HELP! with a personal project. Vosk STT

[IncredibleGeek1]

Body

guideline says give as much info... that's a lot. i've been working on this for exactly a month. (thank God for file properties). instead of this being a LONG winded and extensive comment. cause no one reads those. i'll try to give summary bullet points. PLEASE ask for clarification on ANYTHING. i don't have a normal setup.

Forewarning i don't know how to code I could read python Ish, but most of my code was all created by AI. I just use my IT skills to find working code and integrate to stacking code on top of code. When it doesn't work or AI broke it, or is being stubborn. I revert back to a previous version and then repeat the process all over again. I am kind of learning certain concepts through this process making it easier to discern where to put what.... Sort of. Python is stupidly strict on fricking indenting and syntax. with that out of the way.

1. Commands do not work when the GUI is a module.
2. All features mostly worked whenever the two scripts were together or separate using path to find the main script.
3. Once attempting to put exe compatibility It broke the script. Hence the move to a module where exe and python are no longer needed as compatibility.
4. By compatibility i mean I'll be convert python to a single exe. Hence the need for exe compatibility.
5. Several of the features but not detrimental to the functionality of the project. So don't worry about it, Unless you feel so inclined.
6. Most configuration based things are on json's. To prevent hard coding. That way when compiled it can be configurable still.
7. I do have a words to numbers script However even though it's kind of not working at least it's adequate. It only goes up to like 999. However i wish for to go to at least a million but i also do have json scripts to go all the way up to centillion.
8. I do have three models however i'm working mainly on gigaspeech. Small medium and gigaspeech.
9. My GUI has went through several iterations i am now on dearpygui.

FYI most of whatever i wrote down as the bullet points was all through my program. Yeah i never said dictation and transcripts didn't work i said commands don't. Unless i revert to an older version that still has paths in which case commands semi work.

My project

Please ask any questions if he needed any type of clarification i am more than happy to explain. This is my first major project so i have no idea how you guys want it to be formatted i'm going off of the guideline from up top. i'm trying to be detailed yet concise.

Here are some example logs showing the problem:

Partial: move left
Detected command: move left
{"type": "command", "action": "cmd_move_left", "params": {"key": null, "transcribed_text": ["Some text"]}}

As for the kudos... Is that like a tag like on steam but github's version? If it's just being appreciative than that's understandable. Called common courtesy which isn't so common but whatever. Fixing commands will get a shout out on the read me. Fixing words to numbers same thing. Fixing relative sensitivity same thing. Fixing debug audio which i have explanation on my repo about that, same thing. Separating waveform from the actual other window same thing. For waveform i never actually attempted to, but i know if i do i'll break my damn code again.

Guidelines

• I have read and understood this category's guidelines before making this post.