Create Graph from street file

[Narayan Das]

Hi everyone,

I am new here. I am trying to create a graph from given street file using geopandas, momepy and networkx library. 
With my scripts I can create the Graph, but this graph does not contains any keys so I want to add keys as well in Graph G.
Please have look in my script and let me know where I am doing mistake, If anyone have other idea, I'll be happy to discuss with that as well:- 

Thank you everyone

def construct_graph_from_network_layer(network_layer, col_drop:list=[], length_column:str=None, time_column:str=None, avg_speed:int = 25):

    street = geoapndas.read_file(network_layer)

    street = street.explode().reset_index(drop=True)

    if sum(street.geom_type=='LineString')!=len(street): raise ValueError("Please provide the Proper Network layer, Network layer must have Linestring or Multilinestring type of Geometry!!!")

    if col_drop: street.drop(col_drop, axis=1, inplace=True)

    graph = momepy.gdf_to_nx(street, length='length')

    nodes, edges = nx_to_gdf(graph, points=True, lines=True,spatial_weights=False)  ##Extracting node and edges geodatframe from graph

    nodes['x']= nodes['geometry'].apply(lambda pt: pt.x)

    nodes['y']= nodes['geometry'].apply(lambda pt: pt.y)

    edges.reset_index(inplace=True)

    edges.rename(columns={'node_start':'u', 'node_end':'v','index':'k'}, inplace=True)

    cr1 = edges.estimate_utm_crs()

    if length_column is None:

        length_column= 'length'

        edges[length_column]= edges.to_crs(cr1).length

    meters_per_minute = avg_speed * 1000 / 60  # km per hour to m per minute

    if time_column is None:

        time_column= 'time'

        edges[time_column]= edges[length_column].apply(lambda x: round(x/meters_per_minute, 3))

    edges[length_column] = edges[length_column].apply(lambda x: round(x,2))

    edges= edges[[ 'u', 'v','k']+[x for x in edges.columns if x not in [ 'u', 'v','k']]]

    #Creating Graph and adding node and edges

    G = nx.Graph()

    G.graph['crs'] = 'EPSG:4326'

    for idx, row in nodes.iterrows():

        x = row['x']

        y = row['y']

        G.add_node(idx, x=x, y=y)

    for idx, row in edges.iterrows():

        u = row['u']

        v = row['v']

        k = row['k']

        G.add_edge(u, v, k =k, attr_dict=row[3:].to_dict())

    return G, nodes, edges

[Dan Schult]

It is not clear whether you want an edge attribute, or you want multiple edges between two nodes.

Edge attributes are the way to store information about the edge. 

Using a MultiGraph instead of a Graph is how you store more than one edge between a pair of nodes.

See the tutorial for more info and help with deciding how to configure your graph.

[Rhyd Lewis]

Hello,

The OSMNX library should suit your purposes. A tutorial is here