DOC: Update randn() to use rng.normal()

Updates all references to `randn` to use `normal` from Generator.
I left the matlib versions alone.
In a few spots I added `seed=123`.
I removed the `# may vary` tag from all examples.
I reformatted some longer outputs to use Jupyter style output.

[skip actions] [skip azp] [skip cirrus]

Author: bmwoodruff

numpy/_core/shape_base.py

@@ -418,7 +418,7 @@ def stack(arrays, axis=0, out=None, *, dtype=None, casting="same_kind"):
     Examples
     --------
     >>> rng = np.random.default_rng()
-    >>> arrays = [rng.standard_normal(size=(3,4)) for _ in range(10)]
+    >>> arrays = [rng.normal(size=(3,4)) for _ in range(10)]
     >>> np.stack(arrays, axis=0).shape
     (10, 3, 4)
 

numpy/lib/_function_base_impl.py

@@ -324,7 +324,7 @@ def flip(m, axis=None):
            [[1, 0],
             [3, 2]]])
     >>> rng = np.random.default_rng()
-    >>> A = rng.standard_normal(size=(3,4,5))
+    >>> A = rng.normal(size=(3,4,5))
     >>> np.all(np.flip(A,2) == A[:,:,::-1,...])
     True
     """

numpy/lib/_histograms_impl.py

@@ -973,7 +973,7 @@ def histogramdd(sample, bins=10, range=None, density=None, weights=None):
     Examples
     --------
     >>> rng = np.random.default_rng()
-    >>> r = rng.standard_normal(size=(100,3))
+    >>> r = rng.normal(size=(100,3))
     >>> H, edges = np.histogramdd(r, bins = (5, 8, 4))
     >>> H.shape, edges[0].size, edges[1].size, edges[2].size
     ((5, 8, 4), 6, 9, 5)

numpy/lib/_twodim_base_impl.py

@@ -90,7 +90,7 @@ def fliplr(m):
            [3.,  0.,  0.]])
 
     >>> rng = np.random.default_rng()
-    >>> A = rng.standard_normal(size=(2,3,5))
+    >>> A = rng.normal(size=(2,3,5))
     >>> np.all(np.fliplr(A) == A[:,::-1,...])
     True
 
@@ -144,7 +144,7 @@ def flipud(m):
            [1.,  0.,  0.]])
 
     >>> rng = np.random.default_rng()
-    >>> A = rng.standard_normal(size=(2,3,5))
+    >>> A = rng.normal(size=(2,3,5))
     >>> np.all(np.flipud(A) == A[::-1,...])
     True
 

numpy/linalg/_linalg.py

@@ -275,7 +275,7 @@ def tensorsolve(a, b, axes=None):
     >>> a = np.eye(2*3*4)
     >>> a.shape = (2*3, 4, 2, 3, 4)
     >>> rng = np.random.default_rng()
-    >>> b = rng.standard_normal(size=(2*3, 4))
+    >>> b = rng.normal(size=(2*3, 4))
     >>> x = np.linalg.tensorsolve(a, b)
     >>> x.shape
     (2, 3, 4)
@@ -458,7 +458,7 @@ def tensorinv(a, ind=2):
     >>> ainv.shape
     (8, 3, 4, 6)
     >>> rng = np.random.default_rng()
-    >>> b = rng.standard_normal(size=(4, 6))
+    >>> b = rng.normal(size=(4, 6))
     >>> np.allclose(np.tensordot(ainv, b), np.linalg.tensorsolve(a, b))
     True
 
@@ -468,7 +468,7 @@ def tensorinv(a, ind=2):
     >>> ainv.shape
     (8, 3, 24)
     >>> rng = np.random.default_rng()
-    >>> b = rng.standard_normal(24)
+    >>> b = rng.normal(size=24)
     >>> np.allclose(np.tensordot(ainv, b, 1), np.linalg.tensorsolve(a, b))
     True
 
@@ -982,7 +982,7 @@ def qr(a, mode='reduced'):
     Examples
     --------
     >>> rng = np.random.default_rng()
-    >>> a = rng.standard_normal(size=(9, 6))
+    >>> a = rng.normal(size=(9, 6))
     >>> Q, R = np.linalg.qr(a)
     >>> np.allclose(a, np.dot(Q, R))  # a does equal QR
     True
@@ -1708,12 +1708,9 @@ def svd(a, full_matrices=True, compute_uv=True, hermitian=False):
     Examples
     --------
     >>> rng = np.random.default_rng()
-    >>> a_re = rng.standard_normal(size=(9, 6))
-    >>> a_im = rng.standard_normal(size=(9, 6))
-    >>> a = a_re + 1j*a_im
-    >>> b_re = rng.standard_normal(size=(2, 7, 8, 3))
-    >>> b_im = rng.standard_normal(size=(2, 7, 8, 3))
-    >>> b = b_re + 1j*b_im
+    >>> a = rng.normal(size=(9, 6)) + 1j*rng.normal(size=(9, 6))
+    >>> b = rng.normal(size=(2, 7, 8, 3)) + 1j*rng.normal(size=(2, 7, 8, 3))
+
 
     Reconstruction based on full SVD, 2D case:
 
@@ -2191,7 +2188,7 @@ def pinv(a, rcond=None, hermitian=False, *, rtol=_NoValue):
     ``a+ * a * a+ == a+``:
 
     >>> rng = np.random.default_rng()
-    >>> a = rng.standard_normal(size=(9, 6))
+    >>> a = rng.normal(size=(9, 6))
     >>> B = np.linalg.pinv(a)
     >>> np.allclose(a, np.dot(a, np.dot(B, a)))
     True

numpy/polynomial/hermite.py

@@ -1461,10 +1461,10 @@ def hermfit(x, y, deg, rcond=None, full=False, w=None):
     >>> from numpy.polynomial.hermite import hermfit, hermval
     >>> x = np.linspace(-10, 10)
     >>> rng = np.random.default_rng(seed=123)
-    >>> err = rng.standard_normal(size=(len(x)))/10
+    >>> err = rng.normal(scale=1./10,size=len(x))
     >>> y = hermval(x, [1, 2, 3]) + err
     >>> hermfit(x, y, 2)
-    array([1.02294967, 2.00016403, 2.99994614]) # may vary
+    array([1.02294967, 2.00016403, 2.99994614])
 
     """
     return pu._fit(hermvander, x, y, deg, rcond, full, w)

numpy/polynomial/hermite_e.py

@@ -1388,10 +1388,10 @@ def hermefit(x, y, deg, rcond=None, full=False, w=None):
     >>> from numpy.polynomial.hermite_e import hermefit, hermeval
     >>> x = np.linspace(-10, 10)
     >>> rng = np.random.default_rng(seed=123)
-    >>> err = rng.standard_normal(size=(len(x)))/10
+    >>> err = rng.normal(scale=1./10,size=len(x))
     >>> y = hermeval(x, [1, 2, 3]) + err
     >>> hermefit(x, y, 2)
-    array([1.02284196, 2.00032805, 2.99978457]) # may vary
+    array([1.02284196, 2.00032805, 2.99978457])
 
     """
     return pu._fit(hermevander, x, y, deg, rcond, full, w)

numpy/polynomial/laguerre.py

@@ -1442,10 +1442,10 @@ def lagfit(x, y, deg, rcond=None, full=False, w=None):
     >>> from numpy.polynomial.laguerre import lagfit, lagval
     >>> x = np.linspace(0, 10)
     >>> rng = np.random.default_rng(seed=123)
-    >>> err = rng.standard_normal(size=(len(x)))/10
+    >>> err = rng.normal(scale=1./10,size=len(x))
     >>> y = lagval(x, [1, 2, 3]) + err
     >>> lagfit(x, y, 2)
-    array([1.00578369, 1.99417356, 2.99827656]) # may vary
+    array([1.00578369, 1.99417356, 2.99827656])
 
     """
     return pu._fit(lagvander, x, y, deg, rcond, full, w)

numpy/polynomial/polynomial.py

@@ -1444,16 +1444,16 @@ def polyfit(x, y, deg, rcond=None, full=False, w=None):
     >>> from numpy.polynomial import polynomial as P
     >>> x = np.linspace(-1,1,51)  # x "data": [-1, -0.96, ..., 0.96, 1]
     >>> rng = np.random.default_rng(seed=123)
-    >>> err = rng.standard_normal(size=len(x))
+    >>> err = rng.normal(size=len(x))
     >>> y = x**3 - x + err  # x^3 - x + Gaussian noise
     >>> c, stats = P.polyfit(x,y,3,full=True)
     >>> c # c[0], c[1] approx. -1, c[2] should be approx. 0, c[3] approx. 1
-    array([ 0.23111996, -1.02785049, -0.2241444 ,  1.08405657]) # may vary
+    array([ 0.23111996, -1.02785049, -0.2241444 ,  1.08405657])
     >>> stats # note the large SSR, explaining the rather poor results
-    [array([48.312088]),                                      # may vary
-     4,                                                       # may vary
-     array([1.38446749, 1.32119158, 0.50443316, 0.28853036]), # may vary
-     1.1324274851176597e-14]                                  # may vary
+    [array([48.312088]),
+     4,
+     array([1.38446749, 1.32119158, 0.50443316, 0.28853036]),
+     1.1324274851176597e-14]
 
     Same thing without the added noise
 
@@ -1462,10 +1462,10 @@ def polyfit(x, y, deg, rcond=None, full=False, w=None):
     >>> c # c[0], c[1] ~= -1, c[2] should be "very close to 0", c[3] ~= 1
     array([-6.73496154e-17, -1.00000000e+00,  0.00000000e+00,  1.00000000e+00])
     >>> stats # note the minuscule SSR
-    [array([8.79579319e-31]),                                 # may vary
-     4,                                                       # may vary
-     array([1.38446749, 1.32119158, 0.50443316, 0.28853036]), # may vary
-     1.1324274851176597e-14]                                  # may vary
+    [array([8.79579319e-31]),
+     4,
+     array([1.38446749, 1.32119158, 0.50443316, 0.28853036]),
+     1.1324274851176597e-14]
 
     """
     return pu._fit(polyvander, x, y, deg, rcond, full, w)