progress bar

Makefile

@@ -2,7 +2,7 @@ all: install data train eval
 
 city_distances_full.csv: check generate_data.py
 	@echo "Generating distance data..."
-	@bash -c 'time python generate_data.py -w 8 -c US -s 10000'
+	@bash -c 'time python generate_data.py --country US --workers 8 --chunk-size 8000'
 
 data: city_distances_full.csv
 

generate_data.py

@@ -8,6 +8,7 @@ from functools import lru_cache
 import geonamescache
 import numpy as np
 from geopy.distance import geodesic
+from tqdm import tqdm
 
 MAX_DISTANCE = 20_037.5
 
@@ -115,6 +116,7 @@ def main():
     cities = list(us_cities.values())
     print(f"Num cities: {len(cities)}")
     city_combinations = list(itertools.combinations(cities, 2))
+    # np.random.shuffle(city_combinations)  # will this help or hurt caching? 1.03it/s
     chunk_size = args.chunk_size
     num_chunks = len(city_combinations) // chunk_size + 1
     output_file = args.output_file
@@ -126,8 +128,13 @@ def main():
 
         try:
             executor = concurrent.futures.ProcessPoolExecutor(max_workers=args.workers)
-            for i in range(num_chunks):
+            for i in tqdm(
-                print(f"Processing chunk {i}...")
+                range(num_chunks),
+                total=num_chunks,
+                desc="Processing chunks",
+                ncols=100,
+                bar_format="{l_bar}{bar}{r_bar}",
+            ):
                 chunk = city_combinations[(i * chunk_size) : (i + 1) * chunk_size]
                 futures = {
                     executor.submit(calculate_distance, pair): pair for pair in chunk