rm CondaPkg environment

This commit is contained in:
ton
2023-04-06 13:53:47 +07:00
parent 0a57ed7884
commit c43d949309
3329 changed files with 5725 additions and 447022 deletions

View File

@@ -1,98 +0,0 @@
import numpy as np
import pytest
from skimage._shared.utils import _supported_float_type
from skimage.registration import optical_flow_ilk
from .test_tvl1 import _sin_flow_gen
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
@pytest.mark.parametrize('gaussian', [True, False])
@pytest.mark.parametrize('prefilter', [True, False])
def test_2d_motion(dtype, gaussian, prefilter):
# Generate synthetic data
rnd = np.random.default_rng(0)
image0 = rnd.normal(size=(256, 256))
gt_flow, image1 = _sin_flow_gen(image0)
image1 = image1.astype(dtype, copy=False)
float_dtype = _supported_float_type(dtype)
# Estimate the flow
flow = optical_flow_ilk(image0, image1, gaussian=gaussian,
prefilter=prefilter, dtype=float_dtype)
assert flow.dtype == _supported_float_type(dtype)
# Assert that the average absolute error is less then half a pixel
assert abs(flow - gt_flow).mean() < 0.5
if dtype != float_dtype:
with pytest.raises(ValueError):
optical_flow_ilk(image0, image1, gaussian=gaussian,
prefilter=prefilter, dtype=dtype)
@pytest.mark.parametrize('gaussian', [True, False])
@pytest.mark.parametrize('prefilter', [True, False])
def test_3d_motion(gaussian, prefilter):
# Generate synthetic data
rnd = np.random.default_rng(123)
image0 = rnd.normal(size=(50, 55, 60))
gt_flow, image1 = _sin_flow_gen(image0, npics=3)
# Estimate the flow
flow = optical_flow_ilk(image0, image1, radius=5,
gaussian=gaussian, prefilter=prefilter)
# Assert that the average absolute error is less then half a pixel
assert abs(flow - gt_flow).mean() < 0.5
def test_no_motion_2d():
rnd = np.random.default_rng(0)
img = rnd.normal(size=(256, 256))
flow = optical_flow_ilk(img, img)
assert np.all(flow == 0)
def test_no_motion_3d():
rnd = np.random.default_rng(0)
img = rnd.normal(size=(64, 64, 64))
flow = optical_flow_ilk(img, img)
assert np.all(flow == 0)
def test_optical_flow_dtype():
# Generate synthetic data
rnd = np.random.default_rng(0)
image0 = rnd.normal(size=(256, 256))
gt_flow, image1 = _sin_flow_gen(image0)
# Estimate the flow at double precision
flow_f64 = optical_flow_ilk(image0, image1, dtype='float64')
assert flow_f64.dtype == 'float64'
# Estimate the flow at single precision
flow_f32 = optical_flow_ilk(image0, image1, dtype='float32')
assert flow_f32.dtype == 'float32'
# Assert that floating point precision does not affect the quality
# of the estimated flow
assert abs(flow_f64 - flow_f32).mean() < 1e-3
def test_incompatible_shapes():
rnd = np.random.default_rng(0)
I0 = rnd.normal(size=(256, 256))
I1 = rnd.normal(size=(255, 256))
with pytest.raises(ValueError):
u, v = optical_flow_ilk(I0, I1)
def test_wrong_dtype():
rnd = np.random.default_rng(0)
img = rnd.normal(size=(256, 256))
with pytest.raises(ValueError):
u, v = optical_flow_ilk(img, img, dtype='int')

View File

@@ -1,279 +0,0 @@
import numpy as np
import pytest
from numpy.testing import (assert_almost_equal, assert_array_almost_equal,
assert_array_equal, assert_array_less, assert_equal)
from scipy.ndimage import fourier_shift, shift as real_shift
import scipy.fft as fft
from skimage._shared.testing import fetch
from skimage._shared.utils import _supported_float_type
from skimage.data import camera, brain
from skimage.io import imread
from skimage.registration._masked_phase_cross_correlation import (
_masked_phase_cross_correlation as masked_register_translation,
cross_correlate_masked)
from skimage.registration import phase_cross_correlation
def test_masked_registration_vs_phase_cross_correlation():
"""masked_register_translation should give the same results as
phase_cross_correlation in the case of trivial masks."""
reference_image = camera()
shift = (-7, 12)
shifted = np.real(fft.ifft2(fourier_shift(
fft.fft2(reference_image), shift)))
trivial_mask = np.ones_like(reference_image)
nonmasked_result, *_ = phase_cross_correlation(reference_image, shifted)
masked_result = masked_register_translation(reference_image,
shifted,
reference_mask=trivial_mask,
overlap_ratio=1 / 10)
assert_equal(nonmasked_result, masked_result)
def test_masked_registration_random_masks():
"""masked_register_translation should be able to register translations
between images even with random masks."""
# See random number generator for reproducible results
np.random.seed(23)
reference_image = camera()
shift = (-7, 12)
shifted = np.real(fft.ifft2(fourier_shift(
fft.fft2(reference_image), shift)))
# Random masks with 75% of pixels being valid
ref_mask = np.random.choice(
[True, False], reference_image.shape, p=[3 / 4, 1 / 4])
shifted_mask = np.random.choice(
[True, False], shifted.shape, p=[3 / 4, 1 / 4])
measured_shift = masked_register_translation(reference_image,
shifted,
reference_mask=ref_mask,
moving_mask=shifted_mask)
assert_equal(measured_shift, -np.array(shift))
def test_masked_registration_3d_contiguous_mask():
"""masked_register_translation should be able to register translations
between volumes with contiguous masks."""
ref_vol = brain()[:, ::2, ::2]
offset = (1, -5, 10)
# create square mask
ref_mask = np.zeros_like(ref_vol, dtype=bool)
ref_mask[:-2, 75:100, 75:100] = True
ref_shifted = real_shift(ref_vol, offset)
measured_offset = masked_register_translation(
ref_vol, ref_shifted, reference_mask=ref_mask, moving_mask=ref_mask
)
assert_equal(offset, -np.array(measured_offset))
def test_masked_registration_random_masks_non_equal_sizes():
"""masked_register_translation should be able to register
translations between images that are not the same size even
with random masks."""
# See random number generator for reproducible results
np.random.seed(23)
reference_image = camera()
shift = (-7, 12)
shifted = np.real(fft.ifft2(fourier_shift(
fft.fft2(reference_image), shift)))
# Crop the shifted image
shifted = shifted[64:-64, 64:-64]
# Random masks with 75% of pixels being valid
ref_mask = np.random.choice(
[True, False], reference_image.shape, p=[3 / 4, 1 / 4])
shifted_mask = np.random.choice(
[True, False], shifted.shape, p=[3 / 4, 1 / 4])
measured_shift = masked_register_translation(
reference_image,
shifted,
reference_mask=np.ones_like(ref_mask),
moving_mask=np.ones_like(shifted_mask))
assert_equal(measured_shift, -np.array(shift))
def test_masked_registration_padfield_data():
""" Masked translation registration should behave like in the original
publication """
# Test translated from MATLABimplementation `MaskedFFTRegistrationTest`
# file. You can find the source code here:
# http://www.dirkpadfield.com/Home/MaskedFFTRegistrationCode.zip
shifts = [(75, 75), (-130, 130), (130, 130)]
for xi, yi in shifts:
fixed_image = imread(
fetch(f'registration/tests/data/OriginalX{xi}Y{yi}.png'))
moving_image = imread(
fetch(f'registration/tests/data/TransformedX{xi}Y{yi}.png'))
# Valid pixels are 1
fixed_mask = (fixed_image != 0)
moving_mask = (moving_image != 0)
# Note that shifts in x and y and shifts in cols and rows
shift_y, shift_x = masked_register_translation(
fixed_image, moving_image, reference_mask=fixed_mask,
moving_mask=moving_mask, overlap_ratio=0.1)
# Note: by looking at the test code from Padfield's
# MaskedFFTRegistrationCode repository, the
# shifts were not xi and yi, but xi and -yi
assert_equal((shift_x, shift_y), (-xi, yi))
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
def test_cross_correlate_masked_output_shape(dtype):
"""Masked normalized cross-correlation should return a shape
of N + M + 1 for each transform axis."""
shape1 = (15, 4, 5)
shape2 = (6, 12, 7)
expected_full_shape = tuple(np.array(shape1) + np.array(shape2) - 1)
expected_same_shape = shape1
arr1 = np.zeros(shape1, dtype=dtype)
arr2 = np.zeros(shape2, dtype=dtype)
# Trivial masks
m1 = np.ones_like(arr1)
m2 = np.ones_like(arr2)
float_dtype = _supported_float_type(dtype)
full_xcorr = cross_correlate_masked(
arr1, arr2, m1, m2, axes=(0, 1, 2), mode='full')
assert_equal(full_xcorr.shape, expected_full_shape)
assert full_xcorr.dtype == float_dtype
same_xcorr = cross_correlate_masked(
arr1, arr2, m1, m2, axes=(0, 1, 2), mode='same')
assert_equal(same_xcorr.shape, expected_same_shape)
assert same_xcorr.dtype == float_dtype
def test_cross_correlate_masked_test_against_mismatched_dimensions():
"""Masked normalized cross-correlation should raise an error if array
dimensions along non-transformation axes are mismatched."""
shape1 = (23, 1, 1)
shape2 = (6, 2, 2)
arr1 = np.zeros(shape1)
arr2 = np.zeros(shape2)
# Trivial masks
m1 = np.ones_like(arr1)
m2 = np.ones_like(arr2)
with pytest.raises(ValueError):
cross_correlate_masked(arr1, arr2, m1, m2, axes=(1, 2))
def test_cross_correlate_masked_output_range():
"""Masked normalized cross-correlation should return between 1 and -1."""
# See random number generator for reproducible results
np.random.seed(23)
# Array dimensions must match along non-transformation axes, in
# this case
# axis 0
shape1 = (15, 4, 5)
shape2 = (15, 12, 7)
# Initial array ranges between -5 and 5
arr1 = 10 * np.random.random(shape1) - 5
arr2 = 10 * np.random.random(shape2) - 5
# random masks
m1 = np.random.choice([True, False], arr1.shape)
m2 = np.random.choice([True, False], arr2.shape)
xcorr = cross_correlate_masked(arr1, arr2, m1, m2, axes=(1, 2))
# No assert array less or equal, so we add an eps
# Also could not find an `assert_array_greater`, Use (-xcorr) instead
eps = np.finfo(float).eps
assert_array_less(xcorr, 1 + eps)
assert_array_less(-xcorr, 1 + eps)
def test_cross_correlate_masked_side_effects():
"""Masked normalized cross-correlation should not modify the inputs."""
shape1 = (2, 2, 2)
shape2 = (2, 2, 2)
arr1 = np.zeros(shape1)
arr2 = np.zeros(shape2)
# Trivial masks
m1 = np.ones_like(arr1)
m2 = np.ones_like(arr2)
for arr in (arr1, arr2, m1, m2):
arr.setflags(write=False)
cross_correlate_masked(arr1, arr2, m1, m2)
def test_cross_correlate_masked_over_axes():
"""Masked normalized cross-correlation over axes should be
equivalent to a loop over non-transform axes."""
# See random number generator for reproducible results
np.random.seed(23)
arr1 = np.random.random((8, 8, 5))
arr2 = np.random.random((8, 8, 5))
m1 = np.random.choice([True, False], arr1.shape)
m2 = np.random.choice([True, False], arr2.shape)
# Loop over last axis
with_loop = np.empty_like(arr1, dtype=complex)
for index in range(arr1.shape[-1]):
with_loop[:, :, index] = cross_correlate_masked(arr1[:, :, index],
arr2[:, :, index],
m1[:, :, index],
m2[:, :, index],
axes=(0, 1),
mode='same')
over_axes = cross_correlate_masked(
arr1, arr2, m1, m2, axes=(0, 1), mode='same')
assert_array_almost_equal(with_loop, over_axes)
def test_cross_correlate_masked_autocorrelation_trivial_masks():
"""Masked normalized cross-correlation between identical arrays
should reduce to an autocorrelation even with random masks."""
# See random number generator for reproducible results
np.random.seed(23)
arr1 = camera()
# Random masks with 75% of pixels being valid
m1 = np.random.choice([True, False], arr1.shape, p=[3 / 4, 1 / 4])
m2 = np.random.choice([True, False], arr1.shape, p=[3 / 4, 1 / 4])
xcorr = cross_correlate_masked(arr1, arr1, m1, m2, axes=(0, 1),
mode='same', overlap_ratio=0).real
max_index = np.unravel_index(np.argmax(xcorr), xcorr.shape)
# Autocorrelation should have maximum in center of array
# uint8 inputs will be processed in float32, so reduce decimal to 5
assert_almost_equal(xcorr.max(), 1, decimal=5)
assert_array_equal(max_index, np.array(arr1.shape) / 2)

View File

@@ -1,268 +0,0 @@
import itertools
import warnings
import re
import numpy as np
import pytest
from numpy.testing import assert_allclose
from scipy.ndimage import fourier_shift
import scipy.fft as fft
from skimage import img_as_float
from skimage._shared._warnings import expected_warnings
from skimage._shared.utils import _supported_float_type
from skimage.data import camera, binary_blobs, eagle
from skimage.registration._phase_cross_correlation import (
phase_cross_correlation, _upsampled_dft
)
@pytest.mark.parametrize('normalization', [None, 'phase'])
def test_correlation(normalization):
reference_image = fft.fftn(camera())
shift = (-7, 12)
shifted_image = fourier_shift(reference_image, shift)
# pixel precision
result, _, _ = phase_cross_correlation(reference_image,
shifted_image,
space="fourier",
normalization=normalization)
assert_allclose(result[:2], -np.array(shift))
@pytest.mark.parametrize('normalization', ['nonexisting'])
def test_correlation_invalid_normalization(normalization):
reference_image = fft.fftn(camera())
shift = (-7, 12)
shifted_image = fourier_shift(reference_image, shift)
# pixel precision
with pytest.raises(ValueError):
phase_cross_correlation(reference_image,
shifted_image,
space="fourier",
normalization=normalization)
@pytest.mark.parametrize('normalization', [None, 'phase'])
def test_subpixel_precision(normalization):
reference_image = fft.fftn(camera())
subpixel_shift = (-2.4, 1.32)
shifted_image = fourier_shift(reference_image, subpixel_shift)
# subpixel precision
result, _, _ = phase_cross_correlation(reference_image,
shifted_image,
upsample_factor=100,
space="fourier",
normalization=normalization)
assert_allclose(result[:2], -np.array(subpixel_shift), atol=0.05)
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
def test_real_input(dtype):
reference_image = camera().astype(dtype, copy=False)
subpixel_shift = (-2.4, 1.32)
shifted_image = fourier_shift(fft.fftn(reference_image), subpixel_shift)
shifted_image = fft.ifftn(shifted_image).real.astype(dtype, copy=False)
# subpixel precision
result, error, diffphase = phase_cross_correlation(reference_image,
shifted_image,
upsample_factor=100)
assert result.dtype == _supported_float_type(dtype)
assert_allclose(result[:2], -np.array(subpixel_shift), atol=0.05)
def test_size_one_dimension_input():
# take a strip of the input image
reference_image = fft.fftn(camera()[:, 15]).reshape((-1, 1))
subpixel_shift = (-2.4, 4)
shifted_image = fourier_shift(reference_image, subpixel_shift)
# subpixel precision
result, error, diffphase = phase_cross_correlation(reference_image,
shifted_image,
upsample_factor=20,
space="fourier")
assert_allclose(result[:2], -np.array((-2.4, 0)), atol=0.05)
def test_3d_input():
phantom = img_as_float(binary_blobs(length=32, n_dim=3))
reference_image = fft.fftn(phantom)
shift = (-2., 1., 5.)
shifted_image = fourier_shift(reference_image, shift)
result, error, diffphase = phase_cross_correlation(reference_image,
shifted_image,
space="fourier")
assert_allclose(result, -np.array(shift), atol=0.05)
# subpixel precision now available for 3-D data
subpixel_shift = (-2.3, 1.7, 5.4)
shifted_image = fourier_shift(reference_image, subpixel_shift)
result, error, diffphase = phase_cross_correlation(reference_image,
shifted_image,
upsample_factor=100,
space="fourier")
assert_allclose(result, -np.array(subpixel_shift), atol=0.05)
def test_unknown_space_input():
image = np.ones((5, 5))
with pytest.raises(ValueError):
phase_cross_correlation(
image, image,
space="frank")
def test_wrong_input():
# Dimensionality mismatch
image = np.ones((5, 5, 1))
template = np.ones((5, 5))
with pytest.raises(ValueError):
phase_cross_correlation(template, image)
# Size mismatch
image = np.ones((5, 5))
template = np.ones((4, 4))
with pytest.raises(ValueError):
phase_cross_correlation(template, image)
# NaN values in data
image = np.ones((5, 5))
image[0][0] = np.nan
template = np.ones((5, 5))
with expected_warnings(
[
r"invalid value encountered in true_divide"
+ r"|"
+ r"invalid value encountered in divide"
+ r"|\A\Z"
]
):
with pytest.raises(ValueError):
phase_cross_correlation(template, image, return_error=True)
def test_4d_input_pixel():
phantom = img_as_float(binary_blobs(length=32, n_dim=4))
reference_image = fft.fftn(phantom)
shift = (-2., 1., 5., -3)
shifted_image = fourier_shift(reference_image, shift)
result, error, diffphase = phase_cross_correlation(reference_image,
shifted_image,
space="fourier")
assert_allclose(result, -np.array(shift), atol=0.05)
def test_4d_input_subpixel():
phantom = img_as_float(binary_blobs(length=32, n_dim=4))
reference_image = fft.fftn(phantom)
subpixel_shift = (-2.3, 1.7, 5.4, -3.2)
shifted_image = fourier_shift(reference_image, subpixel_shift)
result, error, diffphase = phase_cross_correlation(reference_image,
shifted_image,
upsample_factor=10,
space="fourier")
assert_allclose(result, -np.array(subpixel_shift), atol=0.05)
@pytest.mark.parametrize("return_error", [True, False, "always"])
@pytest.mark.parametrize("reference_mask", [None, True])
def test_phase_cross_correlation_deprecation(return_error, reference_mask):
# For now, assert that phase_cross_correlation raises a warning that
# returning only shifts is deprecated. In skimage 0.21, this test should be
# updated for the deprecation of the return_error parameter.
should_warn = (
return_error is False
or (return_error != "always" and reference_mask is True)
)
reference_image = np.ones((10, 10))
moving_image = np.ones_like(reference_image)
if reference_mask is True:
# moving_mask defaults to reference_mask, passing moving_mask only is
# not supported, so we don't need to test it
reference_mask = np.ones_like(reference_image)
if should_warn:
msg = (
"In scikit-image 0.21, phase_cross_correlation will start "
"returning a tuple or 3 items (shift, error, phasediff) always. "
"To enable the new return behavior and silence this warning, use "
"return_error='always'."
)
with pytest.warns(FutureWarning, match=re.escape(msg)):
out = phase_cross_correlation(
reference_image=reference_image,
moving_image=moving_image,
return_error=return_error,
reference_mask=reference_mask,
)
assert not isinstance(out, tuple)
else:
with warnings.catch_warnings():
warnings.simplefilter("error")
out = phase_cross_correlation(
reference_image=reference_image,
moving_image=moving_image,
return_error=return_error,
reference_mask=reference_mask,
)
assert isinstance(out, tuple)
assert len(out) == 3
def test_mismatch_upsampled_region_size():
with pytest.raises(ValueError):
_upsampled_dft(
np.ones((4, 4)),
upsampled_region_size=[3, 2, 1, 4])
def test_mismatch_offsets_size():
with pytest.raises(ValueError):
_upsampled_dft(np.ones((4, 4)), 3,
axis_offsets=[3, 2, 1, 4])
@pytest.mark.parametrize(
('shift0', 'shift1'),
itertools.product((100, -100, 350, -350), (100, -100, 350, -350)),
)
def test_disambiguate_2d(shift0, shift1):
image = eagle()[500:, 900:] # use a highly textured part of image
shift = (shift0, shift1)
origin0 = []
for s in shift:
if s > 0:
origin0.append(0)
else:
origin0.append(-s)
origin1 = np.array(origin0) + shift
slice0 = tuple(slice(o, o+450) for o in origin0)
slice1 = tuple(slice(o, o+450) for o in origin1)
reference = image[slice0]
moving = image[slice1]
computed_shift, _, _ = phase_cross_correlation(
reference, moving, disambiguate=True, return_error='always'
)
np.testing.assert_equal(shift, computed_shift)
def test_disambiguate_zero_shift():
"""When the shift is 0, disambiguation becomes degenerate.
Some quadrants become size 0, which prevents computation of
cross-correlation. This test ensures that nothing bad happens in that
scenario.
"""
image = camera()
computed_shift, _, _ = phase_cross_correlation(
image, image, disambiguate=True, return_error='always'
)
assert computed_shift == (0, 0)

View File

@@ -1,117 +0,0 @@
import numpy as np
import pytest
from skimage._shared.utils import _supported_float_type
from skimage.registration import optical_flow_tvl1
from skimage.transform import warp
def _sin_flow_gen(image0, max_motion=4.5, npics=5):
"""Generate a synthetic ground truth optical flow with a sinusoid as
first component.
Parameters:
----
image0: ndarray
The base image to be warped.
max_motion: float
Maximum flow magnitude.
npics: int
Number of sinusoid pics.
Returns
-------
flow, image1 : ndarray
The synthetic ground truth optical flow with a sinusoid as
first component and the corresponding warped image.
"""
grid = np.meshgrid(*[np.arange(n) for n in image0.shape], indexing='ij')
grid = np.stack(grid)
gt_flow = np.zeros_like(grid, dtype=float)
gt_flow[0, ...] = max_motion * np.sin(grid[0]/grid[0].max()*npics*np.pi)
image1 = warp(image0, grid - gt_flow, mode='edge')
return gt_flow, image1
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
def test_2d_motion(dtype):
# Generate synthetic data
rnd = np.random.default_rng(0)
image0 = rnd.normal(size=(256, 256))
gt_flow, image1 = _sin_flow_gen(image0)
image1 = image1.astype(dtype, copy=False)
float_dtype = _supported_float_type(dtype)
# Estimate the flow
flow = optical_flow_tvl1(image0, image1, attachment=5, dtype=float_dtype)
assert flow.dtype == float_dtype
# Assert that the average absolute error is less then half a pixel
assert abs(flow - gt_flow) .mean() < 0.5
if dtype != float_dtype:
with pytest.raises(ValueError):
optical_flow_tvl1(image0, image1, attachment=5, dtype=dtype)
def test_3d_motion():
# Generate synthetic data
rnd = np.random.default_rng(0)
image0 = rnd.normal(size=(100, 100, 100))
gt_flow, image1 = _sin_flow_gen(image0)
# Estimate the flow
flow = optical_flow_tvl1(image0, image1, attachment=10)
# Assert that the average absolute error is less then half a pixel
assert abs(flow - gt_flow) .mean() < 0.5
def test_no_motion_2d():
rnd = np.random.default_rng(0)
img = rnd.normal(size=(256, 256))
flow = optical_flow_tvl1(img, img)
assert np.all(flow == 0)
def test_no_motion_3d():
rnd = np.random.default_rng(0)
img = rnd.normal(size=(64, 64, 64))
flow = optical_flow_tvl1(img, img)
assert np.all(flow == 0)
def test_optical_flow_dtype():
# Generate synthetic data
rnd = np.random.default_rng(0)
image0 = rnd.normal(size=(256, 256))
gt_flow, image1 = _sin_flow_gen(image0)
# Estimate the flow at double precision
flow_f64 = optical_flow_tvl1(image0, image1, attachment=5, dtype=np.float64)
assert flow_f64.dtype == np.float64
# Estimate the flow at single precision
flow_f32 = optical_flow_tvl1(image0, image1, attachment=5, dtype=np.float32)
assert flow_f32.dtype == np.float32
# Assert that floating point precision does not affect the quality
# of the estimated flow
assert np.abs(flow_f64 - flow_f32).mean() < 1e-3
def test_incompatible_shapes():
rnd = np.random.default_rng(0)
I0 = rnd.normal(size=(256, 256))
I1 = rnd.normal(size=(128, 256))
with pytest.raises(ValueError):
u, v = optical_flow_tvl1(I0, I1)
def test_wrong_dtype():
rnd = np.random.default_rng(0)
img = rnd.normal(size=(256, 256))
with pytest.raises(ValueError):
u, v = optical_flow_tvl1(img, img, dtype=np.int64)