Skip to content

Window Semantics

Raul Montoya Cardenas edited this page Jul 29, 2026 · 2 revisions

Window Semantics

Wiki hero

Generated with Grok Build: Grok 4.5 · xAI Imagine (/imagine)

Important for fixtures and interoperability (LIM-41).

spike_count / spike_density

With explicit t_start / t_end, counting uses inclusive endpoints on the sorted timeline (searchsortedfirst / searchsortedlast).

windowed_spike_features

With explicit t_end, windows are half-open [t_start, t_end): spikes with t == window_end are not included (unless the special “include right edge” path when t_end is inferred from data max via nextfloat).

When t_end is omitted and data is non-empty, the stream end is nextfloat(last(times)) so the final spike is included.

Fixtures

test/fixtures/spike_vectors.json encodes these rules. Golden values for density windows and windowed features depend on the half-open vs inclusive distinction.


Last updated: July 29, 2026 Updated by: Grok Build: Grok 4.5 Package tip reference: d2229e2 (main)

Clone this wiki locally