"""
General utilities for astronomical image analysis.
This module provides helper functions for cutouts, statistical metrics,
binning, and generating synthetic images.
"""
from astropy.nddata import Cutout2D
import numpy as np
from astropy.nddata import Cutout2D
from astropy.nddata.utils import NoOverlapError
[docs]
def cutout(data, coords, shape, wcs=None, fill_value=np.nan):
"""
Extract cutouts from data at given coordinates.
Parameters
----------
data : np.ndarray
2D image data.
coords : array-like
List of (x, y) coordinates.
shape : tuple
Shape of the cutout.
wcs : WCS or None, optional
World Coordinate System object.
Returns
-------
np.ndarray
Array of cutout images.
"""
values = []
dummy = np.zeros(shape, dtype=data.dtype)
for coords in coords:
try:
cutout = Cutout2D(
data, coords, shape, wcs=wcs, fill_value=fill_value, mode="partial"
)
except NoOverlapError:
cutout = dummy
values.append(cutout.data)
return np.array(values)
[docs]
def std_diff_metric(fluxes):
"""
Compute the standard deviation of the difference along the last axis.
Parameters
----------
fluxes : np.ndarray
Fluxes array.
Returns
-------
np.ndarray
Standard deviation of the differences.
"""
k = len(list(np.shape(fluxes)))
return np.std(np.diff(fluxes, axis=k - 1), axis=k - 1)
[docs]
def stability_aperture(fluxes):
"""
Compute the mean absolute difference between consecutive fluxes.
Parameters
----------
fluxes : np.ndarray
Fluxes array.
Returns
-------
np.ndarray
Mean absolute difference for each aperture.
"""
lc_c = np.abs(np.diff(fluxes, axis=0))
return np.mean(lc_c, axis=1)
[docs]
def index_binning(x, size):
"""
Bin indices of x into bins of given size.
Parameters
----------
x : array-like
Array to bin.
size : int or float
Bin size.
Returns
-------
list
List of arrays of indices for each bin.
"""
if isinstance(size, float):
bins = np.arange(np.min(x), np.max(x), size)
else:
x = np.arange(0, len(x))
bins = np.arange(0.0, len(x), size)
d = np.digitize(x, bins)
n = np.max(d) + 2
indexes = []
for i in range(0, n):
s = np.where(d == i)
if len(s[0]) > 0:
s = s[0]
indexes.append(s)
return indexes
[docs]
def binned_nanstd(x, bins: int = 12):
"""
Return a function to compute the mean of the standard deviation in bins.
Parameters
----------
x : np.ndarray
Array to bin.
bins : int, optional
Number of bins.
Returns
-------
callable
Function that computes the mean of the standard deviation in bins.
"""
# set binning idxs for white noise evaluation
bins = np.min([x.shape[-1], bins])
n = x.shape[-1] // bins
idxs = np.arange(n * bins)
def compute(f):
return np.nanmean(
np.nanstd(np.array(np.split(f.take(idxs, axis=-1), n, axis=-1)), axis=-1),
axis=0,
)
return compute
[docs]
def fake_image(shape=50, seed=0, stars=30):
"""
Generate a fake image with random stars.
Parameters
----------
shape : int, optional
Size of the image (shape x shape).
seed : int, optional
Random seed.
stars : int, optional
Number of stars.
Returns
-------
tuple
Tuple (image, coords) where image is the generated image and coords are the star positions.
"""
np.random.seed(seed)
image = np.zeros((shape, shape))
coords = (np.random.rand(2, stars) * shape).astype(int).T
for i, j in coords:
image[j, i] = 1.0
return image, coords
[docs]
def share_data(master_files):
"""
Save master calibration arrays to disk as memory-mapped files and return read-only memmap objects.
This function writes each array in `master_files` to a `.array` file using numpy's memmap,
allowing efficient concurrent access from multiple processes. The returned dictionary contains
read-only memmap objects for each calibration file.
Parameters
----------
master_files : dict
Dictionary mapping string keys (e.g., 'bias', 'dark', 'flat') to numpy.ndarray calibration arrays.
Returns
-------
dict
Dictionary mapping the same keys to numpy.memmap objects opened in read-only mode.
"""
get_data = {}
for key, value in master_files.items():
shape = value.shape
dtype = value.dtype
m = np.memmap(f"{key}.array", dtype=dtype, mode="w+", shape=shape)
if value.ndim == 2:
m[:, :] = value[:, :]
else:
m[:] = value[:]
get_data[key] = np.memmap(f"{key}.array", dtype=dtype, mode="r", shape=shape)
del master_files
return get_data
