Allow stripping of sensitive information during github backup process

[patcon]

Re: https://github.com/hyphacoop/worker-coop-scripts#backup2github-command

We have an in-progress script that's intended to backup hackmd and other docs to the more permanent github repo. (not done yet, and not active -- just works locally, but doesn't write to github yet)

Seems that hackmd can have noindex properties, so that even if it's technically public, it won't show up in search indexes: https://github.com/hackmdio/codimd/issues/74

So, assuming that is still an active feature (and respected by search engines), we could have a syntax or notation for stripping certain things from the longterm copy. I could imagine maybe there's

• a notation for anonymizing names, or
• a notation for condensing names down into shorthand acronyms
• a notation to totally redact content so that ██████████████
• a notation to remove content without any indication of it existing (since ppl can trace it back thru yaml front-matter or persistence, and maybe we don't want to hint at where that effort might be fruitful)

Open to feedback!

cc: @uditvira @dcwalk

[patcon]

Some earlier thinking on degrees of anonymization of chat logs, which might bear on the sorts of notations we want to support: https://hackmd.io/3aAIvVu7RsCn4PBInQxDEg#Opt-Out-System

[patcon]

Another idea for processing step:

• Use the config csv (that tells the script what needs backup), and use it to replace any stray hackmd links with their resolved versions in the github repo