AdjacencyCalculator

todo

This module provides all necessary tools to get accesibility matrix from transport graph

class blocksnet.preprocessing.adjacency_calculator.BlockRow(*, geometry: Polygon, index: int)[source]

Bases: BaseRow

geometry: Polygon

class blocksnet.preprocessing.adjacency_calculator.AdjacencyCalculator(*, blocks: GeoDataFrame[BlockRow], graph: MultiDiGraph)[source]

Bases: BaseModel

Class Accessibility calculates accessibility matrix between city blocks. It takes a lot of RAM to calculate one since we have thousands of city blocks.

Methods

get_matrix

blocks: GeoDataFrame[BlockRow]

graph: MultiDiGraph

validate_blocks()[source]

validate_graph()[source]

static _get_nx2nk_idmap(graph: Graph) → dict[source]

This method gets ids from nx graph to place as attribute in nk graph

Attributes

graph: networkx graph

Returns

idmap: dict: map of old and new ids

static _get_nk_attrs(graph: Graph) → dict[source]

This method gets attributes from nx graph to set as attributes in nk graph

Attributes

graph: networkx graph

Returns

idmap: dict: map of old and new attributes

classmethod _convert_nx2nk(graph_nx: MultiDiGraph, idmap: dict | None = None, weight: str = 'weight') → Graph[source]

This method converts networkx graph to networkit graph to fasten calculations.

Attributes

graph_nx: networkx graph idmap: dict

map of ids in old nx and new nk graphs

weight: str: value to be used as a edge’s weight

Returns

graph_nk: nk.Graph: the same graph but now presented in is networkit package Graph class.

_get_nk_distances(nk_dists: Algorithm, loc: Series) → Series[source]

This method calculates distances between blocks using nk SPSP algorithm. The function is called inside apply function.

Attributes

nk_dists: nk.base.Algorithm: Compressed nk graph to compute distances between nodes using SPSP algorithm
loc: pd.Series: Row in the df

Returns

pd.Series with computed distances

get_dataframe() → DataFrame[source]

This methods runs graph to matrix calculations

Returns

accs_matrix: pd.DataFrame: An accessibility matrix that contains time between all blocks in the city