edfcore 0.2.56 → 0.2.58

package/CHANGELOG.md

@@ -6,6 +6,37 @@ 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.
 
+## 0.2.58
+
+- **Added** CHB-MIT to the corpus, closing a gap the README has named since 0.1. It is chosen to be
+  UNLIKE sleep-edfx rather than to add volume: 23 channels at a uniform 256 Hz in one-second
+  records, recorded in 2010 at another institution on other equipment, against 7 channels at mixed
+  rates in 30-second records from 1989. Every real file in this suite came from one dataset until
+  now. Same bit-for-bit parity with pyEDFlib across all 23 channels.
+- It also supplies something no fixture in this project had: **a montage that names one derivation
+  twice.** `T8-P8` appears at index 14 and index 22, and the two carry identical samples — verified
+  against pyEDFlib over the whole hour rather than assumed; I had written the test expecting them
+  to differ. `EdfAmbiguousChannelError` had until now only ever been raised against a fixture
+  written to raise it.
+- That identity makes the refusal matter more, not less. If `getSignal` picked one arbitrarily, no
+  comparison of the returned numbers could reveal which it picked, so a caller would never learn
+  the question had two answers. The error is the only signal there is.
+- Downloaded on demand under the Open Data Commons Attribution License v1.0, hash-verified, never
+  redistributed — the same terms as every other corpus entry.
+
+## 0.2.57
+
+- **Documented** what the corpus work of 0.2.49-0.2.55 actually established. The README's
+  interop section stopped at sample parity, which was the state at 0.2.48 and has not been the
+  state since — annotation onsets, the zero-record-duration scoring file, the 1985-2084 year rule,
+  faithful decimation over 7,950,000 samples, chunk-independence, the memory bound, streaming
+  equivalence, HTTP random access costing under 64 KB for a window twelve hours into a 48 MB file,
+  and `validate` exiting 0 on a real clinical recording were all verified and none of them were
+  claimed. A reader deciding whether to trust this package should not have to reconstruct that
+  from a changelog.
+- Each row names how it is checked rather than only what it asserts, because "we test that" is the
+  kind of claim this project has spent forty releases learning not to make without saying how.
+
 ## 0.2.56
 
 - **Fixed** `tests/README.md`, which said "There are no binary fixtures in this repository" and

package/README.md

@@ -310,14 +310,39 @@ establish.
 chose: a **22-hour clinical polysomnogram** from sleep-edfx, and the teuniz generator files in EDF,
 EDF+ and 24-bit BDF+. Sampled at the start, the middle and the end of every signal — the end window
 being the one that catches record arithmetic drifting with distance, which a sample near record 0
-cannot. Those tests need the corpus and skip without it:
+cannot.
+
+What else those files establish, which synthetic fixtures cannot:
+
+| Claim | Checked how |
+|---|---|
+| Annotation onsets match pyEDFlib | 154 sleep stages from a real scoring file, to the tick |
+| A legally **zero** record duration works | that same file — where `sampleRateHz` is `undefined` and any rate-derived expression yields `NaN` |
+| The 1985–2084 year rule is right | a 1989 recording, against a reader that implements it independently |
+| Decimation keeps every extreme | 7,950,000 samples reduced to 1,000 buckets, compared with an exhaustive reduction |
+| The bucket grid ignores chunk size | 265 chunks versus a handful |
+| Memory is bounded by the chunk | that 22-hour envelope produced under a 512 KiB budget |
+| Streaming equals reading | 42 streamed chunks concatenated and compared sample by sample |
+| **Random access is real** | a 30-second window twelve hours in costs **under 64 KB** over HTTP |
+| `validate` is usable as a CI gate | exits 0 on a real recording from a real sleep lab |
+| The sample scan sees what pyEDFlib sees | whole-signal digital extremes, all five files |
+
+Those tests need the corpus and skip without it. `coverage.test.ts` always runs and says which
+state you are in, because a skipped test and a passing one look identical in a summary line:
 
 ```bash
 npm run corpus:fetch
 ```
 
-**Still open before 1.0.** CHB-MIT, and the edfplus.info conformance files. Those are not claimed
-here until a test produces them.
+**CHB-MIT, as of 0.2.58.** A second real clinical recording, deliberately unlike the first: 23
+channels at a uniform 256 Hz in one-second records, recorded in 2010 at another institution, where
+sleep-edfx is 7 channels at mixed rates in 30-second records from 1989. Same bit-for-bit parity. It
+also supplies something no synthetic fixture here had — a montage that names one derivation twice,
+which is what `EdfAmbiguousChannelError` exists for and had until now only ever been raised against
+a fixture written to raise it.
+
+**Still open before 1.0.** The edfplus.info conformance files. Not claimed here until a test
+produces them.
 
 **Shipped since 0.1.6.** Min/max envelope decimation (`readEnvelope`,
 `readEnvelopeAtResolution`). BioSemi Status-byte helpers (`readTriggers`). Streaming iteration

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 = "0.2.56";
+export declare const VERSION = "0.2.58";
 //# 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 = '0.2.56';
+export const VERSION = '0.2.58';
 //# sourceMappingURL=constants.js.map

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.2.56",
+  "version": "0.2.58",
   "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 = '0.2.56';
+export const VERSION = '0.2.58';