SD-50 Extract data via single step

[adamrtalbot]

This is a preliminary PR for extracting the data using a Python script instead of the shell and JQ stuff.

Currently it offers the following advantages:

• Extract data from all pipeline runs can be performed in the same step, meaning 1 job runner.
• All data can easily be combined into a single output file which can be browsed later.
• There is more control over logic (e.g. if/when is easier). e.g. force delete becomes easier to clear out all runs.
• It can upload a logging file with the Slack message
• Probably easier to maintain long term

Downsides:

• the Tower CLI is missing some key variables, meaning we can't recreate the exact response we were getting before.
• The slackbot requires more elevated permissions (but is easier to use with slack_sdk.
• Errors will be hidden. If a single pipeline fails it may cause the whole pipeline to collapse. The old one was more fault tolerant.
• The input options no longer work properly.

To dos before we can accept PR:

• Increase fault tolerance and logging to give us better feedback
• Include attributes added to Tower CLI in https://github.com/seqeralabs/tower-cli/pull/364
• Incorporate input options (slack, delete, force) which will allow precise control for debugging.

[adamrtalbot]

Closed in favour of #14