Utils#

Utility functions for the Distance Closure package

distanceclosure.utils.prox2dist(p)[source]#

Transforms a non-negative [0,1] proximity to distance in the [0,inf] interval:

\[d = \frac{1}{p} - 1\]

Parameters: p (float) – Proximity value
Returns: d – Distance value
Return type: float

See also

dist2prox

distanceclosure.utils.dist2prox(d)[source]#

Transforms a non-negative integer distance d to a proximity/similarity value in the [0,1] interval:

\[p = \frac{1}{(d+1)}\]

It accepts both dense and sparse matrices.

Parameters: D (matrix) – Distance matrix
Returns: P – Proximity matrix
Return type: matrix

See also

prox2dist

distanceclosure.utils.dict2matrix(d)[source]#

Tranforms a 2D dictionary into a numpy. Usefull when converting Dijkstra results.

Parameters: (dict) (d) –
Returns: m
Return type: Numpy matrix

Warning

If your nodes are identified by names instead of numbers, make sure to keep a mapping.

Examples

>>> d = {0: {0: 0, 1: 1, 2:3}, 1: {0: 1, 1: 0, 2:2}, 2: {0: 3, 1:2, 2:0}}
>>> dict2matrix(d)
    [[ 0 1 3]
     [ 1 0 2]
     [ 3 2 0]]

Note

Uses pandas to accomplish this in a one liner.

See also

matrix2dict

distanceclosure.utils.matrix2dict(m)[source]#

Tranforms a Numpy matrix into a 2D dictionary. Usefull when comparing dense metric and Dijkstra results.

Parameters: (matrix) (m) –
Returns: d (dict)
Return type: 2D dictionary

Examples

>>> m = [[0, 1, 3], [1, 0, 2], [3, 2, 0]]
>>> matrix2dict(m)
    {0: {0: 0, 1: 1, 2:3}, 1: {0: 1, 1: 0, 2:2}, 2: {0: 3, 1:2, 2:0}}

Note

Uses pandas to accomplish this in a one liner.

See also

dict2matrix

distanceclosure.utils.dict2sparse(d)[source]#

Tranforms a 2D dictionary into a Scipy sparse matrix.

Parameters: d (dict) – 2D dictionary
Returns: m – CRS Sparse Matrix
Return type: CSR matrix

Examples

>>> d = {0: {0: 0, 1: 1, 2:3}, 1: {0: 1, 1: 0, 2:2}, 2: {0: 3, 1:2, 2:0}}
>>> dict2sparse(d)
    (0, 1)    1
    (0, 2)    3
    (1, 0)    1
    (1, 2)    2
    (2, 0)    3
    (2, 1)    2

Note

Uses pandas to convert dict into dataframe and then feeds it to the csr_matrix.

See also

dict2matrix, matrix2dict

distanceclosure.utils.from_networkx_to_dijkstra_format(D, weight='weight')[source]#

Converts a NetworkX.Graph object to input variables to be used by cython.dijkstra.

Parameters

D (NetworkX:Graph) – The Distance graph.
weight (string) – The edge property to use as distance weight.

Returns

nodes (list) – List of all nodes converted to sequential numbers.
edges (list) – List of all edges.
neighbors (dict) – Dictionary containing the neighborhood of every node in a fast access format.
dict_int_nodes (dict) – The mapping between original node names and the numeric node names.

Examples

>>> G = nx.path(5)
>>> nx.set_edge_attributes(G, name='distance', values=1)
>>> nodes, edges, neighbors, dict_int_nodes = from_networkx_to_dijkstra_format(G, weight='distance')