API reference#
The API reference collects the documented public Python surface of GeoPrior. It is designed for users who want to move from the narrative workflow pages into the actual modules, classes, helper layers, and configuration surfaces that support reproducible forecasting runs.
This section is not just a flat module listing. It is organized around the main interfaces that users are most likely to touch directly during normal work:
command-line wrappers and staged workflow entry points,
subsidence-model interfaces and supporting model components,
tuning interfaces for calibration and structured experimentation,
utility modules for preprocessing, diagnostics, geospatial handling, and forecast support,
packaged resources and configuration-driven assets,
parameter helpers used to represent physical constants and learnable scalar coefficients.
For a scientific overview of the modeling stack, start with GeoPriorSubsNet and Models overview. For workflow-oriented usage, see Workflow and the examples gallery under Gallery.
Note
The pages collected here emphasize the documented public surface of GeoPrior. Internal helpers and compatibility layers may exist underneath the package, but this reference focuses on the parts of the codebase that users are expected to import, inspect, extend, or call through normal workflows.
Browse the documented command-line wrappers, staged execution modules, and resource pages that support initialization, runs, diagnostics, and exports.
Explore the subsidence model family, supporting mathematical and diagnostic modules, and the interfaces that structure the main GeoPrior forecasting workflow.
Review the documented tuner interfaces used for calibration, configuration search, and structured experimentation.
Understand the utility stack spanning preprocessing, calibration, geospatial handling, forecasting helpers, and supporting model utilities.
Inspect packaged resources, templates, and supporting files that help keep GeoPrior workflows reproducible and configuration-driven.
Learn how scalar physical constants and learnable coefficients are represented through the documented parameter helper layer.
How to read this section#
This part of the documentation can be used in two complementary ways.
First, it can be read as a practical reference. When you need a stable import path, a documented class, or a specific helper function, these pages show the intended public surface and help you locate the right module quickly.
Second, it can be read as an architectural map. Several API layers are closely tied to the scientific workflow, so the reference pages also explain why a module exists, how it fits into the broader forecasting stack, and whether it is typically used directly or through higher-level orchestration.
A good reading path is often:
start with the workflow-facing CLI surface,
move to the subsidence model pages,
review tuner and utility layers as needed,
consult resources and parameter wrappers for reproducibility and physical configuration details.
Tip
If you are new to the package, begin with the workflow and model pages first. If you are writing scripts, extending internals, or building reproducible research pipelines, the utility, resource, and parameter sections will usually become important next.