Milvus

Configure your own Milvus Vector DB instance as data sink.

Configure via UI

Configure via API / Client

Python

  from llama_cloud.types.data_sink_create_params import (
    CloudMilvusVectorStore,
  )

  data_sink = client.data_sinks.create(
      name="my-data-sink",
      component=CloudMilvusVectorStore(
          uri='<uri>',
          collection_name='<collection_name>',
          token='<token>',  # optional
          # embedding dimension
          dim='<dim>'  # optional
      ),
      sink_type="MILVUS",
  )

TypeScript

const dataSink = await client.dataSinks.create({
  name: 'my-data-sink',
  component: {
    uri: '<uri>',
    collection_name: '<collection_name>',
    token: '<token>',  // optional
    // embedding dimension
    dim: '<dim>'  // optional
  },
  sink_type: 'MILVUS',
});

Python (legacy)

  from llama_cloud.types import CloudMilvusVectorStore

  ds = {
      'name': '<your-name>',
      'sink_type': 'MILVUS',
      'component': CloudMilvusVectorStore(
          uri='<uri>',
          collection_name='<collection_name>',
          token='<token>',  # optional
          # embedding dimension
          dim='<dim>'  # optional
      )
  }
  data_sink = client.data_sinks.create_data_sink(request=ds)

TypeScript (legacy)

const ds = {
    'name': 'milvus',
    'sinkType': 'MILVUS',
    'component': {
        'uri': '<uri>',
        'collection_name': '<collection_name>',
        'token': '<token>',  // optional
        // embedding dimension
        'dim': '<dim>'  // optional
    }
}

data_sink = await client.dataSinks.createDataSink({
  projectId: projectId,
  body: ds
})