Update docker images to latest versions

[jay-m-dev]

Update dbmongo, lab, machine and integration and unit tester docker machines to latest compatible versions

[jay-m-dev]

Tested the following package versions as working well together:
machine:

• scikit-learn==0.23.2
• tqdm==4.64.1
• pymongo==4.3.3
• pandas==1.0.5
• numpy==1.19.5
• matplotlib==3.5.3
• mlxtend==0.16.0
• pydot==1.4.2
• requests==2.25.1
• xgboost==1.6.2
• joblib==1.2.0

lab:

• scikit-learn==0.23.2
• tqdm==4.64.1
• pymongo==4.3.3
• pandas==1.0.5
• numpy==1.19.5
• cython==0.29.32
• xgboost==1.6.2
• simplejson==3.17.2
• requests==2.25.1
• joblib==1.2.0
• -e git+https://github.com/lacava/surprise.git@1.1.1.1#egg=scikit-surprise

All but scikit-learn, pandas and numpy have arm64 versions readily available. To make this work on arm64, we'll need to compile wheels for these versions.

Mongo DB can also be updated to version 4.4.18 running on slim-buster
NodeJS can be updated to v16.18.1
These versions of Mongo DB and NodeJS have arm64 versions available.

Versions higher than the ones listed here were found to have breaking changes on the current code base.

[jay-m-dev]

All versions are now updated. The .env file contains variable ARCH that needs to be set to either x64 or arm64 depending on the developer's system. Building on arm64 currently will take a long time until we have the latest wheels updated and we modify the Dockerfiles to pull the wheels.
As the purpose of this issue is to only update versions and this has been accomplished, this issue can be closed. A separate issue has been created (issue #498) to fix the slow build speeds on arm64.