edfcore 0.2.33 → 0.2.34

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -6,6 +6,27 @@ alone does not tell you whether you were affected.
6
6
  edfcore is pre-1.0. Patch releases have carried behaviour changes where the old behaviour was a
7
7
  defect; those are called out below.
8
8
 
9
+ ## 0.2.34
10
+
11
+ - **Added** the golden-value harness this README has withheld a numerical-interop claim for since
12
+ 0.1.0. edfcore now **reproduces pyEDFlib's float64 physical values bit for bit** on EDF and
13
+ 24-bit BDF, across symmetric and asymmetric declared ranges. `scripts/golden/generate.py` writes
14
+ the fixtures with pyEDFlib's own writer, reads them back with pyEDFlib, and records every sample
15
+ as its exact IEEE-754 bit pattern; the test compares with `Object.is`, so one ULP is a failure.
16
+ Nothing under `tests/corpus/golden/` was produced by edfcore, and the goldens are committed so
17
+ CI never needs Python.
18
+
19
+ Every previous test of the pinned scaling expression re-derived that same expression inside the
20
+ test, which proved edfcore agrees with itself and nothing more. This is the first evidence that
21
+ the choice to keep EDFlib's numerically worse form was worth making: substituting the textbook
22
+ `physicalMinimum + (digital - digitalMinimum) * gain` fails on 140 of 256 samples of the
23
+ symmetric fixture — `-492.15686274509807` where pyEDFlib says `-492.156862745098` — and a test
24
+ asserts that the goldens can tell the two apart, so the parity is a real constraint rather than
25
+ a coincidence of the fixtures.
26
+
27
+ Still not claimed, because no test produces it yet: parity with MNE, and validation across the
28
+ public corpora.
29
+
9
30
  ## 0.2.33
10
31
 
11
32
  - **Fixed** annotations being dropped with no diagnostic naming them. `TIMEKEEPING_TAL_NONCONFORMANT`
package/README.md CHANGED
@@ -291,10 +291,17 @@ the reason to trust the reader.
291
291
  **0.1 — reads everything, lies about nothing.** All six variants, the primitive/I/O/convenience
292
292
  layers, `inspectEdf`, `edfcore/validate`, six source adapters, the full diagnostic vocabulary.
293
293
 
294
- **Before 1.0.** Validation against public corpora (sleep-edfx, CHB-MIT, the BioSemi and
295
- edfplus.info test files) and a golden-value harness cross-checking physical values against
296
- pyEDFlib and MNE. Until that harness exists this README makes no numerical-interop claim, and it
297
- will not make one that a test did not produce.
294
+ **Numerical interop, as of 0.2.34.** edfcore reproduces **pyEDFlib's float64 physical values bit
295
+ for bit** on EDF and 24-bit BDF, across symmetric and asymmetric ranges. That is a test, not a
296
+ belief: `scripts/golden/generate.py` writes the fixtures with pyEDFlib's own writer, reads them
297
+ back with pyEDFlib, and records every sample as its exact IEEE-754 bit pattern;
298
+ `tests/corpus/golden-values.test.ts` compares with `Object.is`, so one ULP is a failure. Nothing in
299
+ `tests/corpus/golden/` was produced by edfcore. Substituting the numerically better textbook
300
+ expression fails it on 140 of 256 samples — which is why the EDFlib form is pinned.
301
+
302
+ **Still open before 1.0.** The same harness against MNE, and validation across public corpora
303
+ (sleep-edfx, CHB-MIT, the BioSemi and edfplus.info test files). Those claims are not made here
304
+ until a test produces them.
298
305
 
299
306
  **Shipped since 0.1.6.** Min/max envelope decimation (`readEnvelope`,
300
307
  `readEnvelopeAtResolution`). BioSemi Status-byte helpers (`readTriggers`). Streaming iteration
@@ -109,5 +109,5 @@ export declare const SIGNAL_FIELD_BLOCK_OFFSETS: {
109
109
  readonly reserved: 224;
110
110
  };
111
111
  /** Published package version. Kept in sync with package.json by a test. */
112
- export declare const VERSION = "0.2.33";
112
+ export declare const VERSION = "0.2.34";
113
113
  //# sourceMappingURL=constants.d.ts.map
package/dist/constants.js CHANGED
@@ -79,5 +79,5 @@ export const SIGNAL_FIELD_BLOCK_OFFSETS = {
79
79
  reserved: 224,
80
80
  };
81
81
  /** Published package version. Kept in sync with package.json by a test. */
82
- export const VERSION = '0.2.33';
82
+ export const VERSION = '0.2.34';
83
83
  //# sourceMappingURL=constants.js.map
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "edfcore",
3
- "version": "0.2.33",
3
+ "version": "0.2.34",
4
4
  "description": "Modern, typed, zero-dependency reader for EDF, EDF+, BDF and BDF+ biosignal files. Works in browsers and Node with true random access.",
5
5
  "keywords": [
6
6
  "edf",
package/src/constants.ts CHANGED
@@ -93,4 +93,4 @@ export const SIGNAL_FIELD_BLOCK_OFFSETS = {
93
93
  } as const;
94
94
 
95
95
  /** Published package version. Kept in sync with package.json by a test. */
96
- export const VERSION = '0.2.33';
96
+ export const VERSION = '0.2.34';