flowpath to divide not 1:1 in aggregate output due to pipelines and such.

[dblodgett-usgs]

Some flowpath ids get dropped because we are using divide ids to seed the lookup table here: https://github.com/mikejohnson51/hydrofab/blob/master/R/merge_hydrofabrics.R#L83

This is causing some flowpath ids to be NULL in the output of merge national. I think I have a fix but I don't know what x$cumcount_div is doing in the code linked just above. OK to switch that to x$cumcount_fl @mikejohnson51 ?

[dblodgett-usgs]

This will not work out of the box because we have flowlines that do not realize a catchment and we have some catchment divides that do no have flowpaths.

In the network table, we need all divide and flowline identifiers. We also need a has divide logical and a has flowline logical in the network table. There are very likely going to be catchments that do not have an associated flowpath that flow to a network flowline and others that do not flow to a network feature but still require topology. In that case, we should have an attribute in the network table, "network_type" that can be one of "connector" "network" "internal" "coastal".

[mikejohnson51]

Updates to data model per Nov 1 Technical Meeting:

[mikejohnson51]

Updates to data model per Nov 3 Technical Meeting:

• We are going to change the POI table name to hydrolocations
• The hydrolocations table will contain: {hl_id, mainstem, id, geometry}.
  • hl_id is the old poi_id
• We will split lookup into network_lookup and hydrolocation_lookup
• network_lookup will be identical to the existing lookup but will drop {poi_id, poi_type, and poi_value}
• hydrolocation_lookup will contain: {hl_id, id, hl_reference, hl_link, hl_position, geometry}
  • hl_id is the old poi_id
  • hl_reference is the old poi_type
  • hl_link is the old poi_value
  • hl_position is a constrained vocabulary of {inflow, outflow, along}
  • For now, all POIs will be given a position of outlet