aboutsummaryrefslogtreecommitdiffstats
path: root/fg21sim/utils/fits.py
blob: 4991cd084e5732e3f7c98a6f0da49a288d368087 (plain)
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
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
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
# Copyright (c) 2016 Weitian LI <liweitianux@live.com>
# MIT license

"""
FITS utilities
--------------

read_fits_healpix:
    Read the HEALPix map from a FITS file or a BinTableHDU to 1D array
    in *RING* ordering.

write_fits_healpix:
    Write the HEALPix map to a FITS file with proper header as well
    as the user-provided header.
"""

from datetime import datetime, timezone

import numpy as np
from astropy.io import fits
import healpy as hp


# Column formats for FITS binary table
# Reference:
# http://docs.astropy.org/en/stable/io/fits/usage/table.html#column-creation
FITS_COLUMN_FORMATS = {
    np.dtype("bool"):       "L",
    np.dtype("uint8"):      "B",
    np.dtype("int16"):      "I",
    np.dtype("int32"):      "J",
    np.dtype("int64"):      "K",
    np.dtype("float32"):    "E",
    np.dtype("float64"):    "D",
    np.dtype("complex64"):  "C",
    np.dtype("complex128"): "M",
}


def read_fits_healpix(filename):
    """Read the HEALPix map from a FITS file or a BinTableHDU to 1D array
    in *RING* ordering.

    Parameters
    ----------
    filename : str or `~astropy.io.fits.BinTableHDU`
        Filename of the HEALPix FITS file,
        or an `~astropy.io.fits.BinTableHDU` instance.

    Returns
    -------
    data : 1D `~numpy.ndarray`
        HEALPix data in *RING* ordering with same dtype as input
    header : `~astropy.io.fits.Header`
        Header of the input FITS file

    NOTE
    ----
    This function wraps on `healpy.read_map()`, but set the data type of
    data array to its original value as in FITS file, as well as return
    FITS header as `~astropy.io.fits.Header` instance.
    """
    if isinstance(filename, fits.BinTableHDU):
        hdu = filename
    else:
        # Read the first extended table
        hdu = fits.open(filename)[1]
    # Hack to ignore the dtype byteorder, use native endianness
    dtype = np.dtype(hdu.data.field(0).dtype.type)
    header = hdu.header
    data = hp.read_map(hdu, nest=False, verbose=False)
    return (data.astype(dtype), header)


def write_fits_healpix(filename, hpmap, header=None, clobber=False,
                       checksum=False):
    """Write the HEALPix map to a FITS file with proper header as well
    as the user-provided header.

    This function currently only support one style of HEALPix with the
    following specification:
    - Only one column: I (intensity)
    - ORDERING: RING
    - COORDSYS: G (Galactic)
    - OBJECT: FULLSKY
    - INDXSCHM: IMPLICIT

    Parameters
    ----------
    filename : str
        Filename of the output file to write the HEALPix map data
    hpmap : 1D `~numpy.ndarray`
        1D array containing the HEALPix map data, and the ordering
        scheme should be "RING";
        The data type is preserved in the output FITS file.
    header : `~astropy.io.fits.Header`, optional
        Extra header to be appended to the output FITS
    clobber : bool, optional
        Whether overwrite the existing file?
    checksum : bool, optional
        Whether calculate the checksum for the output file, which is
        recorded as the "CHECKSUM" header keyword.

    NOTE
    ----
    - This function is intended to replace the most common case of
      `healpy.write_map()`, which still uses some deprecated functions of
      `numpy` and `astropy`, meanwhile, it interface/arguments is not very
      handy.
    - This function (currently) only implement the very basic feature of
      the `healpy.write_map()`.
    """
    hpmap = np.asarray(hpmap)
    if hpmap.ndim != 1:
        raise ValueError("Invalid HEALPix data: only support 1D array")
    # Hack to ignore the dtype byteorder, use native endianness
    dtype = np.dtype(hpmap.dtype.type)
    hpmap = hpmap.astype(dtype)
    #
    npix = hpmap.size
    nside = int((npix / 12) ** 0.5)
    colfmt = FITS_COLUMN_FORMATS.get(hpmap.dtype)
    if hpmap.size > 1024:
        hpmap = hpmap.reshape(int(hpmap.size/1024), 1024)
        colfmt = "1024" + colfmt
    #
    hdr = fits.Header()
    # set HEALPix parameters
    hdr["PIXTYPE"] = ("HEALPIX", "HEALPix pixelization")
    hdr["ORDERING"] = ("RING",
                       "Pixel ordering scheme, either RING or NESTED")
    hdr["COORDSYS"] = ("G", "Ecliptic, Galactic or Celestial (equatorial)")
    hdr["NSIDE"] = (nside, "HEALPix resolution parameter")
    hdr["NPIX"] = (npix, "Total number of pixels")
    hdr["FIRSTPIX"] = (0, "First pixel # (0 based)")
    hdr["LASTPIX"] = (npix-1, "Last pixel # (0 based)")
    hdr["INDXSCHM"] = ("IMPLICIT", "Indexing: IMPLICIT or EXPLICIT")
    hdr["OBJECT"] = ("FULLSKY", "Sky coverage, either FULLSKY or PARTIAL")
    #
    hdr["EXTNAME"] = ("HEALPIX", "Name of the binary table extension")
    hdr["CREATOR"] = (__name__, "File creator")
    hdr["DATE"] = (datetime.now(timezone.utc).astimezone().isoformat(),
                   "File creation date")
    # Merge user-provided header
    # NOTE: use the `.extend()` method instead of `.update()` method
    if header is not None:
        hdr.extend(fits.Header(header))
    #
    hdu = fits.BinTableHDU.from_columns([
        fits.Column(name="I", array=hpmap, format=colfmt)
    ], header=hdr)
    hdu.writeto(filename, clobber=clobber, checksum=checksum)