edf2csv 0.5.124 → 0.5.126

package/CHANGELOG.md

@@ -3,6 +3,61 @@
 Notable changes to edf2csv. Versions follow [semantic versioning](https://semver.org); while the
 major version is 0, a minor bump may contain breaking changes.
 
+## 0.5.126
+
+### Fixed: the metadata.json the page explains was a different recording's
+
+output-files prints a whole `metadata.json` as its explanation of the format — the transcript
+someone reads before writing code against those fields. Its `source.path` ends in
+`sleep-study.edf`, which on this site is one specific recording: eight hours of five channels
+at 100, 10 and 1 Hz with an annotation channel.
+
+The sample described three channels at 256, 128 and 1 Hz. 39 MB against 18.7. 412 events
+against 7. Three channel rows against 5. `signal_count` 4 against 6. Started in March 2026
+rather than March 2002. One warning where that conversion raises two — and the field named
+`notes` is the one a reader consults to find out what a warning looks like in the record.
+
+The guard that existed declined to check any of it, on the stated grounds that "the sample
+describes an 8-hour sleep study that is not in this repository". It has been in this repository
+since 0.4.68, as `test/fixtures/sleep-study.mjs`, and is what every `--info` block on the site
+is generated from. The values are now compared against a real conversion of it; the run's own
+fields — the tool version, where the file sat, when it was converted — stay illustrative.
+
+The same page ran `edf2csv sleep-study.edf` over a directory listing of `signals_256hz.csv`,
+`signals_128hz.csv` and `signals_1hz.csv`. That is `recording.edf`, the three-rate example
+sampling-rates uses, and it is now called that.
+
+## 0.5.125
+
+### Fixed: the FAQ's --json example was another recording's summary, warnings included
+
+"How do I check whether a conversion had problems from a script?" runs `--json` on
+`sleep-study.edf` and showed:
+
+```json
+{
+  "files": [
+    { "name": "signals.csv", "rows": 921600 },
+    { "name": "annotations.csv", "rows": 3 },
+    { "name": "channels.csv", "rows": 1 }
+  ],
+  "annotations": 3,
+  "duration_seconds": 3600,
+  "records": 3600,
+  "warnings": []
+}
+```
+
+Every figure belongs to some other file. That recording is eight hours of five channels at
+three rates: three signals files, seven events, five channel rows, 28,800 records. And it
+raises two warnings — the paragraph directly below promises they will be in that array, and
+the same page prints one of them, five answers earlier, about this same file. Someone comparing
+their own output against this block would conclude their conversion had gone wrong.
+
+The block is now what that command writes, `warnings` included, and a test converts the
+recording and holds every `--json` block on the site to it. `elapsed_ms` and `output_dir`
+belong to the run rather than to the recording and stay illustrative.
+
 ## 0.5.124
 
 ### Fixed: a stray space meant four things depending on which option carried it

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edf2csv",
-  "version": "0.5.124",
+  "version": "0.5.126",
   "description": "Convert EDF, EDF+ and BDF biosignal recordings (European Data Format) to CSV from the command line. Local, streaming, and never resamples or alters units.",
   "keywords": [
     "edf",