rename script to adjust certain patterns within a filename in debian

Question

I'm currently a bit stuck, trying to adjust filenames within a folder tree if they have one or both of the following:

1. Remove 'S00' - 'S99' -that part is to be cut out of the filename. s/[sS](0[1-9]|[1-9][0-9])/ could do that I think

2. Any free standing double digits get an 'E' in front of them. Find /\b([0-9]{2})\b/ and substitute E$1 somehow

Example: any of these

• [lorem ipsum] dolor sit amet 07 [consetetur][1080p].mkv

• [lorem ipsum] dolor sit amet S04 07 [consetetur][1080p].mkv

• [lorem ipsum] dolor sit amet S04 E07 [consetetur][1080p].mkv

are changed to

• [lorem ipsum] dolor sit amet E07 [consetetur][1080p].mkv

within their respective subfolders.

Ultimate goal is to automate the process of renaming files as they appear (that latter bit I have to figure out once I'm done with this) so the kodi scraper can recognize episodes

This is how far I got, as you can see, its still a mess as I've found this rename script here a while ago, but I don't know how to properly apply the regex, I just slapped it in

#!/bin/bash

start_dir="//srv/MEDIA2/"
find "$start_dir" -name '*.*' -type f \
  |sort \
  |while read name; do 
     ((i++))
     mv -i "$name" \
#remove season part
       "$(printf 's/[sS](0[1-9]|[1-9][0-9])/' "$(dirname "$name")" $((i)) "$(basename "$name")")"
#prepend E to free standing double digits
       "$(printf 's/\b([0-9]{2})\b/' "$(dirname "$name")" $((i)) "$(basename "$name")")"
    
done

Currently it returns

//home/user/./rnscript.sh: line 14: s/[sS](0[1-9]|[1-9][0-9])/: No such file or directory
//home/user/./rnscript.sh: line 16: s([0-9]{2}/: No such file or directory

If anyone with knowhow in that area could help me out to get this running, it would be greatly appreciated.

Answer 1

Would you please try the following:

#!/bin/bash

shopt -s globstar

start_dir="//srv/MEDIA2/"                       # double // may be meaningless
for name in "$start_dir"/**/*.*; do             # search the directory recursively
    [[ -f $name ]] || continue                  # skip if "$name" is not a file
    newname=$(sed -E 's/[sS](0[1-9]|[1-9][0-9]) +//; s/\b([0-9]{2})\b/E\1/' <<< "$name")
    if [[ "$name" != "$newname" ]]; then        # skip if the filenames are the same
         echo mv -- "$name" "$newname"
    fi
done

Drop echo if the output looks good.
We do not have to sort the files as the variable $i is not used.

[EDIT]

If you want to change S04E02 to E02 as well, please replace the newname= line above to:

newname=$(sed -E 's/[sS](0[1-9]|[1-9][0-9]) *//; s/\b([0-9]{2})\b/E\1/' <<< "$name")

The difference is just change + to *, meaning zero or more whitespaces.