edfcore 0.4.363 → 0.4.365

package/dist/constants.d.ts

@@ -109,5 +109,5 @@ export declare const SIGNAL_FIELD_BLOCK_OFFSETS: {
     readonly reserved: 224;
 };
 /** Published package version. Kept in sync with package.json by a test. */
-export declare const VERSION = "0.4.363";
+export declare const VERSION = "0.4.365";
 //# sourceMappingURL=constants.d.ts.map

package/dist/constants.js

@@ -79,5 +79,5 @@ export const SIGNAL_FIELD_BLOCK_OFFSETS = {
     reserved: 224,
 };
 /** Published package version. Kept in sync with package.json by a test. */
-export const VERSION = '0.4.363';
+export const VERSION = '0.4.365';
 //# sourceMappingURL=constants.js.map

package/docs/CHANGELOG.md

@@ -6,6 +6,40 @@ alone does not tell you whether you were affected.
 edfcore is pre-1.0. Patch releases have carried behaviour changes where the old behaviour was a
 defect; those are called out below.
 
+## 0.4.365
+
+- **Added** a check that the layer table in `AGENTS.md` names the modules that are actually at each
+  layer. `module-layers.test.ts` already enforces that every module declares a layer, that no
+  runtime import goes up one, and that the table names every layer the source declares — but not
+  the contents of any row.
+- AGENTS.md calls the table "a summary of it, not a second definition", and a summary that
+  disagrees with its source is still wrong. This one is easy to disagree with, because two rows
+  split a directory: layer 5 is `io/` and layer 6 carves `io/read.ts` back out, exactly as layers 1
+  and 3 split `tal/`. So a directory claims everything under it except what another row names by
+  file, and both splits are asserted as the point of the rule rather than exceptions to it.
+- Row 7 names nothing — "entry points, and the pure helpers over them" — which supplies the
+  closure that makes the check tight: every module no row claims has to be at layer 7. Only a path
+  counts as a claim, since row 5 also mentions `ByteSource`, which is the type the adapters
+  implement rather than a module being placed.
+
+## 0.4.364
+
+- **Added** checks for three claims in the README's "Design in one page", each of which is an
+  absence and therefore hard to notice. There is no recording-wide rate — a header field holding
+  one would be the obvious convenience and can only ever be right for one channel, so its
+  non-existence is asserted alongside the three per-signal rates the README names.
+- `sampleRateHz` is `undefined` when the record duration is zero, which is legal EDF and which
+  PhysioNet's hypnograms really declare. `Infinity` is what an unguarded division gives, would pass
+  every `typeof` check, and would make `NaN` of every time converted through it — so the check is
+  that it is neither a number nor that number.
+- And there is no gap-filling option, checked as an absence across every option name the public
+  types declare. An option to fill a gap is an option to fabricate samples the amplifier never
+  recorded, and once it exists somebody's config turns it on.
+- Also pinned: that a window is measured on `onsetTicksFromFirstRecord` rather than `onsetTicks`.
+  The fixture puts the two axes in different records on purpose — a half-second offset leaves both
+  inside the same record-aligned chunk, which is how this advice would look correct while being
+  untestable.
+
 ## 0.4.363
 
 - **Corrected** the README's promise about diagnostics and widened 0.4.325's check to cover it. It

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.363",
+  "version": "0.4.365",
   "description": "Modern, typed, zero-dependency reader for EDF, EDF+, BDF and BDF+ biosignal files. Works in browsers and Node with true random access.",
   "keywords": [
     "edf",

package/src/constants.ts

@@ -93,4 +93,4 @@ export const SIGNAL_FIELD_BLOCK_OFFSETS = {
 } as const;
 
 /** Published package version. Kept in sync with package.json by a test. */
-export const VERSION = '0.4.363';
+export const VERSION = '0.4.365';