Feature refactor UI

[jeipollack]

This PR features many changes which I subdivide into major and minor categories:

Major

• Added a run.py module to launch at execution which parses the command line arg specifying the name of the configuration file: configs.yaml
• Added a .env script storing the environment variables required to be set at execution with the dotenv package method load_dotenv
• Provided an .env.example in the repo which the user should update with the appropriate paths to the directories wf-psf repo, the directory for training and test datasets, and the location of the output directory closes #51
• Added a io.py module with methods to automatically make the wf-outputs file structure defined in long-runs instructions
• New UI interface applied for setting training parameters resolves issues #43 and #50 with a main configs.yaml file which should store individual config files: training_config.yaml, eval_config.yaml , plotting_config.yaml and a logging.conf. Only the training_config.yaml and logging.conf have been written and used so far.
• Created the training package which should contain train.py module with a class TrainingParamsHandler for extracting the different parameter subsets for training, psf model, and datasets, and a train method which should do the training once the psf model, training and test datasets have been specified (TO DO)
• Created a psf_model package containing for now a psf_model.py module with a decorator register_psf_class wrapping each PSF model class. At runtime, a dictionary called PSF_CLASSES will get loaded with all of the psf_model classes. The dict key is determined according to the class attribute ids which stores the short model name of the class specified in the training_config.yaml.
• Moved zernikes methods from the utils.py module to a zernikes.py module

Minor Changes

• Reduced script_utils.py to only train the poly model. Have not tested it.
• Corrected minor typos in commented versions
• Created a metrics package and moved the metrics.py module there.
• Updated wf_psf/__init__.py with the train, psf_models and metrics packages

[jeipollack]

oops. I just remembered that I need to write a method to verify that the config parameters are correct.

[sfarrens]

Thanks @jeipollack, I will start reviewing this tomorrow. :)

[jeipollack]

Thanks, @sfarrens. I am reviewing myself, too! I see that I need to tidy up the io.py module. And, since the construction of the output file directory is automatic I was thinking to convert it to internal commands. If you have any thoughts on that, let me know. Thanks!

[jeipollack]

Hi @sfarrens , I pushed doc string style changes for read_config.py.

[jeipollack]

Hi @sfarrens, just want to send you a FYI that I did a few more changes in this branch more related to restructuring the package, but I did not push them. Instead I decided to checkout a new branch to continue development there until this PR can be merged to the dummy_main branch.