Obsdata

class ehtim.obsdata.Obsdata(ra, dec, rf, bw, datatable, tarr, scantable=None, polrep='stokes', source='SgrA', mjd=51544, timetype='UTC', ampcal=True, phasecal=True, opacitycal=True, dcal=True, frcal=True, trial_speedups=False)

A polarimetric VLBI observation of visibility amplitudes and phases (in Jy).

source

The source name

Type:

str

ra

The source Right Ascension in fractional hours

Type:

float

dec

The source declination in fractional degrees

Type:

float

mjd

The integer MJD of the observation

Type:

int

tstart

The start time of the observation in hours

Type:

float

tstop

The end time of the observation in hours

Type:

float

rf

The observation frequency in Hz

Type:

float

bw

The observation bandwidth in Hz

Type:

float

timetype

How to interpret tstart and tstop; either ‘GMST’ or ‘UTC’

Type:

str

polrep

polarization representation, either ‘stokes’ or ‘circ’

Type:

str

tarr

The array of telescope data with datatype DTARR

Type:

numpy.recarray

tkey

A dictionary of rows in the tarr for each site name

Type:

dict

data

the basic data with datatype DTPOL_STOKES or DTPOL_CIRC

Type:

numpy.recarray

scantable

The array of scan information

Type:

numpy.recarray

ampcal

True if amplitudes calibrated

Type:

bool

phasecal

True if phases calibrated

Type:

bool

opacitycal

True if time-dependent opacities correctly accounted for in sigmas

Type:

bool

frcal

True if feed rotation calibrated out of visibilities

Type:

bool

dcal

True if D terms calibrated out of visibilities

Type:

bool

amp

An array of (averaged) visibility amplitudes

Type:

numpy.recarray

bispec

An array of (averaged) bispectra

Type:

numpy.recarray

cphase

An array of (averaged) closure phases

Type:

numpy.recarray

cphase_diag

An array of (averaged) diagonalized closure phases

Type:

numpy.recarray

camp

An array of (averaged) closure amplitudes

Type:

numpy.recarray

logcamp

An array of (averaged) log closure amplitudes

Type:

numpy.recarray

logcamp_diag

An array of (averaged) diagonalized log closure amps

Type:

numpy.recarray

add_all(avg_time=0, return_type='rec', count='max', debias=True, snrcut=0.0, err_type='predicted', num_samples=1000, round_s=0.1)

Adds tables of all all averaged derived quantities self.amp,self.bispec,self.cphase,self.camp,self.logcamp

Parameters:

• avg_time (float) – closure amplitude averaging timescale

• return_type – data frame (‘df’) or recarray (‘rec’)

• debias (bool) – If True, debias the closure amplitude

• count (str) – If ‘min’, return minimal set of closure quantities, if ‘max’ return all closure quantities

• err_type (str) – ‘predicted’ or ‘measured’

• num_samples – number of bootstrap (re)samples if measuring error

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag closure amplitudes with snr lower than this

add_amp(avg_time=0, scan_avg=False, debias=True, err_type='predicted', return_type='rec', round_s=0.1, snrcut=0.0)

Adds attribute self.amp: aan amplitude table with incoherently averaged amplitudes

Parameters:

• avg_time (float) – incoherent integration time in seconds

• scan_avg (bool) – if True, average over scans in self.scans instead of intime

• debias (bool) – if True then apply debiasing

• err_type (str) – ‘predicted’ or ‘measured’

• return_type – data frame (‘df’) or recarray (‘rec’)

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag amplitudes with snr lower than this

add_bispec(avg_time=0, return_type='rec', count='max', snrcut=0.0, err_type='predicted', num_samples=1000, round_s=0.1, uv_min=False)

Adds attribute self.bispec: bispectra table with bispectra averaged for dt

Parameters:

• avg_time (float) – bispectrum averaging timescale

• return_type – data frame (‘df’) or recarray (‘rec’)

• count (str) – If ‘min’, return minimal set of bispectra, if ‘max’ return all bispectra up to reordering

• err_type (str) – ‘predicted’ or ‘measured’

• num_samples – number of bootstrap (re)samples if measuring error

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag bispectra with snr lower than this

add_camp(avg_time=0, return_type='rec', ctype='camp', count='max', debias=True, snrcut=0.0, err_type='predicted', num_samples=1000, round_s=0.1)

Adds attribute self.camp or self.logcamp: closure amplitudes table

Parameters:

• avg_time (float) – closure amplitude averaging timescale

• return_type – data frame (‘df’) or recarray (‘rec’)

• ctype (str) – The closure amplitude type (‘camp’ or ‘logcamp’)

• debias (bool) – If True, debias the closure amplitude

• count (str) – If ‘min’, return minimal set of amplitudes, if ‘max’ return all closure amplitudes up to inverses

• err_type (str) – ‘predicted’ or ‘measured’

• num_samples – number of bootstrap (re)samples if measuring error

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag closure amplitudes with snr lower than this

add_cphase(avg_time=0, return_type='rec', count='max', snrcut=0.0, err_type='predicted', num_samples=1000, round_s=0.1, uv_min=False)

Adds attribute self.cphase: cphase table averaged for dt

Parameters:

• avg_time (float) – closure phase averaging timescale

• return_type – data frame (‘df’) or recarray (‘rec’)

• count (str) – If ‘min’, return minimal set of phases, if ‘max’ return all closure phases up to reordering

• err_type (str) – ‘predicted’ or ‘measured’

• num_samples – number of bootstrap (re)samples if measuring error

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag closure phases with snr lower than this

add_cphase_diag(avg_time=0, return_type='rec', vtype='vis', count='min', snrcut=0.0, err_type='predicted', num_samples=1000, round_s=0.1, uv_min=False)

Adds attribute self.cphase_diag: cphase_diag table averaged for dt

Parameters:

• avg_time (float) – closure phase averaging timescale

• return_type – data frame (‘df’) or recarray (‘rec’)

• vtype (str) – Visibility type (e.g., ‘vis’, ‘llvis’, ‘rrvis’, etc.)

• count (str) – If ‘min’, return minimal set of phases, If ‘max’ return all closure phases up to reordering

• err_type (str) – ‘predicted’ or ‘measured’

• num_samples – number of bootstrap (re)samples if measuring error

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag closure phases with snr lower than this

add_fractional_noise(frac_noise, debias=False)

Add a constant fraction of amplitude at quadrature to thermal noise. Effectively imposes a maximal signal-to-noise ratio. !! this operation is not currently tracked and should be applied with extreme caution!!

Parameters:

• frac_noise (float) – The fraction of noise to add.

• debias (bool) – Whether or not to add frac_noise of debiased amplitudes.

Returns:

An Obsdata object with the inflated noise values.

Return type:

(Obsdata)

add_leakage_noise(Dterm_amp=0.1, min_noise=0.01, debias=False)

Add estimated systematic noise from leakage at quadrature to thermal noise. Requires cross-hand visibilities. !! this operation is not currently tracked and should be applied with extreme caution!!

Parameters:

• Dterm_amp (float) – Estimated magnitude of leakage terms

• min_noise (float) – Minimum fractional systematic noise to add

• debias (bool) – Debias amplitudes before computing fractional noise

Returns:

An Obsdata object with the inflated noise values.

Return type:

(Obsdata)

add_logcamp(avg_time=0, return_type='rec', ctype='camp', count='max', debias=True, snrcut=0.0, err_type='predicted', num_samples=1000, round_s=0.1)

Adds attribute self.logcamp: closure amplitudes table

Parameters:

• avg_time (float) – closure amplitude averaging timescale

• return_type – data frame (‘df’) or recarray (‘rec’)

• ctype (str) – The closure amplitude type (‘camp’ or ‘logcamp’)

• debias (bool) – If True, debias the closure amplitude

• count (str) – If ‘min’, return minimal set of amplitudes, if ‘max’ return all closure amplitudes up to inverses

• err_type (str) – ‘predicted’ or ‘measured’

• num_samples – number of bootstrap (re)samples if measuring error

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag closure amplitudes with snr lower than this

add_logcamp_diag(avg_time=0, return_type='rec', count='min', snrcut=0.0, debias=True, err_type='predicted', num_samples=1000, round_s=0.1)

Adds attribute self.logcamp_diag: logcamp_diag table averaged for dt

Parameters:

• avg_time (float) – diagonal log closure amplitude averaging timescale

• return_type – data frame (‘df’) or recarray (‘rec’)

• debias (bool) – If True, debias the diagonal log closure amplitude

• count (str) – If ‘min’, return minimal set of amplitudes, If ‘max’ return all diagonal log closure amplitudes up to inverses

• err_type (str) – ‘predicted’ or ‘measured’

• num_samples – number of bootstrap (re)samples if measuring error

• round_s (float) – accuracy of datetime object in seconds

• snrcut (float) – flag diagonal log closure amplitudes with snr lower than this

add_scans(info='self', filepath='', dt=0.0165, margin=0.0001, split_subarray=False)

Compute scans and add self.scans to Obsdata object.

Parameters:

• info (str) – ‘self’ to infer from data, ‘txt’ for text file, ‘vex’ for vex schedule file

• filepath (str) – path to txt/vex file with scans info

• dt (float) – minimal time interval between scans in hours

• margin (float) – padding scans by that time margin in hours

avg_coherent(inttime, scan_avg=False, moving=False)

Coherently average data along u,v tracks in chunks of length inttime (sec)

Parameters:

• inttime (float) – coherent integration time in seconds

• scan_avg (bool) – if True, average over scans in self.scans instead of intime

• moving (bool) – averaging with moving window (boxcar width in seconds)

Returns:

Obsdata object containing averaged data

Return type:

(Obsdata)

avg_incoherent(inttime, scan_avg=False, debias=True, err_type='predicted')

Incoherently average data along u,v tracks in chunks of length inttime (sec)

Parameters:

• inttime (float) – incoherent integration time in seconds

• scan_avg (bool) – if True, average over scans in self.scans instead of intime

• debias (bool) – if True, debias the averaged amplitudes

• err_type (str) – ‘predicted’ or ‘measured’

Returns:

Obsdata object containing averaged data

Return type:

(Obsdata)

bispectra(vtype='vis', mode='all', count='min', timetype=False, uv_min=False, snrcut=0.0)

Return a recarray of the equal time bispectra.

Parameters:

• vtype (str) – The visibilty type from which to assemble bispectra (‘vis’, ‘qvis’, ‘uvis’,’vvis’,’rrvis’,’lrvis’,’rlvis’,’llvis’)

• mode (str) – If ‘time’, return phases in a list of equal time arrays, if ‘all’, return all phases in a single array

• count (str) – If ‘min’, return minimal set of bispectra, if ‘max’ return all bispectra up to reordering

• timetype (str) – ‘GMST’ or ‘UTC’

• uv_min (float) – flag baselines shorter than this before forming closure quantities

• snrcut (float) – flag bispectra with snr lower than this

Returns:

A recarray of the bispectra values with datatype DTBIS

Return type:

(numpy.recarry)

bispectra_tri(site1, site2, site3, vtype='vis', timetype=False, snrcut=0.0, method='from_maxset', bs=[], force_recompute=False)

Return complex bispectrum over time on a triangle (1-2-3).

Parameters:

• site1 (str) – station 1 name

• site2 (str) – station 2 name

• site3 (str) – station 3 name

• vtype (str) – The visibilty type from which to assemble bispectra (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’)

• timetype (str) – ‘UTC’ or ‘GMST’

• snrcut (float) – flag bispectra with snr lower than this

• method (str) – ‘from_maxset’ (old, default), ‘from_vis’ (new, more robust)

• bs (list) – optionally pass in the precomputed, time-sorted bispectra

• force_recompute (bool) – if True, recompute bispectra instead of using saved data

Returns:

A recarray of the bispectra on this triangle with datatype DTBIS

Return type:

(numpy.recarry)

bllist(conj=False)

Group the data in a list of same baseline datatables.

Parameters:

conj (bool) – True if tlist_out includes conjugate baselines.

Returns:

a list of data tables containing baseline-partitioned data

Return type:

(list)

c_amplitudes(vtype='vis', mode='all', count='min', ctype='camp', debias=True, timetype=False, snrcut=0.0)

Return a recarray of the equal time closure amplitudes.

Parameters:

• vtype (str) – The visibilty type from which to assemble closure amplitudes (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’)

• ctype (str) – The closure amplitude type (‘camp’ or ‘logcamp’)

• mode (str) – If ‘time’, return amplitudes in a list of equal time arrays, if ‘all’, return all amplitudes in a single array

• count (str) – If ‘min’, return minimal set of amplitudes, if ‘max’ return all closure amplitudes up to inverses

• debias (bool) – If True, debias the closure amplitude

• timetype (str) – ‘GMST’ or ‘UTC’

• snrcut (float) – flag closure amplitudes with snr lower than this

Returns:

A recarray of the closure amplitudes with datatype DTCAMP

Return type:

(numpy.recarry)

c_log_amplitudes_diag(vtype='vis', mode='all', count='min', debias=True, timetype=False, snrcut=0.0)

Return a recarray of the equal time diagonalized log closure amplitudes.

Parameters:

• vtype (str) – The visibilty type (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’) From which to assemble closure amplitudes

• ctype (str) – The closure amplitude type (‘camp’ or ‘logcamp’)

• mode (str) – If ‘time’, return amplitudes in a list of equal time arrays, If ‘all’, return all amplitudes in a single array

• count (str) – If ‘min’, return minimal set of amplitudes, If ‘max’ return all closure amplitudes up to inverses

• debias (bool) – If True, debias the closure amplitude - The individual visibility amplitudes are always debiased.

• timetype (str) – ‘GMST’ or ‘UTC’

• snrcut (float) – flag closure amplitudes with snr lower than this

Returns:

A recarray of diagonalized closure amps with datatype DTLOGCAMPDIAG

Return type:

(numpy.recarry)

c_phases(vtype='vis', mode='all', count='min', ang_unit='deg', timetype=False, uv_min=False, snrcut=0.0)

Return a recarray of the equal time closure phases.

Parameters:

• vtype (str) – The visibilty type from which to assemble closure phases (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’)

• mode (str) – If ‘time’, return phases in a list of equal time arrays, if ‘all’, return all phases in a single array

• count (str) – If ‘min’, return minimal set of phases, if ‘max’ return all closure phases up to reordering

• ang_unit (str) – If ‘deg’, return closure phases in degrees, else return in radians

• timetype (str) – ‘UTC’ or ‘GMST’

• uv_min (float) – flag baselines shorter than this before forming closure quantities

• snrcut (float) – flag bispectra with snr lower than this

Returns:

A recarray of the closure phases with datatype DTCPHASE

Return type:

(numpy.recarry)

c_phases_diag(vtype='vis', count='min', ang_unit='deg', timetype=False, uv_min=False, snrcut=0.0)

Return a recarray of the equal time diagonalized closure phases.

Parameters:

• vtype (str) – The visibilty type (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’) from which to assemble closure phases

• count (str) – If ‘min’, return minimal set of phases, If ‘min-cut0bl’ return minimal set after flagging zero-baselines

• ang_unit (str) – If ‘deg’, return closure phases in degrees, else return in radians

• timetype (str) – ‘UTC’ or ‘GMST’

• uv_min (float) – flag baselines shorter than this before forming closure quantities

• snrcut (float) – flag bispectra with snr lower than this

Returns:

A recarray of diagonalized closure phases (datatype DTCPHASEDIAG),

along with associated triangles and transformation matrices

Return type:

(numpy.recarry)

camp_quad(site1, site2, site3, site4, vtype='vis', ctype='camp', debias=True, timetype=False, snrcut=0.0, method='from_maxset', camps=[], force_recompute=False)

Return closure phase over time on a quadrange (1-2)(3-4)/(1-4)(2-3).

Parameters:

• site1 (str) – station 1 name

• site2 (str) – station 2 name

• site3 (str) – station 3 name

• site4 (str) – station 4 name

• vtype (str) – The visibilty type from which to assemble closure amplitudes (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’)

• ctype (str) – The closure amplitude type (‘camp’ or ‘logcamp’)

• debias (bool) – If True, debias the closure amplitude

• timetype (str) – ‘UTC’ or ‘GMST’

• snrcut (float) – flag closure amplitudes with snr lower than this

• method (str) – ‘from_maxset’ (old, default), ‘from_vis’ (new, more robust)

• camps (list) – optionally pass in the time-sorted, precomputed camps

• force_recompute (bool) – if True, do not use save closure amplitude data

Returns:

A recarray of the closure amplitudes with datatype DTCAMP

Return type:

(numpy.recarry)

chisq(im_or_mov, dtype='vis', pol='I', ttype='nfft', mask=[], **kwargs)

Give the reduced chi^2 of the observation for the specified image and datatype.

Parameters:

• im_or_mov (Image or Movie) – image or movie object on which to test chi^2

• dtype (str) – data type of chi^2 (e.g., ‘vis’, ‘amp’, ‘bs’, ‘cphase’)

• pol (str) – polarization type (‘I’, ‘Q’, ‘U’, ‘V’, ‘LL’, ‘RR’, ‘LR’, or ‘RL’

• mask (arr) – mask of same dimension as im.imvec

• ttype (str) – “fast” or “nfft” or “direct”

• fft_pad_factor (float) – zero pad the image to (fft_pad_factor * image size) in FFT

• conv_func ('str') – The convolving function for gridding; ‘gaussian’, ‘pill’,’cubic’

• p_rad (int) – The pixel radius for the convolving function

• order ('str') – Interpolation order for sampling the FFT

• systematic_noise (float) – adds a fractional systematic noise tolerance to sigmas

• snrcut (float) – a snr cutoff for including data in the chi^2 sum

• debias (bool) – if True then apply debiasing to amplitudes/closure amplitudes

• weighting (str) – ‘natural’ or ‘uniform’

• systematic_cphase_noise (float) – a value in degrees to add to closure phase sigmas

• cp_uv_min (float) – flag short baselines before forming closure quantities

• maxset (bool) – if True, use maximal set instead of minimal for closure quantities

Returns:

image chi^2

Return type:

(float)

cleanbeam(npix, fov, pulse=<function trianglePulse2D>)

Make an image of the observation clean beam.

Parameters:

• npix (int) – The pixel size of the square output image.

• fov (float) – The field of view of the square output image in radians.

• pulse (function) – The function convolved with the pixel values for continuous image.

Returns:

an Image object with the clean beam.

Return type:

(Image)

copy()

Copy the observation object.

Args:

Returns:

a copy of the Obsdata object.

Return type:

(Obsdata)

cphase_tri(site1, site2, site3, vtype='vis', ang_unit='deg', timetype=False, snrcut=0.0, method='from_maxset', cphases=[], force_recompute=False)

Return closure phase over time on a triangle (1-2-3).

Parameters:

• site1 (str) – station 1 name

• site2 (str) – station 2 name

• site3 (str) – station 3 name

• vtype (str) – The visibilty type from which to assemble closure phases (e.g., ‘vis’,’qvis’,’uvis’,’vvis’,’pvis’)

• ang_unit (str) – If ‘deg’, return closure phases in degrees, else return in radians

• timetype (str) – ‘GMST’ or ‘UTC’

• snrcut (float) – flag bispectra with snr lower than this

• method (str) – ‘from_maxset’ (old, default), ‘from_vis’ (new, more robust)

• cphases (list) – optionally pass in the precomputed time-sorted cphases

• force_recompute (bool) – if True, do not use save closure phase tables

Returns:

A recarray of the closure phases with datatype DTCPHASE

Return type:

(numpy.recarry)

data_conj()

Make a data array including all conjugate baselines.

Args:

Returns:

a copy of the Obsdata.data table including all conjugate baselines.

Return type:

(numpy.recarray)

deblur()

Deblur the observation obs by dividing by the Sgr A* scattering kernel.

Args:

Returns:

a new deblurred observation object.

Return type:

(Obsdata)

dirtybeam(npix, fov, pulse=<function trianglePulse2D>, weighting='uniform')

Make an image of the observation dirty beam.

Parameters:

• npix (int) – The pixel size of the square output image.

• fov (float) – The field of view of the square output image in radians.

• pulse (function) – The function convolved with the pixel values for continuous image.

• weighting (str) – ‘uniform’ or ‘natural’

Returns:

an Image object with the dirty beam.

Return type:

(Image)

dirtyimage(npix, fov, pulse=<function trianglePulse2D>, weighting='uniform')

Make the observation dirty image (direct Fourier transform).

Parameters:

• npix (int) – The pixel size of the square output image.

• fov (float) – The field of view of the square output image in radians.

• pulse (function) – The function convolved with the pixel values for continuous image.

• weighting (str) – ‘uniform’ or ‘natural’

Returns:

an Image object with dirty image.

Return type:

(Image)

estimate_noise_rescale_factor(max_diff_sec=0.0, min_num=10, median_snr_cut=0, count='max', vtype='vis', print_std=False)

Estimate a constant noise rescaling factor on all baselines, times, and polarizations. Uses pairwise differences of closure phases relative to the expected scatter. This is useful for AIPS data, which has a missing factor relating ‘weights’ to noise.

Parameters:

• max_diff_sec (float) – The maximum difference of adjacent closure phases (in seconds) If 0, auto-estimates to twice the median scan length.

• min_num (int) – The minimum number of closure phase differences for a triangle to be included in the set of estimators.

• median_snr_cut (float) – Do not include a triangle if its median SNR is below this

• count (str) – If ‘min’, use minimal set of phases, if ‘max’ use all closure phases up to reordering

• vtype (str) – Visibility type (e.g., ‘vis’, ‘llvis’, ‘rrvis’, etc.)

• print_std (bool) – Whether or not to print the std dev. for each closure triangle.

Returns:

The rescaling factor.

Return type:

(float)

filter_subscan_dropouts(perc=0, return_type='rec')

Filtration to drop data and ensure that we only average parts with same timestamp. Potentially this could reduce risk of non-closing errors.

Parameters:

• perc (float) – drop baseline from scan if it has less than this fraction of median baseline observation time during the scan

• return_type (str) – data frame (‘df’) or recarray (‘rec’)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

find_amt_fractional_noise(im, dtype='vis', target=1.0, debias=False, maxiter=200, ftol=1e-20, gtol=1e-20)

Returns the amount of fractional sys error necessary to make the image have a chisq close to the targeted value (1.0)

fit_beam(weighting='uniform', units='rad')

Fit a Gaussian to the dirty beam and return the parameters (fwhm_maj, fwhm_min, theta).

Parameters:

• weighting (str) – ‘uniform’ or ‘natural’.

• units (string) – ‘rad’ returns values in radians, ‘natural’ returns FWHMs in uas and theta in degrees

Returns:

a tuple (fwhm_maj, fwhm_min, theta) of the dirty beam parameters in radians.

Return type:

(tuple)

fit_gauss(flux=1.0, fittype='amp', paramguess=(4.848136811094136e-10, 4.848136811094136e-10, 0.0))

Fit a gaussian to either Stokes I complex visibilities or Stokes I visibility amplitudes.

Parameters:

• flux (float) – total flux in the fitted gaussian

• fitttype (str) – “amp” to fit to visibilty amplitudes

• paramguess (tuble) – initial guess of fit Gaussian (fwhm_maj, fwhm_min, theta)

Returns:

(fwhm_maj, fwhm_min, theta) of the fit Gaussian parameters in radians.

Return type:

(tuple)

flag_UT_range(UT_start_hour=0.0, UT_stop_hour=0.0, flag_type='all', flag_what='', output='kept')

Flag data points within a certain UT range

Parameters:

• UT_start_hour (float) – start of time window

• UT_stop_hour (float) – end of time window

• flag_type (str) – ‘all’, ‘baseline’, or ‘station’

• flag_what (str) – baseline or station to flag

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_anomalous(field='snr', max_diff_seconds=100, robust_nsigma_cut=5, output='kept')

Flag anomalous data points

Parameters:

• field (str) – The quantity to test for

• max_diff_seconds (float) – The moving window size for testing outliers

• robust_nsigma_cut (float) – Outliers further than this from the mean are removed

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_bl(sites, output='kept')

Flag data points that include the specified baseline

Parameters:

• sites (list) – baseline to remove from the data

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_elev(elev_min=0.0, elev_max=90, output='kept')

Flag visibilities for which either station is outside a stated elevation range

Parameters:

• elev_min (float) – Minimum elevation (deg)

• elev_max (float) – Maximum elevation (deg)

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_high_sigma(sigma_cut=0.005, sigma_type='sigma', output='kept')

Flag high sigma (thermal noise on Stoke I) data points

Parameters:

• sigma_cut (float) – remove points with sigma higher than this

• sigma_type (str) – sigma type (sigma, rrsigma, llsigma, etc.)

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_large_fractional_pol(max_fractional_pol=1.0, output='kept')

Flag visibilities for which the fractional polarization is above a specified threshold

Parameters:

• max_fractional_pol (float) – Maximum fractional polarization

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_low_snr(snr_cut=3, output='kept')

Flag low snr data points

Parameters:

• snr_cut (float) – remove points with snr lower than this

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_sites(sites, output='kept')

Flag data points that include the specified sites

Parameters:

• sites (list) – list of sites to remove from the data

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flag_uvdist(uv_min=0.0, uv_max=1000000000000.0, output='kept')

Flag data points outside a given uv range

Parameters:

• uv_min (float) – remove points with uvdist less than this

• uv_max (float) – remove points with uvdist greater than this

• output (str) – returns ‘kept’, ‘flagged’, or ‘both’ (a dictionary)

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

flags_from_file(flagfile, flag_type='station')

Flagging data based on csv file

Parameters:

flagfile – path to csv file with mjds of flagging start / stop time, and optionally baseline / stations

Returns:

a observation object with flagged data points removed

Return type:

(Obsdata)

getClosestScan(time, splitObs=None)

Split observation by scan and grab scan closest to timestamp

Parameters:

• time (float) – Time (GMST) you want to find the scan closest to

• splitObs – a list of Obsdata objects, output from split_obs, to save time

make_hdulist(force_singlepol=False, polrep_out='circ')

Returns an hdulist in the same format as in a saved .uvfits file. :param force_singlepol: if ‘R’ or ‘L’, will interpret stokes I field as ‘RR’ or ‘LL’ :type force_singlepol: str :param polrep_out: ‘circ’ or ‘stokes’: how data should be stored in the uvfits file :type polrep_out: str

Returns:

hdulist (astropy.io.fits.HDUList)

obsdata_args()

“Copy arguments for making a new Obsdata into a list and dictonary

plot_bl(site1, site2, field, debias=False, ang_unit='deg', timetype=False, axis=False, rangex=False, rangey=False, snrcut=0.0, color=(0.11764705882352941, 0.5647058823529412, 1.0), marker='o', markersize=3, label=None, grid=True, ebar=True, axislabels=True, legend=False, show=True, export_pdf='')

Plot a field over time on a baseline site1-site2.

Parameters:

• site1 (str) – station 1 name

• site2 (str) – station 2 name

• field (str) – y-axis field (from FIELDS)

• debias (bool) – If True, debias amplitudes.

• ang_unit (str) – phase unit ‘deg’ or ‘rad’

• timetype (str) – ‘GMST’ or ‘UTC’

• axis (matplotlib.axes.Axes) – add plot to this axis

• rangex (list) – [xmin, xmax] x-axis limits

• rangey (list) – [ymin, ymax] y-axis limits

• color (str) – color for scatterplot points

• marker (str) – matplotlib plot marker

• markersize (int) – size of plot markers

• label (str) – plot legend label

• grid (bool) – Plot gridlines if True

• ebar (bool) – Plot error bars if True

• axislabels (bool) – Show axis labels if True

• legend (bool) – Show legend if True

• show (bool) – Display the plot if true

• export_pdf (str) – path to pdf file to save figure

Returns:

Axes object with data plot

Return type:

(matplotlib.axes.Axes)

plot_camp(site1, site2, site3, site4, vtype='vis', ctype='camp', camps=[], force_recompute=False, debias=False, timetype=False, snrcut=0.0, axis=False, rangex=False, rangey=False, color=(0.11764705882352941, 0.5647058823529412, 1.0), marker='o', markersize=3, label=None, grid=True, ebar=True, axislabels=True, legend=False, show=True, export_pdf='')

Plot closure amplitude over time on a quadrangle (1-2)(3-4)/(1-4)(2-3).

Parameters:

• site1 (str) – station 1 name

• site2 (str) – station 2 name

• site3 (str) – station 3 name

• site4 (str) – station 4 name

• vtype (str) – The visibilty type from which to assemble closure amplitudes (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’)

• ctype (str) – The closure amplitude type (‘camp’ or ‘logcamp’)

• camps (list) – optionally pass in camps so they don’t have to be recomputed

• force_recompute (bool) – if True, recompute camps instead of using stored data

• snrcut (float) – flag closure amplitudes with snr lower than this

• debias (bool) – If True, debias the closure amplitude

• timetype (str) – ‘GMST’ or ‘UTC’

• axis (matplotlib.axes.Axes) – amake_cdd plot to this axis

• rangex (list) – [xmin, xmax] x-axis limits

• rangey (list) – [ymin, ymax] y-axis limits

• color (str) – color for scatterplot points

• marker (str) – matplotlib plot marker

• markersize (int) – size of plot markers

• label (str) – plot legend label

• grid (bool) – Plot gridlines if True

• ebar (bool) – Plot error bars if True

• axislabels (bool) – Show axis labels if True

• legend (bool) – Show legend if True

• show (bool) – Display the plot if True

• export_pdf (str) – path to pdf file to save figure

Returns:

Axes object with data plot

Return type:

(matplotlib.axes.Axes)

plot_cphase(site1, site2, site3, vtype='vis', cphases=[], force_recompute=False, ang_unit='deg', timetype=False, snrcut=0.0, axis=False, rangex=False, rangey=False, color=(0.11764705882352941, 0.5647058823529412, 1.0), marker='o', markersize=3, label=None, grid=True, ebar=True, axislabels=True, legend=False, show=True, export_pdf='')

Plot a closure phase over time on a triangle site1-site2-site3.

Parameters:

• site1 (str) – station 1 name

• site2 (str) – station 2 name

• site3 (str) – station 3 name

• vtype (str) – The visibilty type from which to assemble closure phases (‘vis’,’qvis’,’uvis’,’vvis’,’pvis’)

• cphases (list) – optionally pass in the prcomputed, time-sorted closure phases

• force_recompute (bool) – if True, do not use stored closure phase able

• snrcut (float) – flag closure amplitudes with snr lower than this

• ang_unit (str) – phase unit ‘deg’ or ‘rad’

• timetype (str) – ‘GMST’ or ‘UTC’

• axis (matplotlib.axes.Axes) – add plot to this axis

• rangex (list) – [xmin, xmax] x-axis limits

• rangey (list) – [ymin, ymax] y-axis limits

• color (str) – color for scatterplot points

• marker (str) – matplotlib plot marker

• markersize (int) – size of plot markers

• label (str) – plot legend label

• grid (bool) – Plot gridlines if True

• ebar (bool) – Plot error bars if True

• axislabels (bool) – Show axis labels if True

• legend (bool) – Show legend if True

• show (bool) – Display the plot if True

• export_pdf (str) – path to pdf file to save figure

Returns:

Axes object with data plot

Return type:

(matplotlib.axes.Axes)

plotall(field1, field2, conj=False, debias=False, tag_bl=False, ang_unit='deg', timetype=False, axis=False, rangex=False, rangey=False, xscale='linear', yscale='linear', color=(0.11764705882352941, 0.5647058823529412, 1.0), marker='o', markersize=3, label=None, snrcut=0.0, grid=True, ebar=True, axislabels=True, legend=False, show=True, export_pdf='')

Plot two fields against each other.

Parameters:

• field1 (str) – x-axis field (from FIELDS)

• field2 (str) – y-axis field (from FIELDS)

• conj (bool) – Plot conjuage baseline data points if True

• debias (bool) – If True, debias amplitudes.

• tag_bl (bool) – if True, label each baseline

• ang_unit (str) – phase unit ‘deg’ or ‘rad’

• timetype (str) – ‘GMST’ or ‘UTC’

• axis (matplotlib.axes.Axes) – add plot to this axis

• xscale (str) – ‘linear’ or ‘log’ y-axis scale

• yscale (str) – ‘linear’ or ‘log’ y-axis scale

• rangex (list) – [xmin, xmax] x-axis limits

• rangey (list) – [ymin, ymax] y-axis limits

• color (str) – color for scatterplot points

• marker (str) – matplotlib plot marker

• markersize (int) – size of plot markers

• label (str) – plot legend label

• snrcut (float) – flag closure amplitudes with snr lower than this

• grid (bool) – Plot gridlines if True

• ebar (bool) – Plot error bars if True

• axislabels (bool) – Show axis labels if True

• legend (bool) – Show legend if True

• show (bool) – Display the plot if true

• export_pdf (str) – path to pdf file to save figure

Returns:

Axes object with data plot

Return type:

(matplotlib.axes.Axes)

polchisq(im, dtype='pvis', ttype='nfft', pol_trans=True, mask=[], **kwargs)

Give the reduced chi^2 for the specified image and polarimetric datatype.

Parameters:

• im (Image) – image to test polarimetric chi^2

• dtype (str) – data type of polarimetric chi^2 (‘pvis’,’m’,’pbs’)

• pol (str) – polarization type (‘I’, ‘Q’, ‘U’, ‘V’, ‘LL’, ‘RR’, ‘LR’, or ‘RL’

• mask (arr) – mask of same dimension as im.imvec

• ttype (str) – if “fast” or “nfft” or “direct”

• pol_trans (bool) – True for I,m,chi, False for IQU

• fft_pad_factor (float) – zero pad the image to (fft_pad_factor * image size) in FFT

• conv_func ('str') – The convolving function for gridding; ‘gaussian’, ‘pill’,’cubic’

• p_rad (int) – The pixel radius for the convolving function

• order ('str') – Interpolation order for sampling the FFT

• systematic_noise (float) – adds a fractional systematic noise tolerance to sigmas

• snrcut (float) – a snr cutoff for including data in the chi^2 sum

• debias (bool) – if True then apply debiasing to amplitudes/closure amplitudes

• weighting (str) – ‘natural’ or ‘uniform’

• systematic_cphase_noise (float) – value in degrees to add to closure phase sigmas

• cp_uv_min (float) – flag short baselines before forming closure quantities

• maxset (bool) – if True, use maximal set instead of minimal for closure quantities

Returns:

image chi^2

Return type:

(float)

recompute_uv()

Recompute u,v points using observation times and metadata

Args:

Returns:

New Obsdata object containing the same data with recomputed u,v points

Return type:

(Obsdata)

reorder_baselines(trial_speedups=False)

Reorder baselines to match uvfits convention, based on the telescope array ordering

reorder_baselines_trial_speedups()

Reorder baselines to match uvfits convention, based on the telescope array ordering

reorder_tarr_random(reorder_baselines=True)

Randomly reorder the telescope array.

reorder_tarr_sefd(reorder_baselines=True)

Reorder the telescope array by SEFD minimal to maximum.

reorder_tarr_snr(reorder_baselines=True)

Reorder the telescope array by median SNR maximal to minimal.

res()

Return the nominal resolution (1/longest baseline) of the observation in radians.

Args:

Returns:

normal array resolution in radians

Return type:

(float)

rescale_noise(noise_rescale_factor=1.0)

Rescale the thermal noise on all Stokes parameters by a constant factor. This is useful for AIPS data, which has a missing factor relating ‘weights’ to noise.

Parameters:

noise_rescale_factor (float) – The number to multiple the existing sigmas by.

Returns:

An Obsdata object with the rescaled noise values.

Return type:

(Obsdata)

rescale_zbl(totflux, uv_max, debias=True)

Rescale the short baselines to a new level of total flux.

Parameters:

• totflux (float) – new total flux to rescale to

• uv_max (float) – maximum baseline length to rescale

• debias (bool) – Debias amplitudes before computing original total flux from short bls

Returns:

An Obsdata object with the inflated noise values.

Return type:

(Obsdata)

reverse_taper(fwhm)

Reverse taper the observation with a circular Gaussian kernel

Parameters:

fwhm (float) – real space fwhm size of convolution kernel in radian

Returns:

a new reverse-tapered observation object

Return type:

(Obsdata)

reweight(uv_radius, weightdist=1.0)

Reweight the sigmas based on the local density of uv points

Parameters:

• uv_radius (float) – radius in uv-plane to look for nearby points

• weightdist (float) –

  ??

Returns:

a new reweighted observation object.

Return type:

(Obsdata)

save_oifits(fname, flux=1.0)

Save visibility data to oifits. Polarization data is NOT saved.

Parameters:

• fname (str) – path to output text file

• flux (float) – normalization total flux

save_txt(fname)

Save visibility data to a text file.

Parameters:

fname (str) – path to output text file

save_uvfits(fname, force_singlepol=False, polrep_out='circ')

Save visibility data to uvfits file. :param fname: path to output uvfits file. :type fname: str :param force_singlepol: if ‘R’ or ‘L’, will interpret stokes I field as ‘RR’ or ‘LL’ :type force_singlepol: str :param polrep_out: ‘circ’ or ‘stokes’: how data should be stored in the uvfits file :type polrep_out: str

sourcevec()

Return the source position vector in geocentric coordinates at 0h GMST.

Args:

Returns:

normal vector pointing to source in geocentric coordinates (m)

Return type:

(numpy.array)

split_obs(t_gather=0.0, scan_gather=False)

Split single observation into multiple observation files, one per scan..

Parameters:

• t_gather (float) – Grouping timescale (in seconds). 0.0 indicates no grouping.

• scan_gather – If true, gather data into scans

switch_polrep(polrep_out='stokes', allow_singlepol=True, singlepol_hand='R')

Return a new observation with the polarization representation changed

Parameters:

• polrep_out (str) – the polrep of the output data

• allow_singlepol (bool) – If True, treat single-polarization data as Stokes I when converting from ‘circ’ polrep to ‘stokes’

• singlepol_hand (str) – ‘R’ or ‘L’; determines which parallel-hand is assumed when converting ‘stokes’ to ‘circ’ if only I is present

Returns:

new Obsdata object with potentially different polrep

Return type:

(Obsdata)

switch_timetype(timetype_out='UTC')

Return a new observation with the time type switched

Parameters:

timetype (str) – “UTC” or “GMST”

Returns:

new Obsdata object with potentially different timetype

Return type:

(Obsdata)

taper(fwhm)

Taper the observation with a circular Gaussian kernel

Parameters:

fwhm (float) – real space fwhm size of convolution kernel in radian

Returns:

a new tapered observation object

Return type:

(Obsdata)

tlist(conj=False, t_gather=0.0, scan_gather=False)

Group the data in a list of equal time observation datatables.

Parameters:

• conj (bool) – True if tlist_out includes conjugate baselines.

• t_gather (float) – Grouping timescale (in seconds). 0.0 indicates no grouping.

• scan_gather (bool) – If true, gather data into scans

Returns:

a list of data tables containing time-partitioned data

Return type:

(list)

unpack(fields, mode='all', ang_unit='deg', debias=False, conj=False, timetype=False)

Unpack the data for the whole observation .

Parameters:

• fields (list) – list of unpacked quantities from availalbe quantities in FIELDS

• mode (str) – ‘all’ returns all data in single table, ‘time’ groups output by equal time, ‘bl’ groups by baseline

• ang_unit (str) – ‘deg’ for degrees and ‘rad’ for radian phases

• debias (bool) – True to debias visibility amplitudes

• conj (bool) – True to include conjugate baselines

• timetype (str) – ‘GMST’ or ‘UTC’ changes what is returned for ‘time’

Returns:

unpacked numpy array with data in fields requested

Return type:

(numpy.recarray)

unpack_bl(site1, site2, fields, ang_unit='deg', debias=False, timetype=False)

Unpack the data over time on the selected baseline site1-site2.

Parameters:

• site1 (str) – First site name

• site2 (str) – Second site name

• fields (list) – list of unpacked quantities from available quantities in FIELDS

• ang_unit (str) – ‘deg’ for degrees and ‘rad’ for radian phases

• debias (bool) – True to debias visibility amplitudes

• timetype (str) – ‘GMST’ or ‘UTC’ changes what is returned for ‘time’

Returns:

unpacked numpy array with data in fields requested

Return type:

(numpy.recarray)

unpack_dat(data, fields, conj=False, ang_unit='deg', debias=False, timetype=False)

Unpack the data in a passed data recarray.

Parameters:

• data (numpy.recarray) – data recarray of format DTPOL_STOKES or DTPOL_CIRC

• fields (list) – list of unpacked quantities from availalbe quantities in FIELDS

• conj (bool) – True to include conjugate baselines

• ang_unit (str) – ‘deg’ for degrees and ‘rad’ for radian phases

• debias (bool) – True to debias visibility amplitudes

• timetype (str) – ‘GMST’ or ‘UTC’ changes what is returned for ‘time’

Returns:

unpacked numpy array with data in fields requested

Return type:

(numpy.recarray)

ehtim.obsdata.load_maps(arrfile, obsspec, ifile, qfile=0, ufile=0, vfile=0, src='SgrA', mjd=51544, ampcal=False, phasecal=False)

Read an observation from a maps text file and return an Obsdata object.

Parameters:

• arrfile (str) – path to input array file

• obsspec (str) – path to input obs spec file

• ifile (str) – path to input Stokes I data file

• qfile (str) – path to input Stokes Q data file

• ufile (str) – path to input Stokes U data file

• vfile (str) – path to input Stokes V data file

• src (str) – source name

• mjd (int) – integer observation MJD

• ampcal (bool) – True if amplitude calibrated

• phasecal (bool) – True if phase calibrated

Returns:

Obsdata object loaded from file

Return type:

obs (Obsdata)

ehtim.obsdata.load_obs(fname, polrep='stokes', flipbl=False, remove_nan=False, force_singlepol=None, channel=<built-in function all>, IF=<built-in function all>, allow_singlepol=True, flux=1.0, obsspec=None, ifile=None, qfile=None, ufile=None, vfile=None, src='SgrA', mjd=51544, ampcal=False, phasecal=False)

Smart obs read-in, detects file type and loads appropriately.

Parameters:

• fname (str) – path to input text file

• polrep (str) – load data as either ‘stokes’ or ‘circ’

• flipbl (bool) – flip baseline phases if True.

• remove_nan (bool) – True to remove nans from missing polarizations

• polrep – load data as either ‘stokes’ or ‘circ’

• force_singlepol (str) – ‘R’ or ‘L’ to load only 1 polarization

• channel (list) – list of channels to average in the import. channel=all averages all

• IF (list) – list of IFs to average in the import. IF=all averages all IFS

• flux (float) – normalization total flux

• obsspec (str) – path to input obs spec file

• ifile (str) – path to input Stokes I data file

• qfile (str) – path to input Stokes Q data file

• ufile (str) – path to input Stokes U data file

• vfile (str) – path to input Stokes V data file

• src (str) – source name

• mjd (int) – integer observation MJD

• ampcal (bool) – True if amplitude calibrated

• phasecal (bool) – True if phase calibrated

Returns:

Obsdata object loaded from file

Return type:

obs (Obsdata)

ehtim.obsdata.load_oifits(fname, flux=1.0)

Load data from an oifits file. Does NOT currently support polarization.

Parameters:

• fname (str) – path to input text file

• flux (float) – normalization total flux

Returns:

Obsdata object loaded from file

Return type:

obs (Obsdata)

ehtim.obsdata.load_txt(fname, polrep='stokes')

Read an observation from a text file.

Parameters:

• fname (str) – path to input text file

• polrep (str) – load data as either ‘stokes’ or ‘circ’

Returns:

Obsdata object loaded from file

Return type:

obs (Obsdata)

ehtim.obsdata.load_uvfits(fname, flipbl=False, remove_nan=False, force_singlepol=None, channel=<built-in function all>, IF=<built-in function all>, polrep='stokes', allow_singlepol=True, ignore_pzero_date=True, trial_speedups=False)

Load observation data from a uvfits file.

Parameters:

• fname (str or HDUList) – path to input text file or HDUList object

• flipbl (bool) – flip baseline phases if True.

• remove_nan (bool) – True to remove nans from missing polarizations

• polrep (str) – load data as either ‘stokes’ or ‘circ’

• force_singlepol (str) – ‘R’ or ‘L’ to load only 1 polarization

• channel (list) – list of channels to average in the import. channel=all averages all

• IF (list) – list of IFs to average in the import. IF=all averages all IFS

• remove_nan – whether or not to remove entries with nan data

• ignore_pzero_date (bool) – if True, ignore the offset parameters in DATE field TODO: what is the correct behavior per AIPS memo 117?

Returns:

Obsdata object loaded from file

Return type:

obs (Obsdata)

ehtim.obsdata.merge_obs(obs_List, force_merge=False)

Merge a list of observations into a single observation file.

Parameters:

• obs_List (list) – list of split observation Obsdata objects.

• force_merge (bool) – forces the observations to merge even if parameters are different

Returns:

merged Obsdata object containing all scans in input list

Return type:

mergeobs (Obsdata)