edfcore 0.4.518 → 0.4.520

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.
@@ -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 = "0.4.518";
112
+ export declare const VERSION = "0.4.520";
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 = '0.4.518';
82
+ export const VERSION = '0.4.520';
83
83
  //# sourceMappingURL=constants.js.map
package/docs/CHANGELOG.md CHANGED
@@ -6,6 +6,46 @@ 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
+ ## 0.4.520
10
+
11
+ - **Added** the header-helper agreement checks to the offline suite. `tests/corpus/whole-api.test.ts`
12
+ runs essentially the whole barrel over six real files and asserts the results agree with each
13
+ other; its own docblock says why that is a class of its own — a function can be individually
14
+ correct and still disagree with its neighbour, and six releases of this project were exactly
15
+ that. All of it skips without `npm run corpus:fetch`, so a fresh clone had never run any of it.
16
+ - What the corpus contributes to those properties is not realness. None of them cares where the
17
+ bytes came from; what they need is AWKWARDNESS — a zero record duration, a duplicated channel
18
+ label, a file with no data signal at all, 24-bit samples, a signal with no usable scale. The new
19
+ `tests/support/awkward-files.ts` builds eight such shapes, each named after what makes it
20
+ awkward rather than after the corpus file it stands in for.
21
+ - The properties are the ones that can only fail as a disagreement: `matchSignals` returning a
22
+ different set of channels than `header.dataSignalIndices`, `physicalRangeOf` ordering bounds the
23
+ signal does not declare, `getSignal` answering where `findSignals` found two, the covered
24
+ duration exceeding the span, `summarizeDiagnostics` counting a severity the array does not carry.
25
+ - A final case asserts the shapes really are awkward — that one has no data signals, one has no
26
+ rate anywhere, one repeats a label, one is 24-bit, one has an unscalable signal. Without it the
27
+ run could pass on eight tidy files and prove nothing about the shapes it is named for.
28
+
29
+ ## 0.4.519
30
+
31
+ - **Added** an offline check of the claim `api-helpers.md` makes for `streamRecords`: "a streamed
32
+ chunk and a read chunk are the same object in every respect, diagnostics included". It is what
33
+ lets a caller develop against `readWindow` and switch to streaming for the twenty-two-hour file
34
+ without touching their own downstream code.
35
+ - It was checked in exactly one place, `tests/corpus/large-file.test.ts`, which skips without
36
+ `npm run corpus:fetch` — so on a fresh clone nothing compared the two at all. And that comparison
37
+ is of samples, which is the half that survives a chunk losing its diagnostics, its
38
+ `precededByGap` or its byte accounting.
39
+ - "Every respect" is now taken literally: each streamed chunk is compared field for field with the
40
+ chunk `readRecords` returns for the same records, at four chunk sizes. The fixture makes the
41
+ usually-empty fields non-empty — a gap in the middle so one chunk carries a `precededByGap`, a
42
+ malformed TAL onset so one carries a diagnostic — because a chunk whose optional fields are all
43
+ absent compares equal to a chunk that dropped them, and the run asserts both were populated.
44
+ - The two claims in the same sentence are checked beside it: the chunks arrive in time order, and
45
+ none spans a gap at any chunk size. The second is what makes `durationTicks` mean what a reader
46
+ thinks — `readRecords` will hand back a chunk straddling a discontinuity, because you named the
47
+ records, and a streamed chunk must never be one.
48
+
9
49
  ## 0.4.518
10
50
 
11
51
  - **Added** the other convention `api-types.md` states in bold: "Anything checkable against the
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "edfcore",
3
- "version": "0.4.518",
3
+ "version": "0.4.520",
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 = '0.4.518';
96
+ export const VERSION = '0.4.520';