End of support notice: On October 31, 2025, AWS
will discontinue support for Amazon Lookout for Vision. After October 31, 2025, you will
no longer be able to access the Lookout for Vision console or Lookout for Vision resources.
For more information, visit this
blog post
Setting up your environment
To write client code, your development environment connects remotely to an AWS IoT Greengrass Version 2 core device to which you have deployed a Amazon Lookout for Vision model component and dependencies. Alternatively, you can write code on a core device. For more information, see AWS IoT Greengrass development tools and Develop AWS IoT Greengrass components.
Your client code should use gRPC client to access the Amazon Lookout for Vision Edge Agent. This
section shows how to set up your development environment with gRPC and install third-party
dependencies needed for the DetectAnomalies
example code.
After you finish writing your client code, you create a custom component and deploy the custom component to your edge devices. For more information, see Creating the client application component.
Setting up gRPC
In your development environment, you need a gRPC client that you use in your
code to call the Lookout for Vision Edge Agent API. To do this, you create a gRPC stub by using
a .proto
service definition file for the Lookout for Vision Edge Agent.
Note
You can also get the service definition file from the Lookout for Vision Edge Agent
application bundle. The application bundle is installed when the Lookout for Vision Edge Agent
component is installed as a dependency of the model component. The
application bundle is located at
/greengrass/v2/packages/artifacts-unarchived/aws.iot.lookoutvision.EdgeAgent/
.
Replace edge_agent_version
/lookoutvision_edge_agentedge_agent_version
with version of the Lookout for Vision Edge Agent that
you are using. To get the application bundle, you need to deploy the
Lookout for Vision Edge Agent to a core device.
To set up gRPC
-
Download the zip file, proto.zip. The zip file contains the .proto service definition file (
edge-agent.proto
). -
Unzip the content.
Open a command prompt and navigate to the folder that contains
edge-agent.proto
.Use the following commands to generate the Python client interfaces.
%%bash python3 -m pip install grpcio python3 -m pip install grpcio-tools python3 -m grpc_tools.protoc --proto_path=. --python_out=. --grpc_python_out=. edge-agent.proto
If the commands are successful, the stubs
edge_agent_pb2_grpc.py
andedge_agent_pb2.py
are created in the working directory.Write the client code that uses your model. For more information, see Using a model in your client application component.
Adding third-party dependencies
The DetectAnomalies
example code uses the Pillow
Use the following command to install the Pillow library.
python3 -m pip install Pillow