edfcore 0.4.428 → 0.4.430
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 +36 -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 = "0.4.
|
|
112
|
+
export declare const VERSION = "0.4.430";
|
|
113
113
|
//# sourceMappingURL=constants.d.ts.map
|
package/dist/constants.js
CHANGED
package/docs/CHANGELOG.md
CHANGED
|
@@ -6,6 +6,42 @@ 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
|
+
## 0.4.430
|
|
10
|
+
|
|
11
|
+
- **Added** tests for a patient's date of birth being redacted out of a message that never quotes
|
|
12
|
+
the field. `redactDiagnostic` withholds identification by substituting the field's value out of
|
|
13
|
+
the message by text, and `DATE_IMPLAUSIBLE` defeats every spelling of it: the field says
|
|
14
|
+
`02-MAY-2050`, the message says `2050-05-02`, because it is comparing two dates and prints both
|
|
15
|
+
in one form. No substring of the raw field appears in that sentence. It is redacted anyway, from
|
|
16
|
+
`actual` — which carries whatever the message chose to print.
|
|
17
|
+
- It is also the case with the widest reach. The other identification diagnostics fire on a
|
|
18
|
+
malformed field, so a conformant file never produces them; this one fires on a perfectly
|
|
19
|
+
conformant patient field whose only fault is a year the two-digit header rule resolved into the
|
|
20
|
+
future, which is what a recording made before 1985 or after 2084 looks like. A clinical file with
|
|
21
|
+
nothing wrong with it, printed by a command asked to withhold the patient, and a date of birth in
|
|
22
|
+
the output.
|
|
23
|
+
- The other half is what must survive. Substitution is on the value, never on the code or the rule,
|
|
24
|
+
so the code, the byte offset, the field name, the spec clause and the recording's own start date
|
|
25
|
+
are all still there — and a field nobody asked to withhold is still quoted verbatim.
|
|
26
|
+
|
|
27
|
+
## 0.4.429
|
|
28
|
+
|
|
29
|
+
- **Added** tests for `edfcore validate` exiting 1 because validation failed. The exit-code table on
|
|
30
|
+
`cli.md` gives `1` two meanings in one row — "the file could not be read, or validation failed" —
|
|
31
|
+
and only the first was exercised. `runCli` returns `report.ok ? 0 : 1`, and nothing had ever
|
|
32
|
+
driven it down the `false` side, which is the entire reason the command exists.
|
|
33
|
+
- A CI job gating on conformance branches on that number without reading a word of the output, and
|
|
34
|
+
it can fail in two opposite ways. A gate that never fires passes every recording, including the
|
|
35
|
+
ones the library refused to scale, and nobody investigates a green tick. A gate that always fires
|
|
36
|
+
gets switched off, and the conformance checking goes with it.
|
|
37
|
+
- The second is likelier, and it turns on a boundary now stated as a subject: a warning is not a
|
|
38
|
+
failure. `LABEL_CONVENTION_NONCONFORMANT` is on almost every real recording, CHB-MIT ships a
|
|
39
|
+
duplicated channel label, a file marked EDF+C whose onsets drift is a warning because real
|
|
40
|
+
writers do that, and a zero record duration is legal EDF. A gate rejecting any of those would
|
|
41
|
+
reject the corpus this library was built to read.
|
|
42
|
+
- `header` on the same refused file exits 0, and that contrast is checked. It reports what the
|
|
43
|
+
header says and adjudicates nothing, which is what makes asking for a verdict worth doing.
|
|
44
|
+
|
|
9
45
|
## 0.4.428
|
|
10
46
|
|
|
11
47
|
- **Added** a check that a version number which never reached npm says so at the heading and says
|
package/package.json
CHANGED
package/src/constants.ts
CHANGED