CRON: Delete demo docs

[jvwong]

Should be self-explanatory. Can identify demo docs via DEMO_SECRET.

Related to #995

[gbader]

this should not delete just created docs, so the docs should have a defined lifetime before they are deleted.

[jvwong]

Good point, generally we have a lot of stuff (auto-emailed related-author docs, user-initiated but old/abandoned) along with demos that we'll have to figure out how to handle.

[jvwong]

How to consider docs for automated removal:

i) Trashed docs
Remove immediately

ii) Demo docs
Remove after short period of inactivity

ii) Initiated docs
Remove after long period of inactivity AND

• entries is empty
• article is empty

[jvwong]

There seems to be a bug in the document lastEditedDate type.

In general, it looks like docs with no entries/unedited are type PTYPE<TIME> vs STRING for ones with entries/edited, but there are exceptions (e.g. 2013fc15-ddb3-4f6c-8d83-911dc7af4f1d; 159f2c80-eb33-4ef4-94be-867216479041)

[jvwong]

I'm guessing the time types are reflecting whether this was set in browser (updateLastEditedDate) vs server?

[maxkfranz]

It's NoSQL, so it depends on the value you send to the DB. Regardless of where the inconsistency is, the API (probably the Document) should simply enforce a normalisation.

[jvwong]

It's NoSQL, so it depends on the value you send to the DB. Regardless of where the inconsistency is, the API (probably the Document) should simply enforce a normalisation.

OK I'll look at this.

Back to the main issue - is it better to simply update these docs we don't want as 'trashed' or delete them entirely?

[maxkfranz]

I'm fine leaving the documents marked as 'trashed' in the DB for now. It's not like they're using up that much space, right? As long as demo docs and trashed docs aren't accessible in the BioPAX data exports, I think that's the main thing.

[jvwong]

I like trashing option better. Always easy to clear those out.