Container improvements

[Robinsane]

1. added 3 env variables to make life easier configuring shell_gpt containers:
   IN_CONTAINER: setting this one to true, helps fix some default things that aren't handy when in a container:

• The default for interaction becomes False, since [E]xecute on host doesn't work when you get the output from a container
• Shell_gpt will not check the machine for os and shell (which would be fetched from the container), but will check what the user has configured in following env vars: OS_OUTSIDE_CONTAINER and SHELL_OUTSIDE_CONTAINER

2. adjusted Dockerfile to use new env vars + added an example ollama Dockerfile

[TheR1D]

Could we also update Config and Docker section in README.md please?

[TheR1D]

This should be an argument, so user can setup the image with specific OS and SHELL names.

[Robinsane]

In the upcoming commit I'll leave it as ENV variable that is in the config.py DEFAULT_CONFIG.
Reason being that it might be handy to leave it interchangeable, e.g.:

• All of a sudden you need to work on a server with an os / shell you're very unfamiliar with. You can temporarily overwrite your os and shell name of your often used shellgpt to mock this environment, without having to install on that server / install in a similar virtual local environment
• Similar to above, if I'd like to use shellgpt containers for multiple setups: RHEL, Debian, Ubuntu, powershell, MacOs, ... I can run the same image each time, as long as I overwrite the OS and shell to mock these envs. If I had to make a different image each time that'd be a waste of diskspace and image build time in my opinion.

[ismaelchris]

These settings worked for me:

IN_CONTAINER=false
USE_LITELLM=true

[TheR1D]

Suggested change

	True if cfg.get("SHELL_INTERACTION") != "false" else False,
	cfg.get("SHELL_INTERACTION") == "true",

[Robinsane]

I wrote it this way to ensure the normal default behaviour is set, also when SHELL_INTERACTION contains an unintended / weird value

[TheR1D]

Suggested change

	ENV SHELL_INTERACTION=false
	ENV OVERWRITE_OS_NAME=""
	ENV OVERWRITE_SHELL_NAME=""
	ARG OVERWRITE_OS_NAME=default
	ARG OVERWRITE_SHELL_NAME=default
	ENV SHELL_INTERACTION=false
	ENV OVERWRITE_OS_NAME=${OVERWRITE_OS_NAME}
	ENV OVERWRITE_SHELL_NAME=${OVERWRITE_SHELL_NAME}

[Robinsane]

I believe you misunderstand Docker's ARG:
These are env variables that only exist during the building process of the container image.
Once the image exists, these are gone forever.
Please refer the dockerfile arg documentation for a better, more complete explanation.

ENV sets environment variables that remain, and can be used set as arguments with --env as desired when starting the container