Copy numpy array into Panda multiindex (same size)

I have two matrix: numpy square matrix and a panda multiindexed square matrix. They are the same size. The idea is to get the value from numpy into the multiindex panda matrix to navigate more easily into the data.

My matrix are around 100 000 x 100 000. And my panda matrix has three level of index.

tuples = [('1','A','a'), ('1','A','b'), ('1','A','c'), ('1','B','a'), ('1','B','b'), ('1','B','c'), ('2','A','a'), ('2','A','b'), ('2','B','a')]
index = pd.MultiIndex.from_tuples(tuples, names=['geography', 'product','activity'])
df = pd.DataFrame(index=index, columns=index)

geography                   1                 2      
product                     A        B        A     B
activity                    a  b  c  a  b  c  a  b  a
geography product activity                           
1         A       a         0  0  0  0  0  0  0  0  0
                  b         0  0  0  0  0  0  0  0  0
                  c         0  0  0  0  0  0  0  0  0
          B       a         0  0  0  0  0  0  0  0  0
                  b         0  0  0  0  0  0  0  0  0
                  c         0  0  0  0  0  0  0  0  0
2         A       a         0  0  0  0  0  0  0  0  0
                  b         0  0  0  0  0  0  0  0  0
          B       a         0  0  0  0  0  0  0  0  0

np.random.rand(9,9)
array([[ 0.27302806,  0.33926193,  0.01489047,  0.71959889,  0.43500806,
         0.03607795,  0.03747561,  0.43000199,  0.8091691 ],
       [ 0.96626878,  0.37613022,  0.7739084 ,  0.16724657,  0.01144436,
         0.0107722 ,  0.73513494,  0.13305542,  0.2910334 ],
       [ 0.00622779,  0.93699165,  0.62725798,  0.25009469,  0.14010666,
         0.61826728,  0.72060106,  0.58864557,  0.29375779],
       [ 0.14937979,  0.45269751,  0.68450964,  0.15986812,  0.69879559,
         0.06573519,  0.57504452,  0.49540882,  0.77283616],
       [ 0.60933817,  0.2701683 ,  0.69067959,  0.22806386,  0.79456502,
         0.75107457,  0.2805325 ,  0.27659171,  0.33446821],
       [ 0.82860687,  0.27055835,  0.37684942,  0.18962783,  0.59885119,
         0.31246936,  0.94522335,  0.53487273,  0.00611481],
       [ 0.27683582,  0.23653112,  0.41250374,  0.5024068 ,  0.27621212,
         0.81379001,  0.6704781 ,  0.87521485,  0.04577144],
       [ 0.95516958,  0.21844023,  0.86558273,  0.52300142,  0.91328259,
         0.7587479 ,  0.15201837,  0.15376074,  0.12092142],
       [ 0.36835891,  0.0381736 ,  0.36473176,  0.30510363,  0.19433639,
         0.43431018,  0.00112607,  0.35334684,  0.82307449]])

How I can put the value of the numpy matrix into in the panda multiindex matrix. The two matrix by construction have the same structure, i.e. the numpy matrix is the panda one without label indexes.

I found a dozen of examples to transform multiindex df into numpy array, but not in this way. Only one example of a 3 dimensional numpy array, but mine is not a 3-d np array.