Add drop_na() method to Series class

pyleoclim/core/series.py

@@ -2368,6 +2368,27 @@ def clean(self, verbose=False, keep_log = False):
             new.log += ({len(new.log):'clean', 'verbose': verbose},)
         return new
 
+    def drop_na(self):
+        ''' Remove NaN values while preserving temporal order
+
+        Unlike clean(), this method only removes NaN values without sorting the data.
+
+        Returns
+        -------
+        Series : Series
+            Series object with NaN values removed
+
+        See Also
+        --------
+        clean : Remove NaNs and sort by time
+
+        '''
+        new = self.copy()
+        valid_mask = ~(np.isnan(self.time) | np.isnan(self.value))
+        new.time = self.time[valid_mask]
+        new.value = self.value[valid_mask]
+        return new
+
     def sort(self, verbose=False, ascending = True, keep_log = False):
         ''' Ensure timeseries is set to a monotonically increasing axis.
             If the time axis is prograde to begin with, no transformation is applied.

pyleoclim/tests/test_core_Series.py

@@ -414,6 +414,30 @@ def test_clean(self, gen_normal):
         assert time[len(time) - 1] >= time[0]
         assert len(time) == len(value)
 
+    def test_drop_na(self, gen_normal):
+        """Test drop_na method"""
+
+        # Generate data
+        ts = gen_normal()
+
+        # Add NaN values to test drop_na functionality
+        v = ts.value.copy()
+        v[1] = np.nan
+        v[3] = np.nan
+        ts_with_nans = pyleo.Series(time=ts.time, value=v, dropna=False)
+
+        # Call function for testing
+        ts_clean = ts_with_nans.drop_na()
+
+        # Isolate time and value components
+        time = ts_clean.__dict__["time"]
+        value = ts_clean.__dict__["value"]
+
+        assert len(time) == len(value)
+        assert len(time) == 98  # Original 100 - 2 NaNs = 98
+        assert not np.any(np.isnan(time))
+        assert not np.any(np.isnan(value))
+
 
 
 class TestSeriesAnnualize:   # Written mostly by Claude AI