⚡️ time_integrate uses numpy for trapz. Also handle empty DataFrame

tests/test_math.py

@@ -70,6 +70,18 @@ def test_time_integrate(self):
 
         pd.testing.assert_series_equal(ref, time_integrate(test) / 3600)
 
+        test = pd.DataFrame(
+            {
+                "cpt1": [],
+                "cpt2": [],
+            },
+            index=pd.date_range("2009-01-01 00:00:00", freq="10s", periods=0, tz="UTC"),
+        )
+
+        ref = pd.Series(index=test.columns, dtype=float)
+
+        pd.testing.assert_series_equal(ref, time_integrate(test))
+
     def test_aggregate_time_series(self):
         sim_res = pd.DataFrame(
             {"a": [1, 2], "b": [3, 4]},

tide/math.py

@@ -3,7 +3,6 @@
 import pandas as pd
 import numpy as np
 
-from scipy import integrate
 from collections.abc import Callable
 
 from tide.utils import check_and_return_dt_index_df
@@ -80,8 +79,8 @@ def time_gradient(data: pd.DataFrame | pd.Series) -> pd.DataFrame:
 
 def time_integrate(data: pd.DataFrame | pd.Series) -> pd.Series:
     """
-    Perform time Integration of given time series in X DartaFrame or in a Series.
-    The function computes the integral of each column using `scipy.integrate.trapz`
+    Perform time Integration of given time series in X DataFrame or in a Series.
+    The function computes the integral of each column using `np.trapz`
     function and the time difference between consecutive data points.
 
     Parameters:
@@ -92,14 +91,14 @@ def time_integrate(data: pd.DataFrame | pd.Series) -> pd.Series:
     """
 
     data = check_and_return_dt_index_df(data)
-    chrono = (data.index - data.index[0]).to_series()
-    chrono = chrono.dt.total_seconds()
+    if data.empty:
+        return pd.Series(index=data.columns, dtype=float)
 
-    res_series = pd.Series(dtype="float64")
-    for col in data:
-        res_series[col] = integrate.trapezoid(data[col], chrono)
+    t = (data.index.view("int64") - data.index[0].value) * 1e-9  # seconds
+    y = data.to_numpy()
+    result = np.trapz(y, t, axis=0)
 
-    return res_series
+    return pd.Series(result, index=data.columns)
 
 
 def aggregate_time_series(