update

CHANGES

@@ -21,6 +21,10 @@ The rules for this file:
 Enhancements
   - Add a TI estimator using gaussian quadrature to calculate the free energy.
     (issue #302, PR #304)
+  - Warning issued when the series is `None` for `statistical_inefficiency`
+    (issue #337, PR #304)
+  - ValueError issued when `df` and `series` for `statistical_inefficiency`
+    doesn't have the same length (issue #337, PR #304)
 
 
 22/06/2023 xiki-tempula

environment.yml

@@ -10,3 +10,4 @@ dependencies:
 - scikit-learn
 - pyarrow
 - matplotlib
+- loguru

src/alchemlyb/preprocessing/subsampling.py

@@ -363,6 +363,15 @@ def _prepare_input(df, series, drop_duplicates, sort):
     series : Series
         Formatted Series.
     """
+    if series is None:
+        logger.warning(
+            "The series input is `None`, would not subsample according to statistical inefficiency."
+        )
+
+    elif len(df) != len(series):
+        raise ValueError(
+            f"The length of df ({len(df)}) should be same as the length of series ({len(series)})."
+        )
     if _check_multiple_times(df):
         if drop_duplicates:
             df, series = _drop_duplicates(df, series)

src/alchemlyb/tests/test_preprocessing.py

@@ -544,3 +544,16 @@ def test_statistical_inefficiency(self, caplog, u_nk):
             assert "Running statistical inefficiency analysis." in caplog.text
             assert "Statistical inefficiency:" in caplog.text
             assert "Number of uncorrelated samples:" in caplog.text
+
+
+def test_unequil_input(dHdl):
+    with pytest.raises(ValueError, match="should be same as the length of series"):
+        statistical_inefficiency(dHdl, series=dHdl[:10])
+
+
+def test_series_none(dHdl, caplog):
+    statistical_inefficiency(dHdl, series=None)
+    assert (
+        "The series input is `None`, would not subsample according to statistical inefficiency."
+        in caplog.text
+    )