Zim plugin to inline images in HTML files

htmlimageinliner.py

# Copyright 2022 introt <[email protected]>

# based on windowtitleeditor.py and Zim Desktop Wiki code released under GPLv2

import base64

from pathlib import PurePath

from zim.formats.html import Dumper, html_encode

from zim.plugins import PluginClass
from zim.gui.mainwindow import MainWindowExtension

import logging
logger = logging.getLogger('zim.plugins.htmlimageinliner')


class HtmlImageInlinerPlugin(PluginClass):
	plugin_info = {
		'name': _('HTML Image Inliner'),  # T: plugin name
		'description': _('''\
Enabling this plugin makes the html exporter inline images.
To restore original functionality, disable the plugin.

Limitations:

- Only images attached from the current dir are inlined.
  Other images break.

- Inlined images are still copied into the "page_files"
  directory on export; the files therein can be safely
  deleted.
'''),  # T: plugin description
		'author': 'introt',
	}


class HtmlImageInlinerExtension(MainWindowExtension):
	def __init__(self, plugin, window):
		MainWindowExtension.__init__(self, plugin, window)  # init super

		self.old_dumper = Dumper.dump_img
		Dumper.dump_img = lambda d, t, a, s=None: self.inline_img(d, t, a, s)

		logger.debug('HtmlImageInliner loaded')

	def inline_img(self, dumper, tag, attrib, strings=None):
		l = dumper.linker
		src = l.img(attrib['src'])

		try:
			# XXX: quick hack; instead of improving,
			# please use the time to create a real implementation
			a_dir = PurePath(str(l.notebook.get_attachments_dir(l.source)))
			path = a_dir / PurePath(str(src)).parts[-1]

			with open(path, 'rb') as f:
				src = f'''data:image/png;base64,{base64.b64encode(f.read()).decode('utf-8')}'''
		except:
			pass
		
		opt = ''
		if 'alt' in attrib:
			opt += ' alt="%s"' % html_encode(attrib['alt']).replace('"', '&quot;')
		for o in ('width', 'height'):
			if o in attrib and int(float(attrib[o])) > 0:
				opt += ' %s="%s"' % (o, attrib[o])
		if 'href' in attrib:
			href = dumper.linker.link(attrib['href'])
			return ['<a href="%s"><img src="%s"%s></a>' % (href, src, opt)]
		else:
			return ['<img src="%s"%s>' % (src, opt)]

	def teardown(self):
		Dumper.dump_img = self.old_dumper
		logger.debug('Original dump_img  method restored')

I gather that you're on a Linux system - what's your distribution and how did you install Zim? If Zim is not running in some sandbox it should have access to $PATH, so gedit with or without path should work.

Another way to get the log is to execute Zim from the command line: zim -D > zim.log 2>&1; you'll then find the output file in the current directory, assuming a POSIXy shell.

I'm running linux Pop!_OS 21.04, which is based on Ubuntu 21.04
and have Zim 0.74.3 installed using flatpack.
I cannot find the executable, I get:

(base) alobo@pop-os:~$ which zim
(base) alobo@pop-os:~$ zim
Command 'zim' not found, but can be installed with:
sudo apt install zim
(base) alobo@pop-os:~$ 

As it had not been installed.

But I have been able to locate a lot of zim.log files. The latest in

/home/alobo/.var/app/org.zim_wiki.Zim/cache/tmp/zim-edrucvc6
export-log.txt
zim.log

Anyway, I can load your plugin now, perhaps I needed to reboot.
But I have generated a Test.html with the plugin activated and still get a Test_files
directory with the images:

(base) alobo@pop-os:~/WORK/HSI_Lab/weblabEcodivResponse/Test_files$ ls -l
total 5580
drwxr-xr-x 2 alobo alobo    4096 ene  9 07:47 01_Equipment
-rw-r--r-- 1 alobo alobo  263156 ene  9 07:47 01_Equipment.html
drwxr-xr-x 2 alobo alobo    4096 ene  9 07:47 02_Technical_Notes,_Tests_and_Demos
-rw-r--r-- 1 alobo alobo  126071 ene  9 07:47 02_Technical_Notes,_Tests_and_Demos.html
drwxr-xr-x 2 alobo alobo    4096 ene  9 07:47 03_Protocoles_and_Manuals
-rw-r--r-- 1 alobo alobo    5594 ene  9 07:47 03_Protocoles_and_Manuals.html
-rw-r--r-- 1 alobo alobo    5949 ene  9 07:47 04_In-house_Software.html
-rw-r--r-- 1 alobo alobo    4588 ene  9 07:47 05_DataSets.html
drwxr-xr-x 2 alobo alobo    4096 ene  9 07:47 06_Research_Studies
-rw-r--r-- 1 alobo alobo  366089 ene  9 07:47 06_Research_Studies.html
-rw-rw-r-- 1 alobo alobo   64240 oct  1  2021 figure1.jpg
-rw-rw-r-- 1 alobo alobo   48845 oct  2  2021 figuresmalllab001.jpg
-rw-rw-r-- 1 alobo alobo   46623 oct  1  2021 figuresmalllab.jpg
-rw-r--r-- 1 alobo alobo 3178239 ene 20  2021 LabImagingSpectroscopy_geo3bcn.pdf
-rw-r--r-- 1 alobo alobo   25877 sep  9  2021 pasted_image001.png
-rw-r--r-- 1 alobo alobo   51312 sep  9  2021 pasted_image002.png
-rw-r--r-- 1 alobo alobo  456301 sep  9  2021 pasted_image003.png
-rw-r--r-- 1 alobo alobo   68466 sep  9  2021 pasted_image004.png
-rw-r--r-- 1 alobo alobo  400639 sep  9  2021 pasted_image005.png
-rw-r--r-- 1 alobo alobo  442556 sep 28 19:33 pasted_image006.png
-rw-r--r-- 1 alobo alobo  104586 sep  9  2021 pasted_image.png
drwxrwxr-x 2 alobo alobo    4096 oct  2  2021 _resources

I link both log files:
https://www.dropbox.com/s/kar18g0qyesdi6h/zim.log?dl=0
https://www.dropbox.com/s/m0kjs8qwsm2iq71/export-log.txt?dl=0
and the html file:
https://www.dropbox.com/s/cxbguh3l4w7yy1h/Test.html?dl=0

Thanks a lot for your support.

using flatpack

That would explain gedit not being found; see "Flatpak: Could not find executable when trying to open files with external application" zim-desktop-wiki/zim-desktop-wiki#2169 for more information

I cannot find the executable

According to the Flathub page, flatpak run org.zim_wiki.Zim -D > zim.log 2>&1 should work

Anyway, I can load your plugin now, perhaps I needed to reboot.

Great! The plugin does a monkey patch at runtime so that might indeed be required.

But I have generated a Test.html with the plugin activated and still get a Test_files
directory with the images

https://gist.github.com/introt/fed672dfbea9df399610084be448326b#file-htmlimageinliner-py-L31-L33

find $EXPORTDIR -type f -not -name '*.html' -exec rm {} or similar at your own risk

Hope this helps! I haven't tested exporting multiple pages at once so you might need to use the CLI to export them one by one.

Thanks.
I have deleted the directory Test_files and Test.html shows all images.