ragSaveEmbeddings( )
Saves document embeddings in the RAG vector database.
function ragSaveEmbeddings(params): Promise<{
droppedIndices: number[];
processed: object[];
}>;
| Name | Type | Required? | Description |
|---|
| params | object | ✓ | The parameters for saving embeddings |
| Field | Type | Required? | Default | Description |
|---|
| modelId | string | ✓ | - | The identifier of the embedding model to use |
| documents | string or string[] | ✓ | - | The documents to embed and save (string or array of strings) |
| chunk | boolean | ✓ | - | Whether to chunk the documents before embedding |
| chunkOpts | object | ✗ | - | Options for chunking documents |
| workspace | string | ✗ | - | Optional workspace for isolated storage |
| Field | Type | Required? | Default | Description |
|---|
| chunkSize | number | ✗ | 512 | Size of each chunk |
| chunkOverlap | number | ✗ | 0 | Overlap between chunks |
| chunkStrategy | "character" | "paragraph" | ✗ | "character" | Strategy for chunking |
| splitStrategy | "token" | "character" | "word" | "sentence" | "line" | ✗ | "character" | Strategy for splitting text |
Promise<object> — The processed results and dropped indices
| Field | Type | Description |
|---|
| droppedIndices | number[] | Indices of documents that were dropped |
| processed | object[] | Processed document information |
When the response type is invalid or when the operation fails
