pyird.image package

Submodules

pyird.image.aptrace module

pyird.image.aptrace.aptrace(dat, search_start_row, num_aperture, ign_ord=[], plot=True)

trace apertures by a polynomial function.

Parameters:

  • dat – flat data

  • search_start_row – starting row number to search apertures

  • num_aperture – number of total apertures to be traced

  • ign_ord – orders to be ignored, using when num_aperture is set to a non-default value

  • plot – show figure of traced apertures

Returns:

parameters of a polynomial to trace apertures

Note

the apertures begin its search at search_row, which is the row number of the detector (wavelength direction), and continue in the direction of increasing numbers until it matches the appropriate number of apertures. You may as well change the value of search_row if the aperture trace is failed.

pyird.image.aptrace.cross_section(dat, nrow, num_aperture)

extract cross section in row direction and search peaks.

Parameters:

  • dat – flat data

  • nrow – row number to be extracted

  • num_aperture – number of total apertures to be traced

Returns:

masked data and peaks at the cross section

pyird.image.aptrace.errfunc(coeff, x, y0, data)

calculate error function.

Parameters:

  • coeff – coefficients

  • x – x-array

  • y0 – y-offset

  • data – traced pixel data to be fitted

Returns:

residuals of data and fitting model

pyird.image.aptrace.fit_ord(x, y0, data)

optimize the fitting by using least-square method.

Parameters:

  • x – x-array

  • y0 – y-offset

  • data – traced pixel data to be fitted

Returns:

best fit parameters

pyird.image.aptrace.fitfunc(x, y0, coeff)

Legendre function to trace.

Parameters:

  • x – x-array

  • y0 – y-offset

  • coeff – Legendre polynomial coefficients

Returns:

Legendre function

pyird.image.aptrace.set_aperture(dat, search_start_row, num_aperture, ign_ord=[], plot=True)

search aperture

Parameters:

  • dat – flat data

  • search_start_row – starting row number to search apertures

  • num_aperture – number of total apertures to be traced

  • ign_ord – orders to be ignored

  • plot – show figure of selected apertures

Returns:

peak position in the cross section at search_row

pyird.image.aptrace.trace_pix(dat, search_row, peakind, npix=2048)

trace apertures

Parameters:

  • dat – flat data of an order

  • search_row – row number used to set aperture

  • peakind – aperture (peak position) in the cross section at search_row

  • npix – number of pixels

Returns:

traced pixel data

pyird.image.bias module

Bias subtraction.

  • originally developed by M. Kuzuhara

pyird.image.bias.bias_subtract(channel_cube)

Bias subtraction for channel cube.

Parameters:

channel_cube – channel cube

Returns:

unbiased channel cube bias for channels

Example

>>> #this is equivalent to image_rmbias=bias_subtract_image(im)
>>> channel_cube=image_to_channel_cube(im)
>>> bias_subtracted_channel_cube, channel_bias=bias_subtract(channel_cube)
>>> image_rmbias=channel_cube_to_image(bias_subtracted_channel_cube)
pyird.image.bias.bias_subtract_image(im)

Bias subtraction for image.

Parameters:

im – image

Returns:

bias-subtracted image

Examples

>>> image_rmbias=bias_subtract_image(im)

pyird.image.channel module

pyird.image.channel.bias_region(channel_cube, margin=4)
pyird.image.channel.channel_cube_to_image(channel_cube, revert=True)

conversion of a channel cube to an image.

Parameters:

  • channel_cube – channel_cube

  • revert – if True reverting odd channels

Returns:

2D image

Return type:

image

pyird.image.channel.eopixel_combine(eop)

inverse operation of eopixel_split.

Parameters:

tensor (eo) –

Returns:

channel cube

pyird.image.channel.eopixel_split(channel_cube)

Revert y direction of odd channels.

Parameters:

cube (channel) –

Returns:

eo tensor (2 [e/o] x Nchannel x xsize x ysize)

pyird.image.channel.image_to_channel_cube(image, revert=True, Nch=32)

conversion of an image to a channel cube :param image: 2D image :param revert: if True reverting odd channels :param Nch: Number of the channels of a detector. For IRD Nch=32

Returns:

channel cube (Nch, ch_pix_num, 2048)

pyird.image.channel.revert_channel_cube(channel_cube)

Revert y direction of odd channels.

Parameters:

cube (channel) –

Returns:

channel cube (odd channel revereted)

pyird.image.hotpix module

pyird.image.hotpix.apply_hotpixel_mask(hotpix_mask, rsd, y0, xmin, xmax, coeff, save_path=None)

correct hotpixel.

Parameters:

  • hotpix_mask – mask made from dark

  • rsd – extracted spectrum to be masked

  • y0 – trace infomation

  • xmin – trace infomation

  • xmax – trace infomation

  • coeff – trace infomation

Returns:

masked and interpolated spectrum

pyird.image.hotpix.hotpix_fits_to_dat(fitsfile, save_path)

Convert .fits to .dat so that it can be used in the RV pipeline.

Parameters:

  • fitsfile – .fits file of hot pixel mask

  • save_path – save path for .dat file

Returns:

hot pixel mask in .dat file {pixels, orders, intensity}

pyird.image.hotpix.identify_hotpix(im, threshold=10.0)

Identification of hotpixels using sep.

Parameters:

  • im – image

  • threshold – threshold of sep

Returns:

hot pixel mask obj: sep object

Return type:

hotpix_mask

Notes

See also issue #121 for a comparison among identify_hotpix, identify_hotpix_sigclip, and another method.

pyird.image.hotpix.identify_hotpix_sigclip(im, sigma=4, maxiters=5, frac=0.005)

Identification of hotpixels using sigma clip.

Parameters:

  • im – image

  • sigma – maximum value of sigma exploration range

  • maxiters – maximum value of iterations

  • frac – mask percentage target value (empirically 0.5%)

Returns:

hot pixel mask

Return type:

hotpix_mask

Notes

See also issue #121 for a comparison among identify_hotpix, identify_hotpix_sigclip, and another method.

pyird.image.mask module

pyird.image.mask.trace(trace_func, y0, xmin, xmax, coeff, mask_shape=None, inst='IRD', width=None)

make mask for trace parameters for multiorder.

Parameters:

  • trace_func – trace function

  • x – x-array

  • y0 – y-offset

  • xmin – xmin

  • xmax – xmax

  • coeff – coefficients

  • mask_shape – (optional) shape of mask, c.f. np.shape(image)

  • inst – IRD or REACH

  • width – list of aperture widths ([width_start,width_end])

Returns:

mask image (same shape as im)

Examples

>>> from pyird.image.trace_function import trace_legendre
>>> mask=trace(im, trace_legendre, y0, xmin, xmax, coeff)
pyird.image.mask.trace_from_iraf_trace_file(pathlist, mask_shape=None)

make mask from the iraf trace file.

Parameters:

  • pathlist – trace files path list

  • mask_shape – (optional) shape of mask, c.f. np.shape(image)

Returns:

mask image (same shape as im)

Note

Currently, Only legendre is supported.

pyird.image.oned_extract module

pyird.image.oned_extract.flatten(im, trace_func, y0, xmin, xmax, coeff, inst='IRD', onepix=False, npix=2048, width=None, force_rotate=False)

make mask for trace parameters for multiorder.

Parameters:

  • im – image

  • trace_func – trace function

  • x – x-array

  • y0 – y-offset

  • xmin – xmin

  • xmax – xmax

  • coeff – coefficients

  • inst – instrument (IRD or REACH)

  • onepix – extract the spectrum pixel by pixel in an aperture

  • npix – number of pixels

  • width – list of aperture widths ([width_start,width_end])

  • force_rotate – forces rotating the detector, when the number of the apertures (nap) is not the standard value (i.e. 21 or 51)

Returns:

raw multiorder spectra and multiorder pixel coordinate if onepix is True, return pandas DataFrame of spectra in each pixel

pyird.image.oned_extract.sum_weighted_apertures(im, df_flatn, y0, xmin, xmax, coeff, width, inst)

taking waighted sum of the counts in aperture pixels (refer to hdsis_ecf for HDS/Subaru data)

Parameters:

  • im – image

  • df_flatn – apnormalized flat spectrum

  • y0 – y-offset

  • xmin – xmin

  • xmax – xmax

  • coeff – coefficients

  • inst – instrument (IRD or REACH)

  • onepix – extract the spectrum pixel by pixel in an aperture

  • npix – number of pixels

  • width – list of aperture widths ([width_start,width_end])

Returns:

1D spectrum

pyird.image.operator module

image operator for imcube.

  • imcube has a shape of [N x Ny x Nx], where N is the number of images.

pyird.image.operator.imcombine(imcube, mode='median')

combine images to a image.

Parameters:

  • imcube – imcube

  • mode – median or mean

Returns:

image

pyird.image.pattern_model module

pyird.image.pattern_model.cal_nct(nctrend_im, margin_npixel=4, Ncor=64, sigma=0.1, xscale=32, yscale=64, cube=False)

Calculate the non-common trend model using Gaussian Process.

Parameters:

  • nctrend_im (ndarray) – Input image for calculating non-common trend.

  • margin_npixel (int) – Margin pixels to be excluded.

  • Ncor (int) – Coarsening factor.

  • sigma (float) – Sigma parameter for GP.

  • xscale (int) – Scale factor for X-axis.

  • yscale (int) – Scale factor for Y-axis.

  • cube (bool) – If True, coarsen with cube structure.

Returns:

Non-common trend model.

Return type:

nctrend_model (ndarray)

pyird.image.pattern_model.median_XY_profile(calim0, rm_nct=True, margin_npixel=4)

a simple readout pattern model. (revised by Y.K., May 2023)

Note:

This function assumes model = cmosub_med X-Y + cmo, where cmos_med=cmo-subtracted median and cmo=channel median offset. When rm_cnt=True option corrects non-common trends of channels using a 2D GP. See #10 (https://github.com/prvjapan/pyird/issues/10).

Args:

calim0: masked image for read pattern calibration rm_nct: remove non-common trends of channel using a GP margin_npixel: # of pixels for detector margin

Returns:

model pattern image

pyird.image.pattern_model.median_profile_xy(cal_eotensor_filtered, channel_median_offset, channel_axis=1)

Calculate the median profile of all even/odd channels while preserving the characteristics of each pixel.

Parameters:

  • cal_eotensor_filtered – numpy array, input tensor.

  • channel_median_offset – numpy array, offset to subtract from each channel.

  • channel_axis – int, axis representing the channels (default is 1).

Returns:

numpy array, input tensor with channel median offset subtracted. xyprofile_offset_subtracted_model: numpy array, median profile model preserving pixel profiles.

Return type:

offset_subtracted

pyird.image.pattern_model.pad_nans_with_xy_median(offset_subtracted, channel_median_offset, image_pattern_model_eotensor, channel_axis=1, x_axis=2, y_axis=3)

Pad NaN values in the input tensor with the median profile calculated from x and y axes.

Parameters:

  • offset_subtracted – numpy array, input tensor with subtracted offset.

  • channel_median_offset – numpy array, offset to add back to the channels.

  • image_pattern_model_eotensor – numpy array, tensor with NaNs to be filled.

  • channel_axis – int, axis representing the channels (default is 1).

  • x_axis – int, axis representing the x-axis (default is 2).

  • y_axis – int, axis representing the y-axis (default is 3).

Returns:

numpy array, tensor with NaNs filled.

Return type:

image_pattern_model_eotensor

pyird.image.trace_function module

pyird.image.trace_function.trace_legendre(x, y0, xmin, xmax, coeff)

trace Legendre function.

Parameters:

  • x – x-array

  • y0 – y-offset

  • xmin – xmin

  • xmax – xmax

  • coeff – Legendre polynomial coefficients

Module contents