3. Choose the operation you need

You will learn: when to run, generate, predict, retrain, calibrate or inspect. Start from what you already have.

I have measurements and want a model

Use run. It fits candidates, evaluates them, selects one and creates the final predictor. Start with 1. Run your first complete experiment, then add nodes in 5. Build a pipeline you can explain.

Python SDK calls nirs4all.run(pipeline, dataset, ...). The shared finite native recipe uses Python Core run_pipeline(data, recipe), R nirs4all_run_pipeline, Octave nirs4all.runPipeline, or WASM runBrowserPipeline. Real calls appear together in 10. Run the same recipe in your language.

I want to see which recipes will be tried

Use a generator, which expands choices without fitting. Two preprocessing choices and three component counts make six recipes. Inspect the list before launching 18 fold fits. See the generator node.

A candidate list handles a small comparison. Tuning asks an optimizer to propose settings within a budget. Resume restores a saved compatible search; changing data or search settings starts a different experiment. Begin with fixed alternatives before structural tuning.

I have a saved model and new measurements

Use predict. Reload the predictor, pass raw measurements in the same feature/source order, and keep IDs beside the predicted values. New measurements need no labels. Follow 8. Save a predictor and use it on new observations.

I have new labeled data

Use retrain to build another predictor. Refit uses the selected recipe on its complete original training data; retraining starts a new campaign with new training data. Continuing old neural-network weights is a separate, model-specific option.

I need uncertainty intervals

Use calibrate after fixing the predictor. Predict a separate labeled calibration cohort and learn an allowance from its errors. Future predictions can include intervals. See 6. Measure whether your model answers the question.

I want to compare completed experiments

Open results or a workspace, without starting another fit. Compare held-out scores only when target, units, data and validation protocol match. See 7. Read the result and keep a useful experiment record.

Operation	Needed input	Output	Example or explanation
Generate	Recipe with choices	Candidate recipes	5. Build a pipeline you can explain
Run	Training data and recipe	Scores, winner and predictor	1. Run your first complete experiment
Tune/resume	Data, space and budget/checkpoint	Trials and selected predictor	11. Progress from a baseline to an advanced workflow
Predict	Fitted predictor and compatible X	Estimates and sample IDs	8. Save a predictor and use it on new observations
Retrain	New labeled data and recipe/source model	New experiment	Retrain and Transfer Learning
Calibrate	Fixed predictor and separate labels	Calibration state and intervals	6. Measure whether your model answers the question
Export/load	Completed model or experiment	Durable artifact/reloaded object	10. Run the same recipe in your language
Inspect/compare	Saved scores and predictions	Tables and diagnostic figures	7. Read the result and keep a useful experiment record
Audit robustness	Fixed predictor and defined perturbations	Measured performance changes	Native tuning and conformal calibration

Which settings change the scientific answer?

Setting family	Examples	Why it matters
Data	Targets, sources, groups, partitions	Defines the scientific question
Recipe	Transform order, model, branch/merge	Defines the relation you learn
Evaluation	Folds, metric and aggregation	Defines what a score means
Search	Alternatives, budget, seed	Defines competing recipes
Execution	Engine and installed runtime	Defines executable operations
Storage	Workspace/export destination	Defines what can be inspected or reloaded
Display	Figure visibility and labels	Helps interpretation

The CLI uses groups such as workflow, results, tuning, dataset and workspace; exact flags are in CLI reference.

Checkpoint: “saved model and unlabeled measurements” means load → predict. “Two possible recipes” means compare on shared folds. Continue to 4. Describe your measurements and align your sources.