Improve docker setup & config

[Pwuts]

Background

The current Dockerfile and docker-compose file seem to have unnecessary convolution of configuration and don't contain an optimal configuration out of the box. Also, it is not consistent with the README. This PR attempts to fix that.

This PR either conflicts with or deprecates PRs #1830 and #1199

Changes

• Merge apt-get actions into one layer
• Uncouple app user and python dependencies: install packages globally instead of in /home/appuser
• Create appuser with UID 1000 to match probable permissions on Linux host
• Change workdir to /app (as documented in the README) which is a widely used standard, making it easier to work with for new users
• Improve/fix docker-compose config
  • Use Redis container as memory backend by default
  • Remove unused (and after this PR, redundant) volume mount to /app
  • Mount ./logs/, ./auto_gpt_workspace/ and ./ai_settings.yaml in /app

Documentation

New configuration is consistent with current configuration; no update needed.

Test Plan

No Python code is changed, so testing succeeds if the container builds and runs properly after these changes, which it does.

PR Quality Checklist

• My pull request is atomic and focuses on a single change. (as far as I understand this checkbox)
• I have thoroughly tested my changes with multiple different prompts.
• I have considered potential risks and mitigations for my changes.
• I have documented my changes clearly and comprehensively.
• I have not snuck in any "extra" small tweaks changes

[github-actions]

This pull request has conflicts with the base branch, please resolve those so we can evaluate the pull request.

[github-actions]

Conflicts have been resolved! 🎉 A maintainer will review the pull request shortly.

[ntindle]

May be worth adding
Chrome for selenium in docker to the compose file

https://hub.docker.com/r/selenium/standalone-chrome/

See here for examples

https://github.com/SeleniumHQ/docker-selenium

[Pwuts]

@ntindle currently, Chrome and Firefox along with some other tools are installed in the docker container, see Dockerfile. I'm not convinced this is the way, but also don't have time to dive into it further atm.

Do you think it is worth it to use standalone selenium containers instead of running headless browsers in the auto-gpt container?

[ntindle]

IMO using the selenium headless versions with their drivers makes a lot of sense. Otherwise we’d basically need to duplicate the work they’ve done and will continue to do when things change

[Pwuts]

Could you make a PR to implement that as soon as this one merges? Would be nice :)

[ntindle]

Can do

[Pwuts]

@mikekelly care to review?

[mikekelly]

Quite a bit of this is covering ground from #1199 and will conflict with it

[ntindle]

This combined with a smoke test of the docker container in CI would be top notch

[Pwuts]

@ntindle this PR fixes docker; a draft for a smoke test job is in #3059