Shell script to convert PDF to images and save in sub-folder

Question

I'm trying to write a bash scrip that uses Imagemagick to convert PDF's to images in a given directory (passed as an argument), and then save that conversion into a sub-folder of that directory. I have a script the works perfectly fine for converting and saving into the same directory, but when I try to add the path I'd like to converted files to be saved at the script breaks.

Here's what I have so far that works:

DIR="$@"
EXT="*.pdf"
mkdir -p "${DIR}images"
shopt -s nullglob
for pdffile in "$DIR"*.pdf; do
 convert -density 300 -depth 8 -quality 90 -trim "${pdffile}" "${pdffile%.*}".jpeg
done

Here's what I've tried for saving into a sub-folder. If I add "images/..." however to the last argument in converted like so:

DIR="$@"
EXT="*.pdf"
mkdir -p "${DIR}images"
shopt -s nullglob
for pdffile in "$DIR"*.pdf; do
 convert -density 300 -depth 8 -quality 90 -trim "${pdffile}" "images/${pdffile%.*}".jpeg
done

This is what happens when I execute the script:

$ ls foo
bar.pdf baz.pdf biz.pdf
$ ./pdf-to-img.sh foo/
convert-im6.q16: unable to open image `images/foo/bar.jpeg': No such file or directory @ error/blob.c/OpenBlob/2701.
convert-im6.q16: unable to open image `images/foo/baz.jpeg': No such file or directory @ error/blob.c/OpenBlob/2701.
convert-im6.q16: unable to open image `images/foo/biz.jpeg': No such file or directory @ error/blob.c/OpenBlob/2701.

What am I doing wrong?

• steeldriver: Remember that ${pdffile%.*} will include the DIR prefix. Also please edit your question to show exactly how you're running the script - and the complete (unedited) error.
• Jon.H: Question has been edited to reflect the complete error
• steeldriver: So the error should be obvious now, no? you need to remove the DIR component (in this case foo) from ${pdffile%.*}"
• Jon.H: I've tried that, but unfortunately I'm still getting that error.
• don_crissti: You don't need a loop, you can do this in one go if you use mogrify with -path

Answer 1

Corrected code:

DIR="$@"
EXT="*.pdf"
mkdir -p "${DIR}/images"
shopt -s nullglob
for pdffile in "$DIR"*.pdf; do
 convert -density 300 -depth 8 -quality 90 -trim "${pdffile}" "$(dirname "${pdffile}")/images/$(basename "${pdffile%.*}").jpeg"
done

It should transform this path images/foo/bar.jpeg, which causes error, to this foo/images/bar.jpeg

Explanation:

You have value like this "foo/bar.pdf" in the ${pdffile} variable

dirname ${pdffile} - remove last non-slash component from path, leaving this: "foo"

basename ${pdffile%} - conversely, remove all except file name, leaving this: "bar.pdf"

Finally, inserting subdirctory name $images in between this parts and get needed path for storing images - foo/images/bar.jpeg.

• Jon.H: Thank's for the answer. So the $ in $images was typo and not actually in my script, so it's just images/; sorry about that :( That being said however, I tried your other recommendation but that didn't work. The folder name has spaces in it, and it looks like the spaces are causing a problem. Lets say the real folder name is "PDFs For Courts", the error message I'm getting is: basename: extra operand ‘For’ Try 'basename --help' for more information. convert-im6.q16: no images defined ``../../data/entities . . Courts//.jpeg' @ error/convert.c/ConvertImageCommand/3258.
• MiniMax: If images not variable, dollar sign should be removed. And I add double quotas for spaces handling. Try this string: $(dirname "${pdffile}")/images/$(basename "${pdffile%.*}").jpeg If this will work, I will edit my answer.
• MiniMax: Outer quotas for whole string don't needed in this case, don't add them.
• MiniMax: I have added whole corrected code to my answer.
• Jon.H: I had to add outer quotes for on the output part so that it looked like: "$(dirname "${pdffile}")/images/$(basename "${pdffile%.*}").jpeg" in order to get it to work (otherwise it still choked on the spaces in the directory name) but it works! Thank you.