I-641 Added feedback on reprojection + creating work unit from fits file

[vlnistor]

This PR aims to tackle #641. I use tqdm package for feedback

1. Added feedback during reprojection
2. Added feedback during the creation of a WorkUnit from fits file

[vlnistor]

When it comes to trying to add feedback on creating a WorkUnit from an ImageCollection I couldn't get an example to work:

1. discrete_piles_e2e.ipynb contains an example but requires the lsst stack
2. test_imagecollection.py has a set of mock ImageCollections but they throw when calling toWorkUnit() on them

[vlnistor]

Todo

• [ ] https://github.com/dirac-institute/kbmod/blob/main/src/kbmod/image_collection.py#L531 - Unclear what you meant here
• [x] https://github.com/vlnistor/kbmod/blob/bf9987ae9059c56bc00350bcb87e8c2cd5de38c0/src/kbmod/work_unit.py#L285
• [x] https://github.com/vlnistor/kbmod/blob/bf9987ae9059c56bc00350bcb87e8c2cd5de38c0/src/kbmod/reprojection.py#L131

• [x] https://github.com/vlnistor/kbmod/blob/bf9987ae9059c56bc00350bcb87e8c2cd5de38c0/src/kbmod/reprojection.py#L284 -

  I played with the reproject_demo notebook. I added a time.sleep(3) inside _reproject_images. For a work unit with 3 unique_obstimes_indices, and MAX_PROCESSES set to 1, the progress bar did 3 hops every 3 seconds (as expected). When you set the MAX_PROCESSES to 8 (above the number of parallell jobs) the progress bars hops once (after 3 seconds) as there were enough threads to do the jobs in parallel (again, as expected).

Testing

• [x] Test piping the output to a file and make sure that we can read the TQDM output nicely when not running directly from the command line - You need to redirect stderror for tqdm output to appear (will that be an issue?). It looks similar to this

Reprojecting:   0%|          | 0/3 [00:00]
Reprojecting: 100%|██████████| 3/3 [00:00]

[maxwest-uw]

I think the line numbers may have gotten shifted around, sorry, there should also be tqdm on this line as well in work_unit.py, the loop

for i in range(num_images):
                sci_hdu = hdul[f"SCI_{i}"]

                # Read in the layered image from different extensions.
                img = LayeredImage(
                    sci_hdu.data.astype(np.single),
                    hdul[f"VAR_{i}"].data.astype(np.single),
                    hdul[f"MSK_{i}"].data.astype(np.single),
                    PSF(hdul[f"PSF_{i}"].data),
                    sci_hdu.header["MJD"],
                )

which is the one that actually loads in the images. Everything else looks good!

[vlnistor]

@maxwest-uw that should all be pushed now. Let me know if anything else