edfcore 0.5.5 → 0.5.7
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 +41 -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 = "0.5.
|
|
112
|
+
export declare const VERSION = "0.5.7";
|
|
113
113
|
//# sourceMappingURL=constants.d.ts.map
|
package/dist/constants.js
CHANGED
package/docs/CHANGELOG.md
CHANGED
|
@@ -6,6 +6,47 @@ 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
|
+
## 0.5.7
|
|
10
|
+
|
|
11
|
+
- **Added** the one worked `readWindow` result on `api-reading.md`, built and run. The page shows a
|
|
12
|
+
65-second window over a discontinuous file coming back as two chunks — `{ start: 498, count: 2 }`
|
|
13
|
+
at 498 s and `{ start: 500, count: 3 }` at 560 s with a gap before it — and it is the only place
|
|
14
|
+
the page shows what "one chunk per contiguous run" produces. Every figure in it was prose.
|
|
15
|
+
- The counts are the part worth running. 2 and 3 are not the same number and neither is 5, so a
|
|
16
|
+
reader can see from them that the window did not return every record between its ends. Both fall
|
|
17
|
+
out of the half-open rule: record 499 ends exactly at 500 s and is in, record 503 starts exactly
|
|
18
|
+
at 563 s and is out — the rule that decides whether consecutive epochs partition or overlap. The
|
|
19
|
+
run asserts the excluded record starts exactly at the window's end.
|
|
20
|
+
- `startSeconds: 560` is the other half: sixty seconds of the window lie inside the gap, so the
|
|
21
|
+
second chunk starts 62 seconds after the first while being only 2 records later. A consumer that
|
|
22
|
+
assumed `startSeconds` advanced with the record number would place three records a minute early.
|
|
23
|
+
- The window and both chunks are parsed out of the page, anchored to the printed line rather than
|
|
24
|
+
to the first `startSeconds` on it — an unanchored pattern picked up an earlier example's window
|
|
25
|
+
and built one nothing on the page describes, which passed its own parse and tested a different
|
|
26
|
+
file.
|
|
27
|
+
|
|
28
|
+
## 0.5.6
|
|
29
|
+
|
|
30
|
+
- **Fixed** two lines of the `EdfSignal` block on `api-types.md` describing a different file from
|
|
31
|
+
the other five. The page prints one running example — six one-second records, one 100-sample
|
|
32
|
+
signal — and its test builds that file and checks `index`, `kind`, `samplesPerRecord`,
|
|
33
|
+
`sampleRateHz` and `sampleCount`. It stopped there, and the two lines it stopped before are the
|
|
34
|
+
two that would have failed: `signal.scale` prints a bitValue of `0.1221001221001221` and
|
|
35
|
+
`signal.raw.digitalMinimum` prints `'-2048 '`, while the fixture took the writer's defaults and
|
|
36
|
+
produced `0.015259021896696421` and `'-32768 '`.
|
|
37
|
+
- The page is the specification and the fixture is the reconstruction, so the fixture moved: the
|
|
38
|
+
printed values imply 500 µV over 4,095 digital steps and nothing else, and the conventions
|
|
39
|
+
paragraph at the top of the page introduces the "exposed twice" rule with the same `-2048` as its
|
|
40
|
+
example. Declaring `-250`..`250` over `-2048`..`2047` makes every line of the block true of one
|
|
41
|
+
file, and makes the sentence at the top true of it too.
|
|
42
|
+
- `bitValue` is re-derived through the pinned expression from the declared fields rather than
|
|
43
|
+
compared against the printed float. It is the number a reader checks their own port against, and
|
|
44
|
+
copying the page's sixteen digits into the test would make the two agree without either being
|
|
45
|
+
right.
|
|
46
|
+
- The whole block is now under test, including the raw field with its padding intact — compared
|
|
47
|
+
against the page without collapsing whitespace, since collapsing turns `'-2048 '` into
|
|
48
|
+
`'-2048 '` and quietly passes the check the sentence exists to make.
|
|
49
|
+
|
|
9
50
|
## 0.5.5
|
|
10
51
|
|
|
11
52
|
- **Added** the table under "Files with several sample rates" on `reading-signals.md`, and the
|
package/package.json
CHANGED
package/src/constants.ts
CHANGED