Expected Workflow

VoxKit separates developer-level integration from research workflow configuration. This guide outlines the expected collaboration and deployment workflow.

1. Collaboration: PI and Developers

• •A Principal Investigator (PI) or research lead defines scientific goals and desired workflow.
• •Developers implement low-level capabilities: pipeline stackers, engines (toolkits), analyzers (dataset analysis functions), and a startup routine to download models, datasets, and other assets.

2. Two Levels of Configuration

• •Developer-level: modify code and the API to implement compatible pipelines, toolkits, analyzers, and startup routines (requires coding).
• •Researcher-level: edit configuration JSON files in the bundled `config` folder to define specific workflow steps and guidance (no coding required). A future CLI will simplify this process.

3. Bundling and Post-build Configuration

The app is bundled with developer-provided code and a configuration folder. The JSON configuration can be changed after bundling to adapt to a specific study; most workflow changes do not require code edits.

4. Distribution

• •Distribute the configured executable via a secure network drive or a secure web location.
• •If the generic version satisfies your needs, download it from the website and configure via JSON only—no developer interaction required.

5. Researcher Onboarding

The initial configuration guides users through setup and the steps of the workflow. Once configured and distributed, researchers can run the app without writing code.

Notes

• •Startup routines handle one-time setup tasks such as downloading models and large datasets.
• •Maintain versioned configuration files in the config folder to support reproducible workflows.
• •Developers should provide clear documentation of any custom pipeline components so research staff can configure workflows correctly.