Annotations extras

[PierreGtch]

• closes container for eye tracking related annotation information #12208
• closes [WIP [ENH] metadata attribute for Annotations #12213

Reference issue (if any)

This PR replaces #13213

What does this implement/fix?

The difference with #13213 is that the additional information is a list[dict[str, str | int | float | None]] instead of a dataframe

[PierreGtch]

@drammock I implemented here the recommendations in your #13213 (comment)

[drammock]

just a quick first pass with some high-level stuff, LMK when I should take another look @PierreGtch

[agramfort]

I am onboard with the change although I am not convinced about the word "details"

[PierreGtch]

The failed tests seem unrelated to the PR (Read timed out)

[drammock]

Thanks for sticking with this @PierreGtch. I'm pretty happy with where this ended up. Approving (modulo one docstring typo) but would like to wait on merging until at least one other maintainer has had a look (mostly because I've only read the code, haven't actually had time to play around with the feature interactively, and tomorrow I'm off to a conference for a week so I can't do it soon)

[PierreGtch]

thanks @drammock for your time on this PR!

[PierreGtch]

@sappelhoff @larsoner would you have time to also take a look? (the failed test seems unrelated to the PR)

[larsoner]

Looks good! Just minor stuff from me really. Can you push the next commit with [circle full] in the commit message? That will make CircleCI build all examples, which I think would be a nice sanity check (and save us some possible post-merge pain in case we missed something)

[drammock]

105 CircleCI failures (!). Luckily they're all of 2 types:

1. TypeError: Annotations.append() takes from 4 to 5 positional arguments but 6 were given

• examples/datasets/kernel_phantom.py
• examples/datasets/opm_data.py
• examples/decoding/decoding_spatio_temporal_source.py
• examples/decoding/decoding_time_generalization_conditions.py
• examples/decoding/decoding_unsupervised_spatial_filter.py
• examples/decoding/decoding_xdawn_eeg.py
• examples/decoding/ems_filtering.py
• examples/decoding/linear_model_patterns.py
• examples/inverse/compute_mne_inverse_epochs_in_label.py
• examples/inverse/compute_mne_inverse_raw_in_label.py
• examples/inverse/dics_epochs.py
• examples/inverse/dics_source_power.py
• examples/inverse/evoked_ers_source_power.py
• examples/inverse/mne_cov_power.py
• examples/inverse/multi_dipole_model.py
• examples/inverse/multidict_reweighted_tfmxne.py
• examples/inverse/psf_ctf_vertices_lcmv.py
• examples/io/elekta_epochs.py
• examples/preprocessing/css.py
• examples/preprocessing/define_target_events.py
• examples/preprocessing/eeg_csd.py
• examples/preprocessing/eog_artifact_histogram.py
• examples/preprocessing/eog_regression.py
• examples/preprocessing/find_ref_artifacts.py
• examples/preprocessing/ica_comparison.py
• examples/preprocessing/movement_compensation.py
• examples/preprocessing/movement_detection.py
• examples/preprocessing/muscle_ica.py
• examples/preprocessing/otp.py
• examples/preprocessing/xdawn_denoising.py
• examples/simulation/simulate_evoked_data.py
• examples/simulation/simulate_raw_data.py
• examples/stats/cluster_stats_evoked.py
• examples/stats/fdr_stats_evoked.py
• examples/stats/linear_regression_raw.py
• examples/stats/sensor_permutation_test.py
• examples/time_frequency/compute_csd.py
• examples/time_frequency/compute_source_psd_epochs.py
• examples/time_frequency/source_label_time_frequency.py
• examples/time_frequency/source_power_spectrum_opm.py
• examples/time_frequency/source_power_spectrum.py
• examples/time_frequency/source_space_time_frequency.py
• examples/time_frequency/temporal_whitening.py
• examples/time_frequency/time_frequency_global_field_power.py
• examples/visualization/3d_to_2d.py
• examples/visualization/channel_epochs_image.py
• examples/visualization/eeg_on_scalp.py
• examples/visualization/evoked_whitening.py
• examples/visualization/eyetracking_plot_heatmap.py
• examples/visualization/meg_sensors.py
• examples/visualization/topo_compare_conditions.py
• examples/visualization/topo_customized.py
• tutorials/epochs/10_epochs_overview.py
• tutorials/epochs/15_baseline_regression.py
• tutorials/epochs/20_visualize_epochs.py
• tutorials/epochs/50_epochs_to_data_frame.py
• tutorials/epochs/60_make_fixed_length_epochs.py
• tutorials/evoked/10_evoked_overview.py
• tutorials/evoked/30_eeg_erp.py
• tutorials/evoked/40_whitened.py
• tutorials/forward/20_source_alignment.py
• tutorials/forward/90_compute_covariance.py
• tutorials/intro/10_overview.py
• tutorials/intro/15_inplace.py
• tutorials/intro/20_events_from_raw.py
• tutorials/intro/30_info.py
• tutorials/intro/40_sensor_locations.py
• tutorials/intro/70_report.py
• tutorials/inverse/30_mne_dspm_loreta.py
• tutorials/inverse/50_beamformer_lcmv.py
• tutorials/inverse/70_eeg_mri_coords.py
• tutorials/inverse/80_brainstorm_phantom_elekta.py
• tutorials/inverse/95_phantom_KIT.py
• tutorials/io/70_reading_eyetracking_data.py
• tutorials/machine-learning/50_decoding.py
• tutorials/preprocessing/10_preprocessing_overview.py
• tutorials/preprocessing/15_handling_bad_channels.py
• tutorials/preprocessing/20_rejecting_bad_data.py
• tutorials/preprocessing/30_filtering_resampling.py
• tutorials/preprocessing/35_artifact_correction_regression.py
• tutorials/preprocessing/40_artifact_correction_ica.py
• tutorials/preprocessing/45_projectors_background.py
• tutorials/preprocessing/50_artifact_correction_ssp.py
• tutorials/preprocessing/55_setting_eeg_reference.py
• tutorials/preprocessing/59_head_positions.py
• tutorials/preprocessing/60_maxwell_filtering_sss.py
• tutorials/preprocessing/90_eyetracking_data.py
• tutorials/raw/10_raw_overview.py
• tutorials/raw/20_event_arrays.py
• tutorials/raw/30_annotate_raw.py
• tutorials/raw/40_visualize_raw.py
• tutorials/simulation/70_point_spread.py
• tutorials/simulation/80_dics.py
• tutorials/stats-sensor-space/40_cluster_1samp_time_freq.py
• tutorials/stats-sensor-space/50_cluster_between_time_freq.py
• tutorials/stats-sensor-space/70_cluster_rmANOVA_time_freq.py
• tutorials/stats-sensor-space/75_cluster_ftest_spatiotemporal.py
• tutorials/stats-source-space/20_cluster_1samp_spatiotemporal.py
• tutorials/stats-source-space/60_cluster_rmANOVA_spatiotemporal.py
• tutorials/time-freq/10_spectrum_class.py
• tutorials/time-freq/20_sensors_time_frequency.py
• tutorials/visualization/10_publication_figure.py

2. TypeError: Annotations.init() takes from 4 to 6 positional arguments but 7 were given

• examples/preprocessing/epochs_metadata.py
• tutorials/clinical/20_seeg.py
• tutorials/epochs/40_autogenerate_metadata.py

[larsoner]

Thanks for the quick fix for the bug! For TDD though it's better if we had a test in tests/ somewhere that would have caught that bug / those bugs without needing to run a [circle full] first.... @PierreGtch is it doable to add a test that would have failed before 7406948 ?

[PierreGtch]

@larsoner thanks for the review!
Is this the kind of test you had in mind?

[PierreGtch]

The failed test seems unrelated

[drammock]

@larsoner thanks for the review! Is this the kind of test you had in mind?

Looking at what you had to change in 7406948, I think tests that called __iadd__, _read_annotations_csv, and _read_annotations_fif (either directly, or implicitly via a public API call) should have caught the same failures that running the tutorials caught. My guess is that is what @larsoner meant.

[PierreGtch]

@drammock @larsoner The issue was introduced in the last-minute suggestion from code review: 9e67e3b
And the non-circle-full tests did catch it. Not caught in the initial commit 9e67e3b because tests were aborted due to pre-commit, but in 3621595 we see that ubuntu-pip fails with 928 errors.
So it should be all good to merge

[drammock]

@drammock @larsoner The issue was introduced in the last-minute suggestion from code review: 9e67e3b And the non-circle-full tests did catch it. Not caught in the initial commit 9e67e3b because tests were aborted due to pre-commit, but in 3621595 we see that ubuntu-pip fails with 928 errors. So it should be all good to merge

haha, you're right. I was so focused on the CircleCI results that I didn't notice the copious failures in the pytest suite. I'll revert the new test then (since we already had the needed coverage) and mark for merge-when-green.

Thanks @PierreGtch !

[welcome]

🎉 Congrats on merging your first pull request! 🥳 Looking forward to seeing more from you in the future! 💪