layer-builder

Simple library that supports building AWS Lambda layers as part of your AWS CDK build process. Cross-platform (supports both Win and Mac), targeting both x86 and ARM chips, and based on you existing requirements.txt.

Build from requirements

Using this library, you can build a layer containing the packages defined in a requirements.txt file in your AWS CDK pipeline (or other tools). The packages are installed in a local directory using the appropriate platform target.

Wherever possible we attempt to limit the size of the resulting layer by removing any packages that are already native to Lambda (boto, six, future, etc.) or otherwise unnecessary. additional_commands and cleanup_function parameters can be used to further reduce the size of the package by removing any unnecessary libraries.

from layer_builder import build_from_requirements
from constructs import Construct
from aws_cdk import Stack, aws_lambda as aws_lambda

class LayerStack(Stack):

    def __init__(
        self,
        scope: Construct,
        construct_id: str,
        **kwargs,
    ) -> None:
        super().__init__(scope, construct_id, **kwargs)
        
        build_from_requirements(
            "../../requirements.txt", # The packages to install
            "./build/layer", # Where to place the layer packages
            max_mb=20, # Check if the size of the layer is larger than expected
            clean=True, # Delete previous install
            graviton=True, # Specify targeting ARM architecture
        )
        self.my_layer = aws_lambda.LayerVersion(
            self,
            "my-layer",
            compatible_runtimes=[aws_lambda.Runtime.PYTHON_3_13],
            compatible_architectures=[aws_lambda.Architecture.ARM_64],
            code=aws_lambda.Code.from_asset("./build/layer"),
        )