How to rename misnamed files with their appropriate extensions?

Question

I have a folder with the bulk of files in it of different types, i.e .pdf,.jpg,.png,.tiff etc., but all are named with the extension .JPG. How can I rename all of them with their original extensions? i.e pdf to pdf, tiff to tiff and so on.

I can find the file type by:

file 99.jpg
99.jpg: PDF document, version 1.3

Is there any script or program with which I can do this?

Answer 1

You can use the MIME type found by file:

#! /usr/bin/env bash
for f in ./*.JPG; do
    type=$(file -bi "$f")
    type=${type%%;*}
    base=${f%.*}
    final=
    case "$type" in
        application/pdf) final="$base".pdf ;;
        image/png)       final="$base".png ;;
        image/tiff)      final="$base".tiff ;;
    esac
    if [ "$final" ]; then
        printf '%s --> %s\n' "$f" "$final"
        # mv -f "$f" "$final"
    fi
done

Add more types to the case if you need to (but leave out image/jpeg, since JPEG files already have the right extension). Review the changes, then uncomment the mv line when you're happy with the results.

Answer 2

1. Generate the commands without running them.

   Use mimetype to generate a list of command strings, which is thereafter tweaked by GNU sed's substitute s command:

   cd ~/messed/up/folder/  # go where the files are...
   mimetype -M --output-format 'mv "%f" "%f%m"' *.JPG |
      sed 's#\.[^./"]*/\([^./]*"\)$#\.\1#'
   

2. If some of the file extensions look a little too mime-ish, (i.e. .jpeg instead of .jpg, etc.), then add as many s commands as needed between sed ' and s, for example:

   sed 's/jpeg"$/jpg"/;s#\.[^./"]*/\([^./]*"\)$#\.\1#'
   

3. Once the output looks good, run that with the GNU sed's evaluate e option. (Just put an e before the final '.) So the whole thing might look like:

   cd ~/messed/up/folder/  # go where the files are...
   mimetype -M --output-format 'mv "%f" "%f%m"' *.JPG |
       sed 's/jpeg"$/jpg"/;s#\.[^./"]*/\([^./]*"\)$#\.\1#e'
   ls  # show results