Merge branch 'features/tools' into features/tests

Author: SabineHaas

windpowerlib/modelchain.py

@@ -118,7 +118,8 @@ def rho_hub(self, weather, data_height):
         ----------
         weather : DataFrame or Dictionary
             Containing columns or keys with timeseries for temperature
-            `temp_air` in K and pressure `pressure` in Pa.
+            `temp_air` in K and pressure `pressure` in Pa, as well as
+            optionally the temperature `temp_air_2` in K at a different height.
             If a Dictionary is used the data inside the dictionary has to be of
             the types pandas.Series or numpy.array.
         data_height : Dictionary
@@ -305,8 +306,8 @@ def run_model(self, weather, data_height):
             Containing columns or keys with the timeseries for wind speed
             `v_wind` in m/s, roughness length `z0` in m, temperature
             `temp_air` in K and pressure `pressure` in Pa, as well as
-            optionally wind speed `v_wind_2` in m/s in K at different height
-            for interpolation.
+            optionally wind speed `v_wind_2` in m/s and temperature
+-            `temp_air_2` in K at different height.
             If a Dictionary is used the data inside the dictionary has to be of
             the types pandas.Series or numpy.array.
         data_height : Dictionary

windpowerlib/tools.py

@@ -4,58 +4,47 @@
 
 """
 
-import numpy as np
-
 
 __copyright__ = "Copyright oemof developer group"
 __license__ = "GPLv3"
 
 
-def smallest_difference(value_1, value_2, comp_value, corresp_1, corresp_2):
+def smallest_difference(data_frame, comp_value, column_name):
     r"""
-    Selects the value with the smaller difference to a comparative value.
+    Selects a value with the smallest difference to a comparative value.
 
     Additionally returns a corresponding value. This function is for example
     used in :py:func:`~.modelchain.v_wind_hub` of the
     :class:`~.modelchain.ModelChain` to choose the wind speed data that is
-    close to the hub height of the examined wind turbine. In this case
-    `value_1` and `value_2` are the heights of the corresponding wind speed
-    data sets `corresp_1` and `corresp_2`.
+    close to the hub height of the examined wind turbine. In this case the
+    column of the data frame contains wind speed time series and the indices
+    are the corresponding heights for which these time series apply.
 
     Parameters
     ----------
-    value_1 : float
-        First value of which the difference to `comp_value` will be
-        compared with the difference to `comp_value` of `value_2`.
-    value_2 : float
-        Second value for comparison.
+    data_frame : DataFrame
+        Indices are the values of which the smallest difference to `comp_value`
+        will be found, the corresponding values are in the column
+        specified by `column_name` and they can be floats, pd.Series or
+        np.arrays.
     comp_value : float
         Comparative value.
-    corresp_1 : pd.Series or np.array or float
-        Corresponding value to `value_1`.
-    corresp_2 : pd.Series or np.array or float
-        Corresponding value to `value_2`.
+    column_name : string
+        Name of the column in the `data_frame` that contains the
+        correponding values.
 
     Returns
     -------
     Tuple(float, pd.Series or np.array or float)
-        Value closer to comparing value as float and its corresponding value as
-        float.
+        Closest value to comparative value as float and its corresponding value
+        as float.
 
     """
-    if (value_2 is not None and corresp_2 is not None):
-        if abs(value_1 - comp_value) <= abs(value_2 - comp_value):
-            closest_value = value_1
-        else:
-            closest_value = value_2
-    else:
-        closest_value = value_1
-
-    # Select correponding value
-    if closest_value == value_1:
-        corresp_value = corresp_1
-    else:
-        corresp_value = corresp_2
+    diff = []
+    for index in sorted(data_frame.index):
+        diff.append(abs(comp_value - index))
+    closest_value = sorted(data_frame.index)[diff.index(min(diff))]
+    corresp_value = data_frame[column_name][closest_value]
     return (closest_value, corresp_value)
 
 
@@ -95,29 +84,19 @@ def linear_extra_interpolation(data_frame, requested_height, column_name):
               (height_{requested} - height_1) + value_1
 
     with:
-        :math:`height_2`: largest/smallest value in data frame,
-        :math:`height_1`: second largest/smallest value in data frame,
+        :math:`height_2`: largest/smallest index of data frame,
+        :math:`height_1`: second largest/smallest index of data frame,
         :math:`value_2`: corresponding value to `height_2`,
         :math:`value_1`: correponding value to `height_1`,
-        :math:`height_{requested}` : Height for which the interpolation takes
+        :math:`height_{requested}` : height for which the interpolation takes
         place
 
     """
-    if requested_height > max(data_frame.index):
-        height_2 = max(data_frame.index)
-        value_2 = data_frame[column_name][height_2]
-        height_1 = sorted(data_frame.index)[-2]  # Second largest number
-        value_1 = data_frame[column_name][height_1]
-        interpolant = ((value_2 - value_1) / (height_2 - height_1) *
-                       (requested_height - height_1) + value_1)
-    elif requested_height < min(data_frame.index):
-        height_2 = min(data_frame.index)
-        value_2 = data_frame[column_name][height_2]
-        height_1 = sorted(data_frame.index)[1]  # Second smallest number
-        value_1 = data_frame[column_name][height_1]
-        interpolant = ((value_2 - value_1) / (height_2 - height_1) *
-                       (requested_height - height_1) + value_1)
-    else:
-        interpolant = np.interp(requested_height, data_frame.index,
-                                data_frame[column_name])
+    height_2, value_2 = smallest_difference(data_frame, requested_height,
+                                            column_name)
+    data_frame_2 = data_frame.drop(height_2)
+    height_1, value_1 = smallest_difference(data_frame_2, requested_height,
+                                            column_name)
+    interpolant = ((value_2 - value_1) / (height_2 - height_1) *
+                   (requested_height - height_1) + value_1)
     return interpolant