Extract info from filename for automatic copying

Question

This is my first post on StackOverflow. Be gentle. :)

I have a linux server that I download television shows to and manually move the files to their respective folders for my Plex server to utilize. I would like to automate this process. I have gotten as far as below.

File naming convention Show.Name.SeasonNumberEpisodeNumber.HDTV.X264.etc...

Example: Almost.Human.S01E01.720p.HDTV.X264.mkv

NOTE: The Show name can be different lengths with a . between each word in the name.

I am able to extract the show folder name from the filename.

#!/bin/bash
readonly FILEPATH=~/downloads
readonly SAVEPATH=~/shows

for file in $FILEPATH/*.mkv
do
#Get foldername from filename (everything before .S0 is foldername
        foldername=$(basename "${file%.S0*}" | tr '.' ' ')
#Need to convert extracted season info into folder name ex. S01 = Season 1
#       seasonfolder=$(basename "${file}" | sed -e 's/^[^S0]*//;')

#Copy the file to the path we built.
#Auto-create folder if it doesn't exist?

#       cp $file  "$SAVEPATH/$foldername/#seasonfolder"

done

Issues:

• I would like to also extract the season information from the filename and use it to build the rest of the folder path. Each file has a SxxExx portion that I can use to get the Season info. I would convert, for example, S01 into Season 1 to be part of the folder path.

Resulting Copy command would look something like this (using above filename)

cp Almost.Human.S01E01.720p.HDTV.X264.mkv shows/Almost Human/Season 1

I am not savvy enough with sed or regex to get the syntax right and after lots of searching, no one is quite doing this in any example I can "borrow" from.

Thanks in advance!

UPDATE

Much thanks to Janos! Not only did he provide an excellent solution, but helped me absorb a little more about using regex.

I made a few changes to the final product. After some research about Plex's naming convention requirements, I adjusted the regex to acommodate and also built in a "file exists" check to avoid unnecessary transfers.

Here's the final result that I will be adding to CRON later tonight.

#!/bin/bash
readonly FILEPATH=~/downloads
readonly SAVEPATH=~/shows

for file in $FILEPATH/*.mkv
do
        dfile="$SAVEPATH/$(basename "$file" | sed -e 's/\./ /g' -e 's?\(.*\) [Ss]\([0-9][0-9]\)[Ee]\([0-9][0-9]\) .*?\1/Season \2/\1 - S\2E\3.mkv?')"

        if [ ! -f "$dfile" ]

        then
                cp -v "$file" "$dfile"
                mkdir -p "$(dirname "$dfile")"
        else
                echo "file exists "$dfile""
        fi

done

Answer 1

You could do something like this:

for file in $FILEPATH/*.mkv; do
    # get the destination filename
    dfile="$SAVEPATH/$(basename "$file" | sed -e 's/\./ /g' -e 's?\(.*\) S0\([0-9]\)E\([0-9][0-9]\) .*?\1/Season \2/Episode \3.mkv?')"

    # create the destination directory
    mkdir -p "$(dirname "$dfile")"

    cp "$file" "$dfile"
done

This will create filenames like:

Almost Human/Season 1/Episode 01.mkv

To make filenames like this:

Almost Human/Season 1/Almost Human Episode 01.mkv

then change the sed like this:

sed -e 's/\./ /g' -e 's?\(.*\) S0\([0-9]\)E\([0-9][0-9]\) .*?\1/Season \2/\1 Episode \3.mkv?

The key in this solution is capturing the relevant parts of the name within \(...\), and then making references to these captured matches using \1 for the first \(...\), \2 for the second, and so on.