Very simple command line client for OpenAI API LLMs (ChatGPT 4)
Default config expected in a file called chat_config.json, example:
{ "url":"https://api.openai.com/v1/chat/completions",
"api_key":"ENV_OPENAI_API_KEY",
"model": "gpt-3.5-turbo",
"available_models": [
"gpt-4o-mini","gpt-3.5-turbo"
],
"role":"user",
"temperature":"0.5" }It is possible to store the API key directly in the config file but for security it is recommended to use an environment variable instead.
Set the api_key in the config file to value ENV_OPENAI_API_KEY then and provide an environment variable called OPENAI_API_KEY with the key.
To start chatting in windows open a command shell and type:
set OPENAI_API_KEY=bearer sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
python3 chat.py
User Commands:
!exit: Exit chat!debug: Print last response json from LLM!models: Show available models!T=X: Set temperature to X!model=m: Set model to m (m Integer, see !models command)!context_length=n: Integer, number of previous messages put in the request for context!new_context: Start new dialog!toggle_logging: Switch on/of logging, deafault: On!list_configs: List config files (json) in current folder!load_config <filename>: Load configfile!save_config <filename>: Save current config!print_config:Print current config!help: Print this text