#!/usr/bin/python3

import csv
import io
import json
import zipfile
import urllib.request

shapes = {}

# https://stackoverflow.com/a/5711095
resp = urllib.request.urlopen("https://www.viarail.ca/sites/all/files/gtfs/viarail.zip")
with zipfile.ZipFile(io.BytesIO(resp.read())).open('shapes.txt') as f:
    reader = csv.DictReader(io.TextIOWrapper(f))
    for row in reader:
        # they look like they're probably all in order, but the spec doesn't
        # actually say that they have to be, so we're going to bucket them by
        # route and then sort each bucket's contents to be on the safe side.
        # It's not that much data, and I don't run this download frequently, so
        # the extra CPU cost shouldn't be too outrageous :)
        if row['shape_id'] not in shapes:
            shapes[row['shape_id']] = []
        shapes[row['shape_id']].append(row)

geojson = {
    "type": "FeatureCollection",
    "features": [],
}

for _, shape in shapes.items():
    shape.sort(key=lambda c: int(c['shape_pt_sequence']))
    geojson['features'].append({
            "type": "Feature",
            "geometry": {
                "type": "LineString",
                "coordinates": [
                    [float(l['shape_pt_lon']), float(l['shape_pt_lat'])] for l in shape
                ]
            },
        })

with open('data.geojson', 'w') as f:
    f.write(json.dumps(geojson))