replace SimpleNamespace configs with typed dataclasses and enums

[matthewcornell]

This PR is part of Operational models refactoring ideas #42. It is paired with the idmodels PR [replace SimpleNamespace configs with typed dataclasses and enums #24].

Changes:

• Replace SimpleNamespace configs with typed dataclasses and enums from idmodels.config across all Python models.
• Update idmodels to de6be08.
• Switch Python dependency management from ad-hoc requirements.txt to pip-compile workflow with requirements.in source files; pin Python 3.12 via .python-version.
• Update README to document the new pip-tools workflow.

[lshandross]

The changes themselves LGTM, but I was wondering when you were planning to rebuild the docker containers and if we should wait to do so until after forecasts are submitted for this week so we have time to try running the new version before we submit the next round of forecasts.

[matthewcornell]

The changes themselves LGTM, but I was wondering when you were planning to rebuild the docker containers and if we should wait to do so until after forecasts are submitted for this week so we have time to try running the new version before we submit the next round of forecasts.

@lshandross Great question. I was planning on asking everyone's opinion re: when to deploy the changes. I've tested them locally - both direct main.py calls and local docker build and call - and they're all looking good. That said, functionality hasn't changed yet, so I thought we might defer deployment until after we add more refactorings...

[lshandross]

I'm fine with the current plan of delaying (since that's what we did when there were changes in the pipeline before), but we can also discuss during our meeting about both PRs

[matthewcornell]

Thanks for your comments @lshandross and @trobacker . Here's what I distilled from them:

PLAN:

• class Disease(str, Enum): add RSV
• class PowerTransform(str, Enum): add NONE
• class SARIXRunConfig(RunConfig):
  • move these parameters to the model class?
  • eliminate SARIX and GBQR run config classes if possible
• we may end up having a lot of enums in the configs
  • I think we should keep some kind of a doc for their purpose and intention. E.g. what does "Pooling" mean to a user of the model (even if it may be obvious).
  • documentation on some of the enums within the classes for users?
  • maybe factor in nick's Feature Taxonomy for ML Models
• remove commented code: # model_config = ...

[new via Nick's comment]:

• move the inference/fitting parameters (num_warmup, num_samples, num_chains for SARIX; num_bags, bag_frac_samples for GBQR) from RunConfig to ModelConfig
  • drop SARIXRunConfig and GBQRRunConfig entirely and have a single RunConfig class
• formalize short_run to be a first class feature of model config
  • consider approach outlined in Nick's comment ("put production defaults on the ModelConfig fields, and add a short_run() method that each model subclass defines")

[matthewcornell]

I've updated the code to match requested idmodels changes. Please review 4c01eac .