From e75acc3364f50222c551dbe674df580cd7855912 Mon Sep 17 00:00:00 2001
From: Edward Betts <edward@4angle.com>
Date: Sun, 20 Sep 2026 16:11:14 +0100
Subject: [PATCH] Suppress GWR busy-page error emails and add regression tests

---
 tests/test_update_gwr.py | 57 ++++++++++++++++++++++++++++++++++++++++
 update.py                | 12 +++++++--
 2 files changed, 67 insertions(+), 2 deletions(-)
 create mode 100644 tests/test_update_gwr.py

diff --git a/tests/test_update_gwr.py b/tests/test_update_gwr.py
new file mode 100644
index 0000000..ca02724
--- /dev/null
+++ b/tests/test_update_gwr.py
@@ -0,0 +1,57 @@
+"""Regression tests for GWR update notifications."""
+
+from pathlib import Path
+from unittest.mock import Mock, patch
+
+import flask
+import pytest
+
+from update import update_gwr_advance_ticket_date
+
+
+@pytest.mark.parametrize(
+    ("response_html", "expect_email"),
+    [
+        (
+            "<p>We are sorry that we cannot show you the page you were looking for. "
+            "Our website is extremely busy at the moment.</p>",
+            False,
+        ),
+        (
+            "<p>We are sorry that we cannot show you the page you were looking for.\n"
+            "Our website is   extremely busy at the moment.</p>",
+            False,
+        ),
+        ("<p>Unexpected response without booking dates.</p>", True),
+    ],
+    ids=["busy-page", "busy-page-whitespace", "unexpected-page"],
+)
+def test_gwr_missing_dates_notifications(
+    tmp_path: Path, response_html: str, expect_email: bool
+) -> None:
+    """Only known busy pages skip alerts; failed responses preserve the cache."""
+    cached_html = """
+    <table>
+      <tr><td>Weekdays</td><td>Friday 25 December 2026</td></tr>
+      <tr><td>Saturdays</td><td>Saturday 26 December 2026</td></tr>
+      <tr><td>Sundays</td><td>Sunday 27 December 2026</td></tr>
+    </table>
+    """
+    cache = tmp_path / "advance-tickets.html"
+    cache.write_text(cached_html)
+    config = flask.config.Config(str(tmp_path))
+    config["DATA_DIR"] = str(tmp_path)
+
+    with (
+        patch("update.requests.get", return_value=Mock(text=response_html)),
+        patch("update.agenda.mail.send_mail") as send_mail,
+    ):
+        update_gwr_advance_ticket_date(config)
+
+    if expect_email:
+        send_mail.assert_called_once_with(
+            config, "Error parsing GWR advance ticket booking dates", response_html
+        )
+    else:
+        send_mail.assert_not_called()
+    assert cache.read_text() == cached_html
diff --git a/update.py b/update.py
index 330e398..80bc7a5 100755
--- a/update.py
+++ b/update.py
@@ -14,6 +14,7 @@ from time import time
 # old authlib.jose module.  Importing authlib.deprecate first lets us insert our
 # "ignore" filter at position 0 so it wins the ordering race.
 import authlib.deprecate  # noqa: E402 — must precede the filter below
+
 warnings.filterwarnings("ignore", message="authlib.jose module is deprecated")
 
 import deepdiff
@@ -85,13 +86,20 @@ def update_gwr_advance_ticket_date(config: flask.config.Config) -> None:
     existing_html = open(filename).read()
 
     existing_dates = agenda.gwr.extract_dates(existing_html)
-    assert existing_dates
-    assert list(existing_dates.keys()) == ["Weekdays", "Saturdays", "Sundays"]
+    if existing_dates:
+        assert existing_dates
+        assert list(existing_dates.keys()) == ["Weekdays", "Saturdays", "Sundays"]
 
     new_html = requests.get(agenda.gwr.url).text
 
     new_dates = agenda.gwr.extract_dates(new_html)
     if not new_dates:
+        busy_message = (
+            "We are sorry that we cannot show you the page you were looking for. "
+            "Our website is extremely busy at the moment"
+        )
+        if busy_message in " ".join(new_html.split()):
+            return
         subject = "Error parsing GWR advance ticket booking dates"
         body = new_html
         agenda.mail.send_mail(config, subject, body)