Persist Area Detector viewer state via ConfigSource (TOML ↔ DB round-trip)

[pecomyint]

Problem

All Area Detector viewer toggle states — colormap, autoscale, log scale, crosshair, threshold, freeze, display ROIs, transpose, pixel ordering — are hardcoded defaults that reset on every restart. Users lose their preferred view configuration each session.

Scope — Tier 1: Portable Preferences Only

This issue covers semantic viewer preferences that should round-trip through TOML export/import. Machine-local layout state (window geometry, dock positions, last-typed prefix) is handled separately by #85 via QSettings.

See State Persistence Architecture for the full two-tier design.

Decision rule

"Would I want this setting to travel with my TOML config to another machine?"

• Yes → this issue (ConfigSource)
• No → Persist machine-local layout state via QSettings (Tier 2) #85 (QSettings)

What to persist (Tier 1)

Key	Type	Default	Description
colormap	str	"viridis"	Selected gradient from color bar context menu
autoscale	bool	true	Auto-scale toggle (5%-95% histogram)
log_scale	bool	false	Logarithmic intensity scale
crosshair_visible	bool	false	Crosshair overlay toggle
threshold_enabled	bool	false	Auto threshold toggle
freeze_image	bool	false	Lock/Freeze image toggle
display_rois	bool	true	ROI overlay visibility
transpose	bool	false	Image transpose
pixel_order	str	"C"	Pixel ordering (C or Fortran)
plotting_frequency	int	10	Stats plotting refresh rate

TOML representation

# New section — existing sections unchanged
[viewer_state.area_detector]
colormap = "viridis"
autoscale = true
log_scale = false
crosshair_visible = false
threshold_enabled = false
freeze_image = false
display_rois = true
transpose = false
pixel_order = "C"
plotting_frequency = 10

Implementation

New functions in settings.py

Follow the existing save_detector_prefix() / save_input_channel() pattern:

def load_viewer_state(viewer_key: str = "area_detector") -> dict:
    cfg = ConfigSource(_get_effective_locator()).load()
    return cfg.get("viewer_state", {}).get(viewer_key, {})

def save_viewer_state(viewer_key: str, state: dict) -> bool:
    src = ConfigSource(_get_effective_locator())
    existing = src.load()
    vs = existing.setdefault("viewer_state", {})
    vs[viewer_key] = state
    return src.save({"viewer_state": vs})

Wire into area detector viewer

• On init: call load_viewer_state("area_detector") to set checkbox/colormap values instead of hardcoded defaults
• On each toggle change (existing signal handlers like autoscale_checked(), threshold_checked(), etc.): call save_viewer_state() with updated values
• On closeEvent: bulk save as safety net

The round-trip

• User imports a TOML (with [viewer_state.area_detector]) → it flows into the local DB profile, viewer opens with those preferences
• User changes a toggle in the UI → DB profile is updated immediately via ConfigSource.save()
• User exports → export_profile_to_toml() includes viewer state, producing a complete portable config
• User shares the TOML → recipient loads it and gets detector config AND viewer preferences

The database (dashpva.db) stays local and is not shared. The TOML file is the portable artifact.

Acceptance criteria

• load_viewer_state() and save_viewer_state() added to settings.py
• Area detector viewer reads initial state from config instead of hardcoded defaults
• Toggle changes persist through ConfigSource.save() on each state change
• Viewer state survives application restart
• TOML export includes [viewer_state.area_detector] section
• TOML import with viewer state correctly populates the DB profile and UI
• No changes to ConfigSource internals needed — viewer_state is just another dict key

Related

• Persist machine-local layout state via QSettings (Tier 2) #85 — Tier 2: Machine-local layout state via QSettings