Merge branch 'develop' into 'flag-aggregation'

# Conflicts:
#   CHANGELOG.md

requirements.txt

@@ -9,7 +9,7 @@ matplotlib==3.8.3
 numpy==1.26.4
 outlier-utils==0.0.5
 pyarrow==15.0.0
-pandas==2.1.4
+pandas==2.2.1
 scikit-learn==1.4.1.post1
 scipy==1.12.0
 typing_extensions==4.5.0

saqc/core/translation/basescheme.py

@@ -148,7 +148,8 @@ class MappingScheme(TranslationScheme):
         out = DictOfSeries()
         expected = pd.Index(trans_map.values())
         for field in flags.columns:
-            out[field] = flags[field].replace(trans_map)
+            with pd.option_context("future.no_silent_downcasting", True):
+                out[field] = flags[field].replace(trans_map).infer_objects()
             diff = pd.Index(out[field]).difference(expected)
             if not diff.empty:
                 raise ValueError(

saqc/lib/tools.py

@@ -177,7 +177,24 @@ def periodicMask(
             x[e:e] = True
             return x
 
-    freq = "1" + "mmmhhhdddMMMYYY"[len(season_start)]
+    freq = (
+        "1",
+        "m",
+        "m",
+        "m",
+        "h",
+        "h",
+        "h",
+        "d",
+        "d",
+        "d",
+        "M",
+        "M",
+        "M",
+        "YE",
+        "YE",
+        "YE",
+    )[len(season_start)]
     out = mask.groupby(pd.Grouper(freq=freq)).transform(_selector)
     if invert:
         out = ~out

tests/funcs/test_functions.py

@@ -52,7 +52,7 @@ def test_flagRange(data, field):
     assert all(flagged == expected)
 
 
-def test_flagSeasonalRange(data, field):
+def test_selectTime(data, field):
     data[field].iloc[::2] = 0
     data[field].iloc[1::2] = 50
     nyears = len(data[field].index.year.unique())

tests/funcs/test_outlier_detection.py

@@ -33,7 +33,7 @@ def test_flagMad(spiky_data):
     field, *_ = data.columns
     flags = initFlagsLike(data)
     qc = SaQC(data, flags).flagZScore(
-        field, window="1H", method="modified", thresh=3.5, flag=BAD
+        field, window="1h", method="modified", thresh=3.5, flag=BAD
     )
     flag_result = qc.flags[field]
     test_sum = (flag_result.iloc[spiky_data[1]] == BAD).sum()

tests/fuzzy/lib.py

@@ -104,13 +104,13 @@ def daterangeIndexes(draw, min_size=0, max_size=100):
     max_date = pd.Timestamp("2099-12-31").to_pydatetime()
     start = draw(datetimes(min_value=min_date, max_value=max_date))
     periods = draw(integers(min_value=min_size, max_value=max_size))
-    freq = draw(sampled_from(["D", "H", "T", "min", "S", "L", "ms", "U", "us", "N"]))
+    freq = draw(sampled_from(["D", "h", "min", "s", "ms", "us", "ns"]))
     return pd.date_range(start, periods=periods, freq=freq)
 
 
 @composite
 def frequencyStrings(draw, _):
-    freq = draw(sampled_from(["D", "H", "T", "min", "S", "L", "ms", "U", "us", "N"]))
+    freq = draw(sampled_from(["D", "h", "min", "s", "ms", "us", "ns"]))
     mult = draw(integers(min_value=1, max_value=10))
     value = f"{mult}{freq}"
     return value