#
# gemini_python
#
# gempy.gemini
# pixel_functions.py
# ------------------------------------------------------------------------------
# $Id$
# ------------------------------------------------------------------------------
# The gemini_data_calculations module contains functions that calculate values
# from Gemini data, only get_bias_level() needs to be called
import numpy as np
from datetime import date
from gemini_instruments.gmos import lookup
# ------------------------------------------------------------------------------
[docs]
def get_bias_level(adinput=None, estimate=True):
# Temporarily only do this for GMOS data. It would be better if we could do
# checks on oberving band / type (e.g., OPTICAL / IR) and have this
# function call an appropriate function. This is in place due to the call
# to the function from the primitives_qa module - MS 2014-05-14 see
# Trac #683
try: # the header keyword gets deleted when trimming, so it is there?
adinput.hdr[adinput._keyword_for('overscan_section')]
except KeyError: # no, it isn't!
estimate = True # so we have to estimate
__ALLOWED_TYPES__ = ["GMOS"]
if not set(__ALLOWED_TYPES__).issubset(adinput.tags):
msg = "{}.{} only works for {} data".format(__name__,
"get_bias_level",
__ALLOWED_TYPES__)
raise NotImplementedError(msg)
elif estimate:
bias_level = _get_bias_level_estimate(adinput=adinput)
else:
bias_level = _get_bias_level(adinput=adinput)
return bias_level
def _get_bias_level(adinput=None):
"""
Determine the bias level value from the science extensions of the input
AstroData object. The bias level is equal to the median of the overscan
region. A float or list of floats is returned, depdending on whether the
input is a single-extension slice.
Parameters
----------
adinput: AstroDataGmos
The thing we want to know the bias level of
Returns
-------
float/list
the bias level(s)
"""
# Get the overscan section value using the appropriate descriptor
overscans = adinput.overscan_section()
if not isinstance(overscans, list):
overscans = [overscans]
bias_level = []
# Only measure if *all* extensions return an overscan section
if all(ov is not None for ov in overscans):
# The type of CCD determines the number of contaminated columns in the
# overscan region. Get the pretty detector name value using the
# appropriate descriptor.
detector_name = adinput.detector_name(pretty=True)
if detector_name == "EEV":
nbiascontam = 4
elif detector_name == "e2vDD":
nbiascontam = 5
elif detector_name.startswith("Hamamatsu"):
nbiascontam = 4
else:
nbiascontam = 4
# adinput is always iterable (even if _single); make sure overscans is
for ext, overscan_section in zip(adinput, overscans):
# Turn tuple into list to make mutable
osec = list(overscan_section)
# Don't include columns at edges
if osec[0] == 0:
# Overscan region is on the left
osec[1] -= nbiascontam
osec[0] += 1
else:
# Overscan region is on the right
osec[0] += nbiascontam
osec[1] -= 1
# Extract overscan data. In numpy arrays, y indices come first.
overscan_data = ext.data[osec[2]:osec[3], osec[0]:osec[1]]
bias_level.append(np.median(overscan_data))
# Turn single-element list into a value if sent a single-extension slice
if adinput.is_single:
bias_level = bias_level[0]
else:
bias_level = _get_bias_level_estimate(adinput=adinput)
return bias_level
def _get_bias_level_estimate(adinput=None):
"""
Determine an estiamte of the bias level value from GMOS data.
Parameters
----------
adinput: AstroDataGmos
Image/slice we want to determine the bias level of
Returns
-------
float/list
Estimate of the bias level
"""
# We use OVERSCAN if it exists, otherwise use RAWBIAS if it exists,
# otherwise use the LUT value. To avoid slow slicing later, get the LUT
# values now, even if we're not going to need them
ut_date = adinput.ut_date()
if ut_date >= date(2023, 12, 14):
bias_dict = lookup.gmosampsBias
elif ut_date >= date(2017, 2, 24):
bias_dict = lookup.gmosampsBiasBefore20231214
elif ut_date >= date(2015, 8, 26):
bias_dict = lookup.gmosampsBiasBefore20170224
elif ut_date >= date(2006, 8, 31):
bias_dict = lookup.gmosampsBiasBefore20150826
else:
bias_dict = lookup.gmosampsBiasBefore20060831
read_speed_setting = adinput.read_speed_setting()
gain_setting = adinput.gain_setting()
# This may be a list
ampname = adinput.array_name()
# Create appropriate object (list/value) of LUT bias level(s)
if isinstance(ampname, list):
lut_bias = [bias_dict.get((read_speed_setting, gain_setting, a))
for a in ampname]
else:
lut_bias = bias_dict.get((read_speed_setting, gain_setting, ampname))
# Get the overscan value and the raw bias level from the extension headers
overscan_values = adinput.hdr.get('OVERSCAN')
raw_bias_levels = adinput.hdr.get('RAWBIAS')
try:
bias_level = [o if o is not None else
r if r is not None else b
for o,r,b in zip(overscan_values, raw_bias_levels,
lut_bias)]
except TypeError:
bias_level = overscan_values if overscan_values is not None else \
raw_bias_levels if raw_bias_levels is not None else \
lut_bias
return bias_level