Inspect a Target Pixel File for a Centroid Offset

In this tutorial, we’ll see how to use vetpack to inspect a Target Pixel File (TPF) for a centroid offset with just a few lines of code.

Let’s start by importing vetpack and lightkurve (https://docs.lightkurve.org/), the handy package vetpack relies on for interacting with Kepler, K2, and TESS data products.

import lightkurve as lk
import vetpack as vp

For this example, we’re going to look at the TESS data on WASP-19, a known hot Jupiter host. For context, let’s quickly grab the TESS Sector 9 light curve and take a look at it.

target = 'WASP19'
search_result = lk.search_lightcurvefile(target, mission='TESS')
search_result

SearchResult containing 1 data products.

#	observation	target_name	productFilename	distance
0	TESS Sector 9	35516889	tess2019058134432-s0009-0000000035516889-0139-s_lc.fits	0.0

# Download the single search result and plot the light curve
lcf = search_result[0].download()
lcf.PDCSAP_FLUX.plot()

/Users/brackham/anaconda3/envs/vetpack-dev/lib/python3.8/site-packages/lightkurve/lightcurvefile.py:45: LightkurveWarning: `LightCurveFile.header` is deprecated, please use `LightCurveFile.get_header()` instead.
  warnings.warn("`LightCurveFile.header` is deprecated, please use "

<matplotlib.axes._subplots.AxesSubplot at 0x7fe81e48d340>

Nice light curve! You can clearly see the transits of WASP-19b as well as some out-of-transit variability due to rotation of the active host star WASP-19.

Plotting the TPF

Now, let’s grab the TPF for this dataset. We’ll use lightkurve again.

sr = lk.search_targetpixelfile(target, mission='TESS')
tpf = sr[0].download()

lightkurve has the built-in ability to plot single frames of the TPF.

The default is to plot the first frame, which show significant artifacts due to systematics (scattered light, I think, in this case).

tpf.plot(aperture_mask=tpf.pipeline_mask)

<matplotlib.axes._subplots.AxesSubplot at 0x7fe7d8944760>

We can specify another frame, but that requires as bit of guesswork to find a representative frame if there are a lot of systematics.

tpf.plot(aperture_mask=tpf.pipeline_mask, frame=500)

<matplotlib.axes._subplots.AxesSubplot at 0x7fe81fd3e7c0>

Or we can get around this issue by plotting a median image of the whole TPF with vetpack:

ax = vp.median_image(tpf)
ax.set_title(f'{target} Median Flux')

Text(0.5, 1.0, 'WASP19 Median Flux')

Plotting a difference image of the TPF

Now, let’s use vetpack to plot a in-transit/out-of-transit difference image so that we can visually inspect whether the flux decrement during transit events is indeed located on WASP-19. This, of course, is more useful for vetting candidate planets than analyzing data of confirmed planets like WASP-19b.

We’ll need to specific the mid-transit time, period, and duration of the transit signal of interest. We’ll use the parameters from Exo-FOP TESS for this.

# Parameters from
# https://exofop.ipac.caltech.edu/tess/target.php?id=35516889

t0 = 2458543.621948 - 2457000.  # TBJD
period = 0.788843
duration = 1.624113/24.

Pass them to vetpack.difference_image() along with the TPF object to easily see the result.

ax = vp.difference_image(tpf, t0, period, duration)
ax.set_title(f'{target} In$-$Out Difference Flux')

Text(0.5, 1.0, 'WASP19 In$-$Out Difference Flux')

It does indeed seem that the flux decrement during transit is localized on the location of WASP-19. No surprises there (phew!).

That’s it for now! More functionality is in development. Happy vetting!