This means that you can verify in MATLAB similarity of matrices by checking with the command rref from Symbolic Python (sympy) package that they have the same reduced row echelon form.

>>> import sympy as sy
>>> sy.init_printing(use_unicode=True)   # for pretty printing
>>> A = sy.Matrix([[4, -2], [2, 1]]); A
⎡4  -2⎤
⎢     ⎥
⎣2  1 ⎦
>>> B = sy.Matrix([[3, -2], [1, 2]]); B
⎡3  -2⎤
⎢     ⎥
⎣1  2 ⎦
>>> A.rref() == B.rref()
True
>>> T = sy.Matrix([[1, -1], [1, 0]]); T  # appropriate transformation matrix
⎡1  -1⎤
⎢     ⎥
⎣1  0 ⎦
>>> B == T**(-1) * A * T  # 
True
>>> S = sy.Matrix([[1,1], [0, 2]]); S
⎡1  1⎤
⎢    ⎥
⎣0  2⎦
>>> B == S**(-1) * A * S # transformation matrices are not uniwue
True
>>> P = S * T**(-1); P   # non-trivial transformation matrix that leaves A intact
⎡-1  2⎤
⎢     ⎥
⎣-2  2⎦
>>> A == P**(-1) * A * P
True