[Feature]: Linux Support

[AnoobisE]

Feature Description

Hey there. It would be awesome to have support for Linux.

Motivation / Use Case

Especially now with the growing interest in Gaming, it might be getting more and more interesting for you

Proposed Solution

No response

Alternatives Considered

No response

Additional Context

No response

[Shackless]

Distribution picks so far:

• Arch (Cachy, Manjaro)
• debian (ubuntu)
• Bazzite
• (SteamOS, Mint, Fedora)

[teddybear082]

From what I have learned so far best solution would be to use AppImage format which feels somewhat like the Linux equivalent of pyinstaller where the different distros can all access and use it. Tauri appears to be compatible with AppImage, and I already was able to test a AppImage of core successfully (within reason, that is, that it built and at least loaded without obvious errors).

Some resources: https://v2.tauri.app/distribute/appimage/

https://docs.appimage.org/user-guide/portable-mode.html

build_appimage.sh

Happy to hear from experienced Linux users if I’m on the right track here.

[teddybear082]

Feedback (CachyOS)

• Run using command:

sudo -E WEBKIT_DISABLE_DMABUF_RENDERER=1 ./WingmanAI_3.0.1_amd64.AppImage

Positives

• AppImage loaded just fine, tutorial completed successfully
• Wingman account login fine
• Downloaded local models just fine, recognized models
• Recognized gpu, says it loaded faster whisper with cuda
• Says it loaded pocket-tts model fine
• Benchmark ran fine
• Wingman used LocalAI to announce itself
• Typing to a default wingman generated response as expected
• Selecting an image file in chat works and vision on the loaded image works
• Copy convo to clipboard button works
• Clearing convo history works

Potential Bugs

• Crashes when clicking wrench near a wingman to adjust settings
• Crashes when switching to "General" category (saw error regarding loading autoscreenshot skill, may be certain files for skills do not work on Linux and need OS checks) {However, on reload, reloaded into "General" fine}
• Error in TTS playback: "Error opening RawOutputStream: Invalid sample rate [PaErrorCode -9997] # FIXED in 3.1.0 apparently. Audio working fine, at least when running core from source.
• Drag and drop screenshot does not work
• Links to open audiolibrary or logs do not work # Works with 3.1.0 when running core from source
• General Settings button crashes

[teddybear082]

On the crash when switching to general the error seems to be relate to autoscreenshot: ```Error loading skill 'auto_screenshot': PyGetWindow currently does not support Linux. If you have Xlib knowledge, please contribute! https://github.com/asweigart/pygetwindow
Traceback (most recent call last):
File "/home/teddybear/Github/wingman-ai/wingmen/wingman.py", line 315, in init_skills
skill = ModuleManager.load_skill(
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/teddybear/Github/wingman-ai/services/module_manager.py", line 148, in load_skill
module = import_module(config.module)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.11/importlib/init.py", line 126, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "", line 1204, in _gcd_import
File "", line 1176, in _find_and_load
File "", line 1147, in _find_and_load_unlocked
File "", line 690, in _load_unlocked
File "", line 940, in exec_module
File "", line 241, in _call_with_frames_removed
File "/home/teddybear/Github/wingman-ai/skills/auto_screenshot/main.py", line 6, in
import pygetwindow as gw
File "/home/teddybear/Github/wingman-ai/venv/lib/python3.11/site-packages/pygetwindow/init.py", line 347, in
raise NotImplementedError(
NotImplementedError: PyGetWindow currently does not support Linux. If you have Xlib knowledge, please contribute! https://github.com/asweigart/pygetwindow

[teddybear082]

On switching back to star citizen, client just crashes, no error in terminal when running core from source