export cutout blob metrics

[mkutu]

look into the semifield-cutout blob container and calculate: cutout data filename convention is the semifield image name with "_" appended to the file stem. Each appended integer represents a new cutout for that image. For example: image_id = MD_164345343454 cutout_id = MD_164345343454_0

1. the number of:
   • masks (*_mask.png)
   • cropouts (*.jpg)
   • cutouts (.png)
   • metadata (.json) for each batch

examples of naming mask = MD_164345343454_0_mask.png cropout = MD_164345343454_0.jpg cutout = MD_164345343454_0.png metadata = MD_164345343454_0.json

[mkutu]

• replace "FolderName" with "TimeStamp"
• add columns (n_missing_cropout, n_missing_cutouts, n_missing_masks, n_missing_metadata) for missing file matches in unprocessed_batches.csv (but consider renaming this csv)