CRON: Document update tasks

[jvwong]

Two changes to the cron job for document updates.

1. Skip updates to relatedPapers: Drop this rather heavy, long-running task (see Examining the database data / Root of large data dump size  #1024 (comment)). It's not completely clear that updates to these make any tangible difference (e.g. the most recent papers of interest).

2. Filtering docs to update: Strips out those documents without any paperId, that is, documents that were auto-created to notify referenced authors that their paper was cited by a Biofactoid document (viral email).

In this way, documents will get fresh data for the submitted article, and those submitting a paper not in PubMed at the time will be checked as well.

Some empirical analysis to follow up on (#1024 (comment)):

Condition	Size (MB)	Size (bytes)	Delta	Elapsed Time
Base	104.76	104759158
CRON job (unstable)	246.22	246224854	135%	~24 hours
CRON job (this branch)	105.27	105268838	0.4%	~1 minute

Refs #1024

[maxkfranz]

1. Skip updates to relatedPapers: Drop this rather heavy, long-running task (see Examining the database data / Root of large data dump size  #1024 (comment)). It's not completely clear that updates to these make any tangible difference (e.g. the most recent papers of interest).

It will be worth considering how that process could be reworked generally w.r.t. future efforts re. recency. This is good for now, though we'll have to map this out further in future