Semantic search files - Veryfront

semanticSearchFiles(input: SemanticSearchFilesInput!): SemanticSearchFilesPayload!

Arguments

Argument	Type	Required
`input`	`SemanticSearchFilesInput!`	Yes

Input fields

input

SemanticSearchFilesInput

Field	Type	Required	Description
`queryVector`	`[Float!]!`	Yes	Query vector (embedding of search query)
`dimension`	`EmbeddingDimension!`	Yes	Embedding dimension (must match queryVector length)
`mode`	`SemanticSearchMode`	No	Search mode
`environment`	`SearchEnvironment`	No	Search environment
`branchId`	`ID`	No	Branch ID for PREVIEW environment
`releaseId`	`ID`	No	Release ID for RELEASE environment
`releaseVersion`	`String`	No	Release version for RELEASE environment
`limit`	`Int`	No	Maximum results
`cursor`	`String`	No	Pagination cursor
`options`	`SemanticSearchOptions`	No	Search options

Response

Returns SemanticSearchFilesPayload!. Open the type for its fields and nested types.

Field	Type
`data`	`[SemanticSearchResult!]!`
`pageInfo`	`PageInfo!`
`totalMatches`	`Int!`
`searchTimeMs`	`Int!`

Types marked with ! cannot be null. Lists use [Type] notation. The type index lists all published input and response types.

Example

query SemanticSearchFiles($input: SemanticSearchFilesInput!) {
  semanticSearchFiles(input: $input) {
    searchTimeMs
    totalMatches
    data {
      chunk {
        id
        content
        index
      }
      file {
        id
        checksum
        path
        versionId
      }
      match {
        matchType
        score
      }
    }
    pageInfo {
      hasNextPage
      endCursor
      hasPreviousPage
      startCursor
    }
  }
}

{
  "input": {
    "queryVector": [],
    "dimension": "D768"
  }
}