How to create custom construct library for aws-cdk in python

Question

Recently I have been using aws-cdk to create EC2, VPC and S3 services.

But if I want to create my custom EC2 Library in python(not using JSII) than will be using aws_cdk's aws_ec2 library to actually create the EC2 Instance and a VPC.

The custom library will accept arguments like Instance Name(String) , InstanceType(String) , MachineImage(String) , Subnet Type (String)

Than this arguments will be refer like below:

Disclaimer: Code below Might not be correct

dummy_ec2 = ec2.Instance(self, <InstanceName>, 
                                vpc=<Created_VPC>,
                                instance_type=ec2.InstanceType(<InstanceType>),
                                machine_image=ec2.AmazonLinuxImage(
                                        generation=ec2.AmazonLinuxGeneration.AMAZON_LINUX,
                                        edition=ec2.AmazonLinuxEdition.STANDARD,
                                        virtualization=ec2.AmazonLinuxVirt.HVM,
                                        storage=ec2.AmazonLinuxStorage.GENERAL_PURPOSE
                                ),
                                key_name="demo-key",
                                vpc_subnets=ec2.SubnetSelection(subnet_type=<subnet_type>),
                                role=self.my_role
                                )

Any help Devs how to?

Answer 1

To create a construct, you just have to create a class inheriting from aws_cdk.core.Construct , here's an example:

class InstanceWithVPC(aws_cdk.core.Construct):
    def __init__(self, scope: aws_cdk.core.Construct, id: str, *, instanceType, subnet_type, role, key, **kwargs):
        super().__init__(scope, id)
        self.vpc = ec2.Vpc(...)
        self.instance = ec2.Instance(self, id, 
                                vpc=self.vpc,
                                instance_type=ec2.InstanceType(instanceType),
                                machine_image=ec2.AmazonLinuxImage(
                                        generation=ec2.AmazonLinuxGeneration.AMAZON_LINUX,
                                        edition=ec2.AmazonLinuxEdition.STANDARD,
                                        virtualization=ec2.AmazonLinuxVirt.HVM,
                                        storage=ec2.AmazonLinuxStorage.GENERAL_PURPOSE
                                ),
                                key_name=key,
                                vpc_subnets=ec2.SubnetSelection(subnet_type),
                                role=role
                                )

Of course keep in mind this is only usable in Python, the benefit of using Typescript would be it'd be usable in other languages too which would make more reusable.