Skip to content

feat(fft): preserve capture timestamps and absolute-range editing context (#44) - #6

Merged
KyleHuangFromHZ merged 3 commits into
mainfrom
feature/44-real-time-display
Nov 20, 2026
Merged

KyleHuangFromHZ merged 3 commits into
mainfrom
feature/44-real-time-display

Conversation

@KyleHuangFromHZ

@KyleHuangFromHZ KyleHuangFromHZ commented Nov 19, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Related to crysound/OpenTest#44. FFT records retain trustworthy acquisition timestamps for spectrum/spectrogram frames and pass the mounted chart instance to the host date-range editor.

FFT capture window -> result timestamp -> host clock display/export
workspace Settings -> mounted chart context -> absolute range editor

Relative timing remains available when a source timestamp is absent.

Evidence

  • Before: historical FFT results lacked source-clock metadata, and workspace settings dropped chart context. After: Windows FFT waveform/spectrogram/Waterfall support clock labels, date ranges and exports with the matched host.
  • FFT backend validation previously passed 255 tests; editor-context regression passed five tests and frontend type-check passed. The host/nine-plugin combination built and ran on windows-huang via opentest-run.
  • Host verification: feat(time-axis): complete acquisition time ranges, exports and reports (#44) OpenTest#66 and its docs/verification/issue-44-absolute-range.md / issue-44-review-fixes.md records.
  • Synced current main without conflicts; SonoDAQ runtime changes remain on main and are not introduced by this PR.

Merge Danger

Door: two-way

Blast Radius: FFT

Use the matching OpenTest #66 host SDK/runtime. No hardware driver source is changed. Issue #44 remains open.

@KyleHuangFromHZ
KyleHuangFromHZ requested a review from a team November 19, 2026 02:55
@KyleHuangFromHZ KyleHuangFromHZ changed the title feat: preserve acquisition times in FFT history (#44) feat(fft): preserve capture timestamps and absolute-range editing context (#44) Nov 20, 2026
@KyleHuangFromHZ
KyleHuangFromHZ merged commit 2d198a7 into main Nov 20, 2026
5 checks passed
@KyleHuangFromHZ
KyleHuangFromHZ deleted the feature/44-real-time-display branch November 20, 2026 05:25
KyleHuangFromHZ added a commit to crysound/OpenTest that referenced this pull request Nov 20, 2026
#44) (#66)

## Summary

Related to #44. Complete acquisition-time display, absolute date ranges,
copy/data export and independent report viewports.

```text
trusted capture timestamp -> clock axes / copy / CSV / HDF5 / MAT / reports
s <-> h:m:s             -> auto X range; retain samples and capture origin
fixed calendar range    -> same absolute instants across records
```

- Shared local date/time editor uses decimal 0–1000ms input; no separate
time-zone label. Unknown origins retain seconds; legacy numeric clock
ranges reset to auto.
- Spectrogram/Waterfall use trustworthy frame timestamps. Reports and
archives retain their own viewports; source preferences cannot overwrite
them.
- Unit switches update the existing chart and do not restart Monitor or
briefly show missing clock ticks.

## Evidence

- **Before:** RMS/Peak unit changes disposed the graph and displayed
missing clock labels for about 0.8s. **After:** Windows frame sampling
on waveform/RMS/Peak and historical waveform/Waterfall retained chart
instances and origins with zero missing clock labels.
- Windows opentest-run instance `opentest-44` on `windows-huang`: actual
date/ms edits, repeated measurements, 96000-point CSV comparison,
archived report preview and native Excel export verified. Earlier
HDF5/MAT exports checked every timestamp/clock entry; hardware and
controlled-input limits are recorded separately.
- After syncing main: 548 selected frontend tests, 60 selected Java
tests, frontend test type-check and docs-policy passed. The affected
CI-fixture rerun passed 56 frontend tests and 13 reader tests; full CI
is rerunning on the latest push.
- Final integrated Windows release `20261120T033000Z-bc160024` used
clean host `6bacc40a`, FFT `956e072f` and internal plugins `07d53bce`;
the time-axis implementation is unchanged by later test-fixture repairs
and the independent main reanalysis-path sync.
- Synced main through `69714f2b`; its affected reanalysis/reader checks
passed 59 Java tests, and desktop unit checks passed nine tests with two
Windows-only checks skipped locally.
- Detailed evidence: [absolute
ranges](docs/verification/issue-44-absolute-range.md), [export
repairs](docs/verification/issue-44-review-fixes.md).

## Merge Danger

**Door:** two-way

**Blast Radius:** charts/exports/reports

The optional absolute-range fields preserve seconds compatibility.
Reading legacy clock-only numeric fixed ranges intentionally resets
those ranges to auto. Pair with public FFT PR
crysound/OpenTest-Plugins#6 and the matching internal plugin PR for
source timestamps and mounted editor context. This PR does not close or
change issue #44.

---------

Co-authored-by: 黄凯 <huangkai@huangkaideMacBook-Air.local>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants