Fix missing import for UI and added windows support where tensorrt is not available

[DocGarbanzo]

Fix missing imports and supporting of windows where tensorrt is potentially not working

1. Two imports that are used in the UI were not correctly added to the install files

• pyyaml was missing in conda install files
• pyyaml and plotly were missing in setup.py

2. Bumped the superlinter to v4 because of issue: https://github.com/github/super-linter/issues/2253

3. Also, moved tensorrt imports into function, so loading will not fail unless a tensorrt conversion is called. It seems windows 11 doesn't support tensorrt yet.

4. Bumped version bugfix digit

[Ezward]

Doing Windows conda install using the branch, after successfully recreating the donkey conda environment and activating, my system does not find pip. I'm not sure if installing pip is an explicit part of the installation. I'm running the commands from the Windows Powershell.

conda update -n base -c defaults conda
conda env remove -n donkey
conda env create -f install\envs\windows.yml

That all succeeded, then:

PS C:\Users\emurm\projects\autorope\donkeycar> conda activate donkey
PS C:\Users\emurm\projects\autorope\donkeycar> pip install --user tensorflow==2.2.0
pip : The term 'pip' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if
a path was included, verify that the path is correct and try again.
At line:1 char:1
+ pip install --user tensorflow==2.2.0
+ ~~~
    + CategoryInfo          : ObjectNotFound: (pip:String) [], CommandNotFoundException
    + FullyQualifiedErrorId : CommandNotFoundException
```
It seems the issue is powershell.  If I use the legacy command line and conda activate donkey that worked and I was able to run pip.

So perhaps we need to explictly say to use the legacy command console and not powershell.

[Ezward]

Attempting to install the version of tensorflow-gpu specified in the install instructions failes

(donkey) C:\Users\emurm\projects\autorope\donkeycar>conda install tensorflow-gpu==2.2.0
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.

PackagesNotFoundError: The following packages are not available from current channels:

  - tensorflow-gpu==2.2.0

Current channels:

  - https://repo.anaconda.com/pkgs/main/win-64
  - https://repo.anaconda.com/pkgs/main/noarch
  - https://repo.anaconda.com/pkgs/r/win-64
  - https://repo.anaconda.com/pkgs/r/noarch
  - https://repo.anaconda.com/pkgs/msys2/win-64
  - https://repo.anaconda.com/pkgs/msys2/noarch

To search for alternate channels that may provide the conda package you're
looking for, navigate to

    https://anaconda.org

and use the search bar at the top of the page.

(donkey) C:\Users\emurm\projects\autorope\donkeycar>conda search tensorflow-gpu
Loading channels: done
# Name                       Version           Build  Channel
tensorflow-gpu                 1.8.0      h21ff451_0  pkgs/main
tensorflow-gpu                 1.9.0      hf154084_0  pkgs/main
tensorflow-gpu                1.10.0      hf154084_0  pkgs/main
tensorflow-gpu                1.11.0      h0d30ee6_0  pkgs/main
tensorflow-gpu                1.12.0      h0d30ee6_0  pkgs/main
tensorflow-gpu                1.13.1      h0d30ee6_0  pkgs/main
tensorflow-gpu                1.14.0      h0d30ee6_0  pkgs/main
tensorflow-gpu                1.15.0      h0d30ee6_0  pkgs/main
tensorflow-gpu                 2.0.0      h0d30ee6_0  pkgs/main
tensorflow-gpu                 2.1.0      h0d30ee6_0  pkgs/main
tensorflow-gpu                 2.3.0      he13fc11_0  pkgs/main
tensorflow-gpu                 2.5.0      h17022bd_0  pkgs/main
tensorflow-gpu                 2.6.0      h17022bd_0  pkgs/main

Attempting installation conda install tensorflow-gpu==2.3.0 fails with lot and lots of package conflicts

The fix here to install with pip pip install tensorflow-gpu==2.2.0, so we should change the docs to reflect this.

[DocGarbanzo]

@Ezward - thanks for the thorough testing, that's quite useful to understand potential issues for windows users:

1. pip is explicitly listed in the conda yaml file, so it will be installed in the donkey env. I have no idea why this wouldn't work in powershell, but I have experienced windows as highly non-predictable, so probably no big surprise. Anyway, the docs explicitly state one should run these commands out of the conda prompt, which in my experience opens a standard command prompt in windows, not a powershell, so we should be good.
2. You are right we need to install tensorflow or tensorflow-gpu with pip. The docs are already reflecting this, so I think we should be good on that front, too.
3. I'm checking why the linter CI falls over.... :-/

[DocGarbanzo]

I've bumped the superlinter to version 4 as well, because we are hitting the bug described here: https://github.com/github/super-linter/issues/2253

[Ezward]

Thank you for doing this. I think windows users will appreciate it. I'm a linux user, but I found it much easier to actually get the gpu working in windows than I did in linux.