fixed the constants test

docs/FunctionDescriptions.md

@@ -314,14 +314,13 @@ doi:10.2136/vzj2012.0097.
 ### constant
 
 ```
-constant(eps, length, thmin=None)
+constant(window, eps=0)
 ```
 
-| parameter          | data type | default value | description |
-| ------             | ------    | ------        | ----        |
-| eps                |           |               |             |
-| length             |           |               |             |
-| thmin              |           | `None`        |             |
+| parameter | data type                                                             | default value | description |
+| ------    | ------                                                                | ------        | ----        |
+| window    | [offset string](docs/ParameterDescriptions.md#offset-strings)/integer |               |             |
+| eps       | float                                                                 | 0             |             |
 
 
 ### constants_varianceBased

saqc/funcs/constants_detection.py

@@ -16,33 +16,20 @@ from saqc.lib.tools import (
 # todo: maybe generalize flag_constant to work on non harmonized data as well.
 
 @register("constant")
-def flagConstant(data, field, flagger, eps, length, thmin=None, **kwargs):
-    datacol = data[field]
+def flagConstant(data, field, flagger, eps, window, **kwargs):
 
-    length = (
-        (pd.to_timedelta(length) - data.index.freq).to_timedelta64().astype(np.int64)
-    )
-
-    values = datacol.mask((datacol < thmin) | datacol.isnull()).values.astype(np.int64)
-
-    dates = datacol.index.values.astype(np.int64)
-
-    mask = np.isfinite(values)
-
-    for start_idx, end_idx in slidingWindowIndices(datacol.index, length):
-        mask_chunk = mask[start_idx:end_idx]
-        values_chunk = values[start_idx:end_idx][mask_chunk]
-        dates_chunk = dates[start_idx:end_idx][mask_chunk]
+    window = pd.tseries.frequencies.to_offset(window)
+    def _func(ts):
+        if np.all(ts) and (ts.index[-1] - ts.index[0]) >= window:
+            return ts
+        return ts & False
 
-        # we might have removed dates from the start/end of the
-        # chunk resulting in a period shorter than 'length'
-        # print (start_idx, end_idx)
-        if valueRange(dates_chunk) < length:
-            continue
-        if valueRange(values_chunk) < eps:
-            flagger = flagger.setFlags(field, loc=slice(start_idx, end_idx), **kwargs)
+    diffs = data[field].diff().abs() <= eps
+    mask = diffs.groupby(diffs).apply(_func)
+    # NOTE: the first value of a constant period is missed by the groupby
+    mask[np.where(mask)[0] - 1] = True
 
-    data[field] = datacol
+    flagger = flagger.setFlags(field, mask, **kwargs)
     return data, flagger
 
 

test/funcs/test_constants_detection.py

@@ -19,6 +19,19 @@ def data():
     return constants_data
 
 
+@pytest.mark.parametrize("flagger", TESTFLAGGER)
+def test_flagConstants(data, flagger):
+    data.iloc[5:25] = 200
+    expected = np.arange(5, 25)
+    field, *_ = data.columns
+    flagger = flagger.initFlags(data)
+    data, flagger_result = flagConstant(
+        data, field, flagger, window="15Min", eps=.1,
+    )
+    flags = flagger_result.getFlags(field)
+    assert np.all(flags[expected] == flagger.BAD)
+
+
 @pytest.mark.parametrize("flagger", TESTFLAGGER)
 def test_flagConstants_varianceBased(data, flagger):
     data.iloc[5:25] = 200
@@ -28,16 +41,7 @@ def test_flagConstants_varianceBased(data, flagger):
     data, flagger_result1 = flagConstant_varianceBased(
         data, field, flagger, plateau_window_min="1h"
     )
-    #data, flagger_result2 = flagConstant(
-    #    data, field, flagger, eps=0.1, length='1h'
-    #)
 
     flag_result1 = flagger_result1.getFlags(field)
     test_sum = (flag_result1[expected] == flagger.BAD).sum()
     assert test_sum == len(expected)
-
-    #flag_result2 = flagger_result2.getFlags(field)
-    #test_sum = (flag_result2[expected] == flagger.BAD).sum()
-    #assert test_sum == len(expected)
-
-