included first variant of link_measurement_bs --> currently commented out

measprocess/preprocess.py

@@ -8,7 +8,11 @@ import geopandas as gpd
 import numpy as np
 import pandas as pd
 from tqdm import tqdm
-import tempfile
+from shapely.geometry import Point
+
+import pickle
+import warnings
+
 
 def link_dataframes(
     A: pd.DataFrame, B: pd.DataFrame, ref_col: str, metric=None, verbose=True
@@ -66,3 +70,110 @@ def link_dataframes(
     return combined, np.array(deviations)
 
 
+"""
+def link_measurement_bs( combined, bs_data, used_epsg="EPSG:31287", distance_treshold=1000, band_mapping=None, verbose=True):
+    '''
+    We assume that the elements can be found under a multiindexed dataframe
+    Combined Data should include:
+    A: Band, PCI
+    B: 'Lon.', 'Lat.'
+
+    BS Data should include:
+    technology, physicalcellid, longitude, latitude
+    '''
+
+    if band_mapping is None:
+        band_mapping = {
+            70003: ['LTE1800', 'LTE1805 - 10 MHz'],
+            70020: ['LTE800- 20 MHz', 'LTE801- 10 MHz']
+        }
+
+    '''
+    All Bands which are not in the band mapping can not be considered and will be removed
+    Also the BS data will be filtered for all the valid elements
+    '''
+
+    for band in combined.A.Band.unique():
+        if band not in band_mapping.keys():
+            warnings.warn(f"Band {band} not in Mapping --- Measurements discarded")
+
+    all_technologies = []
+    for _, technology in band_mapping.items():
+        all_technologies = all_technologies + technology
+
+    bs_data = bs_data[bs_data.technology.isin(all_technologies)].reset_index(drop=True)
+    combined = combined[combined.A.Band.isin(band_mapping.keys())].reset_index(drop=True)
+
+    '''
+    Build the Geoseries for the measurements and BS DataFrames
+    This is needed to select the PCIs with the minimum distance
+    '''
+
+    bs_geo_series = gpd.GeoSeries(list(
+        tqdm(
+            (
+                Point(bs_sector.longitude, bs_sector.latitude) for _, bs_sector in bs_data.iterrows()
+            ),
+            total=len(bs_data),
+            disable=(not verbose),
+            leave=False,
+            desc="Build Basestation GeoSeries [1/3]"
+        )
+    )).set_crs("EPSG:4326").to_crs(used_epsg)
+
+    measurement_geo_series = gpd.GeoSeries(list(
+        tqdm(
+            (
+                Point(measurement.B[['Lon.', 'Lat.']]) for _, measurement in combined.iterrows()
+            ),
+            total=len(combined),
+            disable=(not verbose),
+            leave=False,
+            desc="Build Measurement GeoSeries [2/3]"
+        )
+    )).set_crs("EPSG:4326").to_crs(used_epsg)
+
+    '''
+    Link the measurements based on the distance and PCI
+    '''
+
+    dropped_measurements = 0
+    selected_bs_indices, selected_meas_indices = [], []
+    for i, measurement in tqdm(combined.iterrows(), total=combined.shape[0], disable=(not verbose), desc="Select BS [3/3]"):
+
+        bs_candidates = bs_data[
+            (bs_data.physicalcellid == measurement.A.PCI) &
+            (bs_data.technology.isin(band_mapping[measurement.A.Band]))
+        ]
+
+        if not len(bs_candidates):
+            dropped_measurements += 1
+        else:
+            distances = bs_geo_series[bs_candidates.index].distance(
+                measurement_geo_series[i]
+            )
+
+            if distances.min() > distance_treshold:
+                dropped_measurements += 1
+            else:
+                selected_index = distances.index[distances.argmin()]
+                selected_bs_indices.append(
+                    selected_index
+                )
+
+                selected_meas_indices.append(
+                    i
+                )
+
+    new_frame = bs_data.loc[selected_bs_indices]
+    new_frame = new_frame.reset_index(drop=True)
+
+    #Different Lengths?
+    combined = combined.loc[selected_meas_indices]
+    combined = combined.reset_index(drop=True)
+
+    if dropped_measurements > 0:
+        warnings.warn(f"{dropped_measurements} measurements were dropped because no BS was found")
+
+    return pd.concat([combined.A, combined.B, new_frame], keys=['A', 'B', 'C'], axis=1)
+"""