Synthefy
Copyright 2026 Synthefy

This product includes software developed by Synthefy and is licensed under
the Apache License, Version 2.0. The full license text is in the LICENSE file
at the root of this distribution.

------------------------------------------------------------------------------
tabpfn-time-series (PriorLabs)
------------------------------------------------------------------------------
  Source:   https://github.com/PriorLabs/tabpfn-time-series
  Revision: a756ae3 (2026-07-13)
  License:  Apache License, Version 2.0 (see licenses/Apache-2.0.txt)

  Copyright 2025 Prior Labs GmbH

  The model-free time-series feature engineering under
  src/synthefy/nori_ts/tsfeatures/ is vendored from tabpfn-time-series at the
  revision above.

  Modifications by Synthefy (Apache-2.0 section 4(b)):
    - basic_features.py, auto_features.py, and feature_generator_base.py:
      intra-package import paths and distribution-local provenance/ownership
      wording changed; implementation behavior is unchanged.
    - feature_transformer.py: the same import/provenance changes, plus
      `from __future__ import annotations` for Python 3.9. This changes
      annotation representation, not forecasting behavior.
    - data_preparation.py: import/provenance changes, and `generate_test_X`
      takes an explicit `freq` because re-inferring it from a NaN-dropped index
      yields None and fails horizon construction.
    - ts_dataframe.py: implementation body unchanged; only the
      distribution-local provenance/ownership wording changed.

  NO TabPFN model code or weights are included; the TabPFN model carries a
  separate, non-Apache license and is not used or redistributed here.

------------------------------------------------------------------------------
AutoGluon
------------------------------------------------------------------------------
  Source:   https://github.com/autogluon/autogluon
  License:  Apache License, Version 2.0 (see licenses/Apache-2.0.txt)

  Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.

  src/synthefy/nori_ts/tsfeatures/ts_dataframe.py (TimeSeriesDataFrame)
  originates in AutoGluon's timeseries dataset module. It reached this
  distribution via PriorLabs/tabpfn-time-series, which made the modifications
  recorded in the file header; Synthefy retained the implementation body.

------------------------------------------------------------------------------
Chronos (amazon-science)
------------------------------------------------------------------------------
  Source:   https://github.com/amazon-science/chronos-forecasting
  License:  Apache License, Version 2.0 (see licenses/Apache-2.0.txt)

  Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.

  to_gluonts_univariate() in
  src/synthefy/nori_ts/tsfeatures/data_preparation.py is adapted from
  chronos-forecasting's evaluation script.
