Graph utilities¶

Connectivity¶

Connectivity helpers inspect graph components and are useful before routing or OD-matrix calculations. For undirected graphs use connected_components(); for directed graphs choose between weakly_connected_components() and strongly_connected_components().

The largest_* helpers return node-id sets. To actually keep only the largest component, use UrbanGraph.keep_largest_connected_component() or the keep_largest_subgraph parameter in graph builders.

See Connectivity diagnostics for a runnable connectivity workflow.

Editing and transformation¶

The editing helpers return new UrbanGraph objects and do not mutate the source graph unless you use the corresponding UrbanGraph method with inplace=True.

Use subgraph_by_nodes() for node-induced subgraphs, clip_urban_graph() for geometry-based clipping, relabel_urban_graph() for dense node indexes, and join_urban_graphs() for concatenating compatible graph tables.

Use UrbanGraph.to_directed(), UrbanGraph.to_undirected(), and UrbanGraph.simplify_multiedges() for topology transformations.

See Graph operations for editing examples.

Object matching and projection¶

Use nearest_nodes() when an object table only needs a nearest existing graph node id.

Use project_objects2urban_graph() when objects should be inserted into the graph topology. The function prepares an UrbanGraphChanges object with new nodes, new edges, and source edges to delete. Apply the changes with apply_urban_graph_changes(), or use UrbanGraph.project_objects() for the in-memory shortcut.

See Objects, nearest nodes, and graph projection for both workflows, including add_link_edge=True and add_link_edge=False.

Persistence¶

Use write_urban_graph() and read_urban_graph() to store and restore .urbangraph archives. Archives contain node and edge tables and can optionally include the cached adjacency matrix. See UrbanGraph basics.

`connected_components`	Return connected components of an undirected `UrbanGraph`.
`weakly_connected_components`	Return weakly connected components.
`strongly_connected_components`	Return strongly connected components.
`number_connected_components`	Return the number of connected components in an undirected graph.
`number_weakly_connected_components`	Return the number of weakly connected components.
`number_strongly_connected_components`	Return the number of strongly connected components.
`largest_connected_component`	Return the largest connected component of an undirected graph.
`largest_weakly_connected_component`	Return the largest weakly connected component.
`largest_strongly_connected_component`	Return the largest strongly connected component.
`largest_component`	Return the largest component according to the selected connectivity mode.
`relabel_urban_graph`	Relabel node ids to a dense `RangeIndex` and update edge endpoints.
`subgraph_by_nodes`	Return an induced `UrbanGraph` subgraph for the provided node ids.
`clip_urban_graph`	Clip graph nodes by geometry and keep only edges with retained endpoints.
`join_urban_graphs`	Join two compatible `UrbanGraph` objects by concatenating node and edge tables.
`nearest_nodes`	Return the nearest graph node id for each input geometry.
`project_objects2urban_graph`	Prepare graph changes that connect objects to their nearest edges.
`apply_urban_graph_changes`	Apply prepared changes to an `UrbanGraph`.
`read_urban_graph`	Read an `UrbanGraph` from an `.urbangraph` archive.
`write_urban_graph`	Write an `UrbanGraph` to an `.urbangraph` archive.
`validate_graph`	Validate all node, edge, topology and CRS contracts of an `UrbanGraph`.

Optional NetworkX utilities¶

The following helpers are available when NetworkX support is installed.

`graph2gdf`
`gdf2graph`
`keep_largest_nx_component`
`clip_nx_graph`
`read_gml`
`write_gml`
`reproject_graph`