Merge pull request #52 from wind-python/turbine_library_as_csv

Turbine library as csv

Author: birgits

.gitignore

@@ -101,3 +101,5 @@ ENV/
 
 # oedb data dump
 /windpowerlib/data/turbine_data_oedb.h5
+/windpowerlib/data/oedb_power_coefficient_curves.csv
+/windpowerlib/data/oedb_power_curves.csv

doc/whatsnew/v0-1-2.txt

@@ -19,7 +19,7 @@ Bug fixes
 
 Other changes
 #############
-* Make windpowerlib work offline: turbine data from oedb is stored in a hdf5 file for offline usage
+* Make windpowerlib work offline: turbine data from oedb is stored in csv files for offline usage
 * Make :py:func:`~windpowerlib.wind_turbine.get_turbine_types` also accessible via `get_turbine_types()` --> from windpowerlib import get_turbine_types
 
 

tests/test_wind_turbine.py

@@ -38,4 +38,5 @@ def test_get_turbine_data_from_file(self):
                                        file_='not_existent')
 
     def test_get_turbine_types(self):
-        get_turbine_types(print_out=False)
+        get_turbine_types(print_out=True, filter_=True)
+        get_turbine_types(print_out=False, filter_=False)

windpowerlib/turbine_cluster_modelchain.py

@@ -183,8 +183,7 @@ def assign_power_curve(self, weather_df):
         self
 
         """
-
-        # Set turbulence intensity for assigning power curve
+        # Get turbulence intensity from weather if existent
         turbulence_intensity = (
             weather_df['turbulence_intensity'].values.mean() if
             'turbulence_intensity' in

windpowerlib/wind_turbine.py

@@ -171,6 +171,10 @@ def fetch_turbine_data(self, fetch_curve, data_source):
         4200000.0
 
         """
+        if fetch_curve not in ['power_curve', 'power_coefficient_curve']:
+            raise ValueError("'{0}' is an invalid value for ".format(
+                fetch_curve) + "`fetch_curve`. Must be " +
+                             "'power_curve' or 'power_coefficient_curve'.")
         if data_source == 'oedb':
             curve_df, nominal_power = get_turbine_data_from_oedb(
                 turbine_type=self.name, fetch_curve=fetch_curve)
@@ -181,10 +185,6 @@ def fetch_turbine_data(self, fetch_curve, data_source):
             self.power_curve = curve_df
         elif fetch_curve == 'power_coefficient_curve':
             self.power_coefficient_curve = curve_df
-        else:
-            raise ValueError("'{0}' is an invalid value. ".format(
-                             fetch_curve) + "`fetch_curve` must be " +
-                             "'power_curve' or 'power_coefficient_curve'.")
         if self.nominal_power is None:
             self.nominal_power = nominal_power
         return self
@@ -202,15 +202,15 @@ def get_turbine_data_from_file(turbine_type, file_):
 
     Parameters
     ----------
-    turbine_type : string
+    turbine_type : str
         Specifies the turbine type data is fetched for.
-    file_ : string
+    file_ : str
         Specifies the source of the turbine data.
         See the example below for how to use the example data.
 
     Returns
     -------
-    Tuple (pandas.DataFrame, float)
+    tuple(pandas.DataFrame, float)
         Power curve or power coefficient curve (pandas.DataFrame) and nominal
         power (float). Power (coefficient) curve DataFrame contains power
         coefficient curve values (dimensionless) or power curve values in W
@@ -246,11 +246,17 @@ def isfloat(x):
         df = pd.read_csv(file_, index_col=0)
     except FileNotFoundError:
         raise FileNotFoundError("The file '{}' was not found.".format(file_))
-    wpp_df = df[df.turbine_id == turbine_type]
+    # todo: note: this try except statement will be removed in 0.2.0 and only
+    #  the exception will stay. The example power (coefficient) curve files
+    #  will then be adapted
+    try:
+        wpp_df = df[df['turbine_id'] == turbine_type]
+    except KeyError:
+        wpp_df = df[df.index == turbine_type]
     # if turbine not in data file
     if wpp_df.shape[0] == 0:
         pd.set_option('display.max_rows', len(df))
-        logging.info('Possible types: \n{0}'.format(df.turbine_id))
+        logging.info('Possible types: \n{0}'.format(df['turbine_id']))
         pd.reset_option('display.max_rows')
         sys.exit('Cannot find the wind converter type: {0}'.format(
             turbine_type))
@@ -261,73 +267,62 @@ def isfloat(x):
     df = curve_data.transpose().reset_index()
     df.columns = ['wind_speed', 'value']
     df['wind_speed'] = df['wind_speed'].apply(lambda x: float(x))
-    nominal_power = wpp_df['p_nom'].iloc[0]
+    # todo: note: this try except statement will be removed in 0.2.0 and only
+    #  the exception will stay. The example power (coefficient) curve files
+    #  will then be adapted
+    try:
+        nominal_power = wpp_df['p_nom'].iloc[0]
+    except KeyError:
+        nominal_power = float(wpp_df['nominal_power'].iloc[0])
     return df, nominal_power
 
 
 def get_turbine_data_from_oedb(turbine_type, fetch_curve, overwrite=False):
     r"""
-    Fetches data for one wind turbine type from the OpenEnergy Database (oedb).
+    Fetches wind turbine data from the OpenEnergy database (oedb).
 
     If turbine data exists in local repository it is loaded from this file. The
-    file is created when turbine data was loaded from oedb in
+    file is created when turbine data is loaded from oedb in
     :py:func:`~.load_turbine_data_from_oedb`. Use this function with
     `overwrite=True` to overwrite your file with newly fetched data.
-    Use :py:func:`~.check_local_turbine_data` to check
-    weather your local file is up to date.
 
     Parameters
     ----------
-    turbine_type : string
+    turbine_type : str
         Specifies the turbine type data is fetched for.
         Use :py:func:`~.get_turbine_types` to see a table of all wind turbines
-        for which power (coefficient) curve data is provided.
-    fetch_curve : string
+        in oedb containing information about whether power (coefficient) curve
+        data is provided.
+    fetch_curve : str
         Parameter to specify whether a power or power coefficient curve
         should be retrieved from the provided turbine data. Valid options are
         'power_curve' and 'power_coefficient_curve'. Default: None.
-    overwrite : boolean
-        If True local file is overwritten by newly fetch data from oedb, if
+    overwrite : bool
+        If True local file is overwritten by newly fetched data from oedb, if
         False turbine data is fetched from previously saved file.
 
     Returns
     -------
-    Tuple (pandas.DataFrame, float)
+    tuple(pandas.DataFrame, float)
         Power curve or power coefficient curve (pandas.DataFrame) and nominal
         power (float) of one wind turbine type. Power (coefficient) curve
         DataFrame contains power coefficient curve values (dimensionless) or
         power curve values in W with the corresponding wind speeds in m/s.
 
     """
-    # hdf5 filename
     filename = os.path.join(os.path.dirname(__file__), 'data',
-                            'turbine_data_oedb.h5')
-    if os.path.isfile(filename) and not overwrite:
-        logging.debug("Turbine data is fetched from {}".format(filename))
-        with pd.HDFStore(filename) as hdf_store:
-            turbine_data = hdf_store.get('turbine_data')
-    else:
-        turbine_data = load_turbine_data_from_oedb()
-    turbine_data.set_index('turbine_type', inplace=True)
-    # Set `curve` depending on `fetch_curve` to match names in oedb
-    curve = ('cp_curve' if fetch_curve == 'power_coefficient_curve'
-             else fetch_curve)
-    # Select curve and nominal power of turbine type
-    try:
-        df = turbine_data.loc[turbine_type]
-    except KeyError:
-        raise KeyError("Turbine type '{}' not in database. ".format(
-            turbine_type) + "Use 'get_turbine_types()' to see a table of " +
-                       "possible wind turbine types.")
-    if df[curve] is not None:
-        df = pd.DataFrame(df[curve])
+                            'oedb_{}s.csv'.format(fetch_curve))
+    if not os.path.isfile(filename) or overwrite:
+        # Load data from oedb and save to csv file
+        load_turbine_data_from_oedb()
     else:
-        sys.exit("{} of {} not available in ".format(curve, turbine_type) +
-                 "oedb. Use 'get_turbine_types()' to see for which turbine " +
-                 "types power coefficient curves are available.")
-    nominal_power = turbine_data.loc[turbine_type][
-                        'installed_capacity_kw'] * 1000
-    df.columns = ['wind_speed', 'value']
+        logging.debug("Turbine data is fetched from {}".format(filename))
+    # turbine_data = pd.read_csv(filename, index_col=0)
+    df, nominal_power = get_turbine_data_from_file(turbine_type=turbine_type,
+                                                   file_=filename)
+
+    # nominal power and power curve values in W
+    nominal_power = nominal_power * 1000
     if fetch_curve == 'power_curve':
         # power in W
         df['value'] = df['value'] * 1000
@@ -336,64 +331,103 @@ def get_turbine_data_from_oedb(turbine_type, fetch_curve, overwrite=False):
 
 def load_turbine_data_from_oedb():
     r"""
-    Loads turbine data from the OpenEnergy Database (oedb).
+    Loads turbine data from the OpenEnergy database (oedb).
 
-    Turbine data is saved to `filename` for offline usage of windpowerlib.
+    Turbine data is saved to csv files ('oedb_power_curves.csv' and
+    'oedb_power_coefficient_curves.csv') for offline usage of windpowerlib.
+    If the files already exist they are overwritten.
 
     Returns
     -------
-    turbine_data : pd.DataFrame
+    pd.DataFrame
         Contains turbine data of different turbines such as 'manufacturer',
-        'turbine_type', nominal power ('installed_capacity_kw').
+        'turbine_type', 'nominal_power'.
 
     """
     # url of OpenEnergy Platform that contains the oedb
     oep_url = 'http://oep.iks.cs.ovgu.de/'
     # location of data
-    schema = 'model_draft'
-    table = 'openfred_windpower_powercurve'
+    schema = 'supply'
+    table = 'turbine_library'
     # load data
     result = requests.get(
         oep_url + '/api/v0/schema/{}/tables/{}/rows/?'.format(
             schema, table), )
     if not result.status_code == 200:
         raise ConnectionError("Database connection not successful. "
                               "Response: [{}]".format(result.status_code))
-    # extract data to data frame
+    # extract data to dataframe
     turbine_data = pd.DataFrame(result.json())
-    # store data as hdf5
+    # standard file name for saving data
     filename = os.path.join(os.path.dirname(__file__), 'data',
-                            'turbine_data_oedb.h5')
-    with pd.HDFStore(filename) as hdf_store:
-        hdf_store.put('turbine_data', turbine_data)
-    logging.debug("Turbine data is fetched from oedb and saved "
-                  "to {}".format(filename))
+                            'oedb_{}.csv')
+    # get all power (coefficient) curves and save to file
+    # for curve_type in ['power_curve', 'power_coefficient_curve']:
+    for curve_type in ['power_curve', 'power_coefficient_curve']:
+        curves_df = pd.DataFrame(columns=['wind_speed'])
+        for index in turbine_data.index:
+            if (turbine_data['{}_wind_speeds'.format(curve_type)][index]
+                    and turbine_data['{}_values'.format(curve_type)][index]):
+                df = pd.DataFrame(data=[
+                    eval(turbine_data['{}_wind_speeds'.format(curve_type)][
+                             index]),
+                    eval(turbine_data['{}_values'.format(curve_type)][
+                             index])]).transpose().rename(
+                    columns={0: 'wind_speed',
+                             1: turbine_data['turbine_type'][index]})
+                curves_df = pd.merge(left=curves_df, right=df, how='outer',
+                                     on='wind_speed')
+        curves_df = curves_df.set_index('wind_speed').sort_index().transpose()
+        curves_df['turbine_type'] = curves_df.index
+        # add nominal power to power (coefficient) data frame
+        curves_df = pd.merge(
+            left=curves_df, right=turbine_data[['turbine_type',
+                                                'installed_capacity']],
+            on='turbine_type').set_index('turbine_type').rename(
+                columns={'installed_capacity': 'nominal_power'})
+        curves_df.to_csv(filename.format('{}s'.format(curve_type)))
+
     return turbine_data
 
 
-def get_turbine_types(print_out=True):
+def get_turbine_types(print_out=True, filter_=True):
     r"""
-    Get the names of all possible wind turbine types for which the power
-    coefficient curve or power curve is provided in the OpenEnergy Data Base
-    (oedb).
+    Get all wind turbine types provided in the OpenEnergy database (oedb).
+
+    By default only turbine types for which a power coefficient curve or power
+    curve is provided are returned. Set `filter_=False` to see all turbine
+    types for which any data (f.e. hub height, rotor diameter, ...) is
+    provided.
 
     Parameters
     ----------
-    print_out : boolean
+    print_out : bool
         Directly prints a tabular containing the turbine types in column
         'turbine_type', the manufacturer in column 'manufacturer' and
         information about whether a power (coefficient) curve exists (True) or
         not (False) in columns 'has_power_curve' and 'has_cp_curve'.
         Default: True.
+    filter_ : bool
+        If True only turbine types for which a power coefficient curve or
+        power curve is provided in the OpenEnergy database (oedb) are
+        returned. Default: True.
 
     Returns
     -------
-    curves_df : pd.DataFrame
+    pd.DataFrame
         Contains turbine types in column 'turbine_type', the manufacturer in
         column 'manufacturer' and information about whether a power
         (coefficient) curve exists (True) or not (False) in columns
         'has_power_curve' and 'has_cp_curve'.
 
+    Notes
+    -----
+    If the power (coefficient) curve of the desired turbine type (or the
+    turbine type itself) is missing you can contact us via github or
+    [email protected]. You can help us by providing data in the
+    format as shown in
+    `the data base <https://openenergy-platform.org/dataedit/view/model_draft/openfred_windpower_powercurve>`_.
+
     Examples
     --------
     >>> from windpowerlib import wind_turbine
@@ -412,14 +446,17 @@ def get_turbine_types(print_out=True):
     Name: 1, dtype: object
 
     """
-
     df = load_turbine_data_from_oedb()
-    cp_curves_df = df.iloc[df.loc[df['has_cp_curve']].index][
-        ['manufacturer', 'turbine_type', 'has_cp_curve']]
-    p_curves_df = df.iloc[df.loc[df['has_power_curve']].index][
-        ['manufacturer', 'turbine_type', 'has_power_curve']]
-    curves_df = pd.merge(p_curves_df, cp_curves_df, how='outer',
-                         sort=True).fillna(False)
+    if filter_:
+        cp_curves_df = df.loc[df['has_cp_curve']][
+            ['manufacturer', 'turbine_type', 'has_cp_curve']]
+        p_curves_df = df.loc[df['has_power_curve']][
+            ['manufacturer', 'turbine_type', 'has_power_curve']]
+        curves_df = pd.merge(p_curves_df, cp_curves_df, how='outer',
+                             sort=True).fillna(False)
+    else:
+        curves_df = df[['manufacturer', 'turbine_type', 'has_power_curve',
+                        'has_cp_curve']]
     if print_out:
         pd.set_option('display.max_rows', len(curves_df))
         print(curves_df)