Overlapping focal nodes not consolidated; negative resistances returned

[bmcrae]

When we have focal nodes provided in text list format, and some nodes fall in the same cell, 4.0 still solves for each one, which is inefficient (it is solving the same problem multiple times).

One user was confused by this, because the solver produced very small negative resistances between focal nodes that were actually the same node and should thus have had zero resistances.

This is the example reported to me:
https://dl.dropboxusercontent.com/u/1340127/CS/cs_overlapping_nodes.zip

It would be best to consolidate these into a single node (which I think we do for nodes falling in polygons and I think we did in 3.5) and report zero resistances between them in the output matrix, or at least drop redundant focal nodes.

[bmcrae]

Annabel Smith just reported the same problem to the list, and in both her and another user's problem this caused 'nan' values to be returned.

I was wrong in writing that we consolidated redundant points into a single node in 3.5. I replicated the same problem using 3.5 today.

We could save the original points_rc list, prune out redundant points (those that occupy the same cell as another point), solve for the remaining points, and add the redundant points back in to the matrix. @tanmaykm, would you have time to do this? If not, I might have time in July to get back into the code.