

    Example
    {
          "type": "eq",
          "key": "example",
          "value": "example"
        }

  • Compound Filter

    Type: CompoundFilter

    Combine multiple filters using and or or.

    Example
    {
          "type": "and",
          "filters": [
            {
              "type": "eq",
              "key": "example",
              "value": "example"
            }
          ]
        }

A filter to apply based on file attributes.

Example
{
          "type": "eq",
          "key": "example",
          "value": "example"
        }

max_num_results

Type: integer

The maximum number of results to return. This number should be between 1 and 50 inclusive.

Default: 10

Min value: 1

Max value: 50

ranking_options
Type: object

ranker

Type: string

Enable re-ranking; set to none to disable, which can help reduce latency.

Default: auto

Enum: none, auto, default-2024-11-15

score_threshold

Type: number

Default: 0

Min value: 0

Max value: 1

Ranking options for search.

Example
{
          "ranker": "auto",
          "score_threshold": 0
        }

rewrite_query

Type: boolean

Whether to rewrite the natural language query for vector search.

Default: false

ComparisonFilter

A filter used to compare a specified attribute key to a given value using a defined comparison operation.

Name

Description

key

Type: string

The key to compare against the value.

Example: example

type

Type: string

Specifies the comparison operator:

  • eq: equals - ne: not equal - gt: greater than - gte: greater than or equal - lt: less than - lte: less than or equal - in: in - nin: not in

Default: eq

Enum: eq, ne, gt, gte, lt, lte, in, nin

value
One of 4 types

  • Type: string

    Example: example

  • Type: number

  • Type: boolean

  • Type: array
    One of 2 types

    • Type: string

      Example: example

    • Type: number

    Example
    [
          "example"
        ]

The value to compare against the attribute key; supports string, number, or boolean types.

Example: example
Example
{
          "type": "eq",
          "key": "example",
          "value": "example"
        }

CompoundFilter

Combine multiple filters using and or or.

Name

Description

filters
Type: array
One of 2 types

  • Comparison Filter

    Type: ComparisonFilter

    A filter used to compare a specified attribute key to a given value using a defined comparison operation.

    Example
    {
          "type": "eq",
          "key": "example",
          "value": "example"
        }

  • Type: unknown

    Example: null

Array of filters to combine. Items can be ComparisonFilter or CompoundFilter.

Example
[
          {
            "type": "eq",
            "key": "example",
            "value": "example"
          }
        ]

type

Type: string

Type of operation: and or or.

Enum: and, or
Example
{
          "type": "and",
          "filters": [
            {
              "type": "eq",
              "key": "example",
              "value": "example"
            }
          ]
        }

Responses

200 OK

OK

Body

application/json
{
          "object": "vector_store.search_results.page",
          "search_query": [
            "example"
          ],
          "data": [
            {
              "file_id": "example",
              "filename": "example",
              "score": 0,
              "attributes": null,
              "content": [
                {
                  "type": "text",
                  "text": "example"
                }
              ]
            }
          ],
          "has_more": true,
          "next_page": "example"
        }

Name

Description

data

Type: VectorStoreSearchResultItem[]

The list of search result items.

Example
[
          {
            "file_id": "example",
            "filename": "example",
            "score": 0,
            "attributes": null,
            "content": [
              {
                "type": "text",
                "text": "example"
              }
            ]
          }
        ]

has_more

Type: boolean

Indicates if there are more results to fetch.

next_page
Any of 2 types

  • Type: string

    The token for the next page, if any.

    Example: example

  • Type: null

    Example: null

Example: example

object

Type: string

The object type, which is always vector_store.search_results.page

Const: vector_store.search_results.page

Example: example

search_query

Type: string[]

Example
[
          "example"
        ]

VectorStoreFileAttributes

Any of 2 types
  • Type: object

    [additional]
    One of 3 types

    • Type: string

      Max length: 512

      Example: example

    • Type: number

    • Type: boolean

    Example: example

    Set of key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format, and querying for objects via API or the dashboard.

    Max properties: 16

    Example
    {}

  • Type: unknown

    Example: null

Default: null

VectorStoreSearchResultContentObject

Name

Description

text

Type: string

The text content returned from search.

Example: example

type

Type: string

The type of content.

Const: text

Example: example
Example
{
          "type": "text",
          "text": "example"
        }

VectorStoreSearchResultItem

Name

Description

attributes

Type: VectorStoreFileAttributes

Default: null

Example
{}

content

Type: VectorStoreSearchResultContentObject[]

Content chunks from the file.

Example
[
          {
            "type": "text",
            "text": "example"
          }
        ]

file_id

Type: string

The ID of the vector store file.

Example: example

filename

Type: string

The name of the vector store file.

Example: example

score

Type: number

The similarity score for the result.

Min value: 0

Max value: 1
Example
{
          "file_id": "example",
          "filename": "example",
          "score": 0,
          "attributes": null,
          "content": [
            {
              "type": "text",
              "text": "example"
            }
          ]
        }
Предыдущая
Get content
Следующая
Create a vector store file batch