Graph extraction#

The graph building module provides tools to create geospatially enriched graphs from road network data.

class libadalina_core.graph_extraction.MandatoryColumns(value)[source]#: Enum representing mandatory columns required in the input DataFrame.

class libadalina_core.graph_extraction.OneWay(value)[source]#

Enum representing road directions.

Backward = 'backward'#: Goes against the direction of the geometry.

Both = 'both'#: Both directions are allowed.

Forward = 'forward'#: Follows the direction of the geometry.

class libadalina_core.graph_extraction.OpenStreetMapReader(road_types: RoadTypes = RoadTypes.ALL)[source]#

A class to read OpenStreetMap (OSM) data files and convert them into a GeoDataFrame with standardized columns.

from_dataframe(df: DataFrame) → GeoDataFrame[source]#

Convert a pandas DataFrame of OSM data to a GeoDataFrame with the required columns.

Parameters:: df (pd.DataFrame) – The OSM data as a pandas DataFrame. Must contain ‘osm_id’, ‘name’, ‘oneway’, and ‘geometry’ columns.
Returns:: A GeoDataFrame containing only the mandatory columns with standardized names.
Return type:: gpd.GeoDataFrame

read(file_path: str) → GeoDataFrame[source]#

Read an OSM data file and return a GeoDataFrame with the required columns. Accepts CSV, Shapefile, and GeoPackage formats.

Parameters:: file_path (str) – The path to the OSM data file. Must contain ‘osm_id’, ‘name’, ‘oneway’, and ‘geometry’ columns.
Returns:: A GeoDataFrame containing only the mandatory columns with standardized names.
Return type:: gpd.GeoDataFrame

class libadalina_core.graph_extraction.RoadTypes(value)[source]#

Enum representing different types of roads for filtering purposes.

ALL = 'all'#: Keep all roads.

CAR_ONLY = 'only_car'#: Keep only roads accessible by car.

MAIN_ROADS = 'main_roads'#: Keep only main roads (motorways, trunks, primary).

Build a directed graph from a DataFrame of roads. Optionally, enrich the graph with data from another DataFrame using spatial join and aggregation. By default, each edge of the graph is already enriched with distance information.

Parameters:

roads_df (DataFrame) – Road network as a DataFrame. Must contain at least columns of MandatoryColumns enumeration.
name (str) – The name of the graph. Defaults to ‘graph’.
joined_df (DataFrame | None) – DataFrame to join with the roads DataFrame. If None, no join is performed. Defaults to None.
aggregate_functions (list[AggregationFunction]) – List of aggregation functions to apply during the spatial aggregation after the join. If None, no aggregation is performed. Defaults to None. See also AggregationFunction and spatial_aggregation for more details.
buffer_radius_meters (float) – The radius in meters to buffer the road segments before performing the spatial join. If joined_df is None, this parameter is ignored. Defaults to 100 meters.
dataframe_only (bool) – If True, the function stops before building a networkx Graph and returns a Spark DataFrame instead.

Returns:

The resulting directed graph as a networkx DiGraph or a Spark DataFrame if dataframe_only is True.

Return type:

nx.Graph | pyspark.sql.DataFrame

libadalina_core.graph_extraction.graph_to_geopackage(graph: Graph, file_path: str)[source]#

Write a networkx graph to a geopackage.

The function first uses graph_to_pandas to convert the graph into a pandas DataFrame, and then save it to file as a GeoDataFrame in a geopackage format.

Parameters:

graph (networkx.Graph) – The networkx graph to write.
file_path (str) – The path to the geopackage.

libadalina_core.graph_extraction.graph_to_pandas(graph: Graph) → DataFrame[source]#

Convert a networkx graph to a pandas DataFrame.

Parameters:: graph (nx.Graph) – The networkx graph to convert.
Returns:: A pandas DataFrame containing the edges of the graph.
Return type:: pd.DataFrame

libadalina_core.graph_extraction.graph_to_shapefile(graph: Graph, path: str)[source]#

Write a networkx graph to a shapefile.

Parameters:

graph (networkx.Graph) – The networkx graph to write.
path (str) – The path to the shapefile.

libadalina_core.graph_extraction.path_to_geopackage(graph: Graph, path: list, file_path: str)[source]#

Save a path in a networkx graph to a geopackage file.

The function first uses path_to_pandas to convert the path into a pandas DataFrame and then saves it as a file in a geopackage format.

Parameters:

graph (networkx.Graph) – The networkx graph containing the path.
path (list) – A list of nodes representing the path.
file_path (str) – The path to the geopackage file where the edges will be saved.

libadalina_core.graph_extraction.path_to_pandas(graph: Graph, path: list) → DataFrame[source]#

Convert a path in a networkx graph to a pandas DataFrame.

Parameters:

graph (nx.Graph) – The networkx graph containing the path.
path (list) – A list of nodes representing the path.

Returns:

A pandas DataFrame containing the edges of the path.

Return type:

pandas.DataFrame

libadalina_core.graph_extraction.paths_to_geopackage(graph: Graph, paths: list, file_path: str)[source]#

Save paths in a networkx graph to a geopackage file.

The function first uses paths_to_pandas to convert the path into a pandas DataFrame and then saves it as a file in a geopackage format.

Parameters:

graph (networkx.Graph) – The networkx graph containing the path.
paths (list[list]) – A list of paths
file_path (str) – The path to the geopackage file where the edges will be saved.

libadalina_core.graph_extraction.paths_to_pandas(graph: Graph, paths: list) → DataFrame[source]#

Convert paths in a networkx graph to a pandas DataFrame.

Parameters:

graph (nx.Graph) – The networkx graph containing the path.
paths (list) – A list of paths.

Returns:

A pandas DataFrame containing the edges of the paths.

Return type:

pandas.DataFrame