Mini Mode
Mini Mode is a simplified version of the Badger GUI designed for quick, streamlined optimization runs. It removes complexity from the full GUI by consolidating routine creation onto a single page and defaulting to automatic behavior.
Launching Mini Mode
Launch Mini Mode from the command line:
badger -mini
To launch with a template preloaded:
badger -mini -t sphere_test.yaml
The -t (or --template) argument will look for a matching YAML template file in your configured template root directory.
The full GUI is still available via badger -g. The two Modes use separate UI components and do not interfere with each other.
Overview

Mini Mode fits all main controls for creating and running an optimization on a single page without scrolling. The interface has two tabs:
- Routine creation — set up and launch a new optimization
- History — view and navigate past runs
Routine Creation
Selecting a Template

At the top of the page, a dropdown lets you select from available routine templates. When a template is loaded, it populates:
- Variables and their ranges
- Environment parameters
- Selected algorithm
- Constraints and objectives
Environment + VOCS Tab

The main tab combines environment selection and VOCS configuration. Key differences from the full GUI:
- Algorithm selection is on this tab rather than a separate tab
- Algorithm parameters can be edited by expanding the "Parameters" button
- Constraints and Observables tables can be expanded from the bottom of the page (collapsed by default)
For Bayesian generators, n_interpolate_points controls how many points are
measured along the path from the previous point to the newly proposed candidate.
For example, 5 measures five points, including the candidate but excluding the
previous point. Each measurement appears in the live data and archived run and
counts toward the evaluation limit. Pause and stop are checked between points;
resuming continues the remaining points before generating another candidate.
Variable Table

The variable table displays three key columns instead of raw upper/lower bounds:
| Column | Description |
|---|---|
| Saved | The value variables will be reset to if you press reset. Updates each time a new run starts. |
| Current | Live value read from the environment. Updates when Badger writes to variables (dial in, reset, or at each optimization step), or when the table is manually refreshed. |
| Range | Shown as a delta (±) around the current value. Based on the selected range option from the template. |
Range Indicators
- If desired bounds are clipped by hard variable limits, an asterisk (*) appears next to the range value.
- Selected variable rows appear with lighter text; unselected rows use darker gray text.
Adjusting Ranges
- Use the up/down arrows on each row to adjust the range for that variable
- Use the arrows in the header row to adjust ranges for all selected variables at once
- Up scales the range by 1.111×; down scales by 0.9× (ranges adjust proportionally)
- Click the gear icon on a row for more detailed range control via the variable range dialog
Show Checked Only
When "Show checked only" is unchecked, the table scrolls to display the first selected row.
Variable Range Dialog

Clicking the gear icon on a variable row opens a dialog with:
- The current value of the variable
- The current low and high bounds for the variable
- Options for range selection method
- ratio wrt current value: Bounds are calculated as a fraction of the current value on either side of the current value
- ratio wrt full range: Bounds are calculated as a fraction of the full range on either side of the current value
- delta around current value: Bounds are calculated as a delta on either side of the current value
- exact bounds: the input numerical bounds are used.
- A bounds preview bar showing what the resulting bounds will be
- Hard limits displayed (but not editable)
Initial Points

Initial points selection is collapsed by default. Points are automatically chosen from within a subset of the configured variable range.
Running an Optimization
Mini Mode defaults to automatic behavior — there is no manual/automatic Mode toggle. Once your routine is configured:
- Select or load a template
- Verify variables, ranges, and objectives
- Start the run
Variable ranges are automatically set around the current value based on template parameters.
Mini Mode uses a smart run controller that combines play, pause, resume, and restart into a single button. There is no separate pause button — the primary action button changes behavior based on the current routine state and whether you have edited any parameters since the run started.

Play Button Behavior
The play button adapts to context:
| State | Button action | What happens |
|---|---|---|
| Stopped | Start | Launches the routine with the current configuration |
| Running | Pause | Suspends the optimization loop; subprocess state is preserved |
| Paused, no changes | Resume | Continues from where the routine paused |
| Paused, compatible edits (e.g., range changes) | Reload & resume | Stops the subprocess, reloads data, and restarts with the new settings |
| Paused, structural edits (variables/objectives changed) | Restart | Fully restarts the routine with the new VOCS |
The controller decides which branch to take by comparing the current GUI state against the snapshot of the routine that is actively displayed, so you can tweak ranges or algorithm parameters mid-session without manually restarting.
Termination Conditions

The current termination condition is shown directly on the play button as a duration or iteration indicator. Each time you press play, the termination condition is extended, so pressing play repeatedly gives you a consistent "N more iterations per press" behavior rather than resuming indefinitely.
Run Menu
The dropdown next to the play button exposes two actions:
- New Run — forces a fresh routine regardless of current state. Use this when you want to discard the current session and start over.
- Edit Condition — opens a dialog to configure the termination condition (iteration count, time limit, etc.) without starting a new run. Changes take effect the next time you press play.

Pause/Resume Details
Pressing the play button while a routine is running pauses it. The subprocess stays alive and holds its state, so resuming continues from exactly where the routine left off rather than reinitializing. The status bar updates to "Routine name paused" once the subprocess has confirmed the pause, so the indicator always reflects the actual state of the run.
While paused, the last sampled point remains visible in the plots and tables, and you can inspect or edit variable ranges, algorithm parameters, or the termination condition before resuming. The smart run controller (see table above) decides whether the next press of play resumes in place, reloads with your edits, or restarts the routine. It may take a few seconds for the routine to pause after pressing the button, since it waits for an in-progress measurement to finish
History Tab

The History tab lets you browse and load past optimization runs, similar to the History Navigator in the full GUI.