[i] back to grouped 15min time-windows (better results)

FHSTP/01_data_preparation/prep.py

@@ -2,8 +2,8 @@
 # Title: Data Preparation for LINK Configs and Transmissions
 # Author: Fabian Kovac <ds191008@fhstp.ac.at>
 # Team: University of Applied Sciences St. Pölten
-# Version: 3.0
-# Last changed: 2021-10-22
+# Version: 3.1
+# Last changed: 2021-11-03
 #
 
 import sys
@@ -273,7 +273,7 @@ def get_inca_data(datetimes: np.array, lon_a: np.array, lat_a: np.array, lon_b:
         lat_a (np.array): Vector containing latitude values of LINK site a
         lon_b (np.array): Vector containing longitude values of LINK site b
         lat_b (np.array): Vector containing latiotude values of LINK site b
-        length (np.array): Vector containing distance between sites in m
+        length (np.array): Vector containing distance between sites in km
     
     Returns:
         inca_RR (np.array): Vector containing INCA RR data for each transmission
@@ -494,30 +494,26 @@ def prep() -> None:
     
     
     # copy REMOTERXLEVEL to PMIN and PMAX (for aggregation in 15min time window)
-    df_link_window = df_link[['BEGINTIME', 'RADIOLINKID', 'REMOTERXLEVEL']].copy()
-    df_link_window['PMIN'] = df_link_window['REMOTERXLEVEL']
-    df_link_window['PMAX'] = df_link_window['REMOTERXLEVEL']
-    df_link_window['PMEAN'] = df_link_window['REMOTERXLEVEL']
-    df_link_window = df_link_window.drop('REMOTERXLEVEL', axis = 1)
-    _log('Created PMIN, PMAX and PMEAN of REMOTERXLEVEL')
+    df_link['PMIN'] = df_link['REMOTERXLEVEL']
+    df_link['PMAX'] = df_link['REMOTERXLEVEL']
+    _log('Created PMIN, PMAX and mean REMOTERXLEVEL')
 
     # convert 3min windows to 15min windows to get min, max and mean power levels
-    group_cols = [df_link_window['BEGINTIME'].dt.round('15Min'), 'RADIOLINKID']
+    group_cols = [df_link['BEGINTIME'].dt.round('15Min'), 'RADIOLINKID']
     agg_cols = {
         'PMIN' : 'min',
         'PMAX' : 'max',
-        'PMEAN' : 'mean'
+        'REMOTERXLEVEL' : 'mean',
+        'TXLEVEL' : 'mean',
+        'SPEED' : 'median',
+        'CURRRXBITRATE' : 'median',
+        'CURRTXBITRATE' : 'median',
+        'CURRRXPROFILE' : 'median',
+        'CURRTXPROFILE' : 'median'
     }
-    df_link_window = df_link_window.groupby(group_cols).agg(agg_cols).reset_index()
+    df_link = df_link.groupby(group_cols).agg(agg_cols).reset_index()
     _log('Converted 3min windows to 15min windows to get min, max and mean power levels')
     
-    # merge min, max and mean power levels for 15min windows back to link dataframe
-    # only get minutes 00, 15, 30 and 45 (PMIN, PMAX and PMEAN are not NaN --> faster than filtering datetime)
-    df_link = pd.merge(df_link, df_link_window, how = 'left', left_on = ['BEGINTIME', 'RADIOLINKID'], right_on = ['BEGINTIME', 'RADIOLINKID'])
-    del df_link_window
-    _log('Merged min, max and mean power levels for 15min windows back to link dataframe')
-    df_link = df_link.dropna(subset = ['PMIN', 'PMAX', 'PMEAN'])
-    _log('Use only transmissions in minutes 00, 15, 30 and 45')
     
     # convert BEGINTIME to format "%Y%m%d%H%M"
     df_link['BEGINTIME'] = df_link['BEGINTIME'].dt.strftime('%Y%m%d%H%M')
@@ -568,7 +564,6 @@ def prep() -> None:
         'REMOTERXLEVEL' : 'RxLevel',
         'PMIN' : 'RxLevelMin',
         'PMAX' : 'RxLevelMax',
-        'PMEAN' : 'RxLevelMean',
         'TXLEVEL' : 'TxLevel',
         'ATTENUATION' : 'Attn',
         'ATTENUATIONMEAN' : 'AttnMean',