Using the GUI

While the SimulationStudy class offers powerful programmatic control, sometimes you just need a clean visual interface to explore your data without writing scripts. The DigiQual App wraps the core statistical engine into a modern, user-friendly graphical interface, allowing you to run the entire lifecycle—from design to diagnostics to reliability analysis—directly in your browser.

1. How to Access the App

Currently, the DigiQual app is bundled directly with the Python package. Depending on your workflow, there are a few ways to launch it.

Option A: Using uv via CLI (Recommended)

If you use uv for environment management, you can launch the app directly from your terminal as a Command Line Interface (CLI) tool. This is the fastest way to get started.

To run the app instantly without permanently installing it to your environment:

uvx -v digiqual

Note: The -v (verbose) flag is highly recommended here! Without it, uvx runs quietly and the app might appear frozen for a few minutes while it downloads the required statistical packages in the background.

If you plan to use it frequently, you can install it as a persistent global tool on your machine:

uv tool install digiqual
# Once installed, simply type this from any terminal:
digiqual

Linux Users: To ensure the graphical window renders correctly on Linux distributions, you must include the linux-gui extra during installation:

uv tool install "digiqual[linux-gui]"

Option B: From a Python Script

If you are already working inside a Python script or a Jupyter environment, you can launch it locally using the built-in UI function.

# Launch the local application
from digiqual import dq_ui

dq_ui()

This will spin up a local server and automatically open the DigiQual dashboard in your default web browser.

Coming Soon: Standalone Executables We are actively developing standalone, double-click executables for both Windows and macOS. In the near future, you or your colleagues will be able to run the complete DigiQual toolkit as a standard desktop application, without needing to install Python, manage virtual environments, or touch a command line.

2. The App Workflow

The app is divided into three main modules that mirror the programmatic workflow you are already familiar with.

Module 1: Experimental Design

Instead of defining arrays or dictionaries in code, you can use the UI to visually define your experimental parameters.

• Define your input variables (e.g., Length, Angle) and their minimum/maximum ranges.

• Specify the number of initial samples you want.

• Click generate to instantly build a Latin Hypercube Sample (LHS) framework and download the CSV to feed into your external solver (like MATLAB or Abaqus).

Module 2: Simulation Diagnostics

Once your external physics solver has run, you can upload the resulting CSV back into the app. Use the dropdowns to assign your input columns and your outcome column (e.g., “Signal”). The app will automatically run health checks, such as testing for input coverage gaps.

Adaptive Refinement: If the app detects a deliberate gap or flawed data, it will flag an issue. Instead of manually guessing where to add points, the UI provides a remediation tool to automatically generate targeted samples to fill the gap. You can download these new points, run them through your solver, and upload the complete dataset.

Module 3: Reliability Analysis

With a validated dataset, you can construct your Probability of Detection (PoD) curves.

• Select your Parameter of Interest (e.g., “Length”) and set your critical threshold (e.g., 18.0).

• The app handles the model fitting and bootstrapping in the background.

• It automatically visualises the results, displaying both the Signal Response Model (Physics) and the Probability of Detection Curve (Reliability) side-by-side.