numgeo 0.0.2a0

0.0.2a0 Release Notes

High-level changes

Skeletonization

• Skeletonization no longer arbitrarily discards one edge in a pair of edges that directly link the same two nodes in the graph skeleton. (For example, the edge on either side of the hole in Fig. 3c.)
• PolyID field now stores ID of skeletonized polygon in outputs of scipts.skel.process.
  • Previously, this field was named OrigID.
• scipts.skel.process no longer assumes that the same format (shapefile or file geodatabase) is to be used for both the input polygons and the generated outputs.
  • For example, the skeletonization products for polygons stored in a file geodatabase feature class can now be outputted directly to shapefiles, without an error.
  • To output to file GDB feature classes, [write access](#Write access to file geodatabases in gdal) is still independently required.
• graph_only argument added to scipts.skel.process
• Most basic isolation_mode option is now supported, namely, the empty string.
• isolation_mode option "CUT2" is now supported.
  • This option is experimental and may be deprecated in a future version.

Low-level changes

Skeletonization

• isolation_mode option "STORE" is now supported.
• Exposed Skeleton.cut_vor_vert_idxs_array attribute.
• Added Skeleton.get_tail_ends function.
• Added Skeleton.test_poly_contains function.
• trunk_mode option "NONE" is now supported.
• If template_interval < 0, template skeleton is now a (tailed) graph skeleton that is never partitioned.
• Code has been made slightly more extensible.