Search in Vector Store
The vector_store_search function searches for similar documents in a vector store (embedding store) using a vector query.
- Function name:
extensions.com.cloud-apim.llm-extension.vector_store_search
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
provider | string | yes | The embedding store provider id |
payload | object | yes | The payload object |
payload.embedding | object | yes | The embedding to search with |
payload.embedding.vector | array | yes | The vector to search for |
payload.max_results | integer | no | Maximum number of results to return |
payload.min_score | number | no | Minimum similarity score threshold |
Output
Returns the search results as a JSON array of matching documents with their similarity scores.
Example
{
"kind": "call",
"function": "extensions.com.cloud-apim.llm-extension.vector_store_search",
"args": {
"provider": "embedding-store_xxxxx",
"payload": {
"embedding": {
"vector": [0.1, 0.2, 0.3, 0.4]
},
"max_results": 10,
"min_score": 0.5
}
},
"result": "search_results"
}