Survey
- class ehtim.survey.ParameterSet(paramset, params_fixed={})[source]
- paramset
dict containing single parameter set. Each key in dict becomes its own attribute
- Type:
- params_fixed
dict containing non-varying parameters. Each key in dict becomes its own attribute
- Type:
- obs_sc_init
original observation with systematic noise added, reverse taper applied, and extended flux removed
- Type:
- init_img()[source]
Creates initial/prior image. Only gaussian prior option at present, but creates prior attritubute self.initimg using self.zbl and self.prior_fwhm
Args:
Returns:
- load_data()[source]
Loads in uvfits file from self.infile into eht-imaging obs object and averages using self.avg_time Creates self.obs
Args:
Returns:
- make_img()[source]
Reconstructs image with specified parameters (data weights, amount of self-cal, etc) described in paramset dict Creates attributes self.im_out containing the final image and self.caltab containing a corresponding calibration table object for the final image
Args:
Returns:
- preimcal()[source]
Applies pre-imaging calibration to self.obs. This includes flagging sites with no measurements, rescaling short baselines so only compact flux is being imaged, applying a u-v taper, and adding extra sys noise. Creates self.obs_sc which will have self-cal applied in future steps and self.obs_sc_init which is a copy of the initial preimcal observation
Args:
Returns:
- save_statistics()[source]
Saves a csv file with the following statistics: chi^2 closure phase, logcamp, vis wrt the original observation chi^2 vis wrt to original observation with self-cal to final image applied chi^2 closure phase, logcamp, vis wrt the original observation with sys noise and self-cal applied
Args:
Returns:
- ehtim.survey.create_params_fixed(infile, outfile_base, outpath, ground_truth_img='None', save_imgsums=False, save_uvfits=True, save_pdf=False, save_stats=True, save_caltab=True, nproc=1, backend='multiprocessing', ttype='nfft', overwrite=False, selfcal=True, gaintol=[0.02, 0.2], niter_static=3, blurfrac=1, maxit=100, stop=0.0001, fov=128, npixels=64, reverse_taper_uas=5, uv_zblcut=100000000.0, SEFD_error_budget={'AA': 0.1, 'AX': 0.1, 'GL': 0.1, 'LM': 0.1, 'MG': 0.1, 'MM': 0.1, 'PV': 0.1, 'SW': 0.1})[source]
Create a dict for all non-varying survery parameters
- Parameters:
infile (str) – path to input uvfits observation file
outfile_base (str) – name of base filename for all outputs
outpath (str) – path to directory where all outputs should be stored
ground_truth_img (str) – if applicable, path to ground truth fits file
save_imgsums (bool) – save summary pdf for each image
save_uvfits (bool) – save final self-cal observation to uvfits file
save_pdf (bool) – save pdf of each image
save_stats (bool) – save csv file containing statistics for each image
save_caltab (bool) – save a calibration table for each image
nproc (int) – number of parallel processes
backend (str) – either ‘multiprocessing’ or ‘ray’
ttype (str) – “fast” or “nfft” or “direct”
overwrite (bool) – if True, write over existing files with same names - else, skip parameter set
selfcal (bool) – perform self-calibration steps during imaging
gaintol (array) – tolerance for gains to be under/over unity respectively during self-cal
niter_static (int) – number of iterations for each imaging step
blurfrac (int) – factor to blur initial image between iterations
maxit (int) – maximum number of iterations if image does not converge
stop (float) – convergence criterion for imaging
fov (int) – image field of view in uas
npixels (int) – number of image pixels
reverse_taper_uas (int) – fwhm of gassuain in uas to reverse taper observation
uv_zblcut (float) – maximum uv-distance to which is considered short baseline flux
SEFD_error_budget (dict) – SEFD percentage error for each station
- Returns:
dict containing all non-varying survery parameters
- Return type:
(dict)
- ehtim.survey.create_survey_psets(zbl=[0.6], sys_noise=[0.02], avg_time=['scan'], prior_fwhm=[40], sc_phase=[0], xdw_phase=[10], sc_ap=[0], xdw_ap=[1], amp=[0.2], cphase=[1], logcamp=[1], simple=[1], l1=[1], tv=[1], tv2=[1], flux=[1], epsilon_tv=[1e-10])[source]
Create a dataframe given all survey parameters. Default values will create an example dataframe but these values should be adjusted for each specific observation
- Parameters:
zbl (array) – compact flux value (Jy)
sys_noise (array) – percent addition of systematic noise
avg_time (array) – in seconds or ‘scan’ for scan averaging
prior_fwhm (array) – fwhm of gaussian prior in uas
sc_phase (array) – number of iterations to perform phase-only self-cal
xdw_phase (array) – multiplicative factor for data weights after one round of phase-only self-cal
sc_ap (array) – number of iterations to perform amp+phase self-cal
xdw_ap (array) – multiplicative factor for data weights after one round of amp+phase self-cal
amp (array) – data weight to be placed on amplitudes
cphase (array) – data weight to be placed on closure phases
logcamp (array) – data weight to be placed on log closure amplitudes
simple (array) – regularizer weight for relative entropy, favoring similarity to prior image
l1 (array) – regularizer weight for l1 norm, favoring image sparsity
tv (array) – regularizer weight for total variation, favoring sharp edges
tv2 (array) – regularizer weight for total squared variation, favoring smooth edges
flux (array) – regularizer weight for total flux density, favoring final images with flux close to zbl
epsilon_tv (array) – epsilon value used in definition of total variation - rarely need to change this
- Returns:
pandas DataFrame containing all combination of parameter sets along with index values
- Return type:
(DataFrame)