[Python] How can I speed up my CDK Synth step in code pipeline by caching requirements?

Currently, the synth step in my code pipeline looks something like this:

synth=pipelines.ShellStep(
"Synth",
input=pipelines.CodePipelineSource.connection(
self.repository,
self.branch,
connection_arn="arn:aws:codestar-connections:us-east-1:613431292675:connection/db5d023d-5eed-4bb7-808c-df5d5fe5452f",
trigger_on_push=True,
),
commands=[
"cd eval-infra",
"npm install -g aws-cdk", # Installs the cdk cli on Codebuild
"pip install -r requirements.txt", # Instructs Codebuild to install required packages
"npx cdk synth EvalInfraPipeline",
],
primary_output_directory="eval-infra/cdk.out",
),


Unfortunately, this takes almost two minutes to run - and the bulk of this is the npm install and the pip install commands. Two minutes is not a lot in isolation, but when combined with the rest of the build process, it adds up to 10 minute+ build times.

Is it possible to use caching somehow to speed this up? I looked into using a CodeBuildStep instead, with some kind of local or s3 cache, but I couldn't find a working example. Has anyone done something like this before?

Continue reading...