path: root/bin/populate-lut

#!/usr/bin/env python3
# vim:tabstop=4 softtabstop=4 shiftwidth=4 textwidth=160 smarttab expandtab colorcolumn=160
#
# Copyright (C) 2021 Daniel Friesel
#
# SPDX-License-Identifier: BSD-2-Clause

from datetime import datetime, timedelta
from geopy.distance import distance
from progress.bar import Bar

import csv
import json
import numpy as np
import os
import psycopg2
import psycopg2.extras
import requests
import sys


class ProgressBar(Bar):
    sma_window = 500
    suffix = "%(percent).0f%% [%(elapsed_td)s/%(eta_td)s]"


conn = psycopg2.connect(
    dbname=os.getenv("GEOLOOKUP_DBNAME", "geo_to_stations"),
    user=os.getenv("GEOLOOKUP_DBUSER", "geo_to_stations"),
    password=os.getenv("GEOLOOKUP_DBPASS"),
    host=os.getenv("GEOLOOKUP_DBHOST", "localhost"),
)

shape = dict()
stops_by_latlon = dict()

routes_by_shape_id = dict()
trips_by_shape_id = dict()

name_to_eva = dict()
eva_to_name = dict()

# lon + 0.003 ≙ 190 to 220 m
# lat + 0.003 ≙ 333 m
lut_grid_step = 100

try:
    with open("data/iris-stations.json", "r") as f:
        for station in json.load(f):
            name_to_eva[station["name"]] = int(station["eva"])
            eva_to_name[int(station["eva"])] = station["name"]
except FileNotFoundError:
    print(
        "populate-lut requires a list of IRIS stations. Please run the following commands:"
    )
    print()
    print("mkdir -p data")
    print(
        "curl https://git.finalrewind.org/Travel-Status-DE-IRIS/plain/share/stations.json > data/iris-stations.json"
    )
    print()
    sys.exit(1)

try:
    with open("data/nvbw/trips.txt", "r") as f:
        pass
    with open("data/nvbw/shapes.txt", "r") as f:
        pass
    with open("data/nvbw/stop_times.txt", "r") as f:
        pass
except FileNotFoundError:
    print("populate-lut requires GTFS shapes of regional transit lines.")
    print(
        "At present, the best known resource is <https://www.nvbw.de/open-data/fahrplandaten/fahrplandaten-mit-liniennetz>."
    )
    print(
        "(https://www.nvbw.de/fileadmin/user_upload/service/open_data/fahrplandaten_mit_liniennetz/bwspnv.zip)"
    )
    print("Please download and extract it to data/nvbw.")
    sys.exit(1)

print("Loading trips ...")
with open("data/nvbw/trips.txt", "r") as f:
    f.readline()
    cr = csv.reader(f)
    for row in cr:
        route_id, trip_id, service_id, direction_id, block_id, shape_id = row
        if shape_id not in routes_by_shape_id:
            routes_by_shape_id[shape_id] = list()
        routes_by_shape_id[shape_id].append(route_id)
        if shape_id not in trips_by_shape_id:
            trips_by_shape_id[shape_id] = list()
        trips_by_shape_id[shape_id].append(trip_id)

print("Loading stop_times ...")
stops_by_tripid = dict()
with open("data/nvbw/stop_times.txt", "r") as f:
    f.readline()
    cr = csv.reader(f)
    for row in cr:
        (
            trip_id,
            stop_id,
            arrival_time,
            departure_time,
            stop_seq,
            stop_headsign,
            pickup_type,
            dropoff_type,
            dist,
        ) = row
        if trip_id not in stops_by_tripid:
            stops_by_tripid[trip_id] = list()
        stops_by_tripid[trip_id].append((stop_headsign, float(dist)))

print("Loading shapes ...")
with open("data/nvbw/shapes.txt", "r") as f:
    f.readline()
    cr = csv.reader(f)
    prev_lat, prev_lon = None, None
    prev_dist = 0
    for row in cr:
        shape_id, _, lat, lon, dist = row
        if shape_id not in shape:
            shape[shape_id] = list()
            prev_dist = 0
        lat = float(lat)
        lon = float(lon)
        dist = float(dist)
        if dist > prev_dist and dist - prev_dist > lut_grid_step:
            # ensure shape entries are no more than lut_grid_step meters apart
            for i in np.arange(lut_grid_step, dist - prev_dist, lut_grid_step):
                ratio = i / (dist - prev_dist)
                assert 0 <= ratio <= 1
                rel_lat = prev_lat * ratio + lat * (1 - ratio)
                rel_lon = prev_lon * ratio + lon * (1 - ratio)
                shape[shape_id].append((rel_lat, rel_lon, dist))
        shape[shape_id].append((lat, lon, dist))
        prev_dist = dist
        prev_lat = lat
        prev_lon = lon


def add_stops(lat, lon, stops):
    evas = list()
    for stop in stops:
        try:
            evas.append(name_to_eva[stop])
        except KeyError:
            try:
                evas.append(name_to_eva[stop.replace(" (", "(")])
            except KeyError:
                pass

    add_evas(lat, lon, evas)


def add_evas(lat, lon, evas):
    lut_lat_center = round(lat * 1000)
    lut_lon_center = round(lon * 1000)

    for lut_lat in range(lut_lat_center - 0, lut_lat_center + 1):
        for lut_lon in range(lut_lon_center - 0, lut_lon_center + 1):
            if (lut_lat, lut_lon) not in stops_by_latlon:
                stops_by_latlon[(lut_lat, lut_lon)] = set()
            stops_by_latlon[(lut_lat, lut_lon)].update(evas)


# Here be dragons. I don't recall what this code does. It shouldn't be too complicated, though.

num_shapes = len(shape.keys())

for shape_id in ProgressBar("Calculating neighoubrs", max=num_shapes).iter(
    shape.keys()
):
    for trip_id in trips_by_shape_id[shape_id]:
        stops = stops_by_tripid[trip_id]
        first_stop = stops[0]
        last_stop = stops[-1]
        for lat, lon, shape_dist in shape[shape_id]:
            assert first_stop[1] <= shape_dist <= last_stop[1]
            for i, (stop_name, stop_dist) in enumerate(stops):
                if (
                    stop_dist <= shape_dist
                    and i + 1 < len(stops)
                    and stops[i + 1][1] >= shape_dist
                ):
                    add_stops(lat, lon, (stop_name, stops[i + 1][0]))

try:

    class Polyline:
        def __init__(self, json_data):
            self.coordinates = json_data["polyline"]

    with open("data/polydump.json", "r") as f:
        polylines = list(map(Polyline, json.load(f)))

    def add_leg(coordinates, from_eva, to_eva):
        for lat, lon in coordinates:
            add_evas(lat, lon, (from_eva, to_eva))

    for polyline in ProgressBar("Adding polydump data", max=len(polylines)).iter(
        polylines
    ):
        prev_eva = None
        leg = list()
        for coord in polyline.coordinates:
            lat = coord[1]
            lon = coord[0]
            if leg:
                prev_lat = leg[-1][0]
                prev_lon = leg[-1][1]
                prev_dist = distance((prev_lat, prev_lon), (lat, lon)).m
                for i in np.arange(lut_grid_step, prev_dist, lut_grid_step):
                    ratio = i / prev_dist
                    assert 0 <= ratio <= 1
                    rel_lat = prev_lat * ratio + lat * (1 - ratio)
                    rel_lon = prev_lon * ratio + lon * (1 - ratio)
                    leg.append((rel_lat, rel_lon))
            leg.append((lat, lon))
            if len(coord) > 2 and coord[2] != prev_eva:
                if prev_eva:
                    add_leg(leg, prev_eva, coord[2])
                prev_eva = coord[2]
                leg = list()

except FileNotFoundError:
    pass

num_latlons = len(stops_by_latlon.keys())

with conn.cursor() as cur:
    cur.execute("drop table if exists stations")
    cur.execute(
        """create table stations (
        lat integer not null,
        lon integer not null,
        stations jsonb not null,
        primary key (lat, lon)
        )
    """
    )

insert_groups = list()
insert_group = list()
for (lat, lon), stops in stops_by_latlon.items():
    insert_group.append((lat, lon, json.dumps(list(stops))))
    if len(insert_group) >= 100:
        insert_groups.append(insert_group)
        insert_group = list()

insert_groups.append(insert_group)

for insert_group in ProgressBar("Inserting coordinates", max=len(insert_groups)).iter(
    insert_groups
):
    with conn.cursor() as cur:
        psycopg2.extras.execute_values(
            cur, """insert into stations (lat, lon, stations) values %s""", insert_group
        )

conn.commit()