The concepts further-study page is a focused list of topics nSTAT does not implement (population geometry & dimensionality reduction, deep-learning encoders/decoders, spike sorting, vendor-format I/O), with primary references for each. A longer prior version is replaced. All concepts-page content (microelectrode recordings, point-process GLMs, time-rescaling KS, decoding, state-space EM, rhythmic firing, etc.) is unchanged.
docs/concepts/further_study.md — the new
focused page.docs/concepts/from_filters_to_deep_learning.md —
the page now ends at the PPAF → modern-sequence-decoder bridge;
a trailing section pointing to research-frontier pretrained models
is removed.docs/concepts/index.md, population_geometry.md,
self_check.md — cross-references updated, plus
light word-level cleanup.RELEASE_NOTES.md +
docs/changes/2026-05-31-v0.4.3-docs-cleanup.html —
generalize a v0.4.3 historical reference to an internal-only test
filename.docs/concepts/rhythmic_firing_and_clinical_microelectrode.md
— remove a stray </content></invoke> tail
artifact.docs/Examples.md is now a
card-style index mirroring the MATLAB Examples help index; each paper
example links to its figure README. (PR #143)intro.html, extras_summary.html,
whats_new.html, and every per-release
docs/changes/*.html) now share a single clean light
palette and consistent reference tables. (PR #141,
PR #142,
PR #145)docs/concepts/pitfalls_and_faq.md — the
np.random.default_rng(seed) pattern, why simulator
results vary without a seed, and the deterministic-fit vs.
multi-restart-EM split.AGENT_GUIDE.md source layout now lists
examples/tutorials/, examples/extras/, and
docs/concepts/ — surfaces present since v0.4.4 but
missing from the agent-orientation paragraph.Five workflows that were previously running on every push are now
workflow_dispatch-only on this repository: performance-parity,
notebook-full-fidelity, helpfile-check,
readme-check, and nstat-pip-install. The unit
test, docs build, data-integrity, notebook-smoke, and cleanroom-compliance
gates still run automatically on every PR. A new make ci-local
target runs the same check set against the working tree before a push.
(PR #144)
| Check | Result |
|---|---|
make docs-strict (Sphinx -W, clean rebuild) | passes |
| Cross-references to renamed concepts page | all resolve |
Four version-bearing files in sync (tests/test_version_sync.py) | 0.4.5 |
| Public-API surface | unchanged |
None.