It is often preferable to use the primal edges (lines) to visualise metrics on the dual graph (i.e. instead of using points).
This release implements a change where primal edges are automatically copied across to dual graph nodes when using the [`graphs.nx_to_dual()`](https://cityseer.benchmarkurbanism.com/tools/graphs#nx-to-dual) method. Downstream conversion via [`io.network_structure_from_nx()`](https://cityseer.benchmarkurbanism.com/tools/io#network-structure-from-nx) will likewise retain the primal edge geometry for visualisation of metrics written to dual node `GeoDataframe` structures.
4.12.0
Homogenises column naming for gpd metrics columns.
> This is a breaking change.
The overall API remains as before but the `geopandas` column names to which the metrics are unpacked have been changed.
The reason for this change is: - to reduce the verbosity of the column names - to use a consistent naming convention across metrics
Column names can be retrieved from a `geopandas` dataframe using the `columns` attribute and the new names will be self explanatory and largely similar to before.
This simplifies recreation of network structures from cityseer generated node and edges GeoDataFrames. This is useful for situations where the GeoDataFrames have been saved to disk and the network structure needs to be recreated.
4.11.0
Splits the underlying Dijkstra shortest path algorithm into three separate versions for shortest / simplest / segment for performance improvements.
4.10.3
Python 3.12 support
4.9.0
This release generalises the usage of CRS codes so that string or integer format codes can work, as well as non-EPSG codes.
All parameters previously named `to_epsg...` or `from_epsg...` are accordingly now renamed to `to_crs...` and `from_crs...`