ChipUtils.py 10.5 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
import os
import galsim
import ctypes
import numpy as np
from astropy.io import fits
from datetime import datetime

from ObservationSim.Instrument.Chip import Effects as effects
from ObservationSim.Config.Header import generatePrimaryHeader, generateExtensionHeader

try:
    import importlib.resources as pkg_resources
except ImportError:
    # Try backported to PY<37 'importlib_resources'
    import importlib_resources as pkg_resources


def log_info(msg, logger=None):
    if logger:
        logger.info(msg)
    else:
        print(msg, flush=True)

def getChipSLSConf(chipID):
    confFile = ''
    if chipID == 1: confFile = ['CSST_GI2.conf', 'CSST_GI1.conf']
    if chipID == 2: confFile = ['CSST_GV4.conf', 'CSST_GV3.conf']
    if chipID == 3: confFile = ['CSST_GU2.conf', 'CSST_GU1.conf']
    if chipID == 4: confFile = ['CSST_GU4.conf', 'CSST_GU3.conf']
    if chipID == 5: confFile = ['CSST_GV2.conf', 'CSST_GV1.conf']
    if chipID == 10: confFile = ['CSST_GI4.conf', 'CSST_GI3.conf']
    if chipID == 21: confFile = ['CSST_GI6.conf', 'CSST_GI5.conf']
    if chipID == 26: confFile = ['CSST_GV8.conf', 'CSST_GV7.conf']
    if chipID == 27: confFile = ['CSST_GU6.conf', 'CSST_GU5.conf']
    if chipID == 28: confFile = ['CSST_GU8.conf', 'CSST_GU7.conf']
    if chipID == 29: confFile = ['CSST_GV6.conf', 'CSST_GV5.conf']
    if chipID == 30: confFile = ['CSST_GI8.conf', 'CSST_GI7.conf']
    return confFile

40
def generateHeader(chip, ra_cen, dec_cen, img_rot, im_type, pointing_ID, exptime=150., project_cycle=6, run_counter=0, timestamp = 1621915200):
41
42
43
44
45
46
47
48
49
50
51
52
53
54
    datetime_obs = datetime.utcfromtimestamp(timestamp)
    date_obs = datetime_obs.strftime("%y%m%d")
    time_obs = datetime_obs.strftime("%H%M%S")
    h_prim = generatePrimaryHeader(
        xlen=chip.npix_x, 
        ylen=chip.npix_y, 
        pointNum = str(pointing_ID),
        ra=ra_cen, 
        dec=dec_cen, 
        pixel_scale=chip.pix_scale,
        date=date_obs,
        time_obs=time_obs,
        im_type = im_type,
        exptime=exptime,
55
56
        project_cycle=project_cycle,
        run_counter=run_counter,
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
        chip_name=str(chip.chipID).rjust(2, '0')
        )
    h_ext = generateExtensionHeader(
        chip=chip,
        xlen=chip.npix_x, 
        ylen=chip.npix_y, 
        ra=ra_cen, 
        dec=dec_cen,
        pa=img_rot.deg, 
        gain=chip.gain, 
        readout=chip.read_noise, 
        dark=chip.dark_noise, 
        saturation=90000, 
        pixel_scale=chip.pix_scale, 
        pixel_size=chip.pix_size,
        xcen=chip.x_cen,
        ycen=chip.y_cen,
        extName='SCI',
        timestamp = timestamp,
        exptime = exptime,
        readoutTime = chip.readout_time)
    return h_prim, h_ext

80
def outputCal(chip, img, ra_cen, dec_cen, img_rot, im_type, pointing_ID, output_dir, exptime=150., project_cycle=6, run_counter=0, timestamp = 1621915200):
81
82
83
84
85
86
87
88
    h_prim, h_ext = generateHeader(
        chip=chip,
        ra_cen=ra_cen,
        dec_cen=dec_cen,
        img_rot=img_rot,
        im_type=im_type,
        pointing_ID=pointing_ID,
        exptime=exptime,
89
90
        project_cycle=project_cycle,
        run_counter=run_counter,
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
        timestamp=timestamp)
    hdu1 = fits.PrimaryHDU(header=h_prim)
    hdu1.add_checksum()
    hdu1.header.comments['CHECKSUM'] = 'HDU checksum'
    hdu1.header.comments['DATASUM'] = 'data unit checksum'
    hdu2 = fits.ImageHDU(img.array, header=h_ext)
    hdu2.add_checksum()
    hdu2.header.comments['XTENSION'] = 'extension type'
    hdu2.header.comments['CHECKSUM'] = 'HDU checksum'
    hdu2.header.comments['DATASUM'] = 'data unit checksum'
    hdu1 = fits.HDUList([hdu1, hdu2])
    fname = os.path.join(output_dir, h_prim['FILENAME']+'.fits')
    hdu1.writeto(fname, output_verify='ignore', overwrite=True)

def add_sky_background(img, filt, exptime, sky_map=None, tel=None):
    # Add sky background
    if sky_map is None:
        sky_map = filt.getSkyNoise(exptime=exptime)
        sky_map = sky_map * np.ones_like(img.array)
        sky_map = galsim.Image(array=sky_map)
        # Apply Poisson noise to the sky map
        # # (NOTE): only for photometric chips if it utilizes the photon shooting to draw stamps
        # if self.survey_type == "photometric":
        #     sky_map.addNoise(poisson_noise)
    elif img.array.shape != sky_map.shape:
        raise ValueError("The shape img and sky_map must be equal.")
    elif tel is not None: # If sky_map is given in flux
        sky_map = sky_map * tel.pupil_area * exptime
    img += sky_map
    return img, sky_map

def get_flat(img, seed):
    flat_img = effects.MakeFlatSmooth(
                GSBounds=img.bounds, 
                seed=seed)
    flat_normal = flat_img / np.mean(flat_img.array)
    return flat_img, flat_normal

def add_cosmic_rays(img, chip, exptime=150, seed=0):
    cr_map, cr_event_num = effects.produceCR_Map(
        xLen=chip.npix_x, yLen=chip.npix_y, 
        exTime=exptime+0.5*chip.readout_time, 
        cr_pixelRatio=0.003*(exptime+0.5*chip.readout_time)/600.,
        gain=chip.gain, 
        attachedSizes=chip.attachedSizes,
        seed=seed)   # seed: obj-imaging:+0; bias:+1; dark:+2; flat:+3;
    img += cr_map
    cr_map[cr_map > 65535] = 65535
    cr_map[cr_map < 0] = 0
    crmap_gsimg = galsim.Image(cr_map, dtype=np.uint16)
    del cr_map
    return img, crmap_gsimg, cr_event_num

def add_PRNU(img, chip, seed=0):
    prnu_img = effects.PRNU_Img(
        xsize=chip.npix_x, 
        ysize=chip.npix_y, 
        sigma=0.01, 
        seed=seed)
    img *= prnu_img
    return img, prnu_img

def get_poisson(seed=0, sky_level=0.):
    rng_poisson = galsim.BaseDeviate(seed)
    poisson_noise = galsim.PoissonNoise(rng_poisson, sky_level=sky_level)
    return rng_poisson, poisson_noise

def get_base_img(img, read_noise, readout_time, dark_noise, exptime=150.):
    base_level = read_noise**2 + dark_noise*(exptime+0.5*readout_time)
    base_img = base_level * np.ones_like(img.array)
    return base_img

def add_poisson(img, chip, exptime=150., seed=0, sky_level=0., poisson_noise=None, dark_noise=None):
    if poisson_noise is None:
        _, poisson_noise = get_poisson(seed=seed, sky_level=sky_level)
    read_noise = chip.read_noise
    if dark_noise is None:
        dark_noise = chip.dark_noise
    base_img = get_base_img(img=img, read_noise=read_noise, readout_time=chip.readout_time, dark_noise=dark_noise, exptime=exptime)
    img += base_img
    img.addNoise(poisson_noise)
    img -= read_noise**2
    return img, base_img

def add_brighter_fatter(img):
    #Inital dynamic lib
    try:
178
179
        with pkg_resources.files('ObservationSim.Instrument.Chip.libBF').joinpath("libmoduleBF.so") as lib_path:
            print('--1', lib_path)
180
181
            lib_bf = ctypes.CDLL(lib_path)
    except AttributeError:
182
        with pkg_resources.path('ObservationSim.Instrument.Chip.libBF', "libmoduleBF.so") as lib_path:
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
            lib_bf = ctypes.CDLL(lib_path)
    lib_bf.addEffects.argtypes = [ctypes.c_int, ctypes.c_int, ctypes.POINTER(ctypes.c_float), ctypes.POINTER(ctypes.c_float), ctypes.c_int]
    
    # Set bit flag
    bit_flag = 1
    bit_flag = bit_flag | (1 << 2)

    nx, ny = img.array.shape
    nn = nx * ny
    arr_ima= (ctypes.c_float*nn)()
    arr_imc= (ctypes.c_float*nn)()

    arr_ima[:]= img.array.reshape(nn)
    arr_imc[:]= np.zeros(nn)

    lib_bf.addEffects(nx, ny, arr_ima, arr_imc, bit_flag)
199
    img.array[:, :] = np.reshape(arr_imc, [nx, ny])
200
    del arr_ima, arr_imc
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
    return img

def add_inputdark(img, chip, exptime):
    fname = "/share/home/weichengliang/CSST_git/test_new_sim/csst-simulation/ObservationSim/Instrument/data/dark/dark_1000s_example_0.fits"
    hdu = fits.open(fname)
    #ny, nx = img.array.shape
    #inputdark = np.zeros([ny, nx])
    img.array[:, :] += hdu[0].data/hdu[0].header['exptime']*exptime
    hdu.close()
    del inputdark
    return img

def AddPreScan(GSImage, pre1=27, pre2=4, over1=71, over2=80, nsecy = 2, nsecx=8):
    img= GSImage.array
    ny, nx = img.shape
    dx = int(nx/nsecx)
    dy = int(ny/nsecy)

    imgt=np.zeros([int(nsecy*nsecx), int(ny/nsecy+pre2+over2), int(nx/nsecx+pre1+over1)])
    for iy in range(nsecy):
        for ix in range(nsecx):
            if iy % 2 == 0:
                tx = ix
            else:
                tx = (nsecx-1)-ix
            ty = iy 
            chunkidx = int(tx+ty*nsecx) #chunk1-[1,2,3,4], chunk2-[5,6,7,8], chunk3-[9,10,11,12], chunk4-[13,14,15,16]

            imgtemp = np.zeros([int(ny/nsecy+pre2+over2), int(nx/nsecx+pre1+over1)])
            if int(chunkidx/4) == 0:
                imgtemp[pre2:-over2, pre1:-over1] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
                imgt[chunkidx, :, :] = imgtemp
            if int(chunkidx/4) == 1:
                imgtemp[pre2:-over2, over1:-pre1] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
                imgt[chunkidx, :, :] = imgtemp #[:, ::-1]
            if int(chunkidx/4) == 2:
                imgtemp[over2:-pre2, over1:-pre1] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
                imgt[chunkidx, :, :] = imgtemp #[::-1, ::-1]
            if int(chunkidx/4) == 3:
                imgtemp[over2:-pre2, pre1:-over1] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
                imgt[chunkidx, :, :] = imgtemp #[::-1, :]

    imgtx1 = np.hstack(imgt[:nsecx:,       :, :])
    imgtx2 = np.hstack(imgt[:(nsecx-1):-1, :, :])

    newimg = galsim.Image(int(nx+(pre1+over1)*nsecx), int(ny+(pre2+over2)*nsecy), init_value=0)
    newimg.array[:, :] = np.concatenate([imgtx1, imgtx2])

    newimg.wcs = GSImage.wcs
    return newimg

def formatOutput(GSImage, nsecy = 2, nsecx=8):
    img = GSImage.array
    ny, nx = img.shape
    dx = int(nx/nsecx)
    dy = int(ny/nsecy)
    
    imgt = np.zeros([int(nsecx*nsecy), dy, dx])
    for iy in range(nsecy):
        for ix in range(nsecx):
            if iy % 2 == 0:
                tx = ix
            else:
                tx = (nsecx-1)-ix
            ty = iy
            chunkidx = int(tx+ty*nsecx)
            if int(chunkidx/4) == 0:
                imgt[chunkidx, :, :] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
            if int(chunkidx/4) == 1:
                imgt[chunkidx, :, :] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
            if int(chunkidx/4) == 2:
                imgt[chunkidx, :, :] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
            if int(chunkidx/4) == 3:
                imgt[chunkidx, :, :] = img[iy*dy:(iy+1)*dy, ix*dx:(ix+1)*dx]
    
    imgttx0 = np.hstack(imgt[ 0:4:,    :,    :])
    imgttx1 = np.hstack(imgt[ 4:8:,    :, ::-1])
    imgttx2 = np.hstack(imgt[8:12:, ::-1, ::-1])
    imgttx3 = np.hstack(imgt[12:16:,::-1,    :])
    
    newimg = galsim.Image(int(dx*nsecx*nsecy), dy, init_value=0)
    newimg.array[:, :] = np.hstack([imgttx0, imgttx1, imgttx2, imgttx3])
    return newimg