Example Usage¶

The SkyScene object is designed to give you access to many PRFs to make up a field of stars. This will rely on you having gaiaoffline installed and the database for gaiaoffline downloaded and in the right place.

!!! note "You must ensure you have gaiaoffline database installed"

You should follow the [`gaiaoffline` documentation](https://christinahedges.github.io/gaiaoffline/) to ensure your database is installed correctly. This database is 30gb, and you should download the approved one from Zenodo or obtain it from a team member.

To create a SkyScene we three things:

1. The PRF: You should load the PRF you want to be used in the scene. You obtain the PRF from pandoraaperture
2. The WCS: The world coordinate system that we are using. This is determined by where we are pointing. You can obtain the WCS from pandoraref
3. The time: Because stars have proper motion we also need the time, to ensure the catalog has the stars in the right position. You should update the time based on when you're observing with Pandora.

We have convenience methods to load PRFs and WCS using from_pointing which will take an RA, Dec, and theta for the boresight of the telescope and will then load the right reference products.

In [22]:

import pandoraaperture as pa
import matplotlib.pyplot as plt
import pandoraref as pr
import numpy as np
from astropy.time import Time

import pandoraaperture as pa import matplotlib.pyplot as plt import pandoraref as pr import numpy as np from astropy.time import Time

In [23]:

# coordinates for Kepler-10
ra, dec, theta = (285.6794224553767, 50.24130600481639, 85.1230985)

# coordinates for Kepler-10 ra, dec, theta = (285.6794224553767, 50.24130600481639, 85.1230985)

The first kind of SkyScene is the base class and is mostly useful for modeling VISDA full frame images.

In [24]:

scene = pa.SkyScene.from_pointing(ra, dec, theta, time=Time.now())

scene = pa.SkyScene.from_pointing(ra, dec, theta, time=Time.now())

In [25]:

scene

scene

Out[25]:

SkyScene

Let's plot the scene. This models the scene using all the flux values in the catalog.

In [26]:

scene.plot();

scene.plot();

/Users/chedges/Pandora/repos/pandora-aperture/src/pandoraaperture/scene.py:210: FutureWarning: Downcasting object dtype arrays on .fillna, .ffill, .bfill is deprecated and will change in a future version. Call result.infer_objects(copy=False) instead. To opt-in to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`
  cat.phot_bp_mean_flux.fillna(0).values

This looks good, let's take a look at a different kind of SkyScene.

ROISkyScene is for ROIs on the VISDA. The ROIs are specified using their own corners and sizes.

In [27]:

rR, rC = np.mgrid[-500:500:3j, -500:500:3j]

ROI_corners = [(r + 1024 - 25, c + 1024 - 25) for r, c in zip(rR.ravel(), rC.ravel())]
nROIs = len(ROI_corners)
ROI_size = (50, 50)

rR, rC = np.mgrid[-500:500:3j, -500:500:3j] ROI_corners = [(r + 1024 - 25, c + 1024 - 25) for r, c in zip(rR.ravel(), rC.ravel())] nROIs = len(ROI_corners) ROI_size = (50, 50)

In [28]:

scene = pa.ROISkyScene.from_pointing(ra, dec, theta, time=Time.now(), nROIs=nROIs, ROI_size=ROI_size, ROI_corners=ROI_corners)

scene = pa.ROISkyScene.from_pointing(ra, dec, theta, time=Time.now(), nROIs=nROIs, ROI_size=ROI_size, ROI_corners=ROI_corners)

In [29]:

scene.plot(vmax=1000);

scene.plot(vmax=1000);

/Users/chedges/Pandora/repos/pandora-aperture/src/pandoraaperture/scene.py:210: FutureWarning: Downcasting object dtype arrays on .fillna, .ffill, .bfill is deprecated and will change in a future version. Call result.infer_objects(copy=False) instead. To opt-in to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`
  cat.phot_bp_mean_flux.fillna(0).values

This looks good too.

Let's look at the last kind of scene which is the DispersedSkyScene. This is primarily for working with NIRDA data.

In [30]:

scene = pa.DispersedSkyScene.from_pointing(ra, dec, theta, time=Time.now())

scene = pa.DispersedSkyScene.from_pointing(ra, dec, theta, time=Time.now())

/Users/chedges/Pandora/repos/pandora-aperture/src/pandoraaperture/scene.py:219: FutureWarning: Downcasting object dtype arrays on .fillna, .ffill, .bfill is deprecated and will change in a future version. Call result.infer_objects(copy=False) instead. To opt-in to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`
  return cat.j_flux.fillna(0).values * 1 * u.electron / u.second

In [31]:

scene.plot(vmax=1000);

scene.plot(vmax=1000);

/Users/chedges/Pandora/repos/pandora-aperture/src/pandoraaperture/scene.py:219: FutureWarning: Downcasting object dtype arrays on .fillna, .ffill, .bfill is deprecated and will change in a future version. Call result.infer_objects(copy=False) instead. To opt-in to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`
  return cat.j_flux.fillna(0).values * 1 * u.electron / u.second

All SkyScene objects use a PRF, WCS and a catalog of sources to make scenes. Each of these items should be loaded for you by the SkyScene object.

In [32]:

scene.prf

scene.prf

Out[32]:

DispersedPRF [836 elements]

In [33]:

scene.wcs

scene.wcs

Out[33]:

WCS Keywords

Number of WCS axes: 2
CTYPE : 'RA---TAN-SIP' 'DEC--TAN-SIP' 
CRVAL : 285.6794224553767 50.24130600481639 
CRPIX : 2008.0 1024.0 
PC1_1 PC1_2  : 0.08501524470486811 -0.9963796506190656 
PC2_1 PC2_2  : 0.9963796506190656 0.08501524470486811 
CDELT : -0.00033055555555555 0.00033055555555556 
NAXIS : 0  0

In [34]:

scene.cat

scene.cat

Out[34]:

	RA	Dec	source_id	phot_g_mean_flux	phot_bp_mean_flux	phot_rp_mean_flux	j_flux	h_flux	k_flux	teff_gspphot	row	column
0	285.633757	50.231231	2132155154538128768	56236.210944	31355.100882	35614.164609	1540.991844	1572.679262	985.916266	5941.9517	932.076534	1984.175847
1	285.641148	50.233379	2132155085818652928	203377.545737	116295.538335	125442.543287	5250.496097	5448.0384	3137.90674	6142.2866	946.977434	1989.423785
2	285.647999	50.232175	2132155085818653184	42743.629051	17197.44932	36317.625121	2351.799751	3380.960991	2080.846441	4535.444	959.92687	1984.663413
3	285.679214	50.241604	2132155017099178624	807113.660182	430541.476911	539327.990865	24603.695303	27405.653344	16559.225946	5595.7837	1022.675063	2007.933868
4	285.742462	50.244746	2133655472513199360	44080.19357	24707.747994	27664.001865	1191.791655	1314.135538	776.676262	6050.1006	1144.778271	2007.053045
5	285.742767	50.260734	2133655571295536896	35732.399589	14461.998394	30473.324992	1927.526422	2763.377136	1794.072677	4520.6436	1149.299699	2055.215998
6	285.753612	50.237074	2133655403793722752	32228.636685	16795.400158	22203.38556	1128.756697	1296.104922	826.875325	5807.1235	1164.094135	1982.099503
7	285.764802	50.242807	2133655506872938752	3783886.782239	2558851.50372	1839476.5443	58748.981246	50008.077622	30105.097079	9376.287	1186.870348	1997.569689
8	285.786348	50.264355	2133655713031373184	119057.057927	67003.067235	74566.519233	3169.569944	3421.686056	1981.709686	6014.1904	1233.070374	2059.074925
9	285.815075	50.241934	2133655266354774528	47226.165502	25787.084681	30479.605866	1368.359904	1514.398462	890.922715	5815.2905	1281.980343	1986.80266

The object works by creating a design matrix for you.

This design matrix is a Sparse3D object which contains one image per source. When you dot it with the flux values of each source, you get the scene image.

In [35]:

scene.A(delta_pos=(0, 0))

scene.A(delta_pos=(0, 0))

Out[35]:

<(400, 80, 10) Sparse3D array of type float64>

Each SkyScene has an evaluate method which will return the image of the scene.

In [36]:

r, c, im = scene.evaluate()

r, c, im = scene.evaluate()

/Users/chedges/Pandora/repos/pandora-aperture/src/pandoraaperture/scene.py:219: FutureWarning: Downcasting object dtype arrays on .fillna, .ffill, .bfill is deprecated and will change in a future version. Call result.infer_objects(copy=False) instead. To opt-in to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`
  return cat.j_flux.fillna(0).values * 1 * u.electron / u.second

You can move this scene around with the delta_pos argument to shift it by a number of pixels. You should only do this for small shifts.

SkyScene objects enable you to build FITS files containing the information you may need for that scene. You can obtain the following files

Catalog extension¶

This extension shows what targets are likely to fall on a pixel

In [37]:

hdu = scene.get_catalog_hdu()

hdu = scene.get_catalog_hdu()

In [38]:

hdu.data

hdu.data

Out[38]:

FITS_rec([(285.63375724, 50.23123125, '2132155154538128768',   56236.21094419,   31355.10088213,   35614.16460896,  1540.99184352,  1572.67926186,   985.91626641, 5941.9517,  932.07653438, 1984.17584713, 0, 0),
          (285.64114848, 50.23337892, '2132155085818652928',  203377.54573736,  116295.53833483,  125442.5432872 ,  5250.49609725,  5448.03839999,  3137.90673989, 6142.2866,  946.97743365, 1989.42378521, 0, 0),
          (285.64799859, 50.23217547, '2132155085818653184',   42743.62905105,   17197.44932034,   36317.625121  ,  2351.79975073,  3380.96099121,  2080.8464409 , 4535.444 ,  959.9268703 , 1984.66341253, 0, 0),
          (285.67921404, 50.24160444, '2132155017099178624',  807113.66018202,  430541.47691061,  539327.99086458, 24603.69530318, 27405.6533437 , 16559.22594555, 5595.7837, 1022.67506268, 2007.93386828, 0, 0),
          (285.74246182, 50.24474633, '2133655472513199360',   44080.1935701 ,   24707.74799427,   27664.00186461,  1191.79165461,  1314.13553792,   776.67626242, 6050.1006, 1144.77827079, 2007.05304495, 0, 0),
          (285.74276729, 50.26073441, '2133655571295536896',   35732.39958896,   14461.99839408,   30473.32499193,  1927.52642221,  2763.37713634,  1794.0726765 , 4520.6436, 1149.29969872, 2055.21599811, 0, 0),
          (285.75361156, 50.23707355, '2133655403793722752',   32228.63668549,   16795.40015799,   22203.38555981,  1128.75669692,  1296.10492239,   826.87532456, 5807.1235, 1164.09413491, 1982.09950322, 0, 0),
          (285.76480193, 50.24280658, '2133655506872938752', 3783886.78223853, 2558851.50371957, 1839476.5442999 , 58748.9812463 , 50008.07762171, 30105.09707888, 9376.287 , 1186.87034804, 1997.56968904, 0, 0),
          (285.78634785, 50.26435453, '2133655713031373184',  119057.05792735,   67003.06723471,   74566.51923251,  3169.56994443,  3421.68605625,  1981.70968561, 6014.1904, 1233.07037366, 2059.07492471, 0, 0),
          (285.8150753 , 50.24193411, '2133655266354774528',   47226.16550153,   25787.08468124,   30479.60586578,  1368.35990431,  1514.39846237,   890.92271528, 5815.2905, 1281.98034252, 1986.80265976, 0, 0)],
         dtype=(numpy.record, [('RA', '<f8'), ('Dec', '<f8'), ('source_id', 'S19'), ('phot_g_mean_flux', '<f8'), ('phot_bp_mean_flux', '<f8'), ('phot_rp_mean_flux', '<f8'), ('j_flux', '<f8'), ('h_flux', '<f8'), ('k_flux', '<f8'), ('teff_gspphot', '<f8'), ('row', '<f8'), ('column', '<f8'), ('row0', '<i8'), ('column0', '<i8')]))

Aperture extension¶

This extension shows the apertures for a given target. You will need to input the target number to obtain the aperture. You can either use the target index in the catalog, or a SkyCoord to use the target closest to that coordinate.

In [39]:

scene.get_aperture_hdu(1)

scene.get_aperture_hdu(1)

/Users/chedges/Pandora/repos/pandora-aperture/src/pandoraaperture/scene.py:219: FutureWarning: Downcasting object dtype arrays on .fillna, .ffill, .bfill is deprecated and will change in a future version. Call result.infer_objects(copy=False) instead. To opt-in to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`
  return cat.j_flux.fillna(0).values * 1 * u.electron / u.second

Out[39]:

<astropy.io.fits.hdu.compressed.compressed.CompImageHDU at 0x14f713af0>

In [40]:

from astropy.coordinates import SkyCoord
scene.get_aperture_hdu(SkyCoord(ra, dec, unit='deg'))

from astropy.coordinates import SkyCoord scene.get_aperture_hdu(SkyCoord(ra, dec, unit='deg'))

Out[40]:

<astropy.io.fits.hdu.compressed.compressed.CompImageHDU at 0x16e943700>

PRF extension¶

This extension gives you a FITS extension of the PRFs of each source. When combined with the catalog PRF, this gives you all of the information you strictly need for PRF photometry. This does not require a specific target to be input as it gives all PRFs for all possible targets.

In [41]:

scene.get_prf_hdu()

scene.get_prf_hdu()

Out[41]:

<astropy.io.fits.hdu.compressed.compressed.CompImageHDU at 0x16e94c5b0>

Model extension¶

This extension models the whole scene and returns it as an image.

In [42]:

scene.get_model_hdu()

scene.get_model_hdu()

Out[42]:

<astropy.io.fits.hdu.compressed.compressed.CompImageHDU at 0x16e97dc40>

In [ ]:

In [ ]: