DockerCommandLineCodeExecutor support for additional volume mounts, exposed host ports

[andrejpk]

What feature would you like to be added?

I'm using DockerCommandLineCodeExecutor in a project that uses additional system admin tools that need config files and access to specific network ports to operate (e.g. kubectl, helm).

In DockerCommandLineCodeExecutor, add these arguments to the __init function:

        extra_volumes: Dict[str, Dict[str, str]] = {},
        extra_hosts: Dict[str, str] = {},
        init_command: str = ":",

These are all additive to what the docker container has now and if left to defaults, don't break existing uses of DockerCommandLineCodeExecutor

I have my own fork that implements this and will create a PR for this if it looks like a good idea for the project.

Why is this needed?

Enable the agents to:

• work with a richer set of sys admin tools on top of code execution
• add support for a 'project' directory the agents can interact on that's accessible by bash tools and custom scripts

[ekzhu]

Thanks for the suggestion! I think the addition make sense to me. PR welcome!