bristol-eurostar/tests/test_eurostar_scraper.py

97 lines
3 KiB
Python

import json
import pytest
from scraper.eurostar import _hhmm, _parse, timetable_url
# ---------------------------------------------------------------------------
# _hhmm
# ---------------------------------------------------------------------------
def test_hhmm_parses_datetime_string():
assert _hhmm('2026-03-30 09:34:00') == '09:34'
def test_hhmm_none_input():
assert _hhmm(None) is None
def test_hhmm_empty_string():
assert _hhmm('') is None
# ---------------------------------------------------------------------------
# _parse
# ---------------------------------------------------------------------------
def _make_next_data(departures: list) -> str:
data = {
'props': {
'pageProps': {
'pageData': {
'liveDepartures': departures
}
}
}
}
return f'<script id="__NEXT_DATA__" type="application/json">{json.dumps(data)}</script>'
def _departure(dep_dt: str, arr_dt: str) -> dict:
return {
'origin': {'model': {'scheduledDepartureDateTime': dep_dt}},
'destination': {'model': {'scheduledArrivalDateTime': arr_dt}},
}
def test_parse_single_departure():
html = _make_next_data([_departure('2026-03-30 06:01:00', '2026-03-30 09:34:00')])
services = _parse(html, 'Paris Gare du Nord')
assert len(services) == 1
assert services[0] == {
'depart_st_pancras': '06:01',
'arrive_destination': '09:34',
'destination': 'Paris Gare du Nord',
'train_number': '',
}
def test_parse_results_sorted_by_departure():
html = _make_next_data([
_departure('2026-03-30 10:00:00', '2026-03-30 13:00:00'),
_departure('2026-03-30 07:00:00', '2026-03-30 10:00:00'),
])
services = _parse(html, 'Paris Gare du Nord')
assert services[0]['depart_st_pancras'] == '07:00'
assert services[1]['depart_st_pancras'] == '10:00'
def test_parse_skips_entries_with_missing_times():
html = _make_next_data([
_departure(None, '2026-03-30 09:34:00'),
_departure('2026-03-30 08:00:00', None),
_departure('2026-03-30 09:00:00', '2026-03-30 12:00:00'),
])
services = _parse(html, 'Paris Gare du Nord')
assert len(services) == 1
assert services[0]['depart_st_pancras'] == '09:00'
def test_parse_no_next_data_returns_empty():
assert _parse('<html><body>nothing here</body></html>', 'Paris Gare du Nord') == []
def test_parse_empty_departures():
html = _make_next_data([])
assert _parse(html, 'Paris Gare du Nord') == []
def test_timetable_url_uses_station_id_table():
assert timetable_url('Paris Gare du Nord') == (
'https://www.eurostar.com/uk-en/travel-info/timetable/'
'7015400/8727100/london-st-pancras-intl/paris-gare-du-nord'
)
def test_timetable_url_slugifies_destination_name():
assert timetable_url('Rotterdam Centraal') == (
'https://www.eurostar.com/uk-en/travel-info/timetable/'
'7015400/8400530/london-st-pancras-intl/rotterdam-centraal'
)