Skip to content

Commit

Permalink
updated LePhareTableFactory method to be compatible with the nex synt…
Browse files Browse the repository at this point in the history 
…ax. See previous Cigale commit
  • Loading branch information
Wilfried (Mac) committed Nov 23, 2023
1 parent dc22f15 commit 2ed12c5
Showing 1 changed file with 59 additions and 33 deletions.
92 changes: 59 additions & 33 deletions filters.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -454,41 +454,53 @@ def _LePhareTableFactory(self, cleanMethod: CleanMethod = CleanMethod.ZERO, scal
# Compute mean map to scale data
meanMap, _ = self.computeMeanMap(maskVal=0)

dataList = []
stdList = []
data = []
var = []

for pos, filt in enumerate(self.filters):

# Clean and add noise to variance map
data, var = self.cleanAndNoise(filt.data, filt.data2, filt.var, self.mask, cleanMethod=cleanMethod, texp=filt.texp, texpFac=texpFac)

d, v = self.cleanAndNoise(filt.data, filt.data2, filt.var, self.mask,
cleanMethod = cleanMethod,
texp = filt.texp,
texpFac = texpFac,
verbose = filt.verbose
)

# Scale data to have compatible values with LePhare for the flux
data, var = self.scale(data, var, meanMap, factor=scaleFactor)
d, v = self.scale(d, v, meanMap, factor=scaleFactor)

data.append(d)
var.append( v)

# Go to 1D version
data, var, indices = self.arraysTo1D(data, var, indices=True)

# Compute std and convert std and data to mJy unit
dataList = []
stdList = []

mask0 = True
for d, err, filt in zip(data, np.sqrt(var), self.filters):

# Go to 1D version
if pos==0:
data, var, indices = self.arrayTo1D(data, var, indices=True)
else:
data, var = self.arrayTo1D(data, var, indices=False)

# 0 values are cast to NaN otherwise corresponding magnitude would be infinite
mask0 = (np.asarray(data == 0) | np.asarray(var == 0))
data[mask0] = np.nan
var[ mask0] = np.nan
mask0 = mask0 | (np.asarray(data == 0) | np.asarray(var == 0))
d[ mask0] = np.nan
err[mask0] = np.nan

# Compute std instead of variance and go to mag
data, std = countToMag(data, np.sqrt(var), filt.zpt)
# Go to mag
d, err = countToMag(d, err, filt.zpt)

# Cast back pixels with NaN values to -99 mag to specify they are not to be used in the SED fitting
data[mask0] = -99
std[ mask0] = -99
d[ mask0] = -99
err[mask0] = -99

# Append data and std to list
dataList.append(data)
stdList.append( std)
dataList.append(d)
stdList.append( err)

# Redshift column
lf = len(self.filters)
ld = len(data)
ld = len(d)
zs = [self.redshift]*ld

# Compute context (number of filters used - see LePhare documentation) and redshift columns
Expand Down Expand Up @@ -520,7 +532,12 @@ def _CigaleTableFactory(self, cleanMethod: CleanMethod = CleanMethod.ZERO, texpF
for pos, filt in enumerate(self.filters):

# Clean and add noise to variance map
d, v = self.cleanAndNoise(filt.data, filt.data2, filt.var, self.mask, cleanMethod=cleanMethod, texp=filt.texp, texpFac=texpFac)
d, v = self.cleanAndNoise(filt.data, filt.data2, filt.var, self.mask,
cleanMethod = cleanMethod,
texp = filt.texp,
texpFac = texpFac,
verbose = filt.verbose
)
data.append(d)
var.append( v)

Expand All @@ -533,15 +550,15 @@ def _CigaleTableFactory(self, cleanMethod: CleanMethod = CleanMethod.ZERO, texpF

for d, err, filt in zip(data, np.sqrt(var), self.filters):

data, std = countToFlux(d, err, filt.zpt)
d, err = countToFlux(d, err, filt.zpt)

# Append data and std to list
dataList.append(data.to('mJy').value)
stdList.append( std.to( 'mJy').value)
dataList.append(d.to( 'mJy').value)
stdList.append( err.to('mJy').value)

# Redshift column
ll = len(self.filters)
ld = len(data)
ld = len(d)
zs = [self.redshift]*ld

dtypes = [int, float] + [float]*2*ll
Expand Down Expand Up @@ -701,11 +718,13 @@ def clean(data: ndarray, var: ndarray, mask: ndarray, method: CleanMethod = Clea
return data, var

def cleanAndNoise(self, data, data2, var, mask,
cleanMethod: CleanMethod = CleanMethod.ZERO,
texp: Optional[Union[int, float]] = None,
texpFac: int = 0, **kwargs) -> Tuple[ndarray]:
cleanMethod: CleanMethod = CleanMethod.ZERO,
texp : Optional[Union[int, float]] = None,
texpFac : int = 0,
verbose : bool = False,
**kwargs) -> Tuple[ndarray]:
r'''
.. codeauthor:: Wilfried Mercier - IRAP <wilfried.mercier@irap.omp.eu>
.. codeauthor:: Wilfried Mercier - IRAP/LAM <wilfried.mercier@lam.fr>
Clean data and variance maps from bad pixels and add Poisson noise to the variance map.
Expand All @@ -727,11 +746,18 @@ def cleanAndNoise(self, data, data2, var, mask,
:type texp: :python:`int` or :python:`float`
:param texpFac: (**Optional**) factor used to divide the exposition time
:type texpFac: :python:`int`
:param verbose: (**Optional**) whether to print info text or not
:type verbose: :python:`bool`
:returns: cleaned data and cleaned variance map with Poisson noise added
:returns: cleaned data and cleaned variance map (with potentially Poisson noise added)
:rtype: (`ndarray`_, `ndarray`_)
:raises: TypeError if :python:`not isinstance(verbose, bool)`
'''

if not isinstance(verbose, bool):
raise TypeError(f'verbose has type {type(verbose)} but it should be a bool.')

# Clean data and error maps of bad pixels and pixels with negative values
data, var = self.clean(data, var, mask, method=cleanMethod)

Expand Down

0 comments on commit 2ed12c5

Please sign in to comment.