BASH script for sorting files recursively by base filename into folders of the same name

Question

I have the following file structure:

\\HI

• ActionPotential_hi.mp4
• ADHD_hi.mp4
• AlzheimersDisease_hi.mp4
• alzheimers_art_hi.mp4
• artificial_eye_hi.mp4
• more files ...

\\LO

• ActionPotential_lo.mp4
• ADHD_lo.mp4
• AlzheimersDisease_lo.mp4
• alzheimers_art_lo.mp4
• artificial_eye_lo.mp4
• etc.

\\MED

*base_filename*_med.mp4

\\STILLS

*base_filename*_med.jpg

\\CAPTIONS

*base_filename*.adb.xml

\\TRANSCRIPTS

*base_filename*.txt

In order to ingest these into a MarkLogic environment, I need these rearranged to into the following structure, where asset is the base filename.

\\ASSET

• asset_lo.mp4
• asset_med.mp4
• asset_hi.mp4
• asset.txt
• asset.adb.xml
• asset_med.jpg

I would like a bash script to sort these out for me. Suggestions?

Answer 1

find . -type f -print |
while read -r pathname; do
    filename=${pathname##*/}
    case "$filename" in
        *_hi* | *_med* | *_lo*)
            # strip off last underscore and following chars
            new_dirname=${filename%_*} 
            ;;
        *)
            # strip off first dot and following chars
            new_dirname=${filename%%.*} 
            ;;
    esac
    mkdir -p "../$new_dirname"
    echo mv "$pathname" "../$new_dirname/$filename"
done 

Untested. Remove the echo when you're satisfied the mv commands look correct.

I moved the destination directories to the parent dir of CWD because I'm not certain whether find will pick up the newly created directories. Can someone address this point?