wtclean 0.2.8__tar.gz → 0.2.26__tar.gz

{wtclean-0.2.8/wtclean.egg-info → wtclean-0.2.26}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wtclean
-Version: 0.2.8
+Version: 0.2.26
 Summary: Wind turbine SCADA power-curve data cleaning via iterative bin + MAD filtering
 License: BSD-3-Clause
 Keywords: wind-energy,wind-turbine,scada,power-curve,outlier-detection,data-cleaning

{wtclean-0.2.8 → wtclean-0.2.26}/setup.py

@@ -9,7 +9,7 @@ with open(os.path.join(HERE, "README.md"), "r", encoding="utf-8") as fh:
 
 setuptools.setup(
     name="wtclean",
-    version="0.2.8",
+    version="0.2.26",
     description="Wind turbine SCADA power-curve data cleaning via iterative bin + MAD filtering",
     long_description=long_description,
     long_description_content_type="text/markdown",

{wtclean-0.2.8 → wtclean-0.2.26}/test/test_daily_anomaly_detection.py

@@ -84,27 +84,7 @@ class TestDailyAnomalyDetector(unittest.TestCase):
         with self.assertRaises(ValueError):
             _det(df, low_power_ratio=1.5)
         with self.assertRaises(ValueError):
-            DailyAnomalyDetector(
-                "ws", "pw", "ts", df,
-                cut_in_speed=0.0, rated_wind_speed=RATED_WS, rated_power=RATED_PWR
-            )
-
-    def test_output_columns(self):
-        df, _ = _synth()
-        normal, abnormal = _det(df).detect()
-        self.assertEqual(list(normal.columns), list(df.columns))
-        self.assertEqual(list(abnormal.columns), list(df.columns))
-        for frame in (normal, abnormal):
-            for col in ("_row_id", "_date", "_resid"):
-                self.assertNotIn(col, frame.columns)
-
-    def test_single_day_not_all_abnormal(self):
-        df, _ = _synth(days=1)
-        det = _det(df)
-        normal, abnormal = det.detect()
-        self.assertGreater(len(normal), 0)
-        self.assertEqual(len(normal) + len(abnormal), len(df))
-        self.assertEqual(det.abnormal_days_, set())
+            _det(df, bin_width=0)
 
     def test_auto_estimate_params(self):
         df, _ = _synth()

{wtclean-0.2.8 → wtclean-0.2.26}/wtclean/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.1.0"
+__version__ = "0.2.26"
 
 from wtclean.modules.power_curve_filtering import PowerCurveFiltering
 from wtclean.modules.daily_anomaly_detection import DailyAnomalyDetector

{wtclean-0.2.8 → wtclean-0.2.26}/wtclean/modules/daily_anomaly_detection.py

@@ -106,7 +106,7 @@ class DailyAnomalyDetector:
 
         Returns:
             按行划分原始 ``df`` 的正常 / 异常两个子集；被判异常的天可通过
-            ``self.abnormal_days_`` 查看。
+            
         """
         work = self.df.copy()
         work["_row_id"] = np.arange(len(work))
@@ -151,8 +151,6 @@ class DailyAnomalyDetector:
         analysis_mask &= ws <= cut_out_eff
 
         if not analysis_mask.any():
-            self.normal_days_ = set()
-            self.abnormal_days_ = set(work["_date"].dropna())
             return self.df.iloc[0:0].copy(), self.df.copy()
 
         # ---- 整机基准曲线 ----
@@ -182,8 +180,6 @@ class DailyAnomalyDetector:
         day_stats = raw_day_stats.loc[count_mask]
 
         if day_stats.empty:
-            self.normal_days_ = set()
-            self.abnormal_days_ = set(work["_date"].dropna())
             return self.df.iloc[0:0].copy(), self.df.copy()
 
         median_resid = day_stats["median"].median()
@@ -192,9 +188,7 @@ class DailyAnomalyDetector:
             median_resid - k * mad_resid, median_resid + k * mad_resid
         )
         normal_dates = set(day_stats.index[normal_mask])
-        self.normal_days_ = normal_dates
-        self.abnormal_days_ = set(work["_date"].dropna()) - normal_dates
-
+ 
         normal_row_ids = clean_df.loc[
             clean_df["_date"].isin(normal_dates), "_row_id"
         ]
@@ -202,3 +196,4 @@ class DailyAnomalyDetector:
         normal_df = work.loc[is_normal, self.df.columns].copy()
         abnormal_df = work.loc[~is_normal, self.df.columns].copy()
         return normal_df, abnormal_df
+

{wtclean-0.2.8 → wtclean-0.2.26/wtclean.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wtclean
-Version: 0.2.8
+Version: 0.2.26
 Summary: Wind turbine SCADA power-curve data cleaning via iterative bin + MAD filtering
 License: BSD-3-Clause
 Keywords: wind-energy,wind-turbine,scada,power-curve,outlier-detection,data-cleaning