edfcore 0.6.74 → 0.6.76
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/dist/constants.d.ts +1 -1
- package/dist/constants.js +1 -1
- package/docs/CHANGELOG.md +28 -0
- package/package.json +1 -1
- package/src/constants.ts +1 -1
package/dist/constants.d.ts
CHANGED
|
@@ -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 = "0.6.
|
|
112
|
+
export declare const VERSION = "0.6.76";
|
|
113
113
|
//# sourceMappingURL=constants.d.ts.map
|
package/dist/constants.js
CHANGED
package/docs/CHANGELOG.md
CHANGED
|
@@ -6,6 +6,34 @@ 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
|
+
## 0.6.76
|
|
10
|
+
|
|
11
|
+
- **Fixed** the inspector placing each sample at `part.startSeconds + i / rate`, over a rate it
|
|
12
|
+
derived itself as `samplesPerRecord / (recordDurationSeconds || 1)`.
|
|
13
|
+
- That is mistake 3 on `AGENTS.md` — "Do not compute sample indices from `sampleRateHz`. It is
|
|
14
|
+
derived and can be `undefined`" — and the reason `sample-grid.ts` exists: "Every viewer needs
|
|
15
|
+
this and the obvious spelling is wrong." A record duration that does not divide leaves the rate
|
|
16
|
+
irrational in binary (128 samples over 0.3 s is 426.666…), so the trace walks off its own time
|
|
17
|
+
axis as `i` grows. The `|| 1` beside it invented a rate for the legal zero record duration,
|
|
18
|
+
where the library reports none.
|
|
19
|
+
- It now uses `gridSampleStartSeconds`, measured from the chunk's own start so a gap does not
|
|
20
|
+
shift the trace — each chunk is one contiguous run. The test measures the disagreement on that
|
|
21
|
+
awkward duration rather than asserting the call.
|
|
22
|
+
|
|
23
|
+
## 0.6.75
|
|
24
|
+
|
|
25
|
+
- **Fixed** the inspector deciding whether a file needs a complete index by comparing
|
|
26
|
+
`spanSeconds` with `coveredSeconds`. `EdfTimeline` says which fields carry that verdict:
|
|
27
|
+
"`spanTicks === coveredTicks` is the two-probe contiguity verdict, and the ticks are what it
|
|
28
|
+
must be asked of", because the seconds are float64 conversions and two tick counts can round to
|
|
29
|
+
one float once the span is large enough that an ulp exceeds a tick.
|
|
30
|
+
- Asked of the floats, such a file answers "contiguous", `buildRecordIndex` is skipped, and the
|
|
31
|
+
page draws a discontinuous recording on the nominal grid — the silently wrong timeline the
|
|
32
|
+
library exists to refuse, on the page that is its central claim made checkable. The same
|
|
33
|
+
comparison was fixed inside `resolveTimeWindow` in 0.3.4 and the type's docblock records it.
|
|
34
|
+
- The test demonstrates the collision rather than arguing it, and shows the two staying apart at
|
|
35
|
+
an ordinary scale, which is why this could sit there.
|
|
36
|
+
|
|
9
37
|
## 0.6.74
|
|
10
38
|
|
|
11
39
|
- **Fixed** `sample-locate.ts` describing a repeated record onset as something "edfcore reports
|
package/package.json
CHANGED
package/src/constants.ts
CHANGED