[Linux] Run with local ollama

[omnigenous]

I was following a wiki: https://github.com/TheR1D/shell_gpt/wiki/Ollama

But could get it to work with shell_gpt, works fine with ollama run mistral:7b-instruct

Running sgpt --model ollama/mistral:7b-instruct "Who are you?" gives error:

Usage: sgpt [OPTIONS] [PROMPT]
Try 'sgpt --help' for help.
╭─ Error ─────────────────────────────────────────────────────────────────────────────────────────────╮
│ Missing config key: OPENAI_API_KEY                                                                  │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────╯

I tried changing DEFAULT_MODEL to ollama/mistral:7b-instruct, OPENAI_BASE_URL to localhost, OPENAI_USE_FUNCTIONS to false and OPENAI_API_KEY to false in ~/.config/shell_gpt/.sgptrc. Latest change gives long error message so I revert it.

I also saw: #336 (comment)

Tried:

git clone https://github.com/TheR1D/shell_gpt.git
cd shell_gpt
git checkout ollama
python -m venv venv
source venv/bin/activate
pip install -e  .

sgpt --model ollama/mistral:7b-instruct --no-functions "Hi Ollama"

Same "Missing config key: OPENAI_API_KEY" error

[TheR1D]

Hi @omnigenous make sure you have OPENAI_API_KEY set to any value in ~/.config/shell_gpt/.sgptrc.

[omnigenous]

Installing shell-gpt version 1.3.0-1 from aur worked.

And like @TheR1D said in #467 (comment), don't forget to set OPENAI_API_KEY to any value like false in ~/.config/shell_gpt/.sgptrc

Remember to run local ollama ollama serve in background ;)

[gwpl]

Seems like new issue arised: #503