CLI in the workflow#

GeoPrior provides a family-based command-line interface for staged execution, artifact building, and figure generation.

In practice, the CLI is the layer you use when you are ready to do something concrete:

  • run a stage,

  • build an artifact,

  • render a figure,

  • initialize or adjust a configuration.

This page is here to help you place the CLI inside the broader workflow.

Where the CLI fits#

A simple way to think about the documentation is:

  • the User Guide explains the workflow and why each stage matters,

  • the CLI pages show which command to run for a given task,

  • the Examples gallery shows complete worked lessons with outputs, figures, and interpretation.

So when you already know you want to run a command, the best move is to jump into the dedicated CLI pages below. When you want the reasoning, context, and sequence behind the workflow, stay in the User Guide.

Start here for commands#

CLI overview

Start with the main CLI landing page for the overall structure, command families, and reading path.

CLI
CLI introduction

Learn the mental model behind geoprior, geoprior-run, geoprior-build, and geoprior-plot.

Introduction
Shared conventions

See the common CLI patterns once: config installation, runtime overrides, repeated data-loading behavior, and output handling.

Shared conventions
Run family

Go here when you want to execute workflow stages, transfer runs, sensitivity sweeps, or SM3 run-side diagnostics.

Run family
Build family

Go here when you want to materialize datasets, payloads, summaries, metrics, or geospatial side products.

Build family
Plot family

Go here when you want to render figures, maps, uncertainty plots, transfer plots, or supporting diagnostics.

Plot family

How the CLI maps onto the workflow#

The CLI follows the same staged logic as the rest of GeoPrior:

  • Stage 1 prepares the data and writes the artifact set used downstream.

  • Stage 2 trains the model from those prepared artifacts.

  • Stage 3 tunes hyperparameters from the same Stage-1 foundation.

  • Stage 4 runs inference, evaluation, calibration, and export.

  • Stage 5 handles transfer evaluation and related workflows.

That means the CLI is not separate from the workflow. It is the most direct way to operate the workflow once you understand what each stage is for.

A practical reading path#

A good way to move through the docs is:

  1. read Workflow to understand the staged process,

  2. open CLI when you are ready to choose a command,

  3. move into the specific CLI family page that matches your task,

  4. return to the stage pages when you want the scientific or workflow context behind that command,

  5. use Gallery when you want a full worked lesson.