package rerun.sdk_comms.v1alpha1

Get desktop application:
View/edit binary Protocol Buffers messages

Simple buffer for messages between SDKs and viewers. - SDKs produce messages by calling `WriteMessages` - Viewers consume messages by calling `ReadMessages` The buffer is bounded by a memory limit, and will drop the oldest messages when the limit is reached. Whenever `ReadMessages` is called, all buffered messages are sent in the order they were received. The stream will then also yield any new messages passed to `WriteMessages` from any client.

• rpc WriteMessages (stream WriteMessagesRequest, WriteMessagesResponse)

  sdk_comms.proto:20

  TODO(jan): Would it be more efficient to send a "message batch" instead of individual messages? It may allow us to amortize the overhead of the gRPC protocol.

  message WriteMessagesRequest

  sdk_comms.proto:29

  • optional log_msg.v1alpha1.LogMsg log_msg = 1

  message WriteMessagesResponse

  sdk_comms.proto:33

  (message has no fields)

• rpc ReadMessages (ReadMessagesRequest, stream ReadMessagesResponse)

  sdk_comms.proto:21

  message ReadMessagesRequest

  sdk_comms.proto:37

  (message has no fields)

  message ReadMessagesResponse

  sdk_comms.proto:39

  • optional log_msg.v1alpha1.LogMsg log_msg = 1

• rpc WriteTable (WriteTableRequest, WriteTableResponse)

  sdk_comms.proto:23

  message WriteTableRequest

  sdk_comms.proto:45

  • optional common.v1alpha1.TableId id = 1

  • optional common.v1alpha1.DataframePart data = 2

  message WriteTableResponse

  sdk_comms.proto:50

  (message has no fields)

• rpc ReadTables (ReadTablesRequest, stream ReadTablesResponse)

  sdk_comms.proto:24

  message ReadTablesRequest

  sdk_comms.proto:54

  (message has no fields)

  message ReadTablesResponse

  sdk_comms.proto:56

  • optional common.v1alpha1.TableId id = 1

  • optional common.v1alpha1.DataframePart data = 2