Update types & Add more signal generation for unit testing

pyproject.toml

@@ -36,6 +36,7 @@ show-report = true
 
 [tool.ruff]
 line-length = 72
+builtins = ["_"]
 
 [tool.basedpyright]
 venvPath = ".env"

src/science_signal/__init__.py

@@ -25,6 +25,7 @@ def interpolate(*signals: "Signal") -> tuple["Signal", ...]:
     Return each signal with a common frequency/time range (smallest
     range and highest rate)
     """
+    logger.debug(_("interpolate {n} signals").format(n=len(signals)))
     splines = [signal.spline() for signal in signals]
     x_range = interpolate_abciss(*signals)
 

src/science_signal/generator.py

@@ -1,4 +1,4 @@
-from numpy import arange, pi, sin as np_sin
+from numpy import arange, pi, sign, sin as np_sin
 from numpy.random import default_rng
 from science_signal.signal import Signal
 
@@ -8,6 +8,7 @@ def sin(
     rate: float = 10,
     frequency: float = 1,
     amplitude: float = 1,
+    offset: float = 0,
     phase: float = 0,
 ) -> Signal:
     """
@@ -16,10 +17,11 @@ def sin(
     duration: duration of the signal in second
     rate: rate of the signal in Hz (s-1)
     frequency: frequency of the wanted sinus
+    offset: mean of the wanted sinus
     phase: between 0 and 2 pi (if more, congruate)
     """
     x = arange(0, duration, 1 / rate)
-    y = amplitude * np_sin(2 * pi * frequency * x + phase)
+    y = offset + amplitude * np_sin(2 * pi * frequency * x + phase)
 
     return Signal(
         x,
@@ -32,6 +34,7 @@ def cos(
     rate: float = 10,
     frequency: float = 1,
     amplitude: float = 1,
+    offset: float = 0,
     phase: float = 0,
 ) -> Signal:
     """
@@ -42,7 +45,36 @@ def cos(
     frequency: frequency of the wanted sinus
     phase: between 0 and 2 pi (if more, congruate)
     """
-    return sin(duration, rate, frequency, amplitude, phase + pi / 2)
+    return sin(
+        duration, rate, frequency, amplitude, offset, phase + pi / 2
+    )
+
+
+def square(
+    duration: float = 10,
+    rate: float = 10,
+    frequency: float = 1,
+    amplitude: float = 1,
+    offset: float = 0,
+    phase: float = 0,
+) -> Signal:
+    """
+    Generate square signal
+
+    duration: duration of the signal in second
+    rate: rate of the signal in Hz (s-1)
+    frequency: frequency of the wanted square signal
+    phase: between 0 and 2 pi (if more, congruate)
+    """
+    x = arange(0, duration, 1 / rate)
+    y = offset + amplitude * sign(
+        np_sin(2 * pi * frequency * x + phase)
+    )
+
+    return Signal(
+        x,
+        y,
+    )
 
 
 def gaussian_noise(
@@ -56,12 +88,36 @@ def gaussian_noise(
     Generate a gaussian noise signal
 
     duration: duration of the signal in second
-    rate: rate of the signal in hz (s-1)
+    rate: rate of the signal in Hz (s-1)
     sigma: standard deviation of the wanted distribution
     mu: mean of the distribution
+    seed: optional, allowd to specify a seed for testing purpose
     """
     x = arange(0, duration, 1 / rate)
     return Signal(
         x,
         default_rng(seed).normal(mu, sigma, len(x)),
     )
+
+
+def white_noise(
+    duration: float = 10,
+    rate: float = 10,
+    minimum: float = -1,
+    maximum: float = 1,
+    seed: None | int = None,
+) -> Signal:
+    """
+    Generate a white noise signal
+
+    duration: duration of the signal in seconds
+    rate: rate of the signal in hz (s-1)
+    minimum: minimum value of the noise
+    maximum: maximum value of the noise
+    seed: optional, allowd to specify a seed for testing purpose
+    """
+    x = arange(0, duration, 1 / rate)
+    return Signal(
+        x,
+        default_rng(seed).uniform(minimum, maximum, len(x)),
+    )

src/science_signal/signal.py

@@ -1,4 +1,4 @@
-from scipy.interpolate import CubicSpline  # pyright: ignore[reportMissingTypeStubs]
+from scipy.interpolate import CubicSpline  # type: ignore[reportMissingTypeStubs]
 from numpy.typing import ArrayLike, NDArray
 from numpy import (
     append,
@@ -7,14 +7,13 @@ from numpy import (
     linspace,
     array,
     logical_and,
-    logical_or,
     sin,
     where,
     arange,
     zeros,
 )
-from scipy.signal import detrend, welch
-from scipy.fft import rfftfreq, rfft, irfft
+from scipy.signal import detrend, welch # type: ignore[reportMissingTypeStubs]
+from scipy.fft import rfftfreq, rfft, irfft # type: ignore[reportUnknownVariableType]
 from science_signal import interpolate
 
 
@@ -77,7 +76,7 @@ class Signal:
             start = min(self.x)
         if end is None:
             end = max(self.x)
-        indexes = where(logical_and(self.x >= start, self.x <= end))
+        indexes = where(logical_and(self.x >= start, self.x <= end)) # type: ignore[reportOperatorIssue]
         return Signal(
             self.x[indexes],
             self.y[indexes],
@@ -114,30 +113,49 @@ class Signal:
     def filter(
         self, start: None | float = None, end: None | float = None
     ) -> "Signal":
-        freq_x = rfftfreq(len(self), self.sampling)
+        freq_x = rfftfreq(len(self), self.sampling) # type: ignore[reportUnknownVariableType]
         freq_y = rfft(self.y)
 
+        rate = 1 / (freq_x[1] - freq_x[0]) # type: ignore[reportUnknownVariableType]
+
         if start is None:
             start = min(abs(freq_x))
         if end is None:
             end = max(abs(freq_x))
 
-        index_to_remove = where(
-            logical_or(abs(freq_x) <= start, abs(freq_x) >= end)
-        )
-        freq_y[index_to_remove] = 0
+        index_start = where(abs(freq_x) < start) # type: ignore[reportOperatorIssue]
+        index_end = where(abs(freq_x) > end) # type: ignore[reportOperatorIssue]
+
+        if len(index_start) != 0:
+            """
+            damping_start = 10 ** -( # type: ignore[reportUnknownVariableType]
+                arange(1, len(index_start[0]) + 1, 1, dtype=float64)
+                / rate
+            )
+            freq_y[index_start] = freq_y[index_start] * damping_start # type: ignore[reportIndexIssue]
+            """
+            freq_y[index_start] = 0
+        if len(index_end) != 0:
+            """
+            damping_end = 10 ** -( # type: ignore[reportUnknownVariableType]
+                arange(1, len(index_end[0]) + 1, 1, dtype=float64)
+                / rate
+            )
+            freq_y[index_end] = freq_y[index_end] * damping_end # type: ignore[reportAny]
+            """
+            freq_y[index_end] = 0
 
         y = irfft(freq_y)
         return Signal(
             self.x[: len(y)],
-            y[: len(self.x)],
+            y[: len(self.x)], # type: ignore[reportArgumentType]
         )
 
     def psd(self, fft_length: int = 10) -> "Signal":
         """
         Compute psd of a given signal
         """
-        freq, psd = welch(
+        freq, psd = welch( # type: ignore[reportUnknownVariableType]
             self.y,
             self.rate,
             nperseg=fft_length * self.rate,
@@ -192,7 +210,7 @@ class Signal:
                 # suppose same range but different rates
                 return self.operator_signal(
                     Signal(
-                        linspace(self.x[0], self.x[-1], len(others)),
+                        linspace(self.x[0], self.x[-1], len(other)),
                         array(other),
                     ),
                     operator,