User Guide#
The user guide is the operational companion to the scientific foundations and API reference. It explains how GeoPrior-v3 is used in practice: how configuration is organized, how the staged workflow is executed, how diagnostics are interpreted, and how trained workflows lead to inference products and exportable artifacts.
This section is intended for readers who already understand the general project scope and now want to run, inspect, and control the workflow in a structured way. It connects the command-line interface, the configuration system, and the staged execution logic into one narrative so that day-to-day use remains reproducible and transparent.
In broad terms, the user guide answers four practical questions:
What is the overall execution logic of the project?
What does each workflow stage do?
How are configuration, diagnostics, and export paths controlled?
Where should a user look when something does not behave as expected?
Note
The staged workflow is designed to make the project easier to reason about. Rather than hiding all logic behind a single opaque command, GeoPrior-v3 exposes clearly named stages and companion guides for configuration, diagnostics, and output handling.
Reading paths#
Different readers usually arrive here with different goals. The guide is therefore organized so that it can be read either sequentially or by topic.
New workflow users should begin with Workflow, then read Configuration, and finally move through Stage-1, Stage-2, Stage-3, Stage-4, and Stage-5.
CLI-oriented users can pair CLI in the workflow with the stage-specific pages and use Diagnostics when validating intermediate results.
Inference-focused users will usually want Inference and export after the stage pages.
Readers troubleshooting behavior should keep FAQ and Diagnostics close at hand.
Workflow guidance#
Start with the high-level execution logic, the relationship between stages, and the reading order for the operational workflow.
Learn how project behavior is controlled through settings, defaults, and configuration-driven workflow design.
Understand how to inspect intermediate states, read warnings, verify outputs, and evaluate whether a run behaved as expected.
Follow the path from trained or tuned workflows to forecast outputs, artifacts, and downstream export products.
Stage-by-stage execution#
Data preparation, project initialization logic, and the early processing steps that establish the workflow foundation.
Training-oriented execution and the transition from prepared inputs toward calibrated model behavior.
Tuning and search procedures used to refine model configuration and improve workflow robustness.
Inference-time execution, forecast generation, and the practical use of saved workflow outputs.
Transfer, extension, and final workflow products that complete the staged operational path.
Reference and support#
Browse the command-line entry points, command groupings, and how they map onto the staged workflow.
Find concise answers to recurring practical questions about running, configuring, diagnosing, and exporting workflows.