Merge branch 'main' into mnt/fix-test-ui-crash

Author: Siddhazntx

.github/workflows/pypi_release.yml

@@ -48,7 +48,7 @@ jobs:
           python -m build --wheel --sdist --outdir wheelhouse
 
       - name: Store wheels
-        uses: actions/upload-artifact@v6
+        uses: actions/upload-artifact@v7
         with:
           name: wheels
           path: wheelhouse/*
@@ -102,7 +102,7 @@ jobs:
       id-token: write
 
     steps:
-      - uses: actions/download-artifact@v7
+      - uses: actions/download-artifact@v8
         with:
           name: wheels
           path: wheelhouse

pytorch_forecasting/data/__init__.py

@@ -1,8 +1,9 @@
 """
-Datasets, etc. for timeseries data.
+Utilities for time series dataset construction and preprocessing.
 
-Handling timeseries data is not trivial. It requires special treatment.
-This sub-package provides the necessary tools to abstracts the necessary work.
+This subpackage provides dataset classes, normalization and encoding
+utilities, and batching tools required to transform raw time series data
+into model-ready PyTorch datasets.
 """
 
 from pytorch_forecasting.data.encoders import (

pytorch_forecasting/layers/_embeddings/_data_embedding.py

@@ -14,12 +14,19 @@
 class DataEmbedding_inverted(nn.Module):
     """
     Data embedding module for time series data.
-    Args:
-        c_in (int): Number of input features.
-        d_model (int): Dimension of the model.
-        embed_type (str): Type of embedding to use. Defaults to "fixed".
-        freq (str): Frequency of the time series data. Defaults to "h".
-        dropout (float): Dropout rate. Defaults to 0.1.
+
+    Parameters
+    ----------
+    c_in : int
+         Number of input features.
+    d_model : int
+        Dimension of the model.
+    embed_type : str
+        Type of embedding to use. Defaults to "fixed".
+    freq : str
+        Frequency of the time series data. Defaults to "h".
+    dropout : float
+        Dropout rate. Defaults to 0.1.
     """
 
     def __init__(self, c_in, d_model, dropout=0.1):

pytorch_forecasting/layers/_embeddings/_en_embedding.py

@@ -19,11 +19,17 @@ class EnEmbedding(nn.Module):
     """
     Encoder embedding module for time series data. Handles endogenous feature
     embeddings in this case.
-    Args:
-        n_vars (int): Number of input features.
-        d_model (int): Dimension of the model.
-        patch_len (int): Length of the patches.
-        dropout (float): Dropout rate. Defaults to 0.1.
+
+    Parameters
+    ----------
+    n_vars : int
+        Number of input features.
+    d_model : int
+        Dimension of the model.
+    patch_len : int
+        Length of the patches.
+    dropout : float
+        Dropout rate. Defaults to 0.1.
     """
 
     def __init__(self, n_vars, d_model, patch_len, dropout):

pytorch_forecasting/layers/_embeddings/_positional_embedding.py

@@ -14,9 +14,14 @@
 class PositionalEmbedding(nn.Module):
     """
     Positional embedding module for time series data.
-    Args:
-        d_model (int): Dimension of the model.
-        max_len (int): Maximum length of the input sequence. Defaults to 5000."""
+
+    Parameters
+    ----------
+    d_model : int
+        Dimension of the model.
+    max_len : int
+        Maximum length of the input sequence. Defaults to 5000.
+    """
 
     def __init__(self, d_model, max_len=5000):
         super().__init__()

pytorch_forecasting/layers/_embeddings/_sub_nn.py

@@ -13,9 +13,9 @@ def __init__(self, seq_len: int, lag: int, d_model: int, emb_dims: list, device)
         ----------
         seq_len: int
             length of the sequence (sum of past and future steps)
-        lag: (int):
+        lag: int
             number of future step to be predicted
-        hidden_size: int
+        d_model: int
             dimension of all variables after they are embedded
         emb_dims: list
             size of the dictionary for embedding. One dimension for each categorical variable

pytorch_forecasting/layers/_filter/_moving_avg_filter.py

@@ -11,18 +11,19 @@ class MovingAvg(nn.Module):
     """
     Moving Average block for smoothing and trend extraction from time series data.
 
-    A moving average is a smoothing technique that creates a series of average from
+    A moving average is a smoothing technique that creates a series of averages from
     different subsets of a time series.
 
     For example: Given a time series ``x = [x_1, x_2, ..., x_n]``, the moving average
     with a kernel size of `k` calculates the average of each subset of `k` consecutive
     elements, resulting in a new series of averages.
 
-    Args:
-        kernel_size (int):
-            Size of the moving average kernel.
-        stride (int):
-            Stride for the moving average operation, typically set to 1.
+    Parameters
+    ----------
+    kernel_size : int
+        Size of the moving average kernel.
+    stride : int
+        Stride for the moving average operation, typically set to 1.
     """
 
     def __init__(self, kernel_size, stride):