Adding csv archiving

[edmoss345]

Changes to pipeline to accommodate archive functionality

[edmoss345]

Three PRs are connected as this archive process required an update to the SA repo, the pipeline and the toolkit
#20
IDEMSInternational/parenttext-south-africa#7
IDEMSInternational/rapidpro-flow-toolkit#99

[edmoss345]

@istride I've made some changes as we discussed. Have pulled the 'archive' command into its own file so it is not part of the general pipeline. However, have been running into problems with the archive process, doesn't seem to save the csvs correctly when there is a ":" in the name of the sheet, but can't pin down where exactly it's failing and can't find a fix. Example output can be found in the SA repo PR. One for us to discuss

https://github.com/IDEMSInternational/parenttext-south-africa/pull/7/files#diff-82e569e3d1dd3c813c81a8ee5044604d27b37acf72c960ebe1b7883b32f86e8d

[istride]

I think the problem with the colon character (":") is specific to Windows. I managed to create an archive using the code in this PR, on my Linux machine, with colon characters intact. In the short-term, we can just use a Linux system to create the archive - via Docker, a VM or Github Actions.

There is potentially a bigger problem here because I can't guarantee that there aren't characters that would not work in Linux file names. There a couple of ways we might deal with this:

• create a 'safe' file name for each sheet and keep a mapping from the safe name to the real sheet name so that we can restore it when it's needed
• define a set of invalid characters that may not be used in sheet names and make sure that they are not used