pyimcom.analysis

Tools to analyze coadded images.

Classes

OutImage: Wrapper for coadded images (blocks).
NoiseAnal: Analysis of noise frames.
StarsAnal: Analysis of point sources.
_BlkGrp: Abstract base class for groups of blocks (mosiacs or suites).
Mosaic: Wrapper for coadded mosaics (2D arrays of blocks).
Suite: Wrapper for coadded suites (hashed arrays of blocks).

Attributes

ColDescr

Classes

`OutImage`	Wrapper for coadded images (blocks).
`NoiseAnal`	Analysis of noise frames.
`StarsAnal`	Analysis of point sources.
`_BlkGrp`	Abstract base class for groups of blocks (mosiacs or suites).
`Mosaic`	Wrapper for coadded mosaics (2D arrays of blocks).
`Suite`	Wrapper for coadded suites (hashed arrays of blocks).

Module Contents

class OutImage(fpath: str, cfg: pyimcom.config.Config = None, hdu_names: list[str] = None)[source]

Wrapper for coadded images (blocks).

Parameters:

fpath (str or str-like) – Path to the output FITS file.
cfg (Config, optional) – Configuration used for this output image. If provided, no consistency check is performed. If None, it will be extracted from FITS file.
hdu_names (list of str, optional) – List of HDU names of this FITS file. If provided, no consistency check is performed. If None, it will be derived from cfg.

get_hdu_names()[source]: Parse outmaps to get a list of HDU names.

__init__()[source]: Constructor.

get_last_line()[source]: Get last line of a text file.

get_time_consump()[source]: Parse terminal output to get time consumption.

_load_or_save_hdu_list()[source]: Load data from or save data to FITS file.

get_coadded_layer()[source]: Extract a coadded layer from the primary HDU.

get_T_weightmap()[source]: Extract T_weightmap from an additional HDU.

get_mean_coverage()[source]: Compute mean coverage based on T_weightmap.

get_output_map()[source]: Extract an output map from the additional HDUs.

_update_hdu_data()[source]: Update data using data provided by a neighbor.

static get_hdu_names(outmaps: str) → list[str][source]

Parse outmaps to get a list of HDU names.

Parameters:: outmaps (str) – outmaps attribute of a Config instance.
Returns:: A list of HDU names.
Return type:: list of str

fpath = ''[source]

cfg = None[source]

hdu_names = None[source]

static get_last_line(fname: str) → str[source]

Get last line of a text file.

Parameters:: fname (str) – Path to the text file.
Returns:: Last line of the text file.
Return type:: str

get_time_consump() → float[source]

Parse terminal output to get time consumption.

Returns:: Time consumption in seconds (if found), otherwise nan.
Return type:: float

_load_or_save_hdu_list(load_mode: bool = True, save_file: bool = False, auto_to_all: bool = False) → None[source]

Load data from or save data to FITS file.

Parameters:

load_mode (bool, optional) – If True, load data from FITS file (if not already loaded); if False, remove current data from memory (if data exist).
save_file (bool, optional) – Only used when load_mode == False. If (save_file ==) True, save current data to FITS file (overwriting the existing file).
auto_to_all (bool, optional) – Only used when load_mode == False and save_file == True. If (auto_to_all ==) True, change ‘PADSIDES’ from ‘auto’ to ‘all’ in the ‘CONFIG’ HDU of FITS file.

Return type:

None

get_coadded_layer(layer: str, j_out: int = 0) → numpy.array[source]

Extract a coadded layer from the primary HDU.

Parameters:

layer (str) – Name of the layer to be extracted.
j_out (int or None, optional) – Index of the output PSF. If None, return results based on all output PSFs.

Returns:

data – Requested coadded layer. The shape is either (NsideP, NsideP) or (n_out, NsideP, NsideP) (all output PSFs)

Return type:

np.array

get_T_weightmap(flat: bool = False, j_out: int = 0) → numpy.array[source]

Extract T_weightmap from an additional HDU.

Parameters:

flat (bool, optional) – Whether to read the flat version of T_weightmap.
j_out (int or None, optional) – Only used when flat is False. Index of the output PSF. If None, return results based on all output PSFs.

Returns:

data – Requested T_weightmap. Shape is either (n_inimage, n1P, n1P) (1) or (n_out, n_inimage, n1P, n1P) (all output PSFs) or (n_out*n1P, n_inimage*n1P) (flat version)

Return type:

np.array

get_mean_coverage(padding: bool = False) → float[source]

Compute mean coverage based on T_weightmap.

We assume that mean coverage is the same for all output PSFs.

Parameters:: padding (bool, optional) – Whether to include padding postage stamps. The default is False.
Returns:: mean_coverage – Mean coverage based on T_weightmap.
Return type:: float

get_output_map(outmap: str, j_out: int = 0) → numpy.array[source]

Extract an output map from the additional HDUs.

Parameters:

outmap (str) – Name of the output map to be extracted.
j_out (int or None, optional) – Index of the output PSF. If None, return results based on all output PSFs.

Returns:

data – Requested output map. shape is either (NsideP, NsideP) or (n_out, NsideP, NsideP) (all output PSFs).

Return type:

np.array

_update_hdu_data(neighbor: OutImage, direction: str, add_mode: bool = True) → None[source]

Update data using data provided by a neighbor.

This method is developed for postprocessing, i.e., sharing padding postage stamps between adjacent blocks. This method neither loads nor saves hdu_list.

Parameters:

neighbor (OutImage) – Neighboring output image (block) who shares data with “me.”
direction (str) – Which side to update. Must be ‘left’, ‘right’, ‘bottom’, or ‘top’.
add_mode (bool, optional) – If True, update “my” data by adding neighbor’s to “mine;” if False, replace “my” data with neighbor’s.

Return type:

None

class NoiseAnal(outim: OutImage, layer: str)[source]

Analysis of noise frames.

Largely based on diagnostics/noise/noisespecs.py.

Parameters:

outim (OutImage) – Output image to analyze.
layer (str) – Layer name of noise frame to analyze.

__init__()[source]: Constructor.

get_norm()[source]: Get norm for 2D noise power spectrum (classmethod).

azimuthal_average()[source]: Compute radial profile of image (staticmethod)).

_get_wavenumbers()[source]: Calculate wavenumbers for the input image (staticmethod).

__call__()[source]: Analyze specified noise frame of given output image.

clear()[source]: Free up memory space.

AREA[source]

tfr = 3.08[source]

gain = 1.458[source]

ABstd = 3.631e-20[source]

h[source]

m_ab = 23.9[source]

s_in = 0.11[source]

PS1D_COLORS = ['orange', 'darksalmon', 'palevioletred', 'mediumvioletred', 'darkviolet'][source]

PS1D_STYLES = ['solid', 'dotted', 'dashed', 'solid', 'dashdot'][source]

outim[source]

layer[source]

cfg[source]

classmethod get_norm(layer: str, L: int, filtername: str, s_out: float) → float[source]

Get norm for 2D noise power spectrum.

Parameters:

layer (str) – Layer name of noise frame to analyze.
L (int) – Side length of noise frame in px.
filtername (str) – Name of filter used for this output image.
s_out (float) – Output pixel scale in arcsec.

Returns:

Norm for 2D noise power spectrum.

Return type:

float

Notes

For simulated noise frames, dividing by s_in**2 converts from units of S_in^2 to arcsec^2.

static azimuthal_average(image: numpy.array, nradbins: int, rbin: numpy.array = None, ridx: numpy.array = None) → numpy.array[source]

Compute radial profile of image.

Parameters:

image (np.array) – Input image, shape (L, L).
nradbins (int) – Number of radial bins in profile.
rbin (np.array, optional) – “labels” parameter for ndimage utilities. If provided, has shape (L, L). The default is None. If not provided, derive from image.shape.
ridx (np.array, optional) – “index” parameter for ndimage utilities. If provided, has shape (nradbins,). The default is None. If not provided, derive from rbin.

Returns:

radial_mean (np.array) – Mean intensity within each annulus. Main result. Shape is (nradbins,)
radial_err (np.array) – Standard error on the mean: sigma / sqrt(N). Shape is (nradbins,)

static _get_wavenumbers(window_length: int, nradbins: int, rbin: numpy.array = None, ridx: numpy.array = None) → numpy.array[source]

Calculate wavenumbers for the input image.

Parameters:

window_length (int) – the length of one axis of the image.
nradbins (int) – number of radial bins the image should be averaged into
rbin (np.array, optional) – “labels” parameter for ndimage utilities. If provided, shape is (L,L). The default is None. If not provided, derive from image.shape.
ridx (np.array, optional) – “index” parameter for ndimage utilities. If provided, shape is (nradbins,). The default is None. If not provided, derive from rbin.

Returns:

kmean – the wavenumbers for the image, shape (nradbins,)

Return type:

np.array

__call__(padding: bool = False, bin_: bool = True, rbin: numpy.array = None, ridx: numpy.array = None) → None[source]

Analyze specified noise frame of given output image.

Parameters:

padding (bool, optional) – Whether to include padding postage stamps. (to be implemented)
bin (bool, optional) – Whether to bin the 2D spectrum into L/8 x L/8 image. Currently this is ignored, as only bin_ == True is supported.
rbin (np.array, optional, shape : (L, L)) – “labels” parameter for ndimage utilities. The default is None. If not provided, derive from image.shape.
ridx (np.array, optional, shape : (nradbins,)) – “index” parameter for ndimage utilities. The default is None. If not provided, derive from rbin.

Return type:

None

Notes

If the image side length is not a multiple of 8, the extra pixels (L // 8) are clipped.

clear() → None[source]: Free up memory space.

ColDescr[source]

class StarsAnal(outim: OutImage, layer: str = 'gsstar14')[source]

Analysis of point sources.

Largely based on diagnostics/starcube_nonoise.py.

Parameters:

outim (OutImage) – Output image to analyze.
layer (str, optional) – Layer name of injected stars to analyze.

__init__()[source]: Constructor.

__call__()[source]: Analyze given point source frame of given output image.

clear()[source]: Free up memory space.

bd = 40[source]

bd2 = 8[source]

ncol[source]

outim[source]

layer = 'gsstar14'[source]

cfg[source]

__call__(n: int = None, search_radius: float = None, forced_scale: float = None, bdpad: int = None, res: int = None) → None[source]

Analyze given point source frame of given output image.

Parameters:

n (int or None, optional) – Size of output images. If not provided, derive from self.cfg. Same for other parameters.
search_radius (float or None, optional) – Search radius for injected point sources.
forced_scale (float or None, optional) – Forced scale length for star moments.
bdpad (int or None, optional) – Padding region around the edge.
res (int or None, optional) – Resolution of HEALPix grid.

Return type:

None

clear() → None[source]

Free up memory space.

Return type:: None.

class _BlkGrp[source]

Abstract base class for groups of blocks (mosiacs or suites).

__call__()[source]: Run all the analyses below.

get_consump_map()[source]: Get map of time consumption.

get_coverage_map()[source]: Get map of mean coverages.

get_noise_power_spectra()[source]: Analyze noise power spectra of this mosaic.

get_star_catalog()[source]: Analyze injected point sources of this mosaic.

clear()[source]: Free up memory space.

__call__(overwrite: bool = False) → None[source]

Run all the analyses below.

Parameters:: overwrite (bool, optional) – Whether to overwrite existing results.
Return type:: None

get_consump_map(overwrite: bool = False) → None[source]

Get map of time consumption.

Parameters:: overwrite (bool, optional) – Whether to overwrite existing results.
Return type:: None

get_coverage_map(overwrite: bool = False) → None[source]

Get map of mean coverages.

Parameters:: overwrite (bool, optional) – Whether to overwrite existing results.
Return type:: None

get_noise_power_spectra(bins: int = 5, overwrite: bool = False) → None[source]

Analyze noise power spectra of this mosaic.

The output noise power spectra are written to self.cfg.outstem + "_NoisePS.npz". These are in the format:

ps2d_all = 2D power spectrum, [which_noise_layer, ybin, xbin]
ps1d_all = 1D power spectrum, [which_noise_layer, coverage_bin, wavenumber_bin, value_or_err]
wavenumber = 1D array of wavenumbers

Parameters: