GLINT Pipeline’s classes¶

Classes used by the GLINT Data Reduction Software

class glint_classes.ChipProperties(data=None, nbimg=(None, None), transpose=False)¶

Class handling the determination of the properties of the chip.

getRatioCoeff(beam, zeta_coeff)¶

Determines the flux ratio between the different outputs

Parameters:	beam: int, id (1..4) of the intput. zeta_coeff: dic Dictionary in which the flux ratios are stored.
Returns:	zeta_coeff: dic Same dictionary as required in the parameters, with the new entries set by the `beam` parameters.

class glint_classes.File(data=None, nbimg=(None, None), transpose=False)¶

Management of the HDF5 datacube generated by GLINT

Parameters:

data: string (optional): Path to the datacube to load. If None, datacube full of 0 is created, with same dimension as real data (nbimg, 344, 96). In that case, parameter nbimg cannot be None.
nbimg: tup (optional): Load all frames of the datacube at path data from the first to the second-1 element of the tuple. It cannot be (None, None) if mock data is created.
transpose: bol (optional): If True, swappes the 2nd and 3rd axis of the datacube.

binning(arr, binning, axis=0, avg=False)¶

Bin frames together

Parameters:

arr: nd-array: Array containing data to bin
binning: int: Number of frames to bin
axis: int: axis along which the frames are
avg: bol: If True, the method returns the average of the binned frame. Otherwise, return its sum.

Attributes:

Change the attributes

data: ndarray: datacube

cosmeticsFrames(dark, nonoise=False)¶

NOTE FROM 2020-06-26: Deprecated: dark must be a 2d-array full of 0. The cosmetics is performed on the outputs themselves instead of the whole frame. The attributes are meaningless but kept for comptability and further changes.

Remove dark and bias from data frames. Directly acts on data attribute. Compute the variance and standard deviation of the background noise in a signal-free area of the frames.

Parameters:

dark: 2d-array: Average dark
nbimg: tup (optional): Load all frames of the datacube at path data from the first to the second-1 element of the tuple. It cannot be (None, None) if mock data is created.
nonoise: bool: Set to True if data does not have any detector noise (e.g. simulated one). It skips the cosmetics and set estimation of the background variance/std to zero

Attributes:

Change the attributes

data: ndarray: data frame
bg_std: ndarray: Standard deviation of the background of each frame
bg_var: ndarray: Variance of the background of each frame

loadfile(data=None, nbimg=None, transpose=False)¶

Load the datacube when a File-object is created.

Parameters:

data: string (optional): Path to the datacube to load. If None, datacube full of 0 is created, with same dimension as real data (nbimg, 344, 96). In that case, parameter nbimg cannot be None.
nbimg: tup (optional): Load all frames of the datacube at path data from the first to the second-1 element of the tuple. It cannot be (None, None) if mock data is created.
transpose: bol (optional): If True, swappes the 2nd and 3rd axis of the datacube.

Attributes:

Return the attributes

data: ndarray: loaded or created datacube
nbimg: float: number of frames in the data ndarray

class glint_classes.Null(data=None, nbimg=(None, None), transpose=False)¶

Class handling the measurement of the null and photometries from bias-corrected frame.

computeNullDepth()¶

Deprecated. Compute the null depth per spectral channel, per frame, per model, for each output.

Attributes:

Creates the following attributes

nullX: ndarray: Estimated null depth of the Xth null, based on the amplitude attribute.
nullX_err: ndarray: Estimated uncertainty of the estimated Xth null from the amplitude attribute.
null_modelX: ndarray: Estimated null depth of the Xth null, based on the integ_model attribute.
null_modelX_err: ndarray: Estimated uncertainty of the estimated Xth null from the integ_model attribute.
null_windowedX: ndarray: Estimated null depth of the Xth null, based on the integ_windowed attribute.
null_windowedX_err: ndarray: Estimated uncertainty of the estimated Xth null from the integ_windowed attribute.
null_rawX: ndarray: Estimated null depth of the Xth null, based on the raw attribute.
null_raw1_err: ndarray: Estimated uncertainty of the estimated Xth null from the raw attribute.

error_null(null, Iminus, Iplus, Iminus_err, Iplus_err)¶

Deprecated.

Propagate the error on intensities estimations to the null depth. We assume independant and Gaussian distributed values.

Parameters:

null: array, estimated null depths

Iminus: array, intensity from the conventional null output

Iplus: array, intensity from the conventional antinull output

Iminus_err: array, error on the intensity from the conventional antinull output

Iplus_err: array, error on the intensity from the conventional antinull output

Returns:

Error on the null depth

getChannels(channel_pos, sep, spatial_axis, **kwargs)¶

Extract the 16 channels/outputs from frames

Parameters:

channel_pos: list, array-like: Expected position of the arrays
sep: float: Separation (in pixels) between two consecutive channels
spatial_axis: 1d-array: Position-coordinate of each channel
kwargs: optional: Can add the keyword dark associated with the average dark of each channel to perform their cosmetics instead of using the cosmeticsFrames method from class File.

Attributes:

Create the attributes

slices: 4d-darray: Subframes of each channel. Structure as follow: (frame, spectral axis, channel ID, spatial axis)
slices_axes: ndarray: Spatial coordinates of each channel

getIntensities(mode, wl_bounds=(0, <sphinx.ext.autodoc.importer._MockObject object>))¶

Gets the intensity per spectral channel, per frame, per model, for each output.

Parameters:

mode: str,

Select the way the flux is estimated in every outputs:

amplitude uses patterns determined in the script glint_geometric_calibration and a linear least square is performed to get the amplitude of the pattern
model proceeds like amplitude but the integral of the flux is returned
windowed returns a weighted mean as flux of the spectral channel. The weights is the same pattern as the other modes above
raw returns the mean of the flux along the spatial axis over the whole width of the output

wl_bounds:tup, optional

Set the bounds of the bandwidth to keep, in nanometer. Default is to keep all the common spectral channels to all outputs.

Attributes:

pX: ndarray,: Estimated flux in the photometric output X=1..4, from the amplitude attribute.
pX_err: ndarray,: Uncertainty on the estimated flux in the photometric output X
IminusX: ndarray,: Estimated flux in the null output X=1..6, from the amplitude attribute.
IplusX: ndarray,: Estimated flux in the antinull output X=1..6, from the amplitude attribute.
pX_model: ndarray: Estimated flux in the photometric output X=1..4, from the integ_model attribute.
pX_model_err: ndarray: Uncertainty on the estimated flux in the photometric output X
Iminus_modelX: ndarray: Estimated flux in the null output X=1..6, from the integ_model attribute.
Iplus_modelX: ndarray: Estimated flux in the antinull output X=1..6, from the integ_model attribute.
pX_windowed: ndarray: Estimated flux in the photometric output X=1..4, from the integ_windowed attribute.
pX_windowed_err: ndarray: Uncertainty on the estimated flux in the photometric output X
Iminus_windowedX: ndarray: Estimated flux in the null output X=1..6, from the integ_windowed attribute.
Iplus_windowedX: ndarray: Estimated flux in the antinull output X=1..6, from the integ_windowed attribute.
pX_raw: ndarray: Estimated flux in the photometric output X=1..4, from the raw attribute.
pX_raw_err: ndarray: Uncertainty on the estimated flux in the photometric output X
Iminus_rawX: ndarray: Estimated flux in the null output X=1..6, from the raw attribute.
Iplus_rawX: ndarray: Estimated flux in the antinull output X=1..6, from the raw attribute.

getSpectralFlux(spectral_axis, positions, widths, mode_flux, debug=False)¶

Wrapper getting the flux per spectral channel of each output.

Parameters:

spectral_axis: ndarray

Common spectral axis in pixel for every outputs

position: array-like

Positions of each output respect to the column of pixels.

width: array-like

Widths of each output respect to the column of pixels.

mode_flux: string, choose the method to estimate the spectral flux in the outputs among:

amplitude uses patterns determined in the script glint_geometric_calibration and a linear least square is performed to get the amplitude of the pattern
model proceeds like amplitude but the integral of the flux is returned
windowed returns a weighted mean as flux of the spectral channel. The weights is the same pattern as the other modes above
raw returns the mean of the flux along the spatial axis over the whole width of the output. If chosen, other modes are not calculated and positions, widths are not used. Otherwise, the raw mode is used as well as the given mode.

debug: bool

Debug mode. If True, use the python/numpy function _getSpectralFlux which is slow and allow visual check of the well behaviour of the model fitting. Strongly recommended to load only one block and one frame of data and deactivated the save of the final products. If False, use the numba function _getSpectralFluxNumba. For fast and routine use of the measurement of the flux.

Attributes:

Creates the following attributes

raw: ndarray: Estimation of the spectral flux by simply summing slices along spatial axis
raw_err:ndarray, deprecated: Estimation of the uncertainty of the estimation of the raw flux.
amplitude: ndarray: Estimation of the spectral flux as the amplitude of the Gaussian profile fitted by numpy’s linear leastsquare method.
residuals_reg: ndarray: Residuals from the fit which gives amplitude attribute.
amplitude_fit: ndarray: From debug-mode only. Estimation of the spectral flux as the amplitude of the Gaussian profile fitted by the scipy’s curve_fit method, assuming a gaussian profile.
residuals_fit: ndarray: From debug-mode only. Residuals of the fit performed by the scipy’s curve_fit method
cov: scalar: From debug-mode only. Covariance estimated by the scipy’s curve_fit method
amplitude_error: ndarray: Uncertainty of the estimation of amplitude

getTotalFlux()¶: I keep it otherwise some part of code do not work anymore. It is useless though. This method monitores the flux in one spectral channel (column of pixel 56) for the four photometric outputs.

matchSpectralChannels(wl_to_px_coeff, px_to_wl_coeff)¶

All tracks are slightly shifted respect to each other. Need to define the common wavelength to all of them and create a matching map between the spectral channels of every tracks.

Parameters:

wl_to_px_coeff: ndarray: Polynomial coefficients converting wavelength to pixel position along the spectral axis of the frames.
px_to_wl_coeff: ndarray: Polynomial coefficients converting pixel position to wavelength along the spectral axis of the frames.

Attributes:

Creates the following attributes

wl_scale: ndarray: Wavelength scale for each output, in nanometer.
px_scale: ndarray: Wavelength scale for each output, in pixel.

save(path, date)¶

Saves intermediate products for further analyses, into HDF5 file format. The different intensities and null are gathered into dictionaries.

Parameters:

path: str: Path of the file to save. Must contain the name of the file.
date: str: date of the acquisition of the data (YYYY-MM-DD).

Returns:

HDF5 file containing the measured spectral intensities of each output, for each frames, and their uncertainties.

Keywords identifies the nature of the stored data.

Comments into the file contains the following attributes:

date: date of the acquisition of the data;
nbimg: number of frames;
array shape: shape of the data into the data sets.

spectralBinning(wl_min, wl_max, bandwidth, wl_to_px_coeff)¶

Method for keeping or binning a spectral band. It changes the attributes pX, IminusX, IplusX (X=1..6), px_scale and wl_scale of the object.

Parameters:

wl_min: scalar: Lower bound of the bandwidth to keep/bin, in nm
wl_max: scalar: Upper bound of the bandwidth to keep/bin, in nm
bandwidth: scalar or None: Width of the spectrum to bin (in nm), should be lower or equal to the difference between wl_min and wl_max. If it is higher, the whole spectrum is binned. If None, the whole band is binned and the average value is taken.
wl_to_px_coeff: array: Coefficient of conversion from wavelength to pixel position

glint_classes.gaussian(x, A, B, C, loc, sig)¶

Computes a gaussian curve

Parameters:	x: (N,) array Values for which the gaussian is estimated A: float amplitude of the gaussian curve loc: float center of the curve sig: float>0 scale factor of the curve
Returns:	The gaussian curve respect to x values