def run_benchmark():
# Download the test image
filename = os.path.join(os.path.join(path, ".."), "hst_acs_10_col.txt")
if not os.path.isfile(filename):
url_path = "http://astro.dur.ac.uk/~cklv53/files/hst_acs_10_col.txt"
urlretrieve(url_path, filename)
# Load the image
image_pre_cti = np.loadtxt(filename, skiprows=1)
# CTI model parameters
roe = ac.ROE(
dwell_times=[1.0],
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
force_release_away_from_readout=True,
use_integer_express_matrix=False,
)
ccd = ac.CCD(
phases=[
ac.CCDPhase(full_well_depth=1e4,
well_notch_depth=0.0,
well_fill_power=1.0)
],
fraction_of_traps_per_phase=[1.0],
)
traps = [
ac.TrapInstantCapture(density=10.0,
release_timescale=-1.0 / np.log(0.5))
]
express = 5
offset = 0
start = 0
stop = -1
image_post_cti = ac.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=traps,
parallel_express=express,
parallel_offset=offset,
parallel_window_start=start,
parallel_window_stop=stop,
)
def add_cti_to_hst_image(express=1):
# Load the HST image
hdu_list = fits.open(path + "/../examples/acs/jc0a01h8q_raw.fits")
idx_image = 1
input_image = np.array(hdu_list[idx_image].data).astype("float64").T
# Model inputs
trap = ac.Trap(density=1, release_timescale=10)
ccd = ac.CCD(well_notch_depth=0.01,
well_fill_power=0.8,
full_well_depth=84700)
# Select a subset of rows
row_start = 0
row_end = -1
# row_start = 200
# row_end = 400
# Select a subset of columns
column_start = 2662
column_end = column_start + 1
# Refine input image
input_image = input_image[column_start:column_end, row_start:row_end].T
print("%d row(s), %d column(s)" % input_image.shape)
# Add CTI to the column
output_image = ac.add_cti(
image=input_image,
parallel_traps=[trap],
parallel_ccd=ccd,
parallel_express=express,
)
return input_image, output_image
def run_demo():
# Add CTI to a test image, then remove it
image_pre_cti = np.array([
[0.0, 0.0, 0.0, 0.0],
[200.0, 0.0, 0.0, 0.0],
[0.0, 200.0, 0.0, 0.0],
[0.0, 0.0, 200.0, 0.0],
[0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0],
])
roe = ac.ROE(
dwell_times=[1.0],
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
force_release_away_from_readout=True,
use_integer_express_matrix=False,
)
ccd = ac.CCD(
phases=[
ac.CCDPhase(full_well_depth=1e3,
well_notch_depth=0.0,
well_fill_power=1.0)
],
fraction_of_traps_per_phase=[1.0],
)
traps = [
ac.TrapInstantCapture(density=10.0,
release_timescale=-1.0 / np.log(0.5))
]
express = 0
offset = 0
start = 0
stop = -1
print("\n# Test image:")
ac.print_array_2D(image_pre_cti)
print("\n# Add CTI")
image_post_cti = ac.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=traps,
parallel_express=express,
parallel_offset=offset,
parallel_window_start=start,
parallel_window_stop=stop,
serial_roe=roe,
serial_ccd=ccd,
serial_traps=traps,
serial_express=express,
serial_offset=offset,
serial_window_start=start,
serial_window_stop=stop,
verbosity=1,
)
print("\n# Image with CTI added:")
ac.print_array_2D(image_post_cti)
print("\n# Remove CTI")
image_removed_cti = ac.remove_cti(
image=image_post_cti,
n_iterations=4,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=traps,
parallel_express=express,
parallel_offset=offset,
parallel_window_start=start,
parallel_window_stop=stop,
serial_roe=roe,
serial_ccd=ccd,
serial_traps=traps,
serial_express=express,
serial_offset=offset,
serial_window_start=start,
serial_window_stop=stop,
verbosity=1,
)
print("\n# Image with CTI removed:")
ac.print_array_2D(image_removed_cti)
def test__remove_cti__better_removal_with_more_iterations(self):
# Add CTI to a test image, then remove it
image_pre_cti = np.array([
[0.0, 0.0, 0.0, 0.0],
[200.0, 0.0, 0.0, 0.0],
[0.0, 200.0, 0.0, 0.0],
[0.0, 0.0, 200.0, 0.0],
[0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0],
])
roe = ac.ROE(
dwell_times=[1.0],
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
force_release_away_from_readout=True,
use_integer_express_matrix=False,
)
ccd = ac.CCD(
phases=[
ac.CCDPhase(full_well_depth=1e3,
well_notch_depth=0.0,
well_fill_power=1.0)
],
fraction_of_traps_per_phase=[1.0],
)
traps = [
ac.TrapInstantCapture(density=10.0,
release_timescale=-1.0 / np.log(0.5))
]
express = 0
offset = 0
start = 0
stop = -1
# Add CTI
image_add_cti = ac.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=traps,
parallel_express=express,
parallel_offset=offset,
parallel_window_start=start,
parallel_window_stop=stop,
serial_roe=roe,
serial_ccd=ccd,
serial_traps=traps,
serial_express=express,
serial_offset=offset,
serial_window_start=start,
serial_window_stop=stop,
verbosity=0,
)
# Remove CTI
for n_iterations in range(2, 7):
image_remove_cti = ac.remove_cti(
image=image_add_cti,
n_iterations=n_iterations,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=traps,
parallel_express=express,
parallel_offset=offset,
parallel_window_start=start,
parallel_window_stop=stop,
serial_roe=roe,
serial_ccd=ccd,
serial_traps=traps,
serial_express=express,
serial_offset=offset,
serial_window_start=start,
serial_window_stop=stop,
verbosity=0,
)
# Expect better results with more iterations
tolerance = 10**(1 - n_iterations)
assert image_remove_cti == pytest.approx(image_pre_cti,
abs=tolerance)
def test__add_cti__all_trap_types_broadly_similar_results(self):
# Manually set True to make the plot
do_plot = False
# do_plot = True
image_pre_cti = np.zeros((20, 1))
image_pre_cti[2, 0] = 800
trap_ic = ac.TrapInstantCapture(density=10,
release_timescale=-1 / np.log(0.5))
trap_sc = ac.TrapSlowCapture(density=10,
release_timescale=-1 / np.log(0.5),
capture_timescale=0.1)
trap_ic_co = ac.TrapInstantCaptureContinuum(
density=10,
release_timescale=-1 / np.log(0.5),
release_timescale_sigma=0.05)
trap_sc_co = ac.TrapSlowCaptureContinuum(
density=10,
release_timescale=-1 / np.log(0.5),
release_timescale_sigma=0.05,
capture_timescale=0.1,
)
ccd = ac.CCD(phases=[
ac.CCDPhase(
well_fill_power=1, full_well_depth=1000, well_notch_depth=0)
])
roe = ac.ROE(
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
use_integer_express_matrix=True,
)
express = 5
# Standard instant-capture traps
image_post_cti_ic = ac.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=[trap_ic],
parallel_express=express,
verbosity=0,
).T[0]
if do_plot:
pixels = np.arange(len(image_pre_cti))
colours = ["#1199ff", "#ee4400", "#7711dd", "#44dd44"]
plt.figure(figsize=(10, 6))
ax1 = plt.gca()
ax2 = ax1.twinx()
ax1.plot(
pixels,
image_post_cti_ic,
alpha=0.8,
c=colours[0],
label="Instant Capture",
)
# Other trap types
for i, (trap, label) in enumerate(
zip(
[trap_sc, trap_ic_co, trap_sc_co],
[
"Slow Capture", "Instant Capture Continuum",
"Slow Capture Continuum"
],
)):
image_post_cti = ac.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=[trap],
parallel_express=express,
verbosity=0,
).T[0]
if do_plot:
c = colours[i + 1]
ax1.plot(pixels, image_post_cti, alpha=0.8, c=c, label=label)
ax2.plot(
pixels,
(image_post_cti - image_post_cti_ic) / image_post_cti_ic,
alpha=0.8,
ls=":",
c=c,
)
assert image_post_cti == pytest.approx(image_post_cti_ic, rel=0.1)
if do_plot:
ax1.legend(loc="lower left")
ax1.set_yscale("log")
ax1.set_xlabel("Pixel")
ax1.set_ylabel("Counts")
ax2.set_ylabel("Fractional Difference (dotted)")
plt.tight_layout()
plt.show()
def test__add_cti__single_pixel__vary_express_3__compare_old_arctic(self):
# Manually set True to make the plot
do_plot = False
# do_plot = True
image_pre_cti = np.zeros((40, 1))
image_pre_cti[2, 0] = 800
traps = [ac.TrapInstantCapture(density=10, release_timescale=5)]
ccd = ac.CCD(phases=[
ac.CCDPhase(
well_fill_power=0.5, full_well_depth=1000, well_notch_depth=0)
])
roe = ac.ROE(
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
use_integer_express_matrix=True,
)
if do_plot:
pixels = np.arange(len(image_pre_cti))
colours = ["#1199ff", "#ee4400", "#7711dd", "#44dd44", "#775533"]
plt.figure(figsize=(10, 6))
ax1 = plt.gca()
ax2 = ax1.twinx()
for i, (express, image_idl) in enumerate(
zip(
# [2, 40],
[40],
[
# [ # express=2 but different save/restore trap approach
# 0.0000000,
# 0.0000000,
# 773.16687,
# 6.1919436,
# 6.3684354,
# 6.2979879,
# 6.0507884,
# 5.7028670,
# 5.2915287,
# 4.8539200,
# 4.4115725,
# 3.9793868,
# 3.5681694,
# 3.1855009,
# 2.8297379,
# 2.5070403,
# 2.2149866,
# 1.9524645,
# 1.7182100,
# 1.5103321,
# 1.3410047,
# 1.1967528,
# 1.0735434,
# 0.96801299,
# 0.87580109,
# 0.79527515,
# 0.72667038,
# 0.66463155,
# 0.61054391,
# 0.56260395,
# 0.51870286,
# 0.47962612,
# 0.44496229,
# 0.41361856,
# 0.38440439,
# 0.35855818,
# 0.33410615,
# 0.31309450,
# 0.29213923,
# 0.27346680,
# ],
[
0.00000,
0.00000,
773.317,
5.98876,
6.13135,
6.05125,
5.81397,
5.49105,
5.11484,
4.71890,
4.32139,
3.93756,
3.57154,
3.23464,
2.91884,
2.63640,
2.37872,
2.14545,
1.93805,
1.75299,
1.58590,
1.43964,
1.30883,
1.19327,
1.09000,
0.996036,
0.915593,
0.841285,
0.775049,
0.718157,
0.664892,
0.617069,
0.574792,
0.537046,
0.502112,
0.471202,
0.442614,
0.417600,
0.394439,
0.373072,
]
],
)):
image_post_cti = ac.add_cti(
image=image_pre_cti,
parallel_traps=traps,
parallel_ccd=ccd,
parallel_roe=roe,
parallel_express=express,
verbosity=0,
).T[0]
image_idl = np.array(image_idl)
if do_plot:
c = colours[i]
if i == 0:
ax1.plot(
pixels,
image_post_cti,
alpha=0.8,
c=c,
label="%d (py)" % express,
)
ax1.plot(
pixels,
image_idl,
ls="--",
alpha=0.8,
c=c,
label="%d (idl)" % express,
)
ax2.plot(
pixels,
(image_post_cti - image_idl) / image_idl,
alpha=0.8,
ls=":",
c=c,
)
else:
ax1.plot(pixels,
image_post_cti,
alpha=0.8,
c=c,
label="%d" % express)
ax1.plot(pixels, image_idl, alpha=0.8, c=c, ls="--")
ax2.plot(
pixels,
(image_post_cti - image_idl) / image_idl,
alpha=0.8,
ls=":",
c=c,
)
assert image_post_cti == pytest.approx(image_idl, rel=0.03)
if do_plot:
ax1.legend(title="express", loc="lower left")
ax1.set_yscale("log")
ax1.set_xlabel("Pixel")
ax1.set_ylabel("Counts")
ax2.set_ylabel("Fractional Difference (dotted)")
plt.tight_layout()
plt.show()
def test__add_cti__single_pixel__vary_express_2__compare_old_arctic(self):
# Manually set True to make the plot
do_plot = False
# do_plot = True
image_pre_cti = np.zeros((120, 1))
image_pre_cti[102, 0] = 800
traps = [
ac.TrapInstantCapture(density=10,
release_timescale=-1 / np.log(0.5))
]
ccd = ac.CCD(phases=[
ac.CCDPhase(
well_fill_power=0.5, full_well_depth=1000, well_notch_depth=0)
])
roe = ac.ROE(
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
use_integer_express_matrix=True,
)
if do_plot:
pixels = np.arange(len(image_pre_cti))
colours = ["#1199ff", "#ee4400", "#7711dd", "#44dd44", "#775533"]
plt.figure(figsize=(10, 6))
ax1 = plt.gca()
ax2 = ax1.twinx()
for i, (express, image_idl) in enumerate(
zip(
[2, 20],
[
[
0.00000,
0.00000,
42.6722,
250.952,
161.782,
107.450,
73.0897,
50.6914,
35.6441,
25.3839,
18.2665,
13.2970,
9.79078,
7.30555,
5.52511,
4.24364,
3.30829,
2.61813,
2.09710,
1.70752,
],
[
0.00000,
0.00000,
134.103,
163.783,
117.887,
85.8632,
63.6406,
47.9437,
36.6625,
28.4648,
22.4259,
17.9131,
14.4976,
11.8789,
9.84568,
8.25520,
6.98939,
5.97310,
5.14856,
4.47386,
],
],
)):
image_post_cti = ac.add_cti(
image=image_pre_cti,
parallel_traps=traps,
parallel_ccd=ccd,
parallel_roe=roe,
parallel_express=express,
verbosity=0,
).T[0]
image_idl = np.append(np.zeros(100), image_idl)
if do_plot:
c = colours[i]
if i == 0:
ax1.plot(
pixels,
image_post_cti,
alpha=0.8,
c=c,
label="%d (py)" % express,
)
ax1.plot(
pixels,
image_idl,
ls="--",
alpha=0.8,
c=c,
label="%d (idl)" % express,
)
ax2.plot(
pixels,
(image_post_cti - image_idl) / image_idl,
alpha=0.8,
ls=":",
c=c,
)
else:
ax1.plot(pixels,
image_post_cti,
alpha=0.8,
c=c,
label="%d" % express)
ax1.plot(pixels, image_idl, alpha=0.8, c=c, ls="--")
ax2.plot(
pixels,
(image_post_cti - image_idl) / image_idl,
alpha=0.8,
ls=":",
c=c,
)
assert image_post_cti == pytest.approx(image_idl, rel=0.02)
if do_plot:
ax1.legend(title="express", loc="lower left")
ax1.set_yscale("log")
ax1.set_xlabel("Pixel")
ax1.set_xlim(100, 121)
ax1.set_ylabel("Counts")
ax2.set_ylabel("Fractional Difference (dotted)")
plt.tight_layout()
plt.show()
def test__add_cti__single_pixel__vary_express__compare_old_arctic(self):
# Manually set True to make the plot
do_plot = False
# do_plot = True
image_pre_cti = np.zeros((20, 1))
image_pre_cti[2, 0] = 800
# Nice numbers for easy manual checking
traps = [
ac.TrapInstantCapture(density=10,
release_timescale=-1 / np.log(0.5))
]
ccd = ac.CCD(phases=[
ac.CCDPhase(
well_fill_power=1, full_well_depth=1000, well_notch_depth=0)
])
roe = ac.ROE(
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
use_integer_express_matrix=True,
)
if do_plot:
pixels = np.arange(len(image_pre_cti))
colours = ["#1199ff", "#ee4400", "#7711dd", "#44dd44", "#775533"]
plt.figure(figsize=(10, 6))
ax1 = plt.gca()
ax2 = ax1.twinx()
for i, (express, image_idl) in enumerate(
zip(
[1, 2, 5, 10, 20],
[
[
0.00000,
0.00000,
776.000,
15.3718,
9.65316,
5.81950,
3.41087,
1.95889,
1.10817,
0.619169,
0.342489,
0.187879,
0.102351,
0.0554257,
0.0298603,
0.0160170,
0.00855758,
0.00455620,
0.00241824,
0.00128579,
],
[
0.00000,
0.00000,
776.000,
15.3718,
9.65316,
5.81950,
3.41087,
1.95889,
1.10817,
0.619169,
0.348832,
0.196128,
0.109984,
0.0614910,
0.0340331,
0.0187090,
0.0102421,
0.00558406,
0.00303254,
0.00164384,
],
[
0.00000,
0.00000,
776.000,
15.3718,
9.59381,
5.80216,
3.43231,
1.99611,
1.15104,
0.658983,
0.374685,
0.211807,
0.119441,
0.0670274,
0.0373170,
0.0205845,
0.0113179,
0.00621127,
0.00341018,
0.00187955,
],
[
0.00000,
0.00000,
776.160,
15.1432,
9.51562,
5.78087,
3.43630,
2.01144,
1.16452,
0.668743,
0.381432,
0.216600,
0.122556,
0.0689036,
0.0383241,
0.0211914,
0.0116758,
0.00641045,
0.00352960,
0.00195050,
],
[
0.00000,
0.00000,
776.239,
15.0315,
9.47714,
5.77145,
3.43952,
2.01754,
1.17049,
0.673351,
0.384773,
0.218860,
0.124046,
0.0697859,
0.0388253,
0.0214799,
0.0118373,
0.00650488,
0.00358827,
0.00198517,
],
],
)):
image_post_cti = ac.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=traps,
parallel_express=express,
verbosity=0,
).T[0]
image_idl = np.array(image_idl)
if do_plot:
c = colours[i]
if i == 0:
ax1.plot(
pixels,
image_post_cti,
alpha=0.8,
c=c,
label="%d (py)" % express,
)
ax1.plot(
pixels,
image_idl,
ls="--",
alpha=0.8,
c=c,
label="%d (idl)" % express,
)
ax2.plot(
pixels,
(image_post_cti - image_idl) / image_idl,
alpha=0.8,
ls=":",
c=c,
)
else:
ax1.plot(pixels,
image_post_cti,
alpha=0.8,
c=c,
label="%d" % express)
ax1.plot(pixels, image_idl, alpha=0.8, c=c, ls="--")
ax2.plot(
pixels,
(image_post_cti - image_idl) / image_idl,
alpha=0.8,
ls=":",
c=c,
)
assert image_post_cti == pytest.approx(image_idl, rel=0.05)
if do_plot:
ax1.legend(title="express", loc="lower left")
ax1.set_yscale("log")
ax1.set_xlabel("Pixel")
ax1.set_ylabel("Counts")
ax2.set_ylabel("Fractional Difference (dotted)")
plt.tight_layout()
plt.show()
prescan_offset=10)
#
# Noisy image
#
image_pre_cti = image_model + np.random.normal(0, 0.01, image_model.shape)
#image_pre_cti = np.maximum(image_pre_cti,np.zeros(image_pre_cti.shape));
#print(image_pre_cti[0:10,0])
start = time.time_ns()
image_post_cti = arcticpy.add_cti(
image=image_pre_cti,
parallel_traps=parallel_traps,
parallel_ccd=parallel_ccd,
parallel_roe=parallel_roe,
parallel_express=parallel_express,
parallel_prune_n_electrons=parallel_prune_n_electrons,
parallel_prune_frequency=parallel_prune_frequency,
verbosity=0)
print(f"Clocking Time Noisy = {((time.time_ns() - start)/1e9)} s")
#print(image_post_cti[0:10,0])
#
# Noise-free image
#
image_pre_cti = image_model
#print(image_pre_cti[0:10,0])
def test__add_cti__multiple_trap_types_together_or_consecutively(self):
# Manually set True to make the plot
do_plot = False
# do_plot = True
image_pre_cti = np.zeros((20, 1))
image_pre_cti[2, 0] = 800
trap_ic = cti.TrapInstantCapture(density=10,
release_timescale=-1 / np.log(0.5))
trap_sc = cti.TrapSlowCapture(density=10,
release_timescale=-1 / np.log(0.5),
capture_timescale=0.1)
trap_ic_co = cti.TrapInstantCaptureContinuum(
density=10,
release_timescale=-1 / np.log(0.5),
release_timescale_sigma=0.05)
trap_sc_co = cti.TrapSlowCaptureContinuum(
density=10,
release_timescale=-1 / np.log(0.5),
release_timescale_sigma=0.05,
capture_timescale=0.1,
)
ccd = cti.CCD(phases=[
cti.CCDPhase(
well_fill_power=1, full_well_depth=1000, well_notch_depth=0)
])
roe = cti.ROE(
empty_traps_between_columns=True,
empty_traps_for_first_transfers=False,
use_integer_express_matrix=True,
)
express = 5
# Implement all types of traps simultaneously
image_post_cti_simultaneously = cti.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=[trap_ic, trap_sc, trap_ic_co, trap_sc_co],
parallel_express=express,
verbosity=0,
).T[0]
if do_plot:
pixels = np.arange(len(image_pre_cti))
colours = ["#1199ff", "#ee4400", "#7711dd", "#44dd44", "#666666"]
plt.figure(figsize=(10, 6))
ax1 = plt.gca()
ax2 = ax1.twinx()
ax1.plot(
pixels,
image_post_cti_simultaneously,
alpha=0.8,
c=colours[0],
label="All types simultaneously",
)
# Implement trap types one after another
image_post_cti_separately = image_pre_cti
for i, (trap, label) in enumerate(
zip(
[trap_ic, trap_sc, trap_ic_co, trap_sc_co],
[
"After Instant Capture", "After Slow Capture",
"After Instant Capture Continuum",
"After Slow Capture Continuum"
],
)):
image_pre_cti = image_post_cti_separately
image_post_cti_separately = cti.add_cti(
image=image_pre_cti,
parallel_roe=roe,
parallel_ccd=ccd,
parallel_traps=[trap],
parallel_express=express,
verbosity=0,
)
if do_plot:
c = colours[i + 1]
ax1.plot(pixels,
image_post_cti_separately.T[0],
alpha=0.8,
c=c,
label=label)
ax2.plot(
pixels,
(image_post_cti_separately.T[0] -
image_post_cti_simultaneously) /
image_post_cti_simultaneously,
alpha=0.8,
ls=":",
c=c,
)
assert 1 == pytest.approx(1.0001, rel=0.1)
assert image_post_cti_separately.T[0] == pytest.approx(
image_post_cti_simultaneously, rel=0.1)
if do_plot:
ax1.legend(loc="lower left")
ax1.set_yscale("log")
ax1.set_xlabel("Pixel")
ax1.set_ylabel("Counts")
ax2.set_ylabel("Fractional Difference (dotted)")
plt.tight_layout()
plt.show()
traps_in = [
cti.TrapInstantCapture(density=10.0, release_timescale=-1.0 / np.log(0.5))
]
express = 0
# Test input
length = 9
warm_index = 4
pre_cti = np.zeros((length, 1))
pre_cti[warm_index, 0] = 1000
# Add CTI
post_cti = cti.add_cti(
image=pre_cti,
parallel_roe=roe_in,
parallel_ccd=ccd_in,
parallel_traps=traps_in,
parallel_express=express,
verbosity=0,
)
# 1D lines for fitting
pre_cti = pre_cti.T[0]
post_cti = post_cti.T[0]
print("Pre-CTI:", pre_cti)
print("Post-CTI:", post_cti)
# Test noise
noise = np.ones_like(post_cti) * 0.01
# Conversion value for HST ACS
pixel_scale = 0.05