Cleanup

prog/parse.py

-import os
-import json
-import gzip
-from pprint import pprint
-
-data = {}
-fname = "2023-03-25/15/15-17/00000--00250.json.gz"
-
-def load_file(fname):
-    print(fname)
-    a = json.load(gzip.open(fname,  "r"))
-
-    for point in a["features"]:
-        trip = point["properties"]["trip"]
-        gtfs_route = trip['gtfs']['route_id']
-        gtfs_trip = trip['gtfs']['trip_id']
-        data.setdefault(gtfs_route, {})
-        data[gtfs_route].setdefault(gtfs_trip, [])
-        data[gtfs_route][gtfs_trip].append((fname, trip))
-        if gtfs_trip == '13_5106_230218':
-            pprint(fname)
-
-i=0
-main_dir = 'data/2023-03-25/23/'
-for dirname in os.listdir(main_dir):
-    if i == 1: break
-    for fname in os.listdir(main_dir +'/'+ dirname):
-        load_file(main_dir +'/'+ dirname +'/'+ fname)
-        pass
-    i+=1
-
-
-for i in data:
-    print(i, len(data[i]))
-
-print(sum(len(data[i]) for i in data))
-
-#for i in data:
-    #for j in i:
-        #print(len(j))
-        #for k in range(len(j)-1):
-            #if j[k] == j[k+1]:
-                #print(j[k])
-                #print(j[k+1])
-                #print()