Skip to content

Pupil and gaze quality-control workflow

Frozen R source: reference/vignettes/articles/pupil-qc-workflow.Rmd

This page is the Python migration companion for the corresponding gpbiometrics 2.0.0 article. The original R article is retained verbatim in reference/vignettes/; this companion identifies the matching Python API so the scientific workflow can be reproduced without hiding the reference implementation.

Python API crosswalk

The frozen R article calls the following exported functions; all are available under the same names in gpbiometricspy:

  • gp.baseline_correct_gazepoint_pupil(...)
  • gp.create_gazepoint_analysis_decision_log(...)
  • gp.create_gazepoint_qc_supplement(...)
  • gp.create_gazepoint_reproducibility_statement(...)
  • gp.detect_gazepoint_blinks(...)
  • gp.detect_gazepoint_pupil_blinks(...)
  • gp.filter_gazepoint_gaze(...)
  • gp.interpolate_gazepoint_pupil_blinks(...)
  • gp.pipeline_comparison_dashboard(...)
  • gp.plot_gazepoint_missingness(...)
  • gp.profile_gazepoint_export_folder(...)
  • gp.recommend_gazepoint_biometric_exclusions(...)
  • gp.simulate_gazepoint_eye_data(...)
  • gp.smooth_gazepoint_pupil(...)
  • gp.standardize_gazepoint_column_names(...)
  • gp.summarize_gazepoint_export_inventory(...)
  • gp.summarize_gazepoint_missingness(...)
  • gp.summarize_gazepoint_pupil_events(...)
  • gp.summarize_gazepoint_qc_overview(...)
  • gp.validate_gazepoint_format(...)
  • gp.validate_gazepoint_metadata(...)
import gpbiometricspy as gp

# Example entry point from this workflow
# result = gp.baseline_correct_gazepoint_pupil(...)

Interpretation

Use the same conservative physiological interpretation as the R package: derived biometric features are signal-processing outputs and do not directly establish emotion, stress, cognition, preference, health status, or diagnosis.

Executable Python companion

The frozen R call crosswalk above is retained for completeness. The following companion is an executable end-to-end Python workflow using synthetic/public data and the same scientific domain. It is also executed by the test suite.

Run from the repository root:

python examples/tutorials/pupil-qc-workflow.py

from __future__ import annotations
from _shared import *
t=np.arange(-.5,1.01,.05); d=pd.DataFrame({'participant':'P01','trial':'T01','time':t,'pupil':3+.1*np.sin(4*t)}); base=gp.baseline_correct_gazepoint_pupil(d,pupil_col='pupil',time_col='time',trial_cols=['participant','trial'],baseline_window=(-.5,-.1)); smooth=gp.smooth_gazepoint_pupil(base,pupil_cols='pupil',id_cols=['participant','trial'],window=5); clean=gp.clean_gazepoint_pupil_signal(d,pupil_cols=['pupil'],time_col='time',group_cols=['participant','trial']); finish('pupil-qc-workflow',baseline=base,smooth=smooth,clean=clean)

Rendered Python output

These figures are generated from bundled synthetic/public data by scripts/generate_docs_gallery.py using the current Python plotting API.

Pupil and gaze overview

Pupil and gaze overview