style

pyobs/core/cdata.py

@@ -36,7 +36,7 @@ def __init__(self, cov, mask):
         for a in self.mask:
             ia = self.mask.index(a)
             # special case cobs_scalar * obs_vector, then mask is vector but n=1
-            if n==1:
+            if n == 1:
                 self.grad[ia, 0] = 1.0
             else:
                 self.grad[ia, a] = 1.0

pyobs/core/derobs.py

@@ -177,6 +177,6 @@ def error_bias4(x, f):
             x.delta[key].delta,
         )
         dd2 = numpy.sum(d2, axis=1)
-        bias4 += dd2 ** 2 / x.delta[key].n ** 4
+        bias4 += dd2**2 / x.delta[key].n ** 4
 
     return numpy.sqrt(bias4)

pyobs/core/error.py

@@ -104,7 +104,7 @@ def find_opt(self):
 
     def set_opt(self, xopt):
         for a in range(self.size):
-            if isinstance(xopt,(list, numpy.ndarray)):
+            if isinstance(xopt, (list, numpy.ndarray)):
                 if xopt[a] in self.x:
                     i = self.x.index(xopt[a])
                 else:
@@ -131,9 +131,11 @@ def correct_gamma_bias(self):
             # self.cvar[a, :] += self.cvar[a, :] * f
             # Gamma -> Gamma + C/N
             # cumsum(Gamma) -> cumsum(Gamma) + cumsum(C/N)
-            self.cvar[a, :] += numpy.arange(1,len(self.cvar[a,:])+1)*self.var[a,0] / self.N[a]
+            self.cvar[a, :] += (
+                numpy.arange(1, len(self.cvar[a, :]) + 1) * self.var[a, 0] / self.N[a]
+            )
         self.set_opt(self.xopt)
-        
+
     def tauint(self):
         tau = numpy.zeros((self.full_size, 2))
         for a in self.mask:

pyobs/core/ndobs.py

@@ -509,7 +509,7 @@ def __matmul__(self, y):
 
     def reciprocal(self):
         new_mean = numpy.reciprocal(self.mean)
-        g0 = pyobs.gradient(lambda x: -x * (new_mean ** 2), self.mean, gtype="diag")
+        g0 = pyobs.gradient(lambda x: -x * (new_mean**2), self.mean, gtype="diag")
         return pyobs.derobs([self], new_mean, [g0])
 
     def __truediv__(self, y):
@@ -538,7 +538,7 @@ def __rmatmul__(self, y):
         return pyobs.derobs([self], y @ self.mean, [g0])
 
     def __pow__(self, a):
-        new_mean = self.mean ** a
+        new_mean = self.mean**a
         g0 = pyobs.gradient(
             lambda x: a * x * self.mean ** (a - 1), self.mean, gtype="diag"
         )

pyobs/misc/memory.py

@@ -22,8 +22,8 @@
 import sys, os, numpy
 
 book = {}
-MB = 1024.0 ** 2
-GB = 1024.0 ** 3
+MB = 1024.0**2
+GB = 1024.0**3
 
 
 def get_size(obj):

pyobs/misc/random.py

@@ -59,9 +59,8 @@ def sample_boolean(self, N):
         self.state = numpy.random.get_state()
         return r
 
-
     def acrand(self, tau, N, n=1):
-        r = numpy.reshape(self.sample_normal(N*n), (N,n))
+        r = numpy.reshape(self.sample_normal(N * n), (N, n))
 
         if tau > 0.5:
             f = numpy.exp(-1.0 / tau)
@@ -70,19 +69,18 @@ def acrand(self, tau, N, n=1):
             tau = 0.5
         ff = numpy.sqrt(1.0 - f * f)
         rn = numpy.zeros(numpy.shape(r))
-        rn[0,:] = ff * r[0,:]
+        rn[0, :] = ff * r[0, :]
         for i in range(1, N):
-            rn[i,:] = ff * r[i,:] + f * rn[i - 1,:]
+            rn[i, :] = ff * r[i, :] + f * rn[i - 1, :]
         return rn
-
 
     def markov_chain(self, mu, cov, taus, N, couplings=None):
         """
         Create synthetic autocorrelated Monte Carlo data
 
         Parameters:
            mu (float or 1D array): the central value
-           cov (float or array): the target covariance matrix; if a 1-D array is 
+           cov (float or array): the target covariance matrix; if a 1-D array is
               passed, a diagonal covariance matrix is assumed
            taus (float or array): the autocorrelation time(s). Values smaller
            than 0.5 are ignored and set to automatically to 0.5.
@@ -103,38 +101,44 @@ def markov_chain(self, mu, cov, taus, N, couplings=None):
 
         """
         mu = pyobs.array(mu)
-        pyobs.assertion(numpy.ndim(mu)==1, "only 1D arrays are supported")
+        pyobs.assertion(numpy.ndim(mu) == 1, "only 1D arrays are supported")
         na = len(mu)
-        
+
         cov = pyobs.array(cov)
-        pyobs.assertion(numpy.shape(cov)[0]==na, "covariance matrix does not match central values")
-
+        pyobs.assertion(
+            numpy.shape(cov)[0] == na, "covariance matrix does not match central values"
+        )
+
         taus = pyobs.array(taus)
         taus = 0.5 * (taus <= 0.5) + taus * (taus > 0.5)
         nt = len(taus)
         if couplings is None:
-            couplings = pyobs.array(numpy.ones((na,nt)))
+            couplings = pyobs.array(numpy.ones((na, nt)))
         else:
             couplings = pyobs.array(couplings)
-            pyobs.assertion(numpy.shape(couplings)==(na,nt), f"unexpected couplings for {na} values and {nt} modes")
-
-        rn = numpy.zeros((N,na))
-        _c = numpy.stack([couplings]*N)
+            pyobs.assertion(
+                numpy.shape(couplings) == (na, nt),
+                f"unexpected couplings for {na} values and {nt} modes",
+            )
+
+        rn = numpy.zeros((N, na))
+        _c = numpy.stack([couplings] * N)
         for i in range(len(taus)):
-            rn += _c[:,:,i] * self.acrand(taus[i], N, na)
-        
+            rn += _c[:, :, i] * self.acrand(taus[i], N, na)
+
         pref = numpy.sqrt(N / (2 * (couplings**2 @ taus)))
-        
+
         if numpy.ndim(cov) == 1:
             Q = numpy.diag(numpy.sqrt(cov))
         else:
             [w, v] = numpy.linalg.eig(cov)
             Q = numpy.diag(numpy.sqrt(w)) @ v.T
-        
-        if na==1:
+
+        if na == 1:
             return (mu + (pref * rn) @ Q).reshape((N,))
         return mu + (pref * rn) @ Q
 
+
 #     def acrandn(self, mu, cov, tau, N):
 #         """
 #         Create synthetic correlated Monte Carlo 1-D data

pyobs/optimize/interpolation.py

@@ -84,5 +84,5 @@ def __call__(self, x):
             x = numpy.array(x)
         res = pyobs.repeat(self.coeff[0], N)
         for i in range(1, self.k):
-            res += pyobs.repeat(self.coeff[i], N) * (x ** i)
+            res += pyobs.repeat(self.coeff[i], N) * (x**i)
         return res

pyobs/qft/free_scalar.py

@@ -24,7 +24,7 @@
 
 def Gt(t, m, p, L):
     phat = 2 * numpy.sin(numpy.array(p) * numpy.pi / numpy.array(L))
-    om = numpy.arccosh(1.0 + 0.5 * (m ** 2 + numpy.sum(phat ** 2)))
+    om = numpy.arccosh(1.0 + 0.5 * (m**2 + numpy.sum(phat**2)))
     return numpy.exp(-numpy.abs(t) * om) * 0.5 / numpy.sinh(om)
 
 

pyobs/tensor/linalg.py

@@ -195,4 +195,4 @@ def matrix_power(x, a):
        >>> matsqrt @ mat @ matsqrt # return the identity
     """
     [w, v] = eig(x)
-    return v @ pyobs.diag(w ** a) @ pyobs.transpose(v)
+    return v @ pyobs.diag(w**a) @ pyobs.transpose(v)

pyobs/tensor/unary.py

@@ -239,7 +239,7 @@ def arctan(x):
     Examples:
        >>> arctanA = pyobs.arctan(obsA)
     """
-    return __unary(x, numpy.arctan, lambda x: 1 / (1 + x ** 2))
+    return __unary(x, numpy.arctan, lambda x: 1 / (1 + x**2))
 
 
 def cosh(x):
@@ -287,7 +287,7 @@ def arccosh(x):
     Examples:
        >>> B = pyobs.arccosh(obsA)
     """
-    return __unary(x, numpy.arccosh, lambda x: 1.0 / numpy.sqrt(x ** 2 - 1))
+    return __unary(x, numpy.arccosh, lambda x: 1.0 / numpy.sqrt(x**2 - 1))
 
 
 #     new_mean = numpy.arccosh(x.mean)

pyobs/utils.py

@@ -168,7 +168,7 @@ def tex_table(mat, fmt=None):
             else:
                 h += [f"%{fmt[i]}" % mat[a, i]]
                 i += 1
-        outstr += [fr'{" & ".join(h)} \\ ']
+        outstr += [rf'{" & ".join(h)} \\ ']
 
     return outstr
 
@@ -255,7 +255,8 @@ def core(string):
         return numpy.array(out)
     return core(data)
 
+
 def array(x):
-    if numpy.shape(x)==():
+    if numpy.shape(x) == ():
         return numpy.array(x).reshape((1,))
     return numpy.array(x)