Merge pull request #59 from qutech/hotfix/concatenation_with_zero_n_c…

…oeffs

Fix concatenation with zero n_coeffs

filter_functions/pulse_sequence.py

@@ -1281,7 +1281,7 @@ def _concatenate_Hamiltonian(
 
     # Concatenate the coefficients. Place them in the right time segments of
     # the concatenated Hamiltonian.
-    concat_coeffs = np.zeros((len(concat_identifiers), sum(n_dt)), dtype=float)
+    concat_coeffs = np.full((len(concat_identifiers), sum(n_dt)), fill_value=np.nan)
     flat_coeffs = [co for coeff in coeffs for co in coeff]
     for i in range(len(concat_identifiers)):
         # Get the indices in opers (and coeffs) for the i-th unique operator
@@ -1299,17 +1299,18 @@ def _concatenate_Hamiltonian(
         # the remaining segments as usually the sensitivity is constant. If we
         # cannot do this, we have to raise an exception since we cannot know
         # the sensitivities at other moments in time if they are non-trivial.
-        for i, c_coeffs in enumerate(concat_coeffs):
-            zero_mask = (c_coeffs == 0)
-            if zero_mask.any() and not zero_mask.all():
-                nonzero_coeffs = c_coeffs[~zero_mask]
-                constant = (nonzero_coeffs == nonzero_coeffs[0]).all()
-                if constant:
-                    # Fill with constant value
-                    concat_coeffs[i, zero_mask] = nonzero_coeffs[0]
-                else:
-                    raise ValueError('Not all pulses have the same noise operators and ' +
-                                     'non-trivial noise sensitivities so I cannot infer them.')
+        nan_mask = np.isnan(concat_coeffs)
+        test = nan_mask.any(axis=1)
+        for i, (concat_coeff, mask) in enumerate(zip(concat_coeffs[test], nan_mask[test])):
+            nonnan_coeff = concat_coeff[~mask]
+            if (nonnan_coeff == nonnan_coeff[0]).all():
+                # Constant value, use for empty segment
+                concat_coeffs[i, mask] = nonnan_coeff[0]
+            else:
+                raise ValueError('Not all pulses have the same noise operators and ' +
+                                 'non-trivial noise sensitivities so I cannot infer them.')
+    else:
+        concat_coeffs[np.isnan(concat_coeffs)] = 0
 
     return concat_opers, concat_identifiers, concat_coeffs[sort_idx], pulse_identifier_mapping
 
@@ -1482,12 +1483,12 @@ def concatenate_without_filter_function(pulses: Iterable[PulseSequence],
 
     # Compose new control Hamiltonian
     control_values = _concatenate_Hamiltonian(
-        *list(zip(*[tuple(getattr(pulse, key) for key in control_keys) for pulse in pulses])),
+        *zip(*[[getattr(pulse, key) for key in control_keys] for pulse in pulses]),
         kind='control'
     )
     # Compose new control Hamiltonian
     noise_values = _concatenate_Hamiltonian(
-        *list(zip(*[tuple(getattr(pulse, key) for key in noise_keys) for pulse in pulses])),
+        *zip(*[[getattr(pulse, key) for key in noise_keys] for pulse in pulses]),
         kind='noise'
     )
 

tests/test_core.py

@@ -469,6 +469,13 @@ def test_pulse_sequence_attributes_concat(self):
                                     [X, rng.standard_normal(2)]],
                                    [[Z, np.abs(rng.standard_normal(2))]],
                                    [1, 1])
+        pulse_4 = ff.PulseSequence([[Y, rng.standard_normal(2)],
+                                    [X, rng.standard_normal(2)]],
+                                   [[Z, np.ones(2)]],
+                                   [1, 1])
+        pulse_5 = ff.PulseSequence([[Y, np.zeros(5), 'A_0']],
+                                   [[Y, np.zeros(5), 'B_1']],
+                                   1 - rng.random(5))
 
         # Concatenate with different noise opers
         pulses = [testutil.rand_pulse_sequence(2, 1) for _ in range(2)]
@@ -479,6 +486,7 @@ def test_pulse_sequence_attributes_concat(self):
 
         pulse_12 = pulse_1 @ pulse_2
         pulse_21 = pulse_2 @ pulse_1
+        pulse_45 = pulse_4 @ pulse_5
 
         with self.assertRaises(TypeError):
             _ = pulse_1 @ rng.standard_normal((2, 2))
@@ -515,6 +523,16 @@ def test_pulse_sequence_attributes_concat(self):
         self.assertArrayEqual(pulse_12.n_coeffs, [[*z_coeff_1, *z_coeff_2]])
         self.assertArrayEqual(pulse_21.n_coeffs, [[*z_coeff_2, *z_coeff_1]])
 
+        # Make sure zero coefficients are handled correctly
+        self.assertFalse(np.any(np.isnan(pulse_45.c_coeffs)))
+        self.assertFalse(np.any(np.isnan(pulse_45.n_coeffs)))
+        self.assertArrayEqual(pulse_45.c_coeffs,
+                              [[*pulse_4.c_coeffs[0], *np.zeros(5)],
+                               [*pulse_4.c_coeffs[1], *np.zeros(5)]])
+        self.assertArrayEqual(pulse_45.n_coeffs,
+                              [[*pulse_4.n_coeffs[0], *[pulse_4.n_coeffs[0, 0]]*5],
+                               [*[pulse_5.n_coeffs[0, 0]]*2, *pulse_5.n_coeffs[0]]])
+
         omega = np.linspace(-100, 100, 101)
         pulses = (pulse_1, pulse_2, pulse_12, pulse_21)
         for pulse in pulses: