Thesis Sam merge

[SamedVossberg]

Latest updatse (plots, debugging, GUI changes, etc.)

[Copilot]

Pull request overview

This PR merges a broad set of “thesis” updates across analysis notebooks, adds new compact decoding training/pipeline scripts, and substantially extends the Dash visualization layer (new choice filters, improved ensemble plots, and a new session-summary panel), plus introduces a parallelized GLM workflow for single-neuron encoding screens.

Changes:

• Add compact interval decoding pipelines + CLI training scripts (interval-wise and session-wise), including shuffle/bootstrap significance support.
• Introduce a parallel negative-binomial GLM notebook and supporting multiprocessing utilities (with caching) for large-scale single-neuron encoding screens.
• Update Dash plots/wrappers to support R1/R2 choice filters, improve performance/UX of ensemble ridgeline plots, add session summary plot, and tweak session labeling/colormaps; plus minor notebook/session-exclusion updates.

Reviewed changes

Copilot reviewed 31 out of 46 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
scripted_plotting/hmm_actions.ipynb	Notebook metadata update (kernel display name / Python version).
scripted_plotting/animal_6_analysis/single_cell_correlations.ipynb	Update default excluded sessions list used in analysis.
scripted_plotting/animal_6_analysis/regression_reversal.ipynb	Update default excluded sessions list used in analysis.
scripted_plotting/animal_6_analysis/neuron_encoding_check.ipynb	Update excluded sessions list (removes prior inline comment).
scripted_plotting/animal_6_analysis/neuron_classification_trackwise.ipynb	Update excluded sessions list (removes prior inline comment).
scripted_plotting/animal_6_analysis/interval_session_decoding_train.py	New CLI driver to train session-level interval decoders and save compact outputs.
scripted_plotting/animal_6_analysis/interval_session_decoding_pipeline.py	New session-level interval decoding pipeline with shuffle significance and multiprocessing support.
scripted_plotting/animal_6_analysis/interval_decoding_train.py	New CLI driver to train interval decoders (optionally parallel across assemblies) and save outputs.
scripted_plotting/animal_6_analysis/interval_decoding_pipeline.py	New compact interval decoding pipeline with grouped CV, shuffle, and bootstrap significance.
scripted_plotting/animal_6_analysis/ensemble_regression.ipynb	Update default excluded sessions list used in analysis.
scripted_plotting/animal_6_analysis/choice_encoding_glm_parallel.ipynb	New notebook implementing parallelized single-neuron NB-GLM feature screens.
scripted_plotting/animal_6_analysis/choice_encoding_glm_parallel_utils.py	New multiprocessing helpers for parallel GLM screening + FDR postprocessing.
scripted_plotting/animal_6_analysis/.gitignore	Ignore generated analysis artifacts (runs/figures/cache/etc.) within animal_6_analysis.
requirements.txt	Add statsmodels dependency needed by GLM workflows.
ephys_preprocessing/postproc_mea1k_ephys.py	Add TODO note clarifying potential PCA projection formula concern.
dashsrc/plot_components/plots/plot_unit_fr_stability.py	Remove unused zscore import.
dashsrc/plot_components/plots/plot_StayPerformance.py	Add schema-compat column aliases/derivations + session labeling tweaks; adjust unit conversions.
dashsrc/plot_components/plots/plot_SessionKinematics.py	Add grouping options for R1/R2 choice and explicit error on unknown group_by.
dashsrc/plot_components/plots/plot_EnsembleTrialwise.py	Performance/robustness updates: safer index reset, interval_name handling, fewer traces/shapes, improved hover.
dashsrc/plot_components/plots/plot_EnsembleSessionwise.py	Add session summary plot, reduce trace count, improve session ordering/labeling, optional single-trial display.
dashsrc/plot_components/plots/plot_AnimalKinematics.py	Display only session date portion in tick labels.
dashsrc/plot_components/plot_wrappers/wrapper_TrackwiseEnsembleProj.py	Wire new R1/R2 choice filters through wrapper + UI controls.
dashsrc/plot_components/plot_wrappers/wrapper_TrackFiringRate.py	Wire new R1/R2 choice filters through wrapper + UI controls.
dashsrc/plot_components/plot_wrappers/wrapper_EnsembleTrialwise.py	Refactor filtering pipeline, add session-summary subplot, and return both summary + session detail figures.
dashsrc/plot_components/plot_wrappers/wrapper_EnsembleSessionwise.py	Replace event dropdown impl, add session summary graph + single-trial toggle, and refactor filtering helpers.
dashsrc/plot_components/plot_wrappers/data_selection_components.py	Show session “date” only in slider marks (split on underscore).
dashsrc/components/data_loading_controls.py	Add faster session exclusion path using unity_trial_n_rows when present.
dashsrc/components/dashvis_constants.py	Update R1/R2 choice colormaps.
assets/sessionwise_controls.css	New CSS to improve multi-select event dropdown sizing/z-index.
.gitignore	Ignore interval_decoding_runs output and .claude directory.

Comments suppressed due to low confidence (1)

dashsrc/plot_components/plots/plot_StayPerformance.py:79

• _draw_success_rate currently prints debug output on every render, which will spam logs and slow down Dash callbacks. These should be removed or gated behind an explicit debug flag/logger level.

def _draw_success_rate(fig, all_data, double_reward_filter=None):
    print("INNNN")
    print(double_reward_filter)
    if double_reward_filter == [] or double_reward_filter is None or sorted(double_reward_filter) == ['Early R', 'Late R']:
        print("in double")

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.