Add script and update `build-and-test-dbt` workflow to push dbt Python dependencies to S3

[jeancochrane]

This PR adds a new script deploy_dbt_model_dependencies.sh that downloads dependencies for dbt Python models, bundles them into zip archives, and pushes them to S3. It also updates the build-and-test-dbt workflow to use the new script, including the following features:

• If a dbt cache exists, only deploy packages for models that are new or have changed
• If a caller has dispatched the workflow for a specific set of models, only deploy packages for those models
• In all cases, only deploy package bundles that are new or have changed
• Deploy a separate package bundle for each model
• Deploy package bundles to separate S3 locations for dev, CI, and prod, and isolate dev/CI environments by user or branch

Note that this PR leaves out of scope the additional step to refactor the existing reporting.ratio_stats model to use this new dependency management system. That step will be handled in a follow-up issue (https://github.com/ccao-data/data-architecture/issues/439).

Closes https://github.com/ccao-data/data-architecture/issues/417.

Testing

Evidence of build-and-test-dbt pushing packages for changed models to S3 on synchronize event: Link

Evidence of build-and-test-dbt skipping dependencies that have not changed on synchronize event: Link

Evidence of build-and-test-dbt skipping models that have no Python dependencies on workflow_dispatch event: Link

Evidence of cleanup-dbt-resources removing packages from S3 on closed event: Link

[jeancochrane]

Closing this to test the cleanup step.

[jeancochrane]

~@dfsnow Want to give this one more look before I merge?~ Never mind, needs one more change!