Some failing routes need diagnostics

[gregallensworth]

Of the demo, a few of the routes are failing. See the "works in progress" sections.

Unknown yet whether these represent sufficiently-disjointed topology that it cannot find a new segment, or some other nuance.

[gregallensworth]

One significant cause of no-route failures has been found and addressed: The chosen starting segment itself not being eligible for use in the route!

The nearest segment to me may not in fact face N or S. As such, the segment chosen as my starting may not fit the direction filter! We're dead in the water already!

ad3dee8 enhances findNearestSegmentToLatLng() to accept an optional direction flag and will filter by it in selecting the start and end segments. This fixes a whole lot of no-route conditions that had us dead at the very start.

[gregallensworth]

Reference: https://github.com/GreenInfo-Network/east-coast-greenway/issues/19

The break between Kingsland, GA and Darien, GA which prevents any routing connection between Brunswick, GA and any area southward.

[gregallensworth]

Reference: https://github.com/GreenInfo-Network/east-coast-greenway/issues/21

There is a break between two paths in Miami, 661596 and 661598 due to 661598 (M-Path) folding over itself so the endpoint is some distance away. This break causes a complete disconnect in connectivity, so that one cannot route between the north end of Miami and the south end of Miami.

[gregallensworth]

Closing. I have confirmed that the routing failures are not due to routing logic, but data errors that the trails truly do not connect.