[stevevanhooser]

Mirrors VH-Lab/NDI-matlab#1079 (merged).

What

• Progress bars: ndi_session.ingest / ndi_daq_system.ingest report progress through ndi_gui_component_ProgressBarWindow, with one session bar and one bar per DAQ system.
  • On by default; ShowProgress=False turns it off.
  • ProgressBar= / ProgressBarTag= reuse an existing window.
• Parallel per-epoch work: the per-epoch reader and metadata-reader work moves to the new ndi.daq.fun.ingestEpochDocuments. It runs on spawned worker processes when the Parallel.* preferences allow.
  • Database writes stay serial and in epoch order.
  • Readers are sent to workers with _session cleared; a reader that can't be pickled falls back to serial with a warning.
• Supporting ports:
  • ProgressBarWindow gets MATLAB's silent (headless) mode (Silent, isHeadless, silentModeDefault, resolveSilentMode) and per-bar Timeout.
  • New Parallel.OpenPoolIfNone / UseDefaultNumberOfWorkers / NumberOfWorkers preferences and ndi.fun.parallelWorkers. Python has no persistent pool, so the default (OpenPoolIfNone=false) is serial.
• Bridge entries are updated for ndi.daq.system, ndi.session, ndi.preferences, ndi.fun.parallelWorkers and ProgressBarWindow, and added for ndi.daq.fun.ingestEpochDocuments.
• Adds tests/test_ingest_progress_parallel.py (20 tests).

Notes

• Stacked on top: the data writer PR, which mirrors NDI-matlab#1089.
• CI: main has been red since Navigator: DAQ system browser, cloud dataset id, Assign to workspace; faster daqsystem_load #344 on bridge drift from NDI-matlab main (+ndi/+update, +ndi/+gui/+nav, +ndi/+setup, +ndi/+fun/+epoch). That drift is unrelated to this change and will show here too.

---

Generated by Claude Code

[stevevanhooser]

MATLAB/Python bridge completeness fails, but not because of this PR.

It reports four NDI-matlab files with no bridge entry:

• +ndi/+gui/+nav/updateDialog.m
• +ndi/+update/packageStatus.m
• +ndi/+update/private/findCalculators.m
• +ndi/+update/private/pullRepo.m

These were added on NDI-matlab main with the update-dialog work (e.g. 749b1a266). They aren't touched here, and main here has been red since #344. Every daq, session, preferences and ProgressBarWindow entry this PR changes passes.

No fix exists yet. The fix is bridge entries for those four files in src/ndi/gui and src/ndi/update (ported, or not_yet_ported with a decision log). That belongs in its own PR rather than this one. Re-running won't help: the result is deterministic.

---

Generated by Claude Code

[stevevanhooser]

I merged main into this branch (633b17f). The only conflict was in CHANGELOG.md, where #347 and this PR each added an entry; both are kept.

MATLAB/Python bridge completeness will still be red, now for a different reason. #347 fixed the update-dialog entries. What remains is hash drift on two entries:

• ingestEpochDocuments
• mfdaq_reader

NDI-matlab #1090 has since changed both MATLAB files. The port of #1090 is in #346, which is stacked on this PR and moves both entries to NDI-matlab main.

Locally, against NDI-matlab main, everything else passes: 6502 passed, 352 skipped. The checks go green once #346 merges after this one.

---

Generated by Claude Code