subway_gtfs.py

#coding:UTF-8
import csv
import argparse
import datetime
import zipfile
import os.path
import re
import json
import glob
import collections

def write_data(fp, data):
	fields = set()
	for d in data:
		fields.update(d.keys())
	fields = sorted(list(fields))
	w = csv.writer(fp)
	w.writerow(fields)
	for d in data:
		w.writerow(list(map(lambda x:d.get(x,""), fields)))

agency = [
	dict(
		agency_name="神戸市",
		agency_url="http://www.city.kobe.lg.jp/life/access/transport/subway/",
		agency_timezone="Asia/Tokyo",
		agency_lang="ja",
		agency_phone="078-322-5958"
	)
]
routes = [
	dict(
		route_id="seishin",
	 	route_short_name="西神・山手線",
		route_long_name="",
		route_type=1
	),
	dict(
		route_id="kaigan",
		route_short_name="海岸線",
		route_long_name="",
		route_type=1
	)
]
calendar = [
	dict(
		service_id="w",
		monday=1,
		tuesday=1,
		wednesday=1,
		thursday=1,
		friday=1,
		saturday=0,
		sunday=0,
		start_date="20061201"
	),
	dict(
		service_id="h",
		monday=0,
		tuesday=0,
		wednesday=0,
		thursday=0,
		friday=0,
		saturday=1,
		sunday=1,
		start_date="20061201"
	)
]
feed_info = [
	dict(
		feed_publisher_name="Code for Kobe",
		feed_publisher_url="https://github.com/hkwi/kobe-transit",
		feed_lang="ja"
	)
]

with open("dbpedia/subway.json") as fp:
	geo = json.load(fp)

with open("code_for_kobe/translations.txt") as fp:
	translations = [r for r in csv.DictReader(fp)]

if __name__=="__main__":
	ap = argparse.ArgumentParser()
	ap.add_argument("--dst", default="kobe_subway_gtfs")
	ap.add_argument("--src", default="kobe_opendata")
	ap.add_argument("--out", default="kobe_subway_gtfs.zip")
	opts = ap.parse_args()
	
	with open(os.path.join(opts.dst, "agency.txt"), mode="w", encoding="UTF-8") as f:
		write_data(f, agency)
	with open(os.path.join(opts.dst, "routes.txt"), mode="w", encoding="UTF-8") as f:
		write_data(f, routes)
	with open(os.path.join(opts.dst, "translations.txt"), mode="w", encoding="UTF-8") as f:
		write_data(f, translations)
	
	start_date = datetime.datetime.now()
	end_date = start_date + datetime.timedelta(days=31*3)
	with open(os.path.join(opts.dst, "calendar.txt"), mode="w", encoding="UTF-8") as f:
		# 3 カ月前には公示がでているはず…
		for d in calendar:
			d["end_date"]=end_date.strftime("%Y%m%d")
		write_data(f, calendar)
	with open(os.path.join(opts.dst, "feed_info.txt"), mode="w", encoding="UTF-8") as f:
		for d in feed_info:
			d["feed_start_date"]=start_date.strftime("%Y%m%d")
			d["feed_end_date"]=end_date.strftime("%Y%m%d")
		write_data(f, feed_info)
	
	trips = []
	stations = collections.OrderedDict()
	route_stops = collections.OrderedDict()
	for src in glob.glob(os.path.join(opts.src, "*_[h,w]_*.csv")):
		sname = os.path.basename(src)
		m = re.match(r"open_(?P<route>\w+)_(?P<cal>w|h)_(?P<dir>east|west).csv", sname)
		info = m.groupdict()
		
		with open(src, encoding="CP932") as fp:
			for row in csv.reader(fp):
				name = row[0]
				if name[-1] in "発着":
					name = name[:-1]
				
				if name not in ("始発駅","行先駅"):
					stations[name] = True
					route_stops[(name,info["route"])] = True
	
	stop_info = {("谷上","seishin"):{ "駅番号":"S01", "駅構内図":""}}
	stop_name_alias = {
		"ハーバー":"ハーバーランド",
		"中央市場":"中央市場前",
		"運動公園":"総合運動公園",
		"駒ケ林":"駒ヶ林",
	}
	with open(os.path.join(opts.src, "open_kobe_subway_add.csv"), encoding="CP932") as fp:
		for row in csv.DictReader(fp):
			name = row["駅名"]
			if name.endswith("駅"):
				name = name[:-1]
			name = stop_name_alias.get(name, name)
			
			stop_info[(
				name,
				{"西神・山手線":"seishin", "海岸線":"kaigan"}[row["線名"]],
			)] = row
	
	stops = []
	for name in stations.keys():
		stops.append(dict(
			stop_id=name,
			stop_name=name,
			stop_lat=geo[name]["lat"],
			stop_lon=geo[name]["lon"],
			location_type=1,
			wheelchair_boarding=1,
		))
	
	for name, route in route_stops.keys():
		info = stop_info[(name, route)]
		lat = geo[name]["lat"]
		lon = geo[name]["lon"]
		# 新長田はくっついているけれど、S09 と K10 は少し離れている
		if info["駅番号"]=="K10":
			lat = 34.657055
			lon = 135.1441888
		elif info["駅番号"]=="S09":
			lat = 34.657747
			lon = 135.144410
		
		stops.append(dict(
			stop_id=info["駅番号"],
			stop_code=info["駅番号"],
			stop_name=name,
			stop_lat=lat,
			stop_lon=lon,
			stop_url=info["駅構内図"],
			location_type=0,
			parent_station=name,
			wheelchair_boarding=1,
		))
	
	with open(os.path.join(opts.dst, "stops.txt"), mode="w", encoding="UTF-8") as f:
		write_data(f, stops)
	
	stop_times = []
	for src in glob.glob(os.path.join(opts.src, "*_[h,w]_*.csv")):
		src_trips = []
		with open(src, encoding="CP932") as fp:
			r = csv.reader(fp)
			starts = next(r)
			ends = next(r)
			for s,e in zip(starts[1:], ends[1:]):
				src_trips.append(dict(
					start=s,
					end=e,
					stops=[]))
			
			for row in r:
				name = row[0]
				if name[-1] in "発着":
					name = name[:-1]
				
				for i,v in enumerate(row[1:]):
					if not v:
						continue
					src_trips[i]["stops"].append(dict(
						name=name,
						value=v))
		
		sname = os.path.basename(src)
		m = re.match(r"open_(?P<route>\w+)_(?P<cal>w|h)_(?P<dir>east|west).csv", sname)
		if m:
			info = m.groupdict()
			for t in src_trips:
				trip_id = "{0}_{1}_{2}_{3}".format(
					t["start"], 
					info["cal"],
					t["stops"][0]["value"].replace(":","_"),
					info["dir"])
				
				trips.append(dict(
					route_id=info["route"],
					service_id=info["cal"],
					trip_id=trip_id,
					trip_headsign=t["end"],
					direction_id={"east":0,"west":1}[info["dir"]],
				))
				
				for stop_no,stop in enumerate(t["stops"]):
					h,m = tuple(map(int, re.match("(\d+):(\d+)", stop["value"]).groups()))
					if h < 3:
						h += 24
					
					stop_times.append(dict(
						trip_id=trip_id,
						arrival_time="%02d:%02d:00" % (h,m),
						departure_time="%02d:%02d:00" % (h,m),
						stop_id=stop_info[(stop["name"],info["route"])]["駅番号"],
						stop_sequence=stop_no,
					))
	
	with open(os.path.join(opts.dst, "trips.txt"), mode="w", encoding="UTF-8") as f:
		write_data(f, trips)
	
	with open(os.path.join(opts.dst, "stop_times.txt"), mode="w", encoding="UTF-8") as f:
		write_data(f, stop_times)
	
	with zipfile.ZipFile(opts.out, mode="w") as z:
		def write(path):
			z.write(os.path.join(opts.dst, path), path)
		write("agency.txt")
		write("stops.txt")
		write("routes.txt")
		write("trips.txt")
		write("stop_times.txt")
		write("calendar.txt")
		write("feed_info.txt")
#		write("translations.txt")