Skip to content

Commit

Permalink
Added image parsers to all classes that take images
Browse files Browse the repository at this point in the history
Quantity arrays are now acceptable inputs. After parsing, the
attribute is now of type  instead of  as before.
  • Loading branch information
ojustino committed Nov 8, 2022
1 parent 945fb1c commit c015d73
Show file tree
Hide file tree
Showing 3 changed files with 207 additions and 84 deletions.
50 changes: 42 additions & 8 deletions specreduce/background.py
Original file line number Diff line number Diff line change
Expand Up @@ -4,8 +4,9 @@
from dataclasses import dataclass, field

import numpy as np
from astropy.nddata import NDData
from astropy.nddata import NDData, VarianceUncertainty
from astropy import units as u
from specutils import Spectrum1D

from specreduce.extract import _ap_weight_image, _to_spectrum1d_pixels
from specreduce.tracing import Trace, FlatTrace
Expand Down Expand Up @@ -54,6 +55,41 @@ class Background:
disp_axis: int = 1
crossdisp_axis: int = 0

def _parse_image(self):
"""
Convert all accepted image types to a consistently formatted Spectrum1D.
"""

if isinstance(self.image, np.ndarray):
img = self.image
elif isinstance(self.image, u.quantity.Quantity):
img = self.image.value
else: # NDData, including CCDData and Spectrum1D
img = self.image.data

# mask and uncertainty are set as None when they aren't specified upon
# creating a Spectrum1D object, so we must check whether these
# attributes are absent *and* whether they are present but set as None
if getattr(self.image, 'mask', None) is not None:
mask = self.image.mask
else:
mask = np.ma.masked_invalid(img).mask

if getattr(self.image, 'uncertainty', None) is not None:
uncertainty = self.image.uncertainty
else:
uncertainty = VarianceUncertainty(np.ones(img.shape))

unit = getattr(self.image, 'unit', u.Unit('DN')) # or u.Unit()?

spectral_axis = getattr(self.image, 'spectral_axis',
(np.arange(img.shape[self.disp_axis])
if hasattr(self, 'disp_axis')
else np.arange(img.shape[1])) * u.pix)

self.image = Spectrum1D(img * unit, spectral_axis=spectral_axis,
uncertainty=uncertainty, mask=mask)

def __post_init__(self):
"""
Determine the background from an image for subtraction.
Expand Down Expand Up @@ -86,12 +122,9 @@ def _to_trace(trace):
raise ValueError('trace_object.trace_pos must be >= 1')
return trace

if isinstance(self.image, NDData):
# NOTE: should the NDData structure instead be preserved?
# (NDData includes Spectrum1D under its umbrella)
self.image = self.image.data
self._parse_image()

bkg_wimage = np.zeros_like(self.image, dtype=np.float64)
bkg_wimage = np.zeros_like(self.image.data, dtype=np.float64)
for trace in self.traces:
trace = _to_trace(trace)
if (np.any(trace.trace.data >= self.image.shape[self.crossdisp_axis]) or
Expand All @@ -116,9 +149,10 @@ def _to_trace(trace):

self.bkg_wimage = bkg_wimage
if self.statistic == 'average':
self.bkg_array = np.average(self.image, weights=self.bkg_wimage, axis=0)
self.bkg_array = np.average(self.image.data,
weights=self.bkg_wimage, axis=0)
elif self.statistic == 'median':
med_image = self.image.copy()
med_image = self.image.data.copy()
med_image[np.where(self.bkg_wimage) == 0] = np.nan
self.bkg_array = np.nanmedian(med_image, axis=0)
else:
Expand Down
190 changes: 125 additions & 65 deletions specreduce/extract.py
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@

from astropy import units as u
from astropy.modeling import Model, models, fitting
from astropy.nddata import NDData
from astropy.nddata import NDData, VarianceUncertainty

from specreduce.core import SpecreduceOperation
from specreduce.tracing import Trace, FlatTrace
Expand Down Expand Up @@ -149,6 +149,41 @@ class BoxcarExtract(SpecreduceOperation):
def spectrum(self):
return self.__call__()

def _parse_image(self):
"""
Convert all accepted image types to a consistently formatted Spectrum1D.
"""

if isinstance(self.image, np.ndarray):
img = self.image
elif isinstance(self.image, u.quantity.Quantity):
img = self.image.value
else: # NDData, including CCDData and Spectrum1D
img = self.image.data

# mask and uncertainty are set as None when they aren't specified upon
# creating a Spectrum1D object, so we must check whether these
# attributes are absent *and* whether they are present but set as None
if getattr(self.image, 'mask', None) is not None:
mask = self.image.mask
else:
mask = np.ma.masked_invalid(img).mask

if getattr(self.image, 'uncertainty', None) is not None:
uncertainty = self.image.uncertainty
else:
uncertainty = VarianceUncertainty(np.ones(img.shape))

unit = getattr(self.image, 'unit', u.Unit('DN')) # or u.Unit()?

spectral_axis = getattr(self.image, 'spectral_axis',
(np.arange(img.shape[self.disp_axis])
if hasattr(self, 'disp_axis')
else np.arange(img.shape[1])) * u.pix)

self.image = Spectrum1D(img * unit, spectral_axis=spectral_axis,
uncertainty=uncertainty, mask=mask)

def __call__(self, image=None, trace_object=None, width=None,
disp_axis=None, crossdisp_axis=None):
"""
Expand Down Expand Up @@ -267,6 +302,86 @@ class HorneExtract(SpecreduceOperation):
def spectrum(self):
return self.__call__()

def _parse_image(self, variance=None, mask=None, unit=None):
"""
Convert all accepted image types to a consistently formatted Spectrum1D.
Takes some extra arguments exactly as they come from self.__call__() to
handle cases where users specify them as arguments instead of as
attributes of their image object.
"""

if isinstance(self.image, np.ndarray):
img = self.image
elif isinstance(self.image, u.quantity.Quantity):
img = self.image.value
else: # NDData, including CCDData and Spectrum1D
img = self.image.data

# mask is set as None when not specified upon creating a Spectrum1D
# object, so we must check whether it is absent *and* whether it's
# present but set as None
if getattr(self.image, 'mask', None) is not None:
mask = self.image.mask
else:
mask = np.ma.masked_invalid(img).mask

# Process uncertainties, converting to variances when able and throwing
# an error when uncertainties are missing or less easily converted
if (hasattr(self.image, 'uncertainty')
and self.image.uncertainty is not None):
if self.image.uncertainty.uncertainty_type == 'var':
variance = self.image.uncertainty.array
elif self.image.uncertainty.uncertainty_type == 'std':
warnings.warn("image NDData object's uncertainty "
"interpreted as standard deviation. if "
"incorrect, use VarianceUncertainty when "
"assigning image object's uncertainty.")
variance = self.image.uncertainty.array**2
elif self.image.uncertainty.uncertainty_type == 'ivar':
variance = 1 / self.image.uncertainty.array
else:
# other options are InverseVariance and UnknownVariance
raise ValueError("image NDData object has unexpected "
"uncertainty type. instead, try "
"VarianceUncertainty or StdDevUncertainty.")
elif (hasattr(self.image, 'uncertainty')
and self.image.uncertainty is None):
# ignore variance arg to focus on updating NDData object
raise ValueError('image NDData object lacks uncertainty')
else:
if variance is None:
raise ValueError("if image is a numpy or Quantity array, a "
"variance must be specified. consider "
"wrapping it into one object by instead "
"passing an NDData image.")
elif self.image.shape != variance.shape:
raise ValueError("image and variance shapes must match")

if np.any(variance < 0):
raise ValueError("variance must be fully positive")
if np.all(variance == 0):
# technically would result in infinities, but since they're all
# zeros, we can override ones to simulate an unweighted case
variance = np.ones_like(variance)
if np.any(variance == 0):
# exclude such elements by editing the input mask
mask[variance == 0] = True
# replace the variances to avoid a divide by zero warning
variance[variance == 0] = np.nan

variance = VarianceUncertainty(variance)

unit = getattr(self.image, 'unit',
u.Unit(self.unit) if self.unit is not None else u.Unit())

spectral_axis = getattr(self.image, 'spectral_axis',
(np.arange(img.shape[self.disp_axis])
if hasattr(self, 'disp_axis')
else np.arange(img.shape[1])) * u.pix)

self.image = Spectrum1D(img * unit, spectral_axis=spectral_axis,
uncertainty=variance, mask=mask)

def __call__(self, image=None, trace_object=None,
disp_axis=None, crossdisp_axis=None,
bkgrd_prof=None,
Expand Down Expand Up @@ -329,71 +444,16 @@ def __call__(self, image=None, trace_object=None,
mask = mask if mask is not None else self.mask
unit = unit if unit is not None else self.unit

# handle image and associated data based on image's type
if isinstance(image, NDData):
# (NDData includes Spectrum1D under its umbrella)
img = np.ma.array(image.data, mask=image.mask)
unit = image.unit if image.unit is not None else u.Unit()

if image.uncertainty is not None:
# prioritize NDData's uncertainty over variance argument
if image.uncertainty.uncertainty_type == 'var':
variance = image.uncertainty.array
elif image.uncertainty.uncertainty_type == 'std':
# NOTE: CCDData defaults uncertainties given as pure arrays
# to std and logs a warning saying so upon object creation.
# should we remind users again here?
warnings.warn("image NDData object's uncertainty "
"interpreted as standard deviation. if "
"incorrect, use VarianceUncertainty when "
"assigning image object's uncertainty.")
variance = image.uncertainty.array**2
elif image.uncertainty.uncertainty_type == 'ivar':
variance = 1 / image.uncertainty.array
else:
# other options are InverseVariance and UnknownVariance
raise ValueError("image NDData object has unexpected "
"uncertainty type. instead, try "
"VarianceUncertainty or StdDevUncertainty.")
else:
# ignore variance arg to focus on updating NDData object
raise ValueError('image NDData object lacks uncertainty')

else:
if variance is None:
raise ValueError('if image is a numpy array, a variance must '
'be specified. consider wrapping it into one '
'object by instead passing an NDData image.')
elif image.shape != variance.shape:
raise ValueError('image and variance shapes must match')

# check optional arguments, filling them in if absent
if mask is None:
mask = np.ma.masked_invalid(image).mask
elif image.shape != mask.shape:
raise ValueError('image and mask shapes must match.')

if isinstance(unit, str):
unit = u.Unit(unit)
else:
unit = unit if unit is not None else u.Unit()

# create image
img = np.ma.array(image, mask=mask)

if np.all(variance == 0):
# technically would result in infinities, but since they're all zeros
# we can just do the unweighted case by overriding with all ones
variance = np.ones_like(variance)
# parse image and replace optional arguments with updated values
self._parse_image(variance, mask, unit)
variance = self.image.uncertainty.array
unit = self.image.unit

if np.any(variance < 0):
raise ValueError("variance must be fully positive")

if np.any(variance == 0):
# exclude these elements by editing the input mask
img.mask[variance == 0] = True
# replace the variances to avoid a divide by zero warning
variance[variance == 0] = np.nan
# mask any previously uncaught invalid values
or_mask = np.logical_or(mask,
np.ma.masked_invalid(self.image.data).mask)
img = np.ma.masked_array(self.image.data, or_mask)
mask = img.mask

# co-add signal in each image column
ncols = img.shape[crossdisp_axis]
Expand Down
51 changes: 40 additions & 11 deletions specreduce/tracing.py
Original file line number Diff line number Diff line change
Expand Up @@ -5,8 +5,9 @@
import warnings

from astropy.modeling import fitting, models
from astropy.nddata import NDData
from astropy.nddata import NDData, VarianceUncertainty
from astropy.stats import gaussian_sigma_to_fwhm
from astropy import units as u
from scipy.interpolate import UnivariateSpline
from specutils import Spectrum1D
import numpy as np
Expand Down Expand Up @@ -39,9 +40,39 @@ def __getitem__(self, i):
return self.trace[i]

def _parse_image(self):
if isinstance(self.image, Spectrum1D):
# NOTE: should the Spectrum1D structure instead be preserved?
self.image = self.image.data
"""
Convert all accepted image types to a consistently formatted Spectrum1D.
"""

if isinstance(self.image, np.ndarray):
img = self.image
elif isinstance(self.image, u.quantity.Quantity):
img = self.image.value
else: # NDData, including CCDData and Spectrum1D
img = self.image.data

# mask and uncertainty are set as None when they aren't specified upon
# creating a Spectrum1D object, so we must check whether these
# attributes are absent *and* whether they are present but set as None
if getattr(self.image, 'mask', None) is not None:
mask = self.image.mask
else:
mask = np.ma.masked_invalid(img).mask

if getattr(self.image, 'uncertainty', None) is not None:
uncertainty = self.image.uncertainty
else:
uncertainty = VarianceUncertainty(np.ones(img.shape))

unit = getattr(self.image, 'unit', u.Unit('DN')) # or u.Unit()?

spectral_axis = getattr(self.image, 'spectral_axis',
(np.arange(img.shape[self._disp_axis])
if hasattr(self, '_disp_axis')
else np.arange(img.shape[1])) * u.pix)

self.image = Spectrum1D(img * unit, spectral_axis=spectral_axis,
uncertainty=uncertainty, mask=mask)

@property
def shape(self):
Expand Down Expand Up @@ -115,7 +146,7 @@ def set_position(self, trace_pos):
Position of the trace
"""
self.trace_pos = trace_pos
self.trace = np.ones_like(self.image[0]) * self.trace_pos
self.trace = np.ones_like(self.image.data[0]) * self.trace_pos
self._bound_trace()


Expand Down Expand Up @@ -212,12 +243,10 @@ class KosmosTrace(Trace):
def __post_init__(self):
super()._parse_image()

# handle multiple image types and mask uncaught invalid values
if isinstance(self.image, NDData):
img = np.ma.masked_invalid(np.ma.masked_array(self.image.data,
mask=self.image.mask))
else:
img = np.ma.masked_invalid(self.image)
# mask any previously uncaught invalid values
or_mask = np.logical_or(self.image.mask,
np.ma.masked_invalid(self.image.data).mask)
img = np.ma.masked_array(self.image.data, or_mask)

# validate arguments
valid_peak_methods = ('gaussian', 'centroid', 'max')
Expand Down

0 comments on commit c015d73

Please sign in to comment.