Feature/add daily bfg harvester soiltemp4

src/score_hv/harvesters/daily_bfg.py

@@ -25,12 +25,12 @@
 VALID_VARIABLES  = (#'icetk', # sea ice thickness (m)
                     #'lhtfl_ave', # surface latent heat flux (W m^-2)
                     #'prate_ave', # surface precip rate (mm weq. s^-1)
-                    'prateb_ave', # bucket surface precip rate (mm weq. s^-1)
+                    #'prateb_ave', # bucket surface precip rate (mm weq. s^-1)
                     #'pressfc', # surface pressure (Pa)
                     #'snod', # surface snow depth (m)
                     #'soil4', # liquid soil moisture at layer-4 (?)
                     #'soilm', # total column soil moisture content (mm weq.)
-                    #'soilt4', # soil temperature unknown layer 4 (K)
+                    'soilt4', # soil temperature unknown layer 4 (K)
                     #'tg3', # deep soil temperature (K)
                     #'tmp2m', # 2m (surface air) temperature (K)
                     #'tmpsfc', # surface temperature (K)
@@ -183,7 +183,6 @@ def get_data(self):
                 """
                 if statistic == 'mean':
                     value = expected_value
-
                 elif statistic == 'variance':
                     value = -expected_value**2 + np.ma.sum(
                                                    temporal_means**2 * 
@@ -192,10 +191,10 @@ def get_data(self):
 
                 elif statistic == 'maximum':
                     value = np.ma.max(temporal_means)
-                
+
                 elif statistic == 'minimum':
                     value = np.ma.min(temporal_means)
-                
+
                 harvested_data.append(HarvestedData(
                                     self.config.harvest_filenames,
                                     statistic, 
@@ -207,4 +206,4 @@ def get_data(self):
 
         gridcell_area_data.close()
         xr_dataset.close()
-        return harvested_data
+        return harvested_data

tests/test_harvester_daily_bfg_prateb.py → tests/test_harvester_global_soilt4.py

@@ -15,21 +15,20 @@
 from score_hv.yaml_utils import YamlLoader
 from score_hv.harvesters.innov_netcdf import Region, InnovStatsCfg
 
-TEST_DATA_FILE_NAMES = ['bfg_1994010100_fhr09_prateb_control.nc',
-                        'bfg_1994010106_fhr06_prateb_control.nc',
-                        'bfg_1994010106_fhr09_prateb_control.nc',
-                        'bfg_1994010112_fhr06_prateb_control.nc',
-                        'bfg_1994010112_fhr09_prateb_control.nc',
-                        'bfg_1994010118_fhr06_prateb_control.nc',
-                        'bfg_1994010118_fhr09_prateb_control.nc',
-                        'bfg_1994010200_fhr06_prateb_control.nc']
+TEST_DATA_FILE_NAMES = ['bfg_1994010100_fhr09_soilt4_control.nc',
+                        'bfg_1994010106_fhr06_soilt4_control.nc',
+                        'bfg_1994010106_fhr09_soilt4_control.nc',
+                        'bfg_1994010112_fhr06_soilt4_control.nc',
+                        'bfg_1994010112_fhr09_soilt4_control.nc',
+                        'bfg_1994010118_fhr06_soilt4_control.nc',
+                        'bfg_1994010118_fhr09_soilt4_control.nc',
+                        'bfg_1994010200_fhr06_soilt4_control.nc']
 
 DATA_DIR = os.path.join(Path(__file__).parent.parent.resolve(), 'data')
 GRIDCELL_AREA_DATA_PATH = os.path.join(DATA_DIR,
-                                       'gridcell-area' + 
-                                       '_noaa-ufs-gefsv13replay-pds' + 
+                                       'gridcell-area' +
+                                       '_noaa-ufs-gefsv13replay-pds' +
                                        '_bfg_control_1536x768_20231116.nc')
-
 CONFIGS_DIR = 'configs'
 PYTEST_CALLING_DIR = Path(__file__).parent.resolve()
 TEST_DATA_PATH = os.path.join(PYTEST_CALLING_DIR, 'data')
@@ -38,8 +37,8 @@
 
 VALID_CONFIG_DICT = {'harvester_name': hv_registry.DAILY_BFG,
                      'filenames' : BFG_PATH,
-                     'statistic': ['mean', 'variance', 'minimum', 'maximum'],
-                     'variable': ['prateb_ave']}
+                     'statistic': ['mean','variance', 'minimum', 'maximum'],
+                     'variable': ['soilt4']}
 
 def test_gridcell_area_conservation(tolerance=0.001):
 
@@ -56,37 +55,37 @@ def test_gridcell_area_conservation(tolerance=0.001):
 
 def test_variable_names():
     data1 = harvest(VALID_CONFIG_DICT)
-    assert data1[0].variable == 'prateb_ave'
+    assert data1[0].variable == 'soilt4'
 
 def test_global_mean_values(tolerance=0.001):
-    """The value of 3.117e-05 is the mean value of the global means 
-    calculated from eight forecast files:
-
-        bfg_1994010100_fhr09_prateb_control.nc
-        bfg_1994010106_fhr06_prateb_control.nc
-        bfg_1994010106_fhr09_prateb_control.nc
-        bfg_1994010112_fhr06_prateb_control.nc
-        bfg_1994010112_fhr09_prateb_control.nc
-        bfg_1994010118_fhr06_prateb_control.nc
-        bfg_1994010118_fhr09_prateb_control.nc
-        bfg_1994010200_fhr06_prateb_control.nc
+    """ 
+        The value of 288.4441110340619 is the mean value of the 
+        global meas calculated from these eight forecast files:
+
+        bfg_1994010100_fhr09_soilt4_control.nc
+        bfg_1994010106_fhr06_soilt4_control.nc
+        bfg_1994010106_fhr09_soilt4_control.nc
+        bfg_1994010112_fhr06_soilt4_control.nc
+        bfg_1994010112_fhr09_soilt4_control.nc
+        bfg_1994010118_fhr06_soilt4_control.nc
+        bfg_1994010118_fhr09_soilt4_control.nc
+        bfg_1994010200_fhr06_soilt4_control.nc
 
-    When averaged together, these files represent a 24 hour mean. The 
-    average value hard-coded in this test was calculated from 
-    these forecast files using a separate python code.
+        When averaged together, these files represent a 24 hour mean. The 
+        average value hard-coded in this test was calculated from 
+        forecast files using a separate python code.
     """
     data1 = harvest(VALID_CONFIG_DICT)
-    global_mean = 3.1173840683271906e-05
+    global_mean = 288.4441110340619 
     assert data1[0].value <= (1 + tolerance) * global_mean
     assert data1[0].value >= (1 - tolerance) * global_mean
 
 def test_global_mean_values2(tolerance=0.001):
     """Opens each background Netcdf file using the
     netCDF4 library function Dataset and computes the expected value
-    of the provided variable.  In this case prateb_ave.
+    of the provided variable.  In this case soilt4.
     """
     data1 = harvest(VALID_CONFIG_DICT)
-
     gridcell_area_data = Dataset(GRIDCELL_AREA_DATA_PATH)
     norm_weights = gridcell_area_data.variables['area'][:] / np.sum(
                                         gridcell_area_data.variables['area'][:])
@@ -108,85 +107,85 @@ def test_global_mean_values2(tolerance=0.001):
             assert global_mean >= (1 - tolerance) * harvested_tuple.value
 
     gridcell_area_data.close()
-                
+
 def test_gridcell_variance(tolerance=0.001):
     """Opens each background Netcdf file using the
     netCDF4 library function Dataset and computes the variance
-    of the provided variable.  In this case prateb_ave.
+    of the provided variable.  In this case soilt4.
     """
     data1 = harvest(VALID_CONFIG_DICT)
-    
+
     gridcell_area_data = Dataset(GRIDCELL_AREA_DATA_PATH)
     norm_weights = gridcell_area_data.variables['area'][:] / np.sum(
                                         gridcell_area_data.variables['area'][:])
-    
+
     summation = np.ma.zeros(gridcell_area_data.variables['area'].shape)
     for file_count, data_file in enumerate(BFG_PATH):
         test_rootgrp = Dataset(data_file)
-    
+
         summation += test_rootgrp.variables[VALID_CONFIG_DICT['variable'][0]][0]
-        
+
         test_rootgrp.close()
-        
+
     temporal_mean = summation / (file_count + 1)
-    
+
     global_mean = np.ma.sum(norm_weights * temporal_mean)
     variance = np.ma.sum((temporal_mean - global_mean)**2 * norm_weights)
-    
+
     for i, harvested_tuple in enumerate(data1):
         if harvested_tuple.statistic == 'variance':
             assert variance <= (1 + tolerance) * harvested_tuple.value
             assert variance >= (1 - tolerance) * harvested_tuple.value
-            
+
     gridcell_area_data.close()
-    
+
 def test_gridcell_min_max(tolerance=0.001):
     """Opens each background Netcdf file using the
     netCDF4 library function Dataset and computes the maximum
-    of the provided variable.  In this case prateb_ave.
+    of the provided variable.  In this case soilt4.
     """
     data1 = harvest(VALID_CONFIG_DICT)
-    
+
     gridcell_area_data = Dataset(GRIDCELL_AREA_DATA_PATH)
-    
+
     summation = np.ma.zeros(gridcell_area_data.variables['area'].shape)
     for file_count, data_file in enumerate(BFG_PATH):
         test_rootgrp = Dataset(data_file)
-    
+
         summation += test_rootgrp.variables[VALID_CONFIG_DICT['variable'][0]][0]
-        
+
         test_rootgrp.close()
-        
+
     temporal_mean = summation / (file_count + 1)
     minimum = np.ma.min(temporal_mean)
     maximum = np.ma.max(temporal_mean)
-    
-    """The following offline min and max were calculated from an external 
+
+    """The following offline min and max were calculated from an external
     python code
     """
-    offline_min = 0.0
-    offline_max = 0.0043600933
+    offline_min = 223.81133 
+    offline_max = 308.44702
     for i, harvested_tuple in enumerate(data1):
         if harvested_tuple.statistic == 'maximum':
             assert maximum <= (1 + tolerance) * harvested_tuple.value
             assert maximum >= (1 - tolerance) * harvested_tuple.value
-            
+
             assert offline_max <= (1 + tolerance) * harvested_tuple.value
             assert offline_max >= (1 - tolerance) * harvested_tuple.value
-            
-            
+
+
         elif harvested_tuple.statistic == 'minimum':
             assert minimum <= (1 + tolerance) * harvested_tuple.value
             assert minimum >= (1 - tolerance) * harvested_tuple.value
-            
+
             assert offline_min <= (1 + tolerance) * harvested_tuple.value
             assert offline_min >= (1 - tolerance) * harvested_tuple.value
-            
+
     gridcell_area_data.close()
 
 def test_units():
     data1 = harvest(VALID_CONFIG_DICT)
-    assert data1[0].units == "kg/m**2/s"
+    assert data1[0].units == "K"
 
 def test_cycletime():
     """ The hard coded datetimestr 1994-01-01 12:00:00
@@ -202,18 +201,18 @@ def test_cycletime():
 
 def test_longname():
     data1 = harvest(VALID_CONFIG_DICT)
-    var_longname = "bucket surface precipitation rate"
+    var_longname = "soil temperature unknown layer 4"
     assert data1[0].longname == var_longname
 
-def test_precip_harvester():
+def test_soil_moisture_level4_harvester():
     data1 = harvest(VALID_CONFIG_DICT)  
     assert type(data1) is list
     assert len(data1) > 0
     assert data1[0].filenames==BFG_PATH
 
 def main():
     test_gridcell_area_conservation()
-    test_precip_harvester()
+    test_soil_moisture_level4_harvester()
     test_variable_names()
     test_units()
     test_global_mean_values()
@@ -224,4 +223,4 @@ def main():
     test_longname()
 
 if __name__=='__main__':
-    main()
+    main()