import numpy as np
import pandas as pd
import networkx as nx
from scipy.spatial.distance import cdist
from sklearn.neighbors import NearestNeighbors

Given the following array:

A = np.array([[1.25,1.1,0.8,1.23,2.,1.2],
              [1.44,0.9,0.8,1.01,3.,1.2],
              [0.3,2.8,2.8,1.,1.,1.3]])

I build a dataframe of 5 objects with 3 features each, like so:

# creating a list of index names
index_values = ['first', 'second', 'third',
                'fourth', 'fifth','sixth']
columns = ['feat_1', 'feat_2', 'feat_3']

# creating the dataframe
df = pd.DataFrame(data = array, 
                  index = index_values, 
                  columns = columns)
  
# displaying the dataframe
print(df)

        feat_1  feat_2  feat_3
first     1.25    1.44     0.3
second    1.10    0.90     2.8
third     0.80    0.80     2.8
fourth    1.23    1.01     1.0
fifth     2.00    3.00     1.0
sixth     1.20    1.20     1.3

---

Then I calculate the euclidean distances between all objects:

from scipy.spatial.distance import cdist

euclidean = cdist(df, df, 'euclid')

array([[0.        , 2.56204996, 2.61956103, 0.82176639, 1.86711007,
        1.02961158],
       [2.56204996, 0.        , 0.31622777, 1.80803761, 2.90860791,
        1.53297097],
       [2.61956103, 0.31622777, 0.        , 1.86252517, 3.08544972,
        1.60312195],
       [0.82176639, 1.80803761, 1.86252517, 0.        , 2.133776  ,
        0.35637059],
       [1.86711007, 2.90860791, 3.08544972, 2.133776  , 0.        ,
        1.99248588],
       [1.02961158, 1.53297097, 1.60312195, 0.35637059, 1.99248588,
        0.        ]])

---

Now I need to transform distances into similarity, with 1/distances, I guess, and finally use KNN to build a graph with 5 nodes, considering k=3, in order to find the degree of all nodes.

How do I do this last part?