command line commands for image compression?

[tatagi]

Hello!

I prefer not to add files first into library which is not necessary for my use case. Instead, I want to do it out of court, that's why I find command line interface a good option.

having tried loseless compression but it was not giving me the desired result, so I am considering to do lossy compression to size epubs by a great margin.

problem is I don't know what exactly is the command line command to do this , obviously.
I have thousands of files to do this. so batch command would be better.

any help would be grateful.

[djrx]

Depends on your operating system. If you're using Linux/Mac, you can install Homebrew. Once brew is installed, you can install imagemagick

If source is too big
Install ImageMagick using Homebrew

$/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
$brew install imagemagick

### Set the max file size
for i in *.jpg; do
convert "$i" -define jpeg:extent=500kb "${i%.jpg}-new.jpg"
done


### To change just the quality of the file
for i in *.jpg; do
convert "$i" -quality 80% "${i%.jpg}-new.jpg"
done


### Or if no need to keep the original
for i in *.jpg; do
mogrify -quality 80% *.jpg
done


### convert png to jpg
for i in *.png; do
convert "$i" "${I%.png}.jpg"
done

OR

magick mogrify -format jpg *.png
rm *.png

Just copy and paste the whole block of lines for the options you want into a terminal window and go to town

[capink]

Calibre command line ebook-polish only allows for lossless compression of images per the documentation.

If you insist on lossy compression, Editor Chains plugin allows you to use calibre's image compression functionality (among many other editor functions) from the command line:

• Install the plugin from calibre (Calibre main library > preferences > get plugins).
• Open the calibre editor, and then import the chain attached below (Editor > plugins menu > Editor Chains > add chains > right click > import chain).
• From the command line you can type this command to compress images in your ebook (epub and azw3 formats supported):
  
  

  calibre-debug -r "Editor Chains" "Compress Images" path_to_input.epub path_to_output.epub

Note: You can configure the compression ratio by modifying the settings of the chain.

[tatagi]

Originally Posted by djrx

.

thank you. but it is windows 10 that I am using. it seems like imagemagick is useful for img files, not compressing image files that are within epubs directly.

Originally Posted by capink

.

thank you.
oh, btw you're the developer of that plugin. that's nice

could you elaborate it a little further?

for example, all of my epubs are in C:\users\foldername
and some epubs are in its subfolders like
C:\users\foldername\red
C:\users\foldername\blue
etc.

in this case, what is the exact commands for bulk compress to keep new epubs with same name and folder structure?

calibre-debug -r "Editor Chains" "Compress Images" "C:\users\foldername\*.epub" "C:\users\newfoldername\*.epub"

will it work? or can I just specify the folder name so it compresses all the epubs inside?

[isarl]

Originally Posted by tatagi

thank you. but it is windows 10 that I am using. it seems like imagemagick is useful for img files, not compressing image files that are within epubs directly.

Epubs are just zipped directories with a specific set of included files like the container file. If you unzip an ePub, compress the image files, and re-zip the directory, you will have created a new ePub with smaller images. Note that the compressed images should have the exact same filenames, and be in the exact same locations/subdirectories, as the original images (which should be removed before re-zipping).

[capink]

Originally Posted by tatagi

thank you. but it is windows 10 that I am using. it seems like imagemagick is useful for img files, not compressing image files that are within epubs directly.



thank you.
oh, btw you're the developer of that plugin. that's nice

could you elaborate it a little further?

for example, all of my epubs are in C:\users\foldername
and some epubs are in its subfolders like
C:\users\foldername\red
C:\users\foldername\blue
etc.

in this case, what is the exact commands for bulk compress to keep new epubs with same name and folder structure?

calibre-debug -r "Editor Chains" "Compress Images" "C:\users\foldername\*.epub" "C:\users\newfoldername\*.epub"

will it work? or can I just specify the folder name so it compresses all the epubs inside?

No. You will need to write a shell script to achieve this. I am a linux user, so I can easily write a bash script that would work only in linux. Unfortunately, I don't know much about Windows Batch scripts to assist with this. Maybe someone else can help with this.

[BetterRed]

FWIW ==>> How to Install Linux Bash Shell on Windows [Step-by-Step Guide]

BR

[tatagi]

Originally Posted by isarl

.

many thanks. yes that's the most foolproof way but at the same time very tiresome to individually edit files one at a time.

Originally Posted by capink

.

thank you again. at least with your plugin, It doesn't unnecessarily leave caches by opening first file in editor's GUI interface. very useful for that reason alone.

here's another question. could you please make chain action settings to also remove all the embedded fonts in tools-manage fonts - select all fonts - remove selected fonts? if this is combined with compress images settings, it would be much better.

Originally Posted by BetterRed

.

thanks for the tip, I don't know anything about linux but I will see what I can find out .

[capink]

Originally Posted by tatagi

thank you again. at least with your plugin, It doesn't unnecessarily leave caches by opening first file in editor's GUI interface. very useful for that reason alone.

here's another question. could you please make chain action settings to also remove all the embedded fonts in tools-manage fonts - select all fonts - remove selected fonts? if this is combined with compress images settings, it would be much better.

There is already an action called "Remove Files" which can be used to remove any kind of files by specifying the names of the files you want to remove (regex supported), after that you only need to remove any css @font-face rules using a simple regex.

A chain that contains all the above action is attached here. I changed the name of the chain to "Misc Action", so if you want to run it from the command line:

calibre-debug -r "Editor Chains" "Misc Actions" path_to_input.epub path_to_output.epub

P.S. You can run the chain all books in a calibre library if you have the Action Chains plugin installed. So if you want, you can add all your books to temporary calibre library, select all the books there, run the chain and finally save the files to disk again using the save to functionality of calibre.

[tatagi]

Originally Posted by capink

.

that's brilliant thanks a lot you saved me a lot of times and efforts.

In most cases it works like a charm, but this is what I thought.

it's kind of rare issue I can think of, but I tested it with my sample epubs
but the problem is it can delete some files that has .ttf or .otf as part of their name also (e.g "this is the list of .ttf files.html" )

I want to make sure I don't mistakenly delete any crucial files that must be present, regex in this regard is not really fail-proof. any ideas?

[capink]

To be absolutely sure, anchor the regex with $ to make sure the .ttf is the last part of the filename:

(?i).+\.ttf$

[tatagi]

Originally Posted by capink

.

I might be so paranoid but it still won't stop it from deleting folder or some files with no extensions that contain .ttf


I rechecked the plugin's release thread https://www.mobileread.com/forums/sh....php?p=4174583

and removing embedded font using editor's builtin tools via editor chains is maybe not supported. (not in currently available actions)

btw I found a way to bulk compress images using for loop commands + editor chains combo. for those interested :

for /r "C:\Users\foldername" %v in (*.epub) do calibre-debug -r "Editor Chains" "Misc Actions" "%v" "%v.epub"

I have yet to find a way to save the output files in separate folder. if any one knows how, please let me know.

[capink]

Originally Posted by tatagi

I might be so paranoid but it still won't stop it from deleting folder or some files with no extensions that contain .ttf

It will not delete directories. And it only deletes files that end with .ttf extension.

Also, a word of advice: Never ever delete the originals. You have been warned.

[tatagi]

Originally Posted by capink

It will not delete directories. And it only deletes files that end with .ttf extension.

Also, a word of advice: Never ever delete the originals. You have been warned.

I double checked and confirmed it deletes folders

And I've encountered another problem.

If some image files are missing, the entire action is aborted and gives error message :

Editor Chains: Starting chain: Misc Actions
Editor Chains: Misc Actions: starting action No. 1: Compress Images
Editor Chains: action "Compress Images" terminated with exception: 'OEBPS/images/img11.jpg'
Traceback (most recent call last):
  File "calibre_plugins.editor_chains.chains", line 177, in _run_loop
  File "calibre_plugins.editor_chains.actions.compress_images", line 89, in run
  File "calibre\ebooks\oeb\polish\images.py", line 87, in compress_images
KeyError: 'OEBPS/images/img11.jpg'
Traceback (most recent call last):
  File "calibre_plugins.editor_chains.cmdline", line 77, in cmdline_run
  File "calibre_plugins.editor_chains.cmdline", line 53, in run_chain
  File "calibre_plugins.editor_chains.action_chains", line 40, in run_editor_chain
  File "calibre_plugins.editor_chains.chains", line 320, in run
  File "calibre_plugins.editor_chains.chains", line 200, in _run_loop
  File "calibre_plugins.editor_chains.chains", line 177, in _run_loop
  File "calibre_plugins.editor_chains.actions.compress_images", line 89, in run
  File "calibre\ebooks\oeb\polish\images.py", line 87, in compress_images
KeyError: 'OEBPS/images/img11.jpg'

it means, only flawless epub can be properly processed..?

[isarl]

Originally Posted by tatagi

many thanks. yes that's the most foolproof way but at the same time very tiresome to individually edit files one at a time.

Why edit them one at a time? With something like ImageMagick you could batch-convert a bunch. I would likely do it something like this, on Linux. Assumes all your image files are jpegs (with extensions of ".jpg" or ".jpeg") and you want a max filesize of 1MB per image file:

$ unzip -d working_dir "book filename.epub"
$ cd working_dir
$ find . -iname "*.jpg" -o -iname "*.jpeg" -execdir mogrify -define jpeg:extent=1MB '{}' ';'
$ zip -r "book filename.compressed.epub" .
$ mv "book filename.compressed.epub" ..
$ cd ..
$ rm -rf working_dir

Note that ImageMagick's “mogrify” command modifies files in-place. This is (potentially) a DESTRUCTIVE operation!