appmixer.pinecone.database.QueryVectors
1.0.0
Public
| Field | Type | Description |
|---|---|---|
| Index | text | The name of the index. |
| Namespace | text | The namespace of the index. A namespace is a partition within an index. It divides records in an index into separate groups. |
| Vector | text | The query vector. This should be the same length as the dimension of the index being queried. Example: [1.0, 2.0, 3.0]. |
| Top K | number | The number of nearest neighbors to return. The default is 10. The maximum is 10000. |
| Filter | text | A filter expression to apply to the query. You can use vector metadata to limit your search. The filter expression must be a valid JSON. Example: { "genre": { "$eq": "documentary" } }. |
| Include Values | toggle | Include the values of the records in the response. The default is false. |
| Include Metadata | toggle | Include the metadata of the records in the response. The default is false. |
| Result Transformation JSONata | textarea | JSONata transformation expression that will be used to transform the resulting array. This is a convenience feature to aggregate metadata fields across the results to, for example, feed as a context to an LLM model. For example, if the 'text' metadata field is set and the following JSONata expression is used, the result will contain a string that is the concatenation of the 'text' field of the top K results. $join(matches.metadata.(text), "\n").If your metadata contains two fields (e.g. 'text' and 'doc') that you'd like to concatenate, you can use the following expression: $join(matches.metadata.(doc & ";" & text), "\n"). |
| Field | Type | Description |
|---|---|---|
| Result | - | |
| Transformed Result | - |