FAQ¶

General¶

What targets can QuickLook analyze?¶

Any target observed by TESS that has a light curve on MAST. This includes TOI candidates, known exoplanets, eclipsing binaries, variable stars, and arbitrary TIC IDs.

What is the SDE and what values are significant?¶

SDE (Signal Detection Efficiency) is the signal-to-noise metric from the TLS periodogram. Values above ~8 generally indicate a significant periodic transit-like signal. However, astrophysical false positives (eclipsing binaries, contamination) can also produce high SDE values.

Can QuickLook analyze Kepler or K2 data?¶

Not currently. QuickLook is designed for TESS data. Kepler/K2 support may be added in the future.

Errors and troubleshooting¶

`NoDataError: No light curves found`¶

The target has no TESS light curves for the specified pipeline and sector. Try:

A different pipeline (--pipeline qlp or --pipeline tess-spoc)
Removing the sector constraint (--sector -1)
Checking the target name spelling

`No module named 'distutils'` (Python 3.12)¶

The batman-package dependency imports distutils, which was removed in Python 3.12. Install setuptools:

pip install setuptools

This is included automatically if you install quicklook-package >= 1.5.

MAST HTTP 500 errors¶

The MAST archive sometimes experiences outages. These are transient -- wait a few minutes and try again. In CI, network tests automatically mark as xfail when MAST is down.

`I/O operation on closed file` in the web GUI¶

This was caused by running multiple pipeline jobs concurrently. As of v1.5, the GUI serializes all jobs through a single worker queue. If you see this error, make sure you are running the latest version.

`GLS power is NaN`¶

The Generalized Lomb-Scargle periodogram returned NaN, which means the data is unsuitable for GLS (e.g., constant flux, too few data points). QuickLook automatically falls back to astropy's Lomb-Scargle implementation when this happens.

Plots look garbled or axes overlap¶

Try a different matplotlib backend:

export MPLBACKEND=Agg
ql --name TOI-1234 -save

Pipeline parameters¶

How do I choose the window length?¶

The detrending window length controls how aggressive the baseline removal is.

Too short: removes real transit signals along with the trend
Too long: leaves residual systematics in the flattened light curve

The default (None) auto-selects 3x the known transit duration, or 0.5 days if no ephemeris is available. Set --window_length 0 to auto-optimize.

When should I use GP flattening?¶

GP (--flatten_method gp) is useful for complex stellar variability that biweight filtering cannot capture (e.g., multi-mode pulsators). However, it is computationally expensive for short-cadence (2-min) data and is not recommended for routine use.

What does `--mask_ephem` do?¶

It masks known transits (from ExoFOP or custom ephemeris) before detrending. This prevents the detrending algorithm from partially removing the transit signal, which can happen with aggressive window lengths.

Output files¶

How do I read the HDF5 output?¶

import h5py

with h5py.File("TOI-1234_s56_pdcsap_sc_tls.h5", "r") as f:
    period = f["period"][()]
    power = f["power"][()]
    sde = f.attrs["SDE"]
    depth = f.attrs["depth"]
    print(f"Period: {period:.4f} d, SDE: {sde:.1f}, Depth: {depth:.6f}")

Or use the read_tls tool to extract a CSV summary from a directory of H5 files.