Optimization of the control-target gates

[olgOk]

This optimization is basically for the control gates which have three or more control quits.
Right now it is pure brute force solutions.

    """
      Generate and return the tensor of the control gate for any system
      of different number of qubits with the consideration of the given
      control and target qubits.
      Args:
        control: The index of the control qubit
        target: The index of the target qubit
        gate: a type of gate to be generated (X, Y, etc.)
      Returns:
        A tensor of the contol gate
    """
    control_gate = np.eye(2**self.num_qubits, dtype=np.complex128)
    tuples = []
    # Searches for all the numbers up to 2**self.num_qubits such that in 
    # binary representation they have '1' in control-place and '0' in 
    # target place. 
    for i in range(2**self.num_qubits):
      if not (i & (1 << target)) and all(i & (1 << control_qubit) for control_qubit in control):
        swap = i + 2**target
        # Embeds the transformation into the matrix.
        control_gate[i][i] = self.gates[gate][0][0]
        control_gate[i][swap] = self.gates[gate][0][1]
        control_gate[swap][i] = self.gates[gate][1][0]
        control_gate[swap][swap] = self.gates[gate][1][1]
    # If control gate applies Hadamard gate, puts the whole system into
    # superposition.
    if gate == 'H':
      control_gate = control_gate * (1. + 1.j) / np.sqrt(2)
    return control_gate```