Ubuntu – How to batch process JPG images to change their quality with Nautilus-Actions

command linegimpimage processingnautilusnautilus-actions

First of all: Thank you for taking the time to read this and provide some help. Any advice is appreciated.

I am using Nautilus-Actions to add a custom context menu's action to create a copy of a JPG image and reduce its quality (and the file size) while preserving the image's dimensions.

Let's say I have this picture with 2.3 MB (2 261 588 bytes) file size and:

After opening the file in GIMP 2.8 in order to just export a copy and use the JPG Export dialogue box to reduce the quality of the image to 30%:

I get these properties in the image copy (251.8 kB (251 797 bytes) file size):

I am using Nautilus Actions to batch process a lot of images to reproduce the same behaviour via context menu by selecting a bunch of files in Nautilus, right click on any of these selected files, and choose the proper Nautilus Actions action. The result is a file containing the image in the same dimensions as the original, with less quality of course and the file size is a bit greater (275.3 kB (275 265 bytes) file size) but that's not a major problem for me.

The problem is happening when I try to batch process several image files, which results in the first file name cloned into several file names containing a different image each one. With the properties I set in this set of commands:

Basically, each Nautilus Action is using these parameters:

Command: convert

From (imagemagick)

Parameters: %F.jpg -quality 80%% %F-80q.jpg

Which is interpreted as: convert path/to/file1.mid.jpg -quality=80% path/to/file1.mid-80q.jpg

Working directory: %d

So, the result is:

Here we can see each image processed in a different file but just the first file name is used to assign the new names to the copied files. I want a result like:

File1.jpg === File1-80q.jpg
File2.jpg === File2-80q.jpg
File3.jpg === File3-80q.jpg
File4.jpg === File4-80q.jpg

And so on…

I have used Phatch to do such tasks but what I want is to just right click on the selected files and process in a single action.

Is it there a better way to do this? or… Can I improve in some way what I am doing already?

Thank you very much for your support.

P.S. I have seen How can I batch convert images to b/w while preserving the folder structure and Is there a way to batch export SVGs to PNGs? but I am not trying to run this from a terminal but from a context menu. If I am omitting something in this exercise don't hesitate to let me know! I'll appreciate it a lot. Thank you!

Edit

After using muru's answer I find that solution partially does what I expect.

More specifically it solves the file names problem. Nevertheless, I also expected to get smaller file sizes by running the proper convert file.jpg -quality 30% command and it is dropping bigger file sizes as you can see:

Does anyone know if there's something else I should add to the parameters seen in muru's answer? I appreciate a lot your efforts to provide help on this issue.

Update

I am writing this just to let people know I have seen where the problem was with the compression. The file I was trying to compress was already fully compressed so (as mentioned in muru's answer):

The process was just adding overhead without gaining anything.

Best Answer

You cannot use %F twice in a command like that, because %F gets replaced by the names of every file selected. For example, a command of sh with parameters -c 'printf "%%s\n" "$@" > foo' %F %F will create a file named foo with the names of every file selected, twice. Therefore the convert command that actually runs is:

convert file1.jpg file2.jpg ... -quality 80% file1.jpg file2.jpg ... fileN.jpg-80q.jpg

And since the last file is taken to be the output file name, only it will matter.

What you can do is wrap your command in bash -c and run a for loop:

bash -c 'for i; do convert "$i" -quality 80% "${i/%.jpg/-80q.jpg}"; done'

(where I am assuming that all files end with .jpg)

In an action, you will have bash as the command, and for the parameters:

-c 'for i; do convert "$i" -quality 80%% "${i/%%.jpg/-80q.jpg}"; done' - %F

You can make this more complicated to handle any extension, at which point you might as well skip Actions and use scripting.

Related Solutions

Ubuntu – Printing multiple files from Nautilus

I just added a right-click menu option that allows you to select multiple files, select Script/pdf-printer.sh and it'll create PDF files. If you choose another printer name in the script, it should use that instead, but I can't verify that. Older howtos might not work because we've moved to Gnome 3 which has a few changes. In this case, scripts have been moved from ~/.gnome2/nautilus-scripts to ~/.local/share/nautilus/scripts. Otherwise, everything should work as you would expect.

Install http://apt.ubuntu.com/p/cups-pdf. This package provides a PDF printer which prints to the PDF folder in your home folder, $HOME/PDF.
Open Nautilus and create a folder named PDF in your home.
Press Ctrl+H in Nautilus to show hidden files.
Open ~/.local/share/nautilus/scripts.
Create a new file and call it "pdf-printer.sh".

Paste the following:

#!/bin/bash
#
# pdf-printer.sh
#
# Print files from the right-click context menu in Nautilus.
# Place this script in ~/.local/share/nautilus/scripts.
# The printer to use (as shown in the Printer Configuration
# gui or in /etc/cups/printers.conf).
printer=PDF
echo "$NAUTILUS_SCRIPT_SELECTED_FILE_PATHS" | while read file
    do
        lpr -P "$printer" "$file"
done
exit 0

Save the file and close it.
Right-click the file and choose Properties > Permissions. Make sure the box is checked to allow executing the file.

It is possible that you have to restart nautilus for this to work. In that case, you can press Alt+F2 and enter killall nautilus. Other than that, you should now be able to select files, right click and use the script from the "script" menu.

Ubuntu – Comparing two text files

Try this command:

 grep -v -f file2.csv file1.csv > file3.csv

According to grep manual:

  -f FILE, --file=FILE
          Obtain  patterns  from  FILE,  one  per  line.   The  empty file
          contains zero patterns, and therefore matches nothing.   (-f  is
          specified by POSIX.)

  -v, --invert-match
          Invert the sense of matching, to select non-matching lines.  (-v
          is specified by POSIX.)

As Steeldriver said in his comment is better add also -x and -F that:

  -F, --fixed-strings
          Interpret PATTERN as a  list  of  fixed  strings,  separated  by
          newlines,  any  of  which is to be matched.  (-F is specified by
          POSIX.)
  -x, --line-regexp
          Select  only  those  matches  that exactly match the whole line.
          (-x is specified by POSIX.)

So, better command is:

 grep -xvFf file2.csv file1.csv > file3.csv

This command use file2.csv line as pattern and print line of file1.csv that doesn't match (-v).