BrkRaw Viewer is an interactive dataset viewer implemented as a separate CLI plugin for the brkraw command.

The viewer is intentionally maintained outside the BrkRaw core to enable independent development and community contributions around user-facing interfaces.

Scope and intent

BrkRaw Viewer is designed for interactive inspection of Bruker Paravision datasets. It focuses on quick exploration and validation rather than data conversion or analysis.

The goal is to provide practical, researcher-focused features that are useful in everyday workflows, such as quick dataset triage, metadata checks, and lightweight visual QC.

Typical use cases include:

• Browsing studies, scans, and reconstructions
• Verifying scan and reconstruction IDs
• Inspecting acquisition metadata before conversion
• Lightweight visual sanity checks

All data conversion and reproducible workflows are handled by the BrkRaw CLI and Python API.

Why these features exist

Viewer The Viewer tab makes it easy to confirm the right scan and orientation before running a larger workflow.

Registry The Registry reduces repeated filesystem navigation and lets you re-open the current session with a single menu action.

Extensions/hooks Extensions allow modality-specific panels (MRS, BIDS, etc.) to live outside the core viewer so the default install stays lightweight.

Design goal: shared extensibility

brkraw-viewer keeps the BrkRaw design philosophy: extend the ecosystem without changing core logic. The viewer uses the same rules/spec/layout system as the CLI and Python API, and it exposes UI extensions via the brkraw.viewer.hook entry point so new tabs can be added with standalone packages. Viewer hooks can coexist with converter hooks and CLI hooks, so modality-specific logic can flow from conversion into UI without patching the viewer itself.

UI direction

The default viewer targets a tkinter-based implementation.

This choice is intentional: we want a lightweight tool that can be used directly on scanner consoles or constrained environments with minimal dependencies.

More modern GUI frameworks are welcome, but should be developed as separate CLI extensions to keep the default viewer small and easy to install.

Viewer hooks

Viewer extensions are implemented as hooks discovered through brkraw.viewer.hook. Each hook can register a new tab and provide dataset callbacks, enabling feature panels to live outside the core viewer while staying compatible with BrkRaw rules, specs, and converter hooks. See docs/dev/hooks.md for the hook interface and entry point setup.

Installation

For development and testing, install in editable mode:

pip install -e .

Usage

Launch the viewer via the BrkRaw CLI:

brkraw viewer /path/to/bruker/study

Optional arguments allow opening a specific scan or slice:

brkraw viewer /path/to/bruker/study \
    --scan 3 \
    --reco 1

Use an external registry file (instead of ~/.brkraw/config.yaml registry path):

brkraw viewer /path/to/bruker/study --registry ./shared-registry.jsonl

Write registry entries directly to an external JSONL file:

brkraw viewer-registry add /path/to/bruker/study -t ./shared-registry

The viewer can also open .zip or Paravision-exported .PvDatasets archives using Load (folder or archive file).

Update

Recent updates:

• Open folders or archives (.zip / .PvDatasets)
• Viewer: Space (raw/scanner/subject_ras), nibabel RAS display, click-to-set X/Y/Z, optional crosshair + zoom, slicepack/frame sliders only when needed
• Info: rule + spec selection (installed or file), parameter search, lazy Viewer refresh on tab focus
• Registry: add the current session from the + menu when a dataset is loaded
• Convert: BrkRaw layout engine, template + suffix defaults from ~/.brkraw/config.yaml, keys browser (click to add), optional config layout_entries
• Config: edit ~/.brkraw/config.yaml in-app; basic focus/icon UX

This update keeps dependencies minimal and preserves compatibility with the core BrkRaw rule/spec/hook system.

Contributing

We welcome contributions related to:

• New viewer hooks that add modality-specific panels or workflows
• Alternative UI implementations delivered as separate CLI extensions
• fMRI/MRS/BIDS-focused visualization or QC helpers built on hooks
• Multi-dataset session management and registry enhancements
• Performance and memory improvements for large datasets

Contributions should prefer designs where new hooks extend the viewer implicitly through shared BrkRaw abstractions, and where richer UIs are provided as optional CLI extensions rather than increasing the default dependency footprint.

If you are interested in contributing, please start a discussion or open an issue describing your use case and goals.