tl;dr we are deploying the SDK to aws lambdas and ran into a scenario where our primary layer became too large, initially we just stripped individual json files from discovery_cache/documents since this is by far the largest culprit (according to ncdu).

I also noticed that all of these definitions are actually pretty printed json.

I'm suggesting 2 things:

1. Ship a utility to select (and delete other) json definitions that the user wants to use in a deployment.
2. Minify all json contents (see trivial script below), this reduced the package size (on disk) by approximately 20mb (and probably also improves load times although I did not test for that).

fs.readdirSync('.').forEach(f => {
  const data = JSON.parse(fs.readFileSync(f, 'utf8'));
  fs.writeFileSync(f, JSON.stringify(data));
  console.log('processed', f);
});

Environment details

• OS type and version: irrelevant
• Python version: 3.9
• pip version: 22.2.2
• google-api-python-client version: 2.64.0

Steps to reproduce

1. Install package
2. 72mb via ncdu

Code example

n/a

Stack trace

n/a

Making sure to follow these steps will guarantee the quickest resolution possible.

Thanks!