Workspace Shell (Default)¶

Note

The Workspace Shell recently became the default RheoJAX GUI. Fuller, per-mode guides (Fit, Transform, Pipeline) are planned; this page covers orientation — launching, the three modes, and the project file model.

This guide covers the Workspace Shell, the mode-based GUI that rheojax-gui launches by default.

Prerequisites¶

Ensure you have RheoJAX installed (GUI dependencies are included):

uv sync

Verify the installation:

python -c "from PySide6.QtWidgets import QApplication; print('PySide6 OK')"

Launching the Application¶

Command Line¶

The default launch opens the Workspace Shell:

rheojax-gui

Other startup options:

# Open a saved project on startup
rheojax-gui --project analysis.rheojax

# Import raw data into the library on startup (preload only -- this
# does not run a fit or transform automatically)
rheojax-gui --import data.xlsx --protocol relaxation

# Start maximized (useful on high-DPI desktops)
rheojax-gui --maximized

Python API¶

From within a Python script or interactive session:

from rheojax.gui import main
main()

The Workspace Window¶

The Workspace Shell is a single window with a mode toolbar at the top, a dataset library on the left, the active mode’s step sequence in the center, and an inspector panel on the right. A status label in the toolbar shows JAX device and float64 status.

Three Modes¶

Switching modes (toolbar buttons) swaps the center step sequence. Each mode is a linear, forward-unlocking sequence of steps: a step becomes reachable once the previous step is both filled in and valid.

Fit¶

Single-dataset model fitting. Pick a protocol and model, select the dataset to fit, run an NLSQ fit, optionally follow up with a Bayesian NUTS run, visualize the result, and export it. This mirrors the legacy GUI’s Data → Fit → Bayesian → Export workflow, but as one guided sequence instead of separate pages.

Transform¶

Single-dataset transform application. Pick a transform (mastercurve, FFT, derivative, etc.), configure its inputs, run it, visualize the transformed data, and export the result.

Pipeline¶

Batch orchestration across many datasets at once. Assemble a sequence of transform, fit, and export steps, select which datasets in the library to run them against, and click Run All. Pipeline mode has no per-step “run” button by design — running a single step interactively is what Fit and Transform modes are for; Pipeline mode is for repeating a fixed recipe over many datasets in the background.

The File Menu and Projects¶

The File menu provides New, Open…, Save, Save As…, and Close. Save and Open read and write the .rheojax v2 project format: a single archive file capturing the dataset library, fit/transform/ pipeline state, and job history, so a full analysis session can be closed and reopened later.

The window tracks unsaved changes and prompts you to save before starting a new project, opening another one, or closing, if anything is dirty. If background jobs (an NLSQ/NUTS fit, a running pipeline batch) are still in flight, it will also ask whether to cancel them before proceeding.

Troubleshooting¶

GUI Won’t Launch¶

Verify PySide6 is installed:
```
pip show PySide6
```
Check for Qt platform issues (Linux):
```
export QT_QPA_PLATFORM=xcb
```

Try reinstalling:

pip uninstall PySide6
pip install PySide6

Slow Performance¶

Check JAX device (prefer GPU when available)
Reduce data size for initial exploration
Close unnecessary ArviZ diagnostic plots

Display Issues¶

For high-DPI displays, set environment variable:

export QT_AUTO_SCREEN_SCALE_FACTOR=1