Unable to convert with flatpak

[Jopp-gh]

Hello all!
first, thanks for making tools like PyGlossary!
Linux mint cinnamon-user here.

Deb package

The deb version (3.2.1-1.1) doesn't launch at all, I tried to look for more recent versions but there isn't any. Also, I've python 3.10.
Console errors:

$ pyglossary

Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/pyglossary/plugins/babylon_bgl/bgl_reader.py", line 30, in <module>
    GzipFile = __import__(
ModuleNotFoundError: No module named 'pyglossary.plugin_lib.py310'
If things didn't work well with BGL files, install Python 3.5 to 3.7 and try again
Invalid plugin directory: '/home/username/.pyglossary/plugins'
Traceback (most recent call last):
  File "/usr/share/pyglossary/pyglossary.pyw", line 215, in <module>
    log.setVerbosity(args.verbosity)
AttributeError: 'RootLogger' object has no attribute 'setVerbosity'

flatpak package

the flatpak version (5.0.7) instead, is much more recent, launches but throws critical errors and doesn't convert anything, after supplying input/output -paths and -formats:

[CRITICAL] _createReader: readerClass is None

I'm trying to convert from the html AppleDict format. As you can see, dependencies are satisfied:

Requirement already satisfied: lxml in /usr/lib/python3/dist-packages (4.8.0)
Requirement already satisfied: biplist in ./.local/lib/python3.10/site-packages (1.0.3)

later I installed lxml, biplist with admin rights even if this is bad practice but a system-side installation of these dependencies didn't help either!

Steps

Btw, this isn't for me the first time to launch PyGlossary, so far I haven't had luck to convert any dictionary, asking kindly for assistance.

What I did so far:

1. I selected the Body.data as input (PyGlosssary picked did choose automatically the correct format)
2. I generated a simple, empty file as output file
3. choosing Output format: stardict's ifo
4. pressing Convert I get:

[WARNING] Module 'lxml' not found in /app/lib/python3.12/site-packages/pyglossary/plugins/appledict_bin, skipping plugin 'appledict_bin
[CRITICAL] _createReader: readerClass is None

I've no /app at root, furthermore flatpak needs to handle system paths differently because running inside a sandbox.

[Ashvith10]

You need to make use of venv first:

# Assuming that you have already install the latest Python,
# create a virtual environment of the name "pyglossary"
# or whatever other name you prefer.
$ python3 -m venv pyglossary --system-site-packages

# Enter the said environment.
$ source pyglossary/bin/activate

# Now, install pyglossary and the other dependencies
# depending on whether you want to use UI frontends
# (gtk3, gtk4, tkinter), cmd or web server, as well as the
# additional dependencies for the format you need to
# to convert to and from.
$ pip install pyglossary <add other deps like lxml, pyicu, etc here>

# Then, you can run
$ pyglossary --ui-<your preferred ui>

# After you are done, leave the venv environment
$ exit

[Jopp-gh]

Oh, wow thanks a lot for your kind and useful answer! I see, virtual env allow better versions of standard py.

Meanwhile, I uninstalled the outdated versions distributed by my distro, installed pyglossay directly with pip and did convert dictionaries directly to stardict's format without further ado.

So far so good, Goldendict loads my new dictionaries. Only pity, that this conversion didn't apply styling rules, example, all lists are put in the same block together with all longer expansions. Usually this isn't a problem, when the explanations aren't long, still I didn't convert only smaller dicts like Latin. Some dicts are quite voluminous and come with deeper explanations, example oxford en and others.

Therefore, I re-tried another conversion path and followed the tutorial from:
https://github.com/yell0wsuit/appledict2semantic

After converting to xml with pyglossay and extracting the XML file to HTML, I was not unable to conclude with appledict2semantic bc of bugs. Hopefully my bug report is read, to get also text with "beautiful styling".

[Ashvith10]

@Jopp-gh What type of AppleDict are they? Binary? Source? Depending on what they are, you can provide additional flags. Personally, I like using the other command-line format, instead of the interactive one:

pyglossary <input-dictionary> <output-dictionary> \
--read-format=<add-read-format> \
--write-format=<add-write-format> \
--write-options="abc=def; ghi=True"

Check https://github.com/ilius/pyglossary/blob/master/doc/p/appledict_bin.md and https://github.com/ilius/pyglossary/blob/master/doc/p/appledict.md. Depending on which format you have, I think you'll have some degree of control.

I am also having some issues with converting Wordnet to Aard2Slob, but that's probably because of all the available tags:

pyglossary/pyglossary/plugins/aard2_slob/tags.py

Lines 1 to 8 in 9dde242

	t_created_at = "created.at"
	t_label = "label"
	t_created_by = "created.by"
	t_copyright = "copyright"
	t_license_name = "license.name"
	t_license_url = "license.url"
	t_uri = "uri"
	t_edition = "edition"

Only these were allowed:

pyglossary/pyglossary/plugins/aard2_slob/writer.py

Lines 12 to 17 in 9dde242

	from pyglossary.plugins.aard2_slob.tags import (
	t_created_at,
	t_created_by,
	t_label,
	t_uri,
	)

[Jopp-gh]

@Ashvith10

Oh, my 🍎dictionaries are in binary format. I tried running commands both in terminal and the GUI of Pyglossary, but even allowed tags didn't work and threw errors instead.

Such a shame that I can't use:
https://github.com/yell0wsuit/appledict2semantic

The same poster (yell0wsuit) asked Ilius also for a pull to improve the code of Pyglossary for as better, final conversion.
Ilius did the heavyweight lifting, why not add some final touch with a meaningful styling? No idea..

Trying to edit the generated css files of my Stardict (ifo) conversions didn't work so far.
Thanks for trying to help, very appreciated!

Bad example (converted):

Good example (not converted)