The chatgpt-gui from chaoticroman

chatgpt-gui's Introduction

Hello,

I hold a degree in physics with a specialization in scientific computing and computer-controlled experimentation. Since embarking on my professional journey in 2009, I have adopted Python as my primary programming language. Additionally, C and C++ play a pivotal role in my work, especially within the domain of embedded systems, where my expertise with these languages is particularly pronounced.

My professional background encompasses roles in scientific computing at the Czech Academy of Sciences, specifically within the Astrophysical Institute and the Institute of Atmospheric Physics. I have also contributed as a DevOps engineer at Red Hat and have undertaken projects across the medical, aerospace, and industrial sectors.

Presently, I am an independent engineer with a focus on artificial intelligence. I dedicate my time to training and mentoring fellow engineers and providing consultancy services to companies navigating complex challenges.

Should you wish to learn more or get in touch, please visit my homepage at romanpavelka.cz.

chatgpt-gui's People

Contributors

Stargazers

Watchers

chatgpt-gui's Issues

Liniting and style testing shall be part of CI

Add timing info

It would be nice to add timing to Info object to know how long it takes to get the response.

Cli is not handling multiline input

Very nasty if you would like to discuss code snippet.

Switch to GPT-4-32k when context would grow over 8k limit

GPT-4-32k is twice as expensive so let's use GPT-4 with 8k as default and switch when it will get too long only.

Ruff is not really failing on mistake

See: https://github.com/ChaoticRoman/chatgpt-gui/actions/runs/6883273810/job/18723483058

or try unexpected indent in random python file

GUI

CLI should show that request is in the progress.

Switch code review to GPT-4-32k when available

https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4

README issues

README talks about GUI that is still not there.
CLI can mention quitting with CTRL+C and CTRL+D.
API token as an environmental variable is not

History of past conversations

would be nice to save it to user data directory as a JSON
conversation saving code should reside in core.py and conversation created in some ui frontend should be transparent for other frontends as well
you should be able to continue with a converstation, this is going to be ui frontend addition as well