|
Tracks a user-selected point via webcam (OpenCV CSRT tracker) and
streams x_pos / y_pos pixel coordinates as live channels into the
acquisition pipeline. Falls back to a Lissajous figure simulation
when opencv-python is not installed, so the plugin works immediately
without hardware.
- tracker.py: background thread, real camera + sim modes, thread-safe API
- device.py: BaseDevice reading x/y from tracker, wires into engine
- filter.py: custom pixel_to_mm filter (registered in FILTER_CLASSES)
- window.py: floating tool window, live feed, click-to-track ROI selection
- plugin.py: LabPlugin wiring all pieces together, toolbar button,
settings widget, profile save/restore
Also: PluginAction.button_ref_callback lets plugins store a ref to their
toolbar button so they can uncheck it when their window closes.
opencv-python install note added to requirements.txt.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
|
Plugins live in plugins/<name>/ with a manifest.json and a LabPlugin
subclass. Enabled/disabled in Settings > Plugins tab; state persists in
plugins/enabled.json and in .labdaq profiles (loading a profile
enables/disables plugins to match). Plugins can contribute toolbar
buttons, BaseDevice instances (auto-wired into acquisition pipeline),
custom signal filter classes, a settings widget, and profile state via
get_save_state/apply_save_state. Load errors now report the exact
failing stage (file not found, class missing, on_load crash).
Includes example plugin skeleton and docs/plugin-development.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
