Extend the ability to support third party LLM through customizable API URL

[vaderyang]

Add api_url , max_tokens into settings. These settings extends the aichat to support third party LLM services which using OPENAI-like API such as llama-cpp-python[server], LocalAI and etc.

Example:
start local API service first, below is an example for aichat with a local codellama on a MacBook laptop

python3 -m pip install llama-cpp-python
python3 -m llama_cpp.server --n_gpu_layers 1 --model ~/models/codellama_13b.gguf --port 3000

aichat -I http://localhost:3000/v1/chat/completions --max_tokens 2048

[vaderyang]

if this project is dead, should I branch from this and grow my own version?

[jacksongoode]

I would support it @vaderyang !

[sigoden]

see #150