Source code for acdc.analysis.analyze_correction

import os
import itertools
import numpy as np
import argparse
from matplotlib import pyplot as plt
dirname = os.path.dirname(__file__)
stylesheet = os.path.join(dirname, "niceplot.mplstyle")
plt.style.use(stylesheet)
import glob
from astropy.io import fits
from scipy import stats
import asdf

from acdc.superdark.cos_fuv_superdark import Superdark
from acdc.utils.utils import get_binning_pars, bin_coords, unbin_image

LOCAL_REFDIR = "/grp/hst/cdbs/lref"
LOCAL_DARKDIR = "/astro/sveash/cos_dark/final_superdarks"
PHA_INCLUSIVE = [2, 23]
PHA_INCL_EXCL = [2, 24]
# Colorblind-safe palette below
#COLORS = ["#1b9e77", "#d95f02", "#7570b3", "#e7298a", "#66a61e", "#a6cee3",
#          "#1f78b4", "#b2df8a", "#33a02c", "#fb9a99", "#e31a1c", "#fdbf6f",
#          "#ff7f00", "#cab2d6", "#6a3d9a", "#ffff99", "#b15928", "#a6cee3"]
COLORS = ["#9970ab", "#5aae61", "#d95f02", "#e7298a", "#1bddf2", "#1f78b4", 
          "#fb9a99", "#fdbf6f", "#ffe44b", "#b15928", "#cab2d6", "#b2df8a", 
          "#000000", "#7a7a7a", "#911cff", "#a6cee3"]

[docs] class CosImage(): def __init__(self, filename, arr, segment, cenwave, hv, exptime, binning=None, data_binned=None): self.filename = filename self.data = arr self.segment = segment self.cenwave = cenwave self.hv = hv self.exptime = exptime self.binning = binning self.data_binned = data_binned @classmethod def from_flt(cls, flt_file): data = fits.getdata(flt_file) hdr0 = fits.getheader(flt_file, 0) hdr1 = fits.getheader(flt_file, 1) segment = hdr0["segment"] self = cls(data, flt_file, segment, hdr0["cenwave"], hdr1[f"HVLEVEL{segment[-1]}"], hdr1["exptime"]) return self @classmethod def from_asdf(cls, asdf_file): S = Superdark.from_asdf(asdf_file) if S.pha_bins != PHA_INCL_EXCL: S.bin_superdark(1, 1, pha_bins=PHA_INCL_EXCL) if S.bin_x != 1 or S.bin_y != 1: data_binned = S.superdarks[0] data = unbin_image(data_binned, S.bin_x, S.bin_y, S.xstart, S.ystart, S.xend, S.yend) else: data_binned = None data = S.superdarks[0] keys = ["bin_pha", "bin_x", "bin_y", "xstart", "xend", "ystart", "yend", "phastart", "phaend"] binning = {} for k in keys: binning[k] = getattr(S, k) self = cls(data, asdf_file, S.segment, None, S.hv, S.total_exptime, binning, data_binned) def get_background_regions(self, xtractab, cenwave): xtract = fits.getdata(xtractab) inds = np.where((xtract["segment"] == self.segment) & (xtract["cenwave"] == cenwave) & (xtract["aperture"] == "PSA")) xtract_data = xtract[inds] b1 = xtract_data["b_bkg1"] b2 = xtract_data["b_bkg2"] height1 = xtract_data["b_hgt1"] height2 = xtract_data["b_hgt1"] m = xtract_data["slope"] smoothsize = xtract_data["bwidth"] x = np.arange(0, 16384) bkg1_mid = m*x + b1 bkg1_y1 = bkg1_mid + height1/2 bkg1_y0 = bkg1_mid - height1/2 bkg2_mid = m*x + b2 bkg2_y1 = bkg2_mid + height2/2 bkg2_y0 = bkg2_mid - height2/2 bkg1_y0 = np.round(bkg1_y0) bkg1_y0 = bkg1_y0.astype(int) bkg2_y0 = np.round(bkg2_y0) bkg2_y0 = bkg2_y0.astype(int) bkg1_y1 = np.round(bkg1_y1) bkg1_y1 = bkg1_y1.astype(int) bkg2_y1 = np.round(bkg2_y1) bkg2_y1 = bkg2_y1.astype(int) # self.bkg1_xstart = bkg1_x0 # self.bkg1_xend = bkg1_x1 self.bkg1_ystart = bkg1_y0 self.bkg1_yend = bkg1_y1 # self.bkg2_xstart = bkg2_x0 # self.bkg2_xend = bkg2_x1 self.bkg2_ystart = bkg2_y0 self.bkg2_yend = bkg2_y1
[docs] def sum_bkg_region(data, bkg_lo, bkg_hi): assert np.shape(data)[1] == len(bkg_lo) == len(bkg_hi), "Data X-dimension must match limits of background regions" datasum = [] for i in range(len(bkg_lo)): colsum = np.sum(data[bkg_lo[i]:bkg_hi[i], i]) datasum.append(colsum) datasum = np.array(datasum) return datasum
[docs] def sum_region_ylims(data_2d, y0, y1, x0=None, x1=None): """ Sum a region of a 2D image. *NOTE*: The upper limits (y1, x1) are EXCLUSIVE, as is typical in python. E.g. specifying y0=9, y1=11 will sum across rows 9 and 10 only. Args: data_2d (2D np.array): 2-dimensional array to sum over. y0 (int): Lower row limit. y1 (int): Upper row limit (EXCLUSIVE). x0 (int): Lower column limit. x1 (int): Upper column limit (EXCLUSIVE). """ if x0 is None: x0 = 0 if x1 is None: sh = np.shape(data_2d) x1 = sh[1] datasum = np.sum(data_2d[y0:y1, x0:x1], axis=0) return datasum
[docs] def smooth_array(data, smoothsize=100): bin_edges = np.arange(0, len(data), smoothsize) binned_y, edges, binnumber = stats.binned_statistic(np.arange(len(data)), data, statistic="mean", bins=bin_edges) binned_x = bin_edges[:-1] + (smoothsize/2.) return binned_y, binned_x
[docs] def get_data(flt_custom, flt_default, superdark_custom, dark_dir=LOCAL_DARKDIR, ref_dir=LOCAL_REFDIR): data_custom = fits.getdata(flt_custom) data_default = fits.getdata(flt_default) segment = fits.getval(flt_custom, "segment") cenwave = fits.getval(flt_custom, "cenwave") hv = fits.getval(flt_custom, f"HVLEVEL{segment[-1]}", 1) xtractab_file0 = fits.getval(flt_custom, "xtractab") xtractab_file1 = xtractab_file0.split("$")[1] xtractab_file = os.path.join(ref_dir, xtractab_file1) xtract = fits.getdata(xtractab_file) inds = np.where((xtract["segment"] == segment) & (xtract["cenwave"] == cenwave) & (xtract["aperture"] == "PSA")) xtract_data = xtract[inds] b1 = xtract_data["b_bkg1"] b2 = xtract_data["b_bkg2"] height1 = xtract_data["b_hgt1"] height2 = xtract_data["b_hgt1"] m = xtract_data["slope"] smoothsize = xtract_data["bwidth"] x = np.arange(0, 16384) bkg1_mid = m*x + b1 bkg1_y1 = bkg1_mid + height1/2 bkg1_y0 = bkg1_mid - height1/2 bkg2_mid = m*x + b2 bkg2_y1 = bkg2_mid + height2/2 bkg2_y0 = bkg2_mid - height2/2 bkg1_y0 = np.round(bkg1_y0) bkg1_y0 = bkg1_y0.astype(int) bkg2_y0 = np.round(bkg2_y0) bkg2_y0 = bkg2_y0.astype(int) bkg1_y1 = np.round(bkg1_y1) bkg1_y1 = bkg1_y1.astype(int) bkg2_y1 = np.round(bkg2_y1) bkg2_y1 = bkg2_y1.astype(int) superdark_lo = glob.glob(os.path.join(dark_dir, f"*{segment}*{hv}*quiescent*asdf")) superdark_lo = [x for x in superdark_lo if "binned" not in x] superdark_hi = glob.glob(os.path.join(dark_dir, f"*{segment}*{hv}*active*asdf")) superdark_hi = [x for x in superdark_hi if "binned" not in x] assert len(superdark_lo) != 0, \ f"No quiescent superdarks found for {segment} {hv}" assert len(superdark_hi) != 0, \ f"No active superdarks found for {segment} {hv}" assert len(superdark_lo) == 1, \ f"Too many quiescent superdarks found for {segment} {hv}" assert len(superdark_hi) == 1, \ f"Too many active superdarks found for {segment} {hv}" superdark_lo = superdark_lo[0] superdark_hi = superdark_hi[0] S_lo = Superdark.from_asdf(superdark_lo) S_lo.bin_superdark(1, 1, pha_bins=[1,31]) #S_lo.bin_superdark(1, 1, pha_bins=PHA_INCL_EXCL) S_hi = Superdark.from_asdf(superdark_hi) S_hi.bin_superdark(1, 1, pha_bins=[1,31]) #S_hi.bin_superdark(1, 1, pha_bins=PHA_INCL_EXCL) lo_exptime = S_lo.total_exptime hi_exptime = S_hi.total_exptime S_lo_flt0 = S_lo.superdarks[0] / lo_exptime S_hi_flt0 = S_hi.superdarks[0] / hi_exptime assert np.shape(S_lo_flt0) == np.shape(S_hi_flt0), \ "Quiescent and active superdark dimensions do not match" xstart = S_lo.bin_xstart xend = S_lo.bin_xend ystart = S_lo.bin_ystart yend = S_lo.bin_yend S_lo_flt = np.zeros(16777216).reshape(1024, 16384) S_hi_flt = np.zeros(16777216).reshape(1024, 16384) S_lo_flt[ystart:yend, xstart:xend] = S_lo_flt0 S_hi_flt[ystart:yend, xstart:xend] = S_hi_flt0 bkg1_sum_lo = sum_bkg_region(S_lo_flt, bkg1_y0, bkg1_y1) bkg1_lo,_ = smooth_array(bkg1_sum_lo) bkg2_sum_lo = sum_bkg_region(S_lo_flt, bkg2_y0, bkg2_y1) bkg2_lo,_ = smooth_array(bkg2_sum_lo) bkg1_sum_hi = sum_bkg_region(S_hi_flt, bkg1_y0, bkg1_y1) bkg1_hi,_ = smooth_array(bkg1_sum_hi) bkg2_sum_hi = sum_bkg_region(S_hi_flt, bkg2_y0, bkg2_y1) bkg2_hi,_ = smooth_array(bkg2_sum_hi) bkg1_sum_custom = sum_bkg_region(data_custom, bkg1_y0, bkg1_y1) bkg1_custom,_ = smooth_array(bkg1_sum_custom) bkg2_sum_custom = sum_bkg_region(data_custom, bkg2_y0, bkg2_y1) bkg2_custom,_ = smooth_array(bkg2_sum_custom) bkg1_sum_default = sum_bkg_region(data_default, bkg1_y0, bkg1_y1) bkg1_default,_ = smooth_array(bkg1_sum_default) bkg2_sum_default = sum_bkg_region(data_default, bkg2_y0, bkg2_y1) bkg2_default, binned_x = smooth_array(bkg2_sum_default) af = asdf.open(superdark_custom) pha_str = f"pha{PHA_INCL_EXCL[0]}-{PHA_INCL_EXCL[1]}" dark_im = af.tree[pha_str] dark_im_flt0 = dark_im / af.tree["total_exptime"] binning = get_binning_pars(af) dark_im_flt_perpixel = dark_im_flt0 / binning["bin_y"] / binning["bin_x"] dark_im_flt = np.zeros(16777216).reshape(1024, 16384) xs = np.arange(binning["xstart"], binning["xend"], binning["bin_x"]) ys = np.arange(binning["ystart"], binning["yend"], binning["bin_y"]) for i in range(len(xs)-1): for j in range(len(ys)-1): dark_im_flt[ys[j]:ys[j+1], xs[i]:xs[i+1]] = dark_im_flt_perpixel[j,i] dark_x = np.arange(16384) bkg1_dark = sum_bkg_region(dark_im_flt, bkg1_y0, bkg1_y1) bkg2_dark = sum_bkg_region(dark_im_flt, bkg2_y0, bkg2_y1) return binned_x, bkg1_default, bkg2_default, bkg1_custom, bkg2_custom, \ bkg1_lo, bkg2_lo, bkg1_hi, bkg2_hi, bkg1_dark, bkg2_dark, dark_x
#return np.arange(16384), bkg1_sum_default, bkg2_sum_default, bkg1_sum_custom, bkg2_sum_custom, bkg1_sum_lo, bkg2_sum_lo, bkg1_sum_hi, bkg2_sum_hi
[docs] def plot_data(x, bkg1_default, bkg2_default, bkg1_custom, bkg2_custom, bkg1_lo, bkg2_lo, bkg1_hi, bkg2_hi, bkg1_dark, bkg2_dark, dark_x, targname, flt_default): fig, axes0 = plt.subplots(3, 1, figsize=(9, 12)) fig.subplots_adjust(hspace=0.3) segment = fits.getval(flt_default, "segment") rootname = fits.getval(flt_default, "rootname").lower() hv = fits.getval(flt_default, f"HVLEVEL{segment[-1]}", 1) date = fits.getval(flt_default, "date-obs", 1) axes = axes0.flatten() bkg1 = {"default": bkg1_default, "custom": bkg1_custom, "lo": bkg1_lo, "hi": bkg1_hi, "superdark": bkg1_dark} bkg2 = {"default": bkg2_default, "custom": bkg2_custom, "lo": bkg2_lo, "hi": bkg2_hi, "superdark": bkg2_dark} av = {"default": (bkg1_default + bkg2_default) / 2., "custom": (bkg1_custom + bkg2_custom) / 2., "lo": (bkg1_lo + bkg2_lo) / 2., "hi": (bkg1_hi + bkg2_hi) / 2., "superdark": (bkg1_dark + bkg2_dark) / 2.} dicts = [bkg1, bkg2, av] titles = ["Background Region 1", "Background Region 2", "Average"] alpha = 0.7 for i in range(3): ax = axes[i] ax.hlines(y=0, xmin=-1000, xmax=17384, colors="lightgrey", linestyles="--") d = dicts[i] ax.set_ylabel("Counts/s") ax.plot(x, d["default"], COLORS[0], alpha=alpha, lw=1.7, label="Default FLT") ax.plot(x, d["custom"], COLORS[1], alpha=alpha, lw=1.7, label="Custom FLT", zorder=100) ax.plot(x, d["lo"], COLORS[2], alpha=alpha, lw=1.7, label="Lo Dark") ax.plot(x, d["hi"], COLORS[3], alpha=alpha, lw=1.7, label="Hi Dark") ax.plot(dark_x, d["superdark"], COLORS[4], alpha=alpha, lw=1.7, label="Custom Dark") ax.set_ylim(-.0001, 0.0004) ax.set_xlim(-500, 16884) ax.set_title(titles[i]) axes[0].legend() figname = f"{rootname}_{segment}_flt_comp.png" title = f"{segment}; HV={hv}; {date} ({rootname})" fig.suptitle(title, size=20, y=.94) fig.savefig(figname) print(f"Wrote {figname}")
[docs] def compare_backgrounds(flt_default, flt_custom, targname, superdark_custom): binned_x, bkg1_default, bkg2_default, bkg1_custom, bkg2_custom, bkg1_lo, \ bkg2_lo, bkg1_hi, bkg2_hi, bkg1_dark, bkg2_dark, dark_x = get_data(flt_custom, flt_default, superdark_custom) plot_data(binned_x, bkg1_default, bkg2_default, bkg1_custom, bkg2_custom, \ bkg1_lo, bkg2_lo, bkg1_hi, bkg2_hi, bkg1_dark, bkg2_dark, dark_x, targname, flt_default)
if __name__ == "__main__": parser = argparse.ArgumentParser() parser.add_argument("-d", "--default", help="Name of default CalCOS FLT file") parser.add_argument("-c", "--custom", help="Name of custom-corrected FLT file") parser.add_argument("-t", "--targname", help="Name of target") parser.add_argument("-s", "--superdark_custom", help="Name of combined superdark from ACDC") args = parser.parse_args() compare_backgrounds(args.default, args.custom, args.targname, args.superdark_custom)