move to logging from printing

Author: soodoku

allocator/__init__.py

@@ -1 +1,53 @@
+"""
+Allocator package for clustering and routing optimization.
+"""
 
+import logging
+import sys
+
+
+def setup_logging(level=logging.INFO):
+    """
+    Set up logging configuration for the allocator package.
+    
+    Args:
+        level: Logging level (DEBUG, INFO, WARNING, ERROR)
+    """
+    # Create formatter
+    formatter = logging.Formatter(
+        '%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+        datefmt='%Y-%m-%d %H:%M:%S'
+    )
+    
+    # Get root logger for allocator package
+    logger = logging.getLogger('allocator')
+    logger.setLevel(level)
+    
+    # Remove existing handlers to avoid duplicates
+    for handler in logger.handlers[:]:
+        logger.removeHandler(handler)
+    
+    # Console handler
+    console_handler = logging.StreamHandler(sys.stdout)
+    console_handler.setLevel(level)
+    console_handler.setFormatter(formatter)
+    logger.addHandler(console_handler)
+    
+    return logger
+
+
+def get_logger(name):
+    """
+    Get a logger instance for a specific module.
+    
+    Args:
+        name: Module name (typically __name__)
+    
+    Returns:
+        Logger instance
+    """
+    return logging.getLogger(f'allocator.{name}')
+
+
+# Set up default logging
+setup_logging()

allocator/distance_matrix.py

@@ -7,6 +7,9 @@
 import time
 
 import numpy as np
+
+from allocator import get_logger
+logger = get_logger(__name__)
 import requests
 import googlemaps
 import utm
@@ -117,7 +120,7 @@ def osrm_distance_matrix(X, Y=None, chunksize=MAX_DISTANCE_MATRIX_SIZE,
             count += 1
             r = requests.get(url)
             if r.status_code != 200:
-                print(("OSRM Table API request error: {0:s}".format(r.text)))
+                logger.error(f"OSRM Table API request error: {r.text}")
                 break
             dm = r.json()['durations']
             arr = np.array(dm)
@@ -129,7 +132,7 @@ def osrm_distance_matrix(X, Y=None, chunksize=MAX_DISTANCE_MATRIX_SIZE,
             o = c
         else:
             o = np.concatenate((o, c), axis=0)
-    print(("Total API requests: {0:d}".format(count)))
+    logger.info(f"Total API requests: {count}")
     return o
 
 
@@ -182,8 +185,7 @@ def google_distance_matrix(X, Y=None, api_key=None, duration=True):
             try:
                 matrix = gmaps.distance_matrix(sources, destinations)
             except Exception as e:
-                print(("Google Distance Matrix API error: {0!s}"
-                      .format(e)))
+                logger.error(f"Google Distance Matrix API error: {e}")
                 return o
             time.sleep(1)
             """
@@ -218,7 +220,7 @@ def google_distance_matrix(X, Y=None, api_key=None, duration=True):
             o = c
         else:
             o = np.concatenate((o, c), axis=0)
-    print(("Total API requests: {0:d}, elements: {1:d}".format(count, nXY)))
+    logger.info(f"Total API requests: {count}, elements: {nXY}")
     return o
 
 
@@ -287,20 +289,20 @@ def google_distance_matrix(X, Y=None, api_key=None, duration=True):
          (100.92367939299999, 12.9881022409)]
 
     o = osrm_distance_matrix(A, B, 10)
-    print((o.sum()))
+    logger.debug(f"Matrix sum: {o.sum()}")
     o = osrm_distance_matrix(A, B, 20)
-    print((o.sum()))
+    logger.debug(f"Matrix sum: {o.sum()}")
     o = osrm_distance_matrix(A, B, MAX_DISTANCE_MATRIX_SIZE)
-    print((o.sum()))
+    logger.debug(f"Matrix sum: {o.sum()}")
 
     api_key = os.environ.get('API_KEY', None)
     if api_key:
         o = google_distance_matrix(A, B, api_key=api_key)
-        print((o.sum()))
+        logger.debug(f"Matrix sum: {o.sum()}")
         # Test with origins over 25 elements
         o = osrm_distance_matrix(A[:30], B[:3], MAX_DISTANCE_MATRIX_SIZE)
-        print((o.sum()))
+        logger.debug(f"Matrix sum: {o.sum()}")
         o = google_distance_matrix(A[:30], B[:3], api_key=api_key)
-        print((o.sum()))
+        logger.debug(f"Matrix sum: {o.sum()}")
     else:
-        print("Please set Google API key to environment variable `API_KEY`")
+        logger.warning("Please set Google API key to environment variable `API_KEY`")

allocator/shortest_path_gm.py

@@ -13,6 +13,9 @@
 
 import pandas as pd
 
+from allocator import get_logger
+logger = get_logger(__name__)
+
 import googlemaps
 import polyline
 
@@ -40,7 +43,7 @@ def main(argv=sys.argv[1:]):
 
     args = parser.parse_args(argv)
 
-    print(args)
+    logger.debug(f"Arguments: {args}")
 
     df = pd.read_csv(args.input)
 
@@ -57,7 +60,7 @@ def main(argv=sys.argv[1:]):
 
     output = []
     for i, l in enumerate(sorted(df.assigned_points.unique())):
-        print(("Google Direction API request for #{:d}...".format(l)))
+        logger.info(f"Google Direction API request for #{l}...")
         start = None
         adf = df.loc[(df.assigned_points == l), ['start_lat', 'start_long']]
         nodes = adf.to_records(index=False).tolist()
@@ -66,7 +69,7 @@ def main(argv=sys.argv[1:]):
         tour = []
         N = len(nodes)
         if N >= 25:
-            print("WARN: The maximum allowed location is 24 including origin and destination")
+            logger.warning("The maximum allowed location is 24 including origin and destination")
         else:
             start = ', '.join([str(x) for x in nodes[0]])
             waypoints = [', '.join([str(x) for x in n]) for n in nodes[1:]]
@@ -83,8 +86,6 @@ def main(argv=sys.argv[1:]):
                     t_dist = 0
                     t_time = 0
                     for i, leg in enumerate(legs):
-                        if i == 0:
-                            sloc = leg['start_location']
                         d = leg['distance']['value']
                         t_dist += d
                         t = leg['duration']['value']
@@ -94,7 +95,7 @@ def main(argv=sys.argv[1:]):
                 t_min = int(t_time / 60 + 1)
                 tour = [0] + [(i + 1) for i in wp] + [0]
             except Exception as e:
-                print(('ERROR: {0!s}'.format(e)))
+                logger.error(f'ERROR: {e}')
         B = df.loc[df.assigned_points == l,
                    ['segment_id', 'start_lat', 'start_long']]
         B.reset_index(drop=True, inplace=True)
@@ -129,7 +130,7 @@ def main(argv=sys.argv[1:]):
             if args.save_plot:
                 fn, fe = os.path.splitext(args.save_plot)
                 fname = "{:s}-{:d}{:s}".format(fn, l, fe)
-                print(("Plotting to file '{:s}'".format(fname)))
+                logger.debug(f"Plotting to file '{fname}'")
                 fig.savefig(fname)
             plt.close()
 
@@ -145,7 +146,7 @@ def main(argv=sys.argv[1:]):
                        ';'.join([str(p) for p in new_path])])
 
     # save output to file
-    print(("Save the output file to '{:s}'".format(args.output)))
+    logger.info(f"Saved output file to '{args.output}'")
     odf = pd.DataFrame(output, columns=['worker_id', 'distance', 'duration',
                                         'n', 'path_order'])
     odf.to_csv(args.output, index=False)

allocator/shortest_path_ortools.py

@@ -10,7 +10,9 @@
 import argparse
 
 import pandas as pd
-import numpy as np
+
+from allocator import get_logger
+logger = get_logger(__name__)
 
 from random import randint
 
@@ -79,7 +81,7 @@ def ortools_tsp(A, args):
     cost = 0
     if assignment:
         # Solution cost.
-        print((assignment.ObjectiveValue()))
+        logger.info(f"Objective value: {assignment.ObjectiveValue()}")
         cost = assignment.ObjectiveValue()
         # Inspect solution.
         # Only one route here; otherwise iterate from 0 to
@@ -93,9 +95,9 @@ def ortools_tsp(A, args):
             node = assignment.Value(routing.NextVar(node))
         route += '0'
         path.append(0)
-        print(route)
+        logger.debug(f"Route: {route}")
     else:
-        print('No solution found.')
+        logger.error('No solution found.')
 
     return cost, path
 
@@ -116,8 +118,7 @@ def do_save_map(args, label, C):
         points = polyline.decode(out['routes'][0]['geometry'])
         cost = int(float(out['routes'][0]['distance'] / 1000.0))
         duration = out['routes'][0]['duration']
-        print(("Route map distance: {:.1f}, duration: {:.1f}"
-              .format(cost, duration)))
+        logger.info(f"Route map distance: {cost:.1f}, duration: {duration:.1f}")
         # FIXME: unused
         # legs = out['routes'][0]['legs']
         # waypoints = out['waypoints']
@@ -142,10 +143,10 @@ def do_save_map(args, label, C):
 
         fn, fe = os.path.splitext(args.save_map)
         fname = "{:s}-{:d}{:s}".format(fn, label, fe)
-        print(("Save map HTML to file '{:s}'".format(fname)))
+        logger.debug(f"Saved map HTML to file '{fname}'")
         map_osm.save(fname)
     except Exception as e:
-        print(("Error: Cannot save map to file ({!s})".format(e)))
+        logger.error(f"Error: Cannot save map to file ({e})")
 
 
 def main(argv=sys.argv[1:]):
@@ -179,7 +180,7 @@ def main(argv=sys.argv[1:]):
 
     args = parser.parse_args(argv)
 
-    print(args)
+    logger.debug(f"Arguments: {args}")
 
     df = pd.read_csv(args.input)
 
@@ -195,7 +196,7 @@ def main(argv=sys.argv[1:]):
     output = []
     total_cost = 0
     for i, l in enumerate(sorted(df.assigned_points.unique())):
-        print(("Search TSP path for #{:d}...".format(l)))
+        logger.info(f"Search TSP path for #{l}...")
         adf = df.loc[df.assigned_points == l, ['start_long', 'start_lat']]
         A = adf.values
         # FIXME: OSRM distance matrix actually isn't distance but it's duration
@@ -211,7 +212,7 @@ def main(argv=sys.argv[1:]):
         title = ('TSP: {:d}, Cost: {:0.1f}, N: {:d} ({:s})'
                  .format(l, cost, N, args.distance_func.title()))
 
-        print(title)
+        logger.info(title)
 
         if args.plot or args.save_plot:
             fig = plt.figure(figsize=(16, 16))
@@ -236,7 +237,7 @@ def main(argv=sys.argv[1:]):
             if args.save_plot:
                 fn, fe = os.path.splitext(args.save_plot)
                 fname = "{:s}-{:d}{:s}".format(fn, l, fe)
-                print(("Plotting to file '{:s}'".format(fname)))
+                logger.debug(f"Plotting to file '{fname}'")
                 fig.savefig(fname)
             plt.close()
 
@@ -253,10 +254,10 @@ def main(argv=sys.argv[1:]):
         new_path = path[pos:] + path[:pos]
         output.append([l, cost, N, ';'.join([str(p) for p in new_path])])
 
-    print(("Total cost: {0:.1f}".format(total_cost)))
+    logger.info(f"Total cost: {total_cost:.1f}")
 
     # save output to file
-    print(("Save the output file to '{:s}'".format(args.output)))
+    logger.info(f"Saved output file to '{args.output}'")
     odf = pd.DataFrame(output, columns=['worker_id', 'cost', 'n',
                                         'path_order'])
     odf.to_csv(args.output, index=False)

allocator/shortest_path_osrm.py

@@ -11,6 +11,9 @@
 
 import pandas as pd
 
+from allocator import get_logger
+logger = get_logger(__name__)
+
 import requests
 import polyline
 
@@ -28,7 +31,7 @@ def osrm_trip(coords, osrm_base_url=None):
     a = ';'.join([','.join(str(c) for c in b) for b in coords])
     if osrm_base_url is None:
         if len(coords) > 100:
-            print("ERROR: Maximum locations for public OSRM server is 100")
+            logger.error("Maximum locations for public OSRM server is 100")
             return points, cost, duration, tour
         url = ('http://router.project-osrm.org/trip/v1/driving/' + a +
                '?overview=full')
@@ -49,10 +52,9 @@ def osrm_trip(coords, osrm_base_url=None):
                 tour.append(w['waypoint_index'])
         else:
             data = r.json()
-            print(("OSRM ERROR code={0!s}, message={0!s}"
-                  .format(data['code'], data['message'])))
+            logger.error(f"OSRM ERROR code={data['code']}, message={data['message']}")
     except Exception as e:
-        print(("ERROR: {0!s}".format(e)))
+        logger.error(f"ERROR: {e}")
     return points, cost, duration, tour
 
 
@@ -76,7 +78,7 @@ def main(argv=sys.argv[1:]):
 
     args = parser.parse_args(argv)
 
-    print(args)
+    logger.debug(f"Arguments: {args}")
 
     df = pd.read_csv(args.input)
 
@@ -101,8 +103,7 @@ def main(argv=sys.argv[1:]):
         else:
             N = len(coords)
 
-        print(('TSP: {:d}, Cost: {:d}, Duration: {:0.1f}, N: {:d}'
-              .format(l, cost, duration, N)))
+        logger.info(f'TSP: {l}, Cost: {cost}, Duration: {duration:.1f}, N: {N}')
 
         total_distance += cost
         total_duration += duration
@@ -150,14 +151,13 @@ def main(argv=sys.argv[1:]):
 
             fn, fe = os.path.splitext(args.save_map)
             fname = "{:s}-{:d}{:s}".format(fn, l, fe)
-            print(("Save map HTML to file '{:s}'".format(fname)))
+            logger.debug(f"Saved map HTML to file '{fname}'")
             map_osm.save(fname)
 
-    print(("Total distance: {0:.1f}, duration: {1:.1f}".format(total_distance,
-                                                              total_duration)))
+    logger.info(f"Total distance: {total_distance:.1f}, duration: {total_duration:.1f}")
 
     # save output to file
-    print(("Save the output file to '{:s}'".format(args.output)))
+    logger.info(f"Saved output file to '{args.output}'")
     odf = pd.DataFrame(output, columns=['worker_id', 'distance', 'duration', 'n',
                                         'path_order'])
     odf.to_csv(args.output, index=False)

allocator/tests/test_cluster_kahip.py

@@ -20,7 +20,7 @@
 def kahip_available():
     """Check if KaHIP dependencies are available"""
     try:
-        import kahipwrapper
+        import kahipwrapper  # noqa: F401
         return True
     except ImportError:
         return False