feat: shorten package import path from pytorch_forecasting to ptf with legacy redirect

Author: harshsomankar123-tech

.github/ISSUE_TEMPLATE/bug_report.md

@@ -39,7 +39,7 @@ Add any other context about the problem here.
 <!--
 Please run the following code snippet and paste the output here:
 
-from pytorch_forecasting import show_versions; show_versions()
+from ptf import show_versions; show_versions()
 -->
 
 </details>

.gitignore

@@ -146,4 +146,4 @@ lightning_logs
 .DS_Store
 
 # data
-pytorch_forecasting/data/*.parquet
+ptf/data/*.parquet

CHANGELOG.md

@@ -143,7 +143,7 @@ Release focusing on:
 * Tuner import change due to a Lightning breaking change. Lightning v2.6 introduced a breaking change in its checkpoint loading behavior, which caused unpickling errors during weight loading in `pytorch-forecasting` (see #2000).
 To address this, `pytorch-forecasting` now provides its own `Tuner` wrapper that exposes the required `weights_only` argument when calling `lr_find()`.
 
-  * When using `pytorch-forecasting > 1.5.0` with `lightning > 2.5`, please use `pytorch_forecasting.tuning.Tuner` in place of `lightning.pytorch.tuner.Tuner`. See #2000 for details.
+  * When using `pytorch-forecasting > 1.5.0` with `lightning > 2.5`, please use `ptf.tuning.Tuner` in place of `lightning.pytorch.tuner.Tuner`. See #2000 for details.
 
 ### Maintenance
 
@@ -563,7 +563,7 @@ Maintenance update widening compatibility ranges and consolidating dependencies:
 ### Changed
 
 - Dropping Python 3.6 support, adding 3.10 support (#479)
-- Refactored dataloader sampling - moved samplers to pytorch_forecasting.data.samplers module (#479)
+- Refactored dataloader sampling - moved samplers to ptf.data.samplers module (#479)
 - Changed transformation format for Encoders to dict from tuple (#949)
 
 ### Contributors

README.md

@@ -83,7 +83,7 @@ import lightning.pytorch as pl
 from lightning.pytorch.loggers import TensorBoardLogger
 from lightning.pytorch.callbacks import EarlyStopping, LearningRateMonitor
 # import dataset, network to train and metric to optimize
-from pytorch_forecasting import TimeSeriesDataSet, TemporalFusionTransformer, QuantileLoss
+from ptf import TimeSeriesDataSet, TemporalFusionTransformer, QuantileLoss
 from lightning.pytorch.tuner import Tuner
 
 # load data: this is pandas dataframe with at least a column for

docs/source/_ext/model_overview.py

@@ -1,7 +1,7 @@
 """Sphinx extension to auto-generate model overview table from registry.
 
 Registers the ``.. model-overview::`` directive which queries the
-``pytorch_forecasting._registry.all_objects`` registry, extracts model
+``ptf._registry.all_objects`` registry, extracts model
 tags, and renders an RST table comparing model capabilities.
 
 """
@@ -82,7 +82,7 @@ def _get_model_rows():
     Each row maps column header to display value. Includes both v1 and v2
     models; the Version column distinguishes them.
     """
-    from pytorch_forecasting._registry import all_objects
+    from ptf._registry import all_objects
 
     tag_keys = [
         "info:name",

docs/source/api_v2.rst

@@ -5,7 +5,7 @@ API v2
     Please note that the v2 modules are currently in active-development and is in beta right now, so please use this API with caution.
     See v1 documentation :doc:`here <api>` - it is stable and can be used in the production pipelines.
 
-.. currentmodule:: pytorch_forecasting
+.. currentmodule:: ptf
 
 We are currently developing version 2 of PyTorch Forecasting. The primary objective of this redesign is to improve the software architecture and provide a more intuitive workflow for developers and data scientists.
 

docs/source/conf.py

@@ -25,7 +25,7 @@
 sys.path.insert(0, str(PROJECT_PATH))  # noqa
 sys.path.insert(0, str(SOURCE_PATH / "_ext"))  # custom Sphinx extensions
 
-import pytorch_forecasting  # isort:skip
+import ptf  # isort:skip
 
 # -- Project information -----------------------------------------------------
 
@@ -89,7 +89,7 @@ def skip(app, what, name, obj, skip, options):
 
 apidoc_output_folder = SOURCE_PATH.joinpath("api")
 
-PACKAGES = [pytorch_forecasting.__name__]
+PACKAGES = [ptf.__name__]
 
 
 def get_by_name(string: str):
@@ -128,7 +128,7 @@ def get_items(self, names):
                     if hasattr(obj, "__module__"):
                         mod_name = obj.__module__
                         t = f"{mod_name}.{t}"
-                    if t.startswith("pytorch_forecasting"):
+                    if t.startswith("ptf"):
                         new_names.append(t)
         new_items = super().get_items(sorted(new_names))
         return new_items

docs/source/data.rst

@@ -8,7 +8,7 @@ Data
    | Try it out before release: :doc:`v2 Data Pipeline <data_v2>` | :doc:`v2 API Reference <api_v2>`
    | **Caution: v2 is WIP and unstable. Not yet production-ready.**
 
-.. currentmodule:: pytorch_forecasting.data
+.. currentmodule:: ptf.data
 
 Loading data for timeseries forecasting is not trivial - in particular if covariates are included and values are missing.
 PyTorch Forecasting provides the :py:class:`~timeseries.TimeSeriesDataSet` which comes with a :py:meth:`~timeseries.TimeSeriesDataSet.to_dataloader`
@@ -25,7 +25,7 @@ Time series data set
 The time series dataset is the central data-holding object in PyTorch Forecasting. It primarily takes
 a pandas DataFrame along with some metadata. See the :ref:`tutorial on passing data to models <passing-data>` to learn more it is coupled to models.
 
-.. autoclass:: pytorch_forecasting.data.timeseries.TimeSeriesDataSet
+.. autoclass:: ptf.data.timeseries.TimeSeriesDataSet
    :noindex:
    :members: __init__
 
@@ -34,7 +34,7 @@ Details
 
 See the API documentation for further details on available data encoders and the :py:class:`~timeseries.TimeSeriesDataSet`:
 
-.. currentmodule:: pytorch_forecasting
+.. currentmodule:: ptf
 
 .. autosummary::
    :toctree: api

docs/source/data_v2.rst

@@ -5,7 +5,7 @@ Data v2
     Please note that the v2 modules are currently in active-development and is in beta right now, so please use this API with caution.
     See complete documentation for v2 API :doc:`here <api_v2>` and stable v1 documentation :doc:`here <api>`.
 
-.. currentmodule:: pytorch_forecasting
+.. currentmodule:: ptf
 
 Loading and managing time series data for deep learning can be complex, especially when handling varying sequence lengths, multiple covariates, and categorical encodings.
 
@@ -22,7 +22,7 @@ The D1 Layer is the foundational data ingestion layer. Its primary responsibilit
 
 Unlike the v1 dataset, the D1 layer does not handle complex preprocessing or batching logic, keeping it lightweight and highly modular.
 
-.. autoclass:: pytorch_forecasting.data.timeseries._timeseries_v2.TimeSeries
+.. autoclass:: ptf.data.timeseries._timeseries_v2.TimeSeries
    :noindex:
    :members: __init__
 
@@ -43,7 +43,7 @@ Because different forecasting architectures require specific input shapes and st
 
 Each model is optimally designed to be compatible with one or more specific DataModules. You can easily verify which DataModule pairs correctly with your chosen model by checking the compatibility overview table in the **:doc:`v2 Models <models_v2>`** documentation.
 
-.. autoclass:: pytorch_forecasting.data.data_module._tslib_data_module.TslibDataModule
+.. autoclass:: ptf.data.data_module._tslib_data_module.TslibDataModule
    :noindex:
    :members: __init__
 
@@ -53,7 +53,7 @@ API Reference
 
 See the detailed API documentation for the V2 data classes below:
 
-.. currentmodule:: pytorch_forecasting
+.. currentmodule:: ptf
 
 .. autosummary::
    :toctree: api

docs/source/faq.rst

@@ -1,7 +1,7 @@
 FAQ
 ====
 
-.. currentmodule:: pytorch_forecasting
+.. currentmodule:: ptf
 
 Common issues and answers. Other places to seek help from: