Figure out how best to regularly clean up folder images from N drive

[beckerah]

Things to consider:

How do I figure out which images are folders?
I can run the barcode-guid matching script. If this isn't producing good enough results, I can run the species-ocr barcode reader on the N drive images as well. This produces an output.json file, which I already have a script to parse.
I can grab the original output.json file from the AU computer and pull out the image names and barcodes. This might be useful for comparing against barcode-guid matching db, to check that all barcodes were correctly interpreted and added to the db. Though it might be faster to just run the images on the N drive through the species-ocr barcode reader.

I think it makes sense to move all folder images to a specified directory first, then I can just do a quick glance to make sure they are actually folders before deleting them.

[beckerah]

Draft skeleton workflow:

1. Run barcode-guid matching script
2. Compare barcode-guid db with species-web export for relevant dates
3. If discrepancy is over determined threshold (30?), run species-ocr barcode reader on images and update barcode-guid db with additional found barcodes
4. Move all images with no found barcode into a temp folder (tifs and jsons in one temp folder, jpegs in another). Create log file for each time images are moved. (Temp folders are currently on N drive at N:\SCI-SNM-DigitalCollections\DaSSCo\MASTER_IMAGE_STORE\temp_au_folder_images)
5. Check to make sure no specimen images. (If specimen image found, check if barcode is in species-web export or on incorrectly ID'd spreadsheet. Move back to original folder - all 3 associated files.)
6. Delete all files (tifs, jsons, jpegs)

[beckerah]

A thought: Would it be possible to have species-ocr move all folder images to a separate folder after uploading to web? If that were the case, then we could run the ingestion client on the original session folder, meaning that only specimen images would be left there to be ingested.

[beckerah]

This relates to #567