infra: use apt python3.12 for base image and decouple cudaq image

[rmshaffer]

Issue #, if available:

Description of changes:

Two improvements here:

1. Switches the base image to apt-installed python3.12, which reduces build time by ~50% and image size by ~10%.

2. Makes the cudaq image independent of the base image, so that changes to both images can now be built in parallel; no need to wait for a new base image to be published to ECR before rebuilding the cudaq image.

Testing done:

Tested the container builds locally and ran tests.

Merge Checklist

Put an x in the boxes that apply. You can also fill these out after creating the PR. If you're unsure about any of them, don't hesitate to ask. We're here to help! This is simply a reminder of what we are going to look for before merging your pull request.

General

• I have read the CONTRIBUTING doc
• I used the commit message format described in CONTRIBUTING
• I have updated any necessary documentation, including READMEs and API docs (if appropriate)

Tests

• I have added tests that prove my fix is effective or that my feature works (if appropriate)
• I have checked that my tests are not configured for a specific region or account (if appropriate)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[speller26]

Do these improvements carry over to the other images?

[rmshaffer]

Do these improvements carry over to the other images?

The tensorflow and pytorch images are based on SageMaker DLC images, which already have most dependencies. So this particular optimization only matters for the base and cudaq images.

[dtrenev]

Change makes sense to me. Haven't manually built the images to check, but I expect a "simple" MPI job to work with the BASE image and a CUDA-Q MPI job (e.g. the ones in cuda-q notebook 6) to work with the CUDAQ image.