Evaluate technical requirements for using https://github.com/sebastian-nagel/sfm-facebook-harvester. Is the referenced fork of SFM something we can bring back to the "trunk" by making whatever changes will allow the harvester to plug in?
Consider data/policy concerns
Consider GW (and beyond) use cases
Consider tests that would be needed to ensure that future changes to SFM don't break the harvester that may be maintained in a different repository, by a different organization.
Since it does not rely on the Graph API but scrapes it requires continuous support as changes in the FB website, corresponding changes in the facebook-scraper library require integration and monitoring, moreover being blocked is a concern
Data/policy issues may also vary depending on the country as well as usage (scientific vs. commercial)