greengrass v2 ml component

hello I created aws greengrass v2 ml component.

I want to run the ml model on Raspberry Pi with greengrass installed. So I wrote the ml component and deployed it. The result I expect is that the ml code I wrote is executed on Raspberry Pi.

This is a custom component. I referred to https://docs.aws.amazon.com/greengrass/v2/developerguide/ml-customization.html this link.

But there is a problem. If I deployed the components, say that the Deploy is complete. But there's an error. Looking at the log, it seems that there is a problem opening the virtual environment(venv), or it is because the tensorflow is not downloaded

I want to write a script that opens the virtual environment properly in the recipe of the component. The recipe I wrote looks like this. { "RecipeFormatVersion": "2020-01-25", "ComponentName": "com.example.jamesML", "ComponentVersion": "1.0.5", "ComponentType": "aws.greengrass.generic", "ComponentDescription": "Capstone Design james machine learning.", "ComponentPublisher": "Me", "ComponentConfiguration": { "DefaultConfiguration": { "accessControl": { "aws.greengrass.ipc.mqttproxy": { "com.example.jamesML:mqttproxy:1": { "policyDescription": "Allows access to publish via topic ml/dlr/image-classification.", "operations": [ "aws.greengrass#PublishToIoTCore" ], "resources": [ "ml/dlr/image-classification" ] } } } } }, "Manifests": [ { "Platform": { "os": "linux", "architecture": "arm" }, "Lifecycle": { "setEnv": { "Script": "cd venv&&activate venv" }, "run": { "RequiresPrivilege": "true", "script": "pip install opencv-python;pip install tensorflow==2.3;python3 james.py" } }, "Artifacts": [ { "Uri": "s3://greengrass-sagemaker-0930/james_ml.zip", "Digest": "CY6f7pUyMRrgbxbKoDMig3GWQCJ4LvKyA5xnhGWGhlY=", "Algorithm": "SHA-256", "Unarchive": "ZIP", "Permission": { "Read": "OWNER", "Execute": "NONE" } } ] } ], "Lifecycle": {} } Can you solve the problem here? Is this part I wrote wrong? "setEnv": { "Script": "cd venv&&activate venv" }, I want to know the answer. Please. Please.

This is my greengrass v2 component log (Raspberry Pi) 2022-10-13T10:01:44.004Z [INFO] (pool-2-thread-28) com.example.jamesML: shell-runner-start. {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=STARTING, command=["pip install opencv-python;pip install tensorflow==2.3;python3 james.py"]} 2022-10-13T10:01:45.326Z [INFO] (Copier) com.example.jamesML: stdout. Looking in indexes: https://pypi.org/simple, https://www.piwheels.org/simple. {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:45.329Z [INFO] (Copier) com.example.jamesML: stdout. Requirement already satisfied: opencv-python in /usr/local/lib/python3.9/dist-packages (4.6.0.66). {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:45.373Z [INFO] (Copier) com.example.jamesML: stdout. Requirement already satisfied: numpy>=1.14.5 in /usr/local/lib/python3.9/dist-packages (from opencv-python) (1.23.4). {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:45.901Z [WARN] (Copier) com.example.jamesML: stderr. WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv. {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:47.470Z [INFO] (Copier) com.example.jamesML: stdout. Looking in indexes: https://pypi.org/simple, https://www.piwheels.org/simple. {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:48.909Z [WARN] (Copier) com.example.jamesML: stderr. ERROR: Could not find a version that satisfies the requirement tensorflow==2.3 (from versions: none). {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:48.910Z [WARN] (Copier) com.example.jamesML: stderr. ERROR: No matching distribution found for tensorflow==2.3. {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:49.161Z [WARN] (Copier) com.example.jamesML: stderr. python3: can't open file 'james.py': [Errno 2] No such file or directory. {scriptName=services.com.example.jamesML.lifecycle.run.script, serviceName=com.example.jamesML, currentState=RUNNING} 2022-10-13T10:01:49.169Z [INFO] (Copier) com.example.jamesML: Run script exited. {exitCode=2, serviceName=com.example.jamesML, currentState=RUNNING}