.. This file is generated by scripts/render_dataset_docs.py. Do not edit by hand.

TCBench Alpha
=============

Synthetic-backed storm-track benchmark adapter aligned to the TCBench Alpha ecosystem.

Overview
--------

TCBench Alpha is the public storm adapter used by several tropical cyclone baselines on the shared track-intensity evaluator.

The current implementation is synthetic-backed, but it preserves the task, metric, and reporting surface used by the shared tropical cyclone benchmark.

At a Glance
-----------

.. list-table::
   :widths: 28 72
   :stub-columns: 1

   * - Provider
     - TCBench Alpha ecosystem surfaced through a PyHazards adapter
   * - Hazard Family
     - Tropical Cyclone
   * - Source Role
     - Track Benchmark
   * - Coverage
     - Benchmark-aligned tropical cyclone track and intensity samples
   * - Geometry
     - Storm-track history sequences
   * - Spatial Resolution
     - Storm-centered best-track sequences
   * - Temporal Resolution
     - Historical track windows with forecast horizons
   * - Update Cadence
     - Generated locally for smoke and benchmark-alignment runs
   * - Period of Record
     - Synthetic-backed benchmark adapter
   * - Formats
     - PyTorch tensors via the dataset registry
   * - Registry Entry
     - ``tcbench_alpha``

Data Characteristics
--------------------

- Storm-history sequences with future latitude, longitude, and intensity targets.
- Registry-backed benchmark adapter rather than a raw external benchmark ingestion path.
- Intended for benchmark-linked track-intensity forecasting runs.

Typical Use Cases
~~~~~~~~~~~~~~~~~

- Tropical Cyclone MLP, SAF-Net, and TCIF-fusion smoke tests.
- Shared tropical cyclone benchmark runs.
- Regression checks for storm-track baselines.

Access
------

Use the links below to access the upstream source or its public documentation.

- `TCBench Alpha repository <https://github.com/msgomez06/TCBench_Alpha>`_

PyHazards Usage
---------------

Use this adapter when you want the public TCBench Alpha-aligned storm surface exposed by the tropical cyclone benchmark.

Registry Workflow
~~~~~~~~~~~~~~~~~

Primary dataset name: ``tcbench_alpha``

.. code-block:: python

   from pyhazards.datasets import load_dataset

   data = load_dataset(
       "tcbench_alpha",
       micro=True,
       history=6,
       horizon=5,
   ).load()

   train = data.get_split("train")
   print(train.inputs.shape, train.targets.shape)

Related Coverage
~~~~~~~~~~~~~~~~

**Benchmarks:** :doc:`Tropical Cyclone Benchmark </benchmarks/tropical_cyclone_benchmark>`, :doc:`TCBench Alpha </benchmarks/tcbench_alpha>`

**Representative Models:** :doc:`Tropical Cyclone MLP </modules/models_tropicalcyclone_mlp>`, :doc:`SAF-Net </modules/models_saf_net>`, :doc:`TCIF-fusion </modules/models_tcif_fusion>`

Inspection Workflow
-------------------

This dataset is currently surfaced as a registry-backed benchmark adapter,
so there is no standalone inspection CLI documented for it.

Notes
-----

- This is a synthetic-backed benchmark adapter rather than a full TCBench Alpha ingestion pipeline.

Reference
---------

- `TCBench Alpha <https://github.com/msgomez06/TCBench_Alpha>`_.