Python numpy.common_type() Examples

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def common_type(*arrays):
    """Return a scalar type which is common to the input arrays.

    .. seealso:: :func:`numpy.common_type`
    """
    if len(arrays) == 0:
        return numpy.float16

    default_float_dtype = numpy.dtype('float64')
    dtypes = []
    for a in arrays:
        if a.dtype.kind == 'b':
            raise TypeError('can\'t get common type for non-numeric array')
        elif a.dtype.kind in 'iu':
            dtypes.append(default_float_dtype)
        else:
            dtypes.append(a.dtype)

    return functools.reduce(numpy.promote_types, dtypes).type 

def test_df_arith_2d_array_collike_broadcasts(self,
                                                  all_arithmetic_operators):
        # GH#23000
        opname = all_arithmetic_operators

        arr = np.arange(6).reshape(3, 2)
        df = pd.DataFrame(arr, columns=[True, False], index=['A', 'B', 'C'])

        collike = arr[:, [1]]  # shape --> (nrows, 1)
        assert collike.shape == (df.shape[0], 1)

        exvals = {True: getattr(df[True], opname)(collike.squeeze()),
                  False: getattr(df[False], opname)(collike.squeeze())}

        dtype = None
        if opname in ['__rmod__', '__rfloordiv__']:
            # Series ops may return mixed int/float dtypes in cases where
            #   DataFrame op will return all-float.  So we upcast `expected`
            dtype = np.common_type(*[x.values for x in exvals.values()])

        expected = pd.DataFrame(exvals, columns=df.columns, index=df.index,
                                dtype=dtype)

        result = getattr(df, opname)(collike)
        tm.assert_frame_equal(result, expected) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def random_MPS(L, d, chimax, func=randmat.standard_normal_complex, bc='finite', form='B'):
    site = Site(charges.LegCharge.from_trivial(d))
    chi = [chimax] * (L + 1)
    if bc == 'finite':
        for i in range(L // 2 + 1):
            chi[i] = chi[L - i] = min(chi[i], d**i)
    Bs = []
    for i in range(L):
        B = func((d, chi[i], chi[i + 1]))
        B /= np.sqrt(chi[i + 1]) * d
        Bs.append(B)
    dtype = np.common_type(*Bs)
    psi = MPS.from_Bflat([site] * L, Bs, bc=bc, dtype=dtype, form=None)
    if form is not None:
        psi.canonical_form()
        psi.convert_form(form)
    return psi 

def dot_generalized(a, b):
    a = asarray(a)
    if a.ndim >= 3:
        if a.ndim == b.ndim:
            # matrix x matrix
            new_shape = a.shape[:-1] + b.shape[-1:]
        elif a.ndim == b.ndim + 1:
            # matrix x vector
            new_shape = a.shape[:-1]
        else:
            raise ValueError("Not implemented...")
        r = np.empty(new_shape, dtype=np.common_type(a, b))
        for c in itertools.product(*map(range, a.shape[:-2])):
            r[c] = dot(a[c], b[c])
        return r
    else:
        return dot(a, b) 

def test_df_arith_2d_array_collike_broadcasts(self,
                                                  all_arithmetic_operators):
        # GH#23000
        opname = all_arithmetic_operators

        arr = np.arange(6).reshape(3, 2)
        df = pd.DataFrame(arr, columns=[True, False], index=['A', 'B', 'C'])

        collike = arr[:, [1]]  # shape --> (nrows, 1)
        assert collike.shape == (df.shape[0], 1)

        exvals = {True: getattr(df[True], opname)(collike.squeeze()),
                  False: getattr(df[False], opname)(collike.squeeze())}

        dtype = None
        if opname in ['__rmod__', '__rfloordiv__']:
            # Series ops may return mixed int/float dtypes in cases where
            #   DataFrame op will return all-float.  So we upcast `expected`
            dtype = np.common_type(*[x.values for x in exvals.values()])

        expected = pd.DataFrame(exvals, columns=df.columns, index=df.index,
                                dtype=dtype)

        result = getattr(df, opname)(collike)
        tm.assert_frame_equal(result, expected) 

def test_common_type(self):
        self.check(np.common_type) 

def sparse_transform(m, *args):
    """
    Performs a sparse transform of a dense tensor.
    Args:
        m (ndarray): a tensor to transform;
        *args: alternating indexes and bases to transform into;

    Returns:
        The transformed tensor.
    """
    result = m
    for i, (index, basis) in enumerate(zip(args[::2], args[1::2])):

        if len(basis.shape) != 2:
            raise ValueError("Transform {:d} is not a matrix: shape = {}".format(
                i, repr(basis.shape)
            ))
        if result.shape[index] != basis.shape[0]:
            raise ValueError("Dimension mismatch of transform {:d}: m.shape[{:d}] = {:d} != basis.shape[0] = {:d}".format(
                i, index, result.shape[index], basis.shape[0],
            ))
        if "getcol" not in dir(basis):
            raise ValueError("No 'getcol' in the transform matrix {:d}: not a CSC sparse matrix?")

        result_shape = result.shape[:index] + (basis.shape[1],) + result.shape[index + 1:]
        new_result = numpy.zeros(result_shape, numpy.common_type(*(
            args[1::2] + (m,)
        )))
        for b2 in range(basis.shape[1]):
            slice_b2 = (slice(None),) * index + (b2,)
            col = basis.getcol(b2)
            for b1 in col.nonzero()[0]:
                slice_b1 = (slice(None),) * index + (b1,)
                new_result[slice_b2] += col[b1, 0] * result[slice_b1]
        result = new_result

    return result 

def get_sparse_ov_transform(oo, vv):
    """
    Retrieves a sparse `ovov` transform out of sparse `oo` and `vv` transforms.
    Args:
        oo (ndarray): the transformation in the occupied space;
        vv (ndarray): the transformation in the virtual space;

    Returns:
        The resulting matrix representing the sparse transform in the `ov` space.
    """
    i, a = oo.shape
    j, b = vv.shape

    # If the input is dense the result is simply
    # return (oo[:, numpy.newaxis, :, numpy.newaxis] * vv[numpy.newaxis, :, numpy.newaxis, :]).reshape(i*j, a*b)

    result_data = numpy.zeros(oo.nnz * vv.nnz, dtype=numpy.common_type(oo, vv))
    result_indices = numpy.zeros(len(result_data), dtype=int)
    result_indptr = numpy.zeros(a * b + 1, dtype=int)

    ptr_counter = 0
    for i_a in range(a):
        oo_col = oo.getcol(i_a)
        assert tuple(oo_col.indptr.tolist()) == (0, len(oo_col.data))
        i_i, oo_col_v = oo_col.indices, oo_col.data
        for i_b in range(b):
            vv_col = vv.getcol(i_b)
            assert tuple(vv_col.indptr.tolist()) == (0, len(vv_col.data))
            i_j, vv_col_v = vv_col.indices, vv_col.data

            data_length = len(i_i) * len(i_j)
            result_indices[ptr_counter:ptr_counter + data_length] = ((i_i * j)[:, numpy.newaxis] + i_j[numpy.newaxis, :]).reshape(-1)
            result_data[ptr_counter:ptr_counter + data_length] = (oo_col_v[:, numpy.newaxis] * vv_col_v[numpy.newaxis, :]).reshape(-1)
            result_indptr[i_a * b + i_b] = ptr_counter

            ptr_counter += data_length

    result_indptr[-1] = ptr_counter
    return sparse.csc_matrix((result_data, result_indices, result_indptr)) 

def remove_phase_difference(v1, v2, axis=0, threshold=1e-5):
    """Removes the phase difference between two vectors."""
    dtype = numpy.common_type(numpy.asarray(v1), numpy.asarray(v2))
    v1, v2 = numpy.array(v1, dtype=dtype), numpy.array(v2, dtype=dtype)
    v1, v2 = pull_dim(v1, axis), pull_dim(v2, axis)
    v2 /= phase_difference(v1, v2, threshold=threshold)[:, numpy.newaxis]
    return v1, v2 

def remove_phase_difference(v1, v2, axis=0, threshold=1e-5):
    """Removes the phase difference between two vectors."""
    dtype = numpy.common_type(numpy.asarray(v1), numpy.asarray(v2))
    v1, v2 = numpy.array(v1, dtype=dtype), numpy.array(v2, dtype=dtype)
    v1, v2 = pull_dim(v1, axis), pull_dim(v2, axis)
    v2 /= phase_difference(v1, v2, threshold=threshold)[:, numpy.newaxis]
    return v1, v2 

def remove_phase_difference(v1, v2, axis=0, threshold=1e-5):
    """Removes the phase difference between two vectors."""
    dtype = numpy.common_type(numpy.asarray(v1), numpy.asarray(v2))
    v1, v2 = numpy.array(v1, dtype=dtype), numpy.array(v2, dtype=dtype)
    v1, v2 = pull_dim(v1, axis), pull_dim(v2, axis)
    v2 /= phase_difference(v1, v2, threshold=threshold)[:, numpy.newaxis]
    return v1, v2 

def test_array_function_common_type(self, xp):
        return numpy.common_type(xp.arange(2, dtype='f8'),
                                 xp.arange(2, dtype='f4')) 

def dtype(self):
        """Returns the dtype that should be returned by ``to_array``"""
        return np.common_type(*tuple(self._lt)) 

def nulp_diff(x, y, dtype=None):
    """For each item in x and y, return the number of representable floating
    points between them.

    Parameters
    ----------
    x : array_like
        first input array
    y : array_like
        second input array
    dtype : dtype, optional
        Data-type to convert `x` and `y` to if given. Default is None.

    Returns
    -------
    nulp : array_like
        number of representable floating point numbers between each item in x
        and y.

    Examples
    --------
    # By definition, epsilon is the smallest number such as 1 + eps != 1, so
    # there should be exactly one ULP between 1 and 1 + eps
    >>> nulp_diff(1, 1 + np.finfo(x.dtype).eps)
    1.0
    """
    import numpy as np
    if dtype:
        x = np.array(x, dtype=dtype)
        y = np.array(y, dtype=dtype)
    else:
        x = np.array(x)
        y = np.array(y)

    t = np.common_type(x, y)
    if np.iscomplexobj(x) or np.iscomplexobj(y):
        raise NotImplementedError("_nulp not implemented for complex array")

    x = np.array(x, dtype=t)
    y = np.array(y, dtype=t)

    if not x.shape == y.shape:
        raise ValueError("x and y do not have the same shape: %s - %s" %
                         (x.shape, y.shape))

    def _diff(rx, ry, vdt):
        diff = np.array(rx-ry, dtype=vdt)
        return np.abs(diff)

    rx = integer_repr(x)
    ry = integer_repr(y)
    return _diff(rx, ry, t) 

def nulp_diff(x, y, dtype=None):
    """For each item in x and y, return the number of representable floating
    points between them.

    Parameters
    ----------
    x : array_like
        first input array
    y : array_like
        second input array
    dtype : dtype, optional
        Data-type to convert `x` and `y` to if given. Default is None.

    Returns
    -------
    nulp : array_like
        number of representable floating point numbers between each item in x
        and y.

    Examples
    --------
    # By definition, epsilon is the smallest number such as 1 + eps != 1, so
    # there should be exactly one ULP between 1 and 1 + eps
    >>> nulp_diff(1, 1 + np.finfo(x.dtype).eps)
    1.0
    """
    import numpy as np
    if dtype:
        x = np.array(x, dtype=dtype)
        y = np.array(y, dtype=dtype)
    else:
        x = np.array(x)
        y = np.array(y)

    t = np.common_type(x, y)
    if np.iscomplexobj(x) or np.iscomplexobj(y):
        raise NotImplementedError("_nulp not implemented for complex array")

    x = np.array(x, dtype=t)
    y = np.array(y, dtype=t)

    if not x.shape == y.shape:
        raise ValueError("x and y do not have the same shape: %s - %s" %
                         (x.shape, y.shape))

    def _diff(rx, ry, vdt):
        diff = np.array(rx-ry, dtype=vdt)
        return np.abs(diff)

    rx = integer_repr(x)
    ry = integer_repr(y)
    return _diff(rx, ry, t)