Re-factor detect_supernova and adopt snappl

[wmwv]

Refactor and adopt to snappl framework.

1. [ ] Standardize Filename access
2. [x] Use Image class for data arrays
3. [x] Use Image class for PSF
4. [ ] Use Image class for WCS #40, #45
5. [x] Adopt a config
6. [x] Use logger

[wmwv]

@rknop Does snappl have a way to abstract out the loading of the RomanTDS files?

E.g., I'd like to replace things like

   INPUT_IMAGE_PATTERN = (
        SIMS_DIR
        + "/RomanTDS/images/simple_model/{band}/{pointing}/Roman_TDS_simple_model_{band}_{pointing}_{sca}.fits.gz"
    )

and more broadly the entire block of similar class variables defined starting at

https://github.com/Roman-Supernova-PIT/detect_supernova/blob/main/detect_supernova/pipeline.py#L36

with something more standardized, but I can't find something already existing in snappl. Is it in future snappl?

[wmwv]

Will not notably improve filename standardization in this Issue. This is now clearly internal to sidecar. It will take inputs from the snpitdb (or manual) and computing its own subtractions and files, and then the output is back to snpitdb in the form of detections. The internal file usage and structure don't have to conform to anything more snappl-/SNPIT wide standards. They could still be done better, but it doesn't matter for the larger Photometry pipeline.

[wmwv]

WCS needs a bit more work. See #40, #45