You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardexpand all lines: Extensions/Mongo/README.md
+37-1
Original file line number
Diff line number
Diff line change
@@ -20,11 +20,47 @@ Source and sink settings require both `ConnectionString` and `DatabaseName` para
20
20
21
21
### Sink
22
22
23
+
```json
24
+
{
25
+
"ConnectionString": "",
26
+
"DatabaseName: "",
27
+
"Collection": ""
28
+
}
29
+
```
30
+
31
+
# MongoDB Vector Extension (Beta)
32
+
33
+
The MongoDB Vector extension is a Sink only extension that builds on the MongoDB extension by providing additional capabilities for generating embeddings using Azure OpenAI APIs.
34
+
35
+
> **Note**: When specifying the MongoDB Vector extension as the Sink property in configuration, utilize the name **MongoDB-Vector(beta)**.
36
+
37
+
## Settings
38
+
39
+
The settings are based on the MongoDB extension settings with additional parameters for generating embeddings.
40
+
41
+
### Additional Sink Settings
42
+
43
+
The sink settings require the following additional parameters:
44
+
45
+
-`GenerateEmbedding`: If set to true, the sink will generate embeddings for the records before writing them to the database. The sink requires the `OpenAIUrl`, `OpenAIKey`, and `OpenAIDeploymentModel` parameters to be set. Following paramaters are required if this is true
46
+
-`OpenAIUrl`: The URL of the OpenAI API
47
+
-`OpenAIKey`: The API key for the OpenAI API
48
+
-`OpenAIDeploymentModel`: The deployment model to use for the OpenAI API
49
+
-`SourcePropEmbedding`: The property in the source data that should be used to generate the embeddings
50
+
-`DestPropEmbedding`: New property name that will be added to the source data with the generated embeddings
0 commit comments