Github # 112 replace asserts with exceptions

Author: rlabbe

filterpy/__init__.py

@@ -14,4 +14,4 @@
 for more information.
 """
 
-__version__ = "1.2.5"
+__version__ = "1.3"

filterpy/changelog.txt

@@ -1,4 +1,4 @@
-Version 1.2.6
+Version 1.3
 =============
 
 * #113 added plotting of 3D covariance ellipsoid with plot_3d_covariance
@@ -11,7 +11,6 @@ Version 1.2.6
 
 
 
-
 Version 1.2.5
 =============
 

filterpy/common/discretization.py

@@ -73,7 +73,9 @@ def Q_discrete_white_noise(dim, dt=1., var=1., block_size=1):
     John Wiley & Sons, 2001. Page 274.
     """
 
-    assert dim == 2 or dim == 3 or dim == 4
+    if not (dim == 2 or dim == 3 or dim == 4):
+        assert ValueError("dim must be between 2 and 4")
+
     if dim == 2:
         Q = [[.25*dt**4, .5*dt**3],
              [ .5*dt**3,    dt**2]]
@@ -130,7 +132,9 @@ def Q_continuous_white_noise(dim, dt=1., spectral_density=1.,
            [0.        , 0.        , 0.        , 0.        , 0.005     , 0.1       ]])
     """
 
-    assert dim == 2 or dim == 3 or dim == 4
+    if not (dim == 2 or dim == 3 or dim == 4):
+        assert ValueError("dim must be between 2 and 4")
+
     if dim == 2:
         Q = [[(dt**3)/3., (dt**2)/2.],
              [(dt**2)/2.,    dt]]

filterpy/common/kinematic.py

@@ -30,7 +30,8 @@ def kinematic_state_transition(order, dt):
     step `dt`.
     """
 
-    assert order >= 0 and int(order) == order, "order must be an int >= 0"
+    if not(order >= 0 and int(order) == order):
+        raise ValueError("order must be an int >= 0")
 
     # hard code common cases for computational efficiency
     if order == 0:
@@ -136,9 +137,12 @@ def kinematic_kf(dim, order, dt=1., dim_z=1, order_by_dim=True):
         [x y z x' y' z' x'' y'' z'']
     """
 
-    assert dim >= 1
-    assert order >= 0
-    assert dim_z >= 1
+    if dim < 1:
+        raise ValueError("dim must be >= 1")
+    if order < 0:
+        raise ValueError("order must be >= 0")
+    if dim_z < 1:
+        raise ValueError("dim_z must be >= 1")
 
     dim_x = order + 1
 

filterpy/gh/gh_filter.py

@@ -93,9 +93,8 @@ def __init__(self, x0, dt, order, g, h=None, k=None):
             difference between the measurement and the prediction
         """
 
-
-        assert order >= 0
-        assert order <= 2
+        if order < 0 or order > 2:
+            raise ValueError('order must be between 0 and 2')
 
         if np.isscalar(x0):
             self.x = np.zeros(order+1)
@@ -241,10 +240,6 @@ def __init__(self, x, dx, dt, g, h):
             filter h gain parameter.
         """
 
-        assert np.isscalar(dt)
-        assert np.isscalar(g)
-        assert np.isscalar(h)
-
         self.x = x
         self.dx = dx
         self.dt = dt
@@ -498,10 +493,6 @@ def __init__(self, x, dx, ddx, dt, g, h, k):
             filter k gain parameter.
         """
 
-        assert np.isscalar(dt)
-        assert np.isscalar(g)
-        assert np.isscalar(h)
-
         self.x = x
         self.dx = dx
         self.ddx = ddx
@@ -899,16 +890,16 @@ def critical_damping_parameters(theta, order=2):
     July, 1975
 
     """
-    assert theta >= 0
-    assert theta <= 1
+    if theta < 0 or theta > 1:
+        raise ValueError('theta must be between 0 and 1')
 
     if order == 2:
         return (1. - theta**2, (1. - theta)**2)
 
     if order == 3:
         return (1. - theta**3, 1.5*(1.-theta**2)*(1.-theta), .5*(1 - theta)**3)
 
-    raise Exception('bad order specified: {}'.format(order))
+    raise ValueError('bad order specified: {}'.format(order))
 
 
 def benedict_bornder_constants(g, critical=False):

filterpy/kalman/IMM.py

@@ -70,7 +70,8 @@ def __init__(self, filters, mu, M):
 
         """
 
-        assert len(filters) > 1
+        if len(filters) < 1:
+            raise ValueError('filters must contain at least one filter')
 
         self.filters = filters
         self.mu = mu

filterpy/kalman/ensemble_kalman_filter.py

@@ -118,7 +118,11 @@ def fx(x, dt):
                 f.update(np.asarray([z]))
         """
 
-        assert dim_z > 0
+        if dim_z <= 0:
+            raise ValueError('dim_z must be greater than zero')
+
+        if N <= 0:
+            raise ValueError('N must be greater than zero')
 
         self.dim_x = len(x)
         self.dim_z = dim_z
@@ -152,9 +156,10 @@ def initialize(self, x, P):
             covariance of the state
         """
 
-        assert x.ndim == 1
-        self.sigmas = multivariate_normal(mean=x, cov=P, size=self.N)
+        if x.ndim != 1:
+            raise ValueError('x must be a 1D array')
 
+        self.sigmas = multivariate_normal(mean=x, cov=P, size=self.N)
         self.x = x
         self.P = P
 

filterpy/kalman/information_filter.py

@@ -101,9 +101,12 @@ class InformationFilter(object):
 
     def __init__(self, dim_x, dim_z, dim_u=0, compute_log_likelihood=True):
 
-        assert dim_x > 0
-        assert dim_z > 0
-        assert dim_u >= 0
+        if dim_z < 1:
+            raise ValueError('dim_x must be 1 or greater')
+        if dim_z < 1:
+            raise ValueError('dim_x must be 1 or greater')
+        if dim_u < 0:
+            raise ValueError('dim_x must be 0 or greater')
 
         self.dim_x = dim_x
         self.dim_z = dim_z

filterpy/kalman/kalman_filter.py

@@ -207,9 +207,12 @@ def __init__(self, dim_x, dim_z, dim_u=0, compute_log_likelihood=True):
             off.
         """
 
-        assert dim_x > 0
-        assert dim_z > 0
-        assert dim_u >= 0
+        if dim_z < 1:
+            raise ValueError('dim_x must be 1 or greater')
+        if dim_z < 1:
+            raise ValueError('dim_x must be 1 or greater')
+        if dim_u < 0:
+            raise ValueError('dim_x must be 0 or greater')
 
         self.dim_x = dim_x
         self.dim_z = dim_z
@@ -297,9 +300,6 @@ def update(self, z, R=None, H=None):
         # predict new x with residual scaled by the kalman gain
         self.x = self.x + dot(self.K, self.y)
 
-        if self.x.ndim == 2:
-            assert self.x.shape[0] == self.dim_x and self.x.shape[1] == 1
-
         # P = (I-KH)P(I-KH)' + KRK'
         I_KH = self._I - dot(self.K, H)
         self.P = dot(dot(I_KH, self.P), I_KH.T) + dot(dot(self.K, R), self.K.T)
@@ -490,7 +490,6 @@ def test_matrix_dimensions(self, z=None, H=None, R=None, F=None, Q=None):
                "Shape of F must be ({},{}), but is {}".format(
                    self.dim_x, self.dim_x, F.shape)
 
-
         assert np.ndim(H) == 2, \
                "Shape of H must be (dim_z, {}), but is {}".format(
                    P.shape[0], shape(H))
@@ -816,7 +815,9 @@ def rts_smoother(self, Xs, Ps, Fs=None, Qs=None):
 
         """
 
-        assert len(Xs) == len(Ps)
+        if len(Xs) != len(Ps):
+            raise ValueError('length of Xs and Ps must be the same')
+
         n = Xs.shape[0]
         dim_x = Xs.shape[1]
 
@@ -992,8 +993,9 @@ def log_likelihood_of(self, z):
 
     @alpha.setter
     def alpha(self, value):
-        assert np.isscalar(value)
-        assert value > 0.
+
+        if not np.isscalar(value) or value <= 0:
+            raise ValueError('alpha must be a float greater than 0')
 
         self._alpha_sq = value**2
 
@@ -1457,7 +1459,9 @@ def rts_smoother(Xs, Ps, Fs, Qs):
         (x, P, K, pP) = rts_smoother(mu, cov, kf.F, kf.Q)
     """
 
-    assert len(Xs) == len(Ps)
+    if len(Xs) != len(Ps):
+        raise ValueError('length of Xs and Ps must be the same')
+
     n = Xs.shape[0]
     dim_x = Xs.shape[1]
 

filterpy/kalman/mmae.py

@@ -70,11 +70,14 @@ def __init__(self, filters, p, dim_x, H=None):
         dim_x : float
             number of random variables in the state X
 
-        H :
+        H : Measurement matrix
         """
 
-        assert len(filters) == len(p)
-        assert dim_x > 0
+        if len(filters) != len(p):
+            raise ValueError('length of filters and p must be the same')
+
+        if dim_x < 1:
+            raise ValueError('dim_x must be >= 1')
 
         self.filters = filters
         self.p = np.asarray(p)

filterpy/kalman/sigma_points.py

@@ -139,8 +139,9 @@ def sigma_points(self, x, P):
             Ordered by Xi_0, Xi_{1..n}, Xi_{n+1..2n}
         """
 
-        assert self.n == np.size(x), "expected size {}, but size is {}".format(
-            self.n, np.size(x))
+        if self.n != np.size(x):
+            raise ValueError("expected size(x) {}, but size is {}".format(
+                self.n, np.size(x)))
 
         n = self.n
 
@@ -314,7 +315,10 @@ def sigma_points(self, x, P):
 
         """
 
-        assert self.n == np.size(x)
+        if self.n != np.size(x):
+            raise ValueError("expected size(x) {}, but size is {}".format(
+                self.n, np.size(x)))
+
         n = self.n
 
         if np.isscalar(x):
@@ -461,8 +465,9 @@ def sigma_points(self, x, P):
             Ordered by Xi_0, Xi_{1..n}
         """
 
-        assert self.n == np.size(x), "expected size {}, but size is {}".format(
-            self.n, np.size(x))
+        if self.n != np.size(x):
+            raise ValueError("expected size(x) {}, but size is {}".format(
+                self.n, np.size(x)))
 
         n = self.n
 

filterpy/kalman/square_root.py

@@ -94,9 +94,12 @@ def __init__(self, dim_x, dim_z, dim_u=0):
             Kalman Filtering. Wiley and sons, 2012.
         """
 
-        assert dim_x > 0
-        assert dim_z > 0
-        assert dim_u >= 0
+        if dim_z < 1:
+            raise ValueError('dim_x must be 1 or greater')
+        if dim_z < 1:
+            raise ValueError('dim_x must be 1 or greater')
+        if dim_u < 0:
+            raise ValueError('dim_x must be 0 or greater')
 
         self.dim_x = dim_x
         self.dim_z = dim_z

filterpy/memory/fading_memory.py

@@ -101,9 +101,8 @@ class FadingMemoryFilter(object):
 
     def __init__(self, x0, dt, order, beta):
 
-
-        assert order >= 0
-        assert order <= 2
+        if order < 0 or order > 2:
+            raise ValueError('order must be between 0 and 2')
 
         if np.isscalar(x0):
             self.x = np.zeros(order+1)