From b2b22916641f6ac862b09aca283c24db883ca101 Mon Sep 17 00:00:00 2001
From: Edward Betts <edward@4angle.com>
Date: Sun, 29 Oct 2023 14:59:03 +0000
Subject: [PATCH] Add GWR advance ticket availability date

Closes: #33
---
 agenda/__init__.py   | 41 ++++++++++++++++++++++++++++++++++++++++-
 templates/index.html |  8 ++++++++
 2 files changed, 48 insertions(+), 1 deletion(-)

diff --git a/agenda/__init__.py b/agenda/__init__.py
index 79a9589..25ab8bc 100644
--- a/agenda/__init__.py
+++ b/agenda/__init__.py
@@ -2,6 +2,7 @@ import configparser
 import json
 import operator
 import os
+import re
 import typing
 import warnings
 from datetime import date, datetime, timedelta, timezone
@@ -54,6 +55,43 @@ access_key = config.get("exchangerate", "access_key")
 data_dir = config.get("data", "dir")
 
 
+def extract_weekday_date(html: str) -> date | None:
+    """Furthest date of GWR advance ticket booking."""
+    # Compile a regular expression pattern to match the relevant table row
+    pattern = re.compile(
+        r"<tr>\s*<td>Weekdays</td>\s*<td>(.*?)</td>\s*</tr>", re.DOTALL
+    )
+
+    # Search the HTML for the pattern
+    if not (match := pattern.search(html)):
+        return None
+    date_str = match.group(1)
+
+    # If the year is missing, use the current year
+    if not date_str[-1].isdigit():
+        date_str += f" {date.today().year}"
+
+    return datetime.strptime(date_str, "%A %d %B %Y").date()
+
+
+def get_gwr_advance_tickets_page_html(ttl: int = 3600) -> str:
+    """Get advance-tickets web page HTML with cache."""
+    filename = os.path.join(data_dir, "advance-tickets.html")
+    url = "https://www.gwr.com/your-tickets/choosing-your-ticket/advance-tickets"
+    mtime = os.path.getmtime(filename) if os.path.exists(filename) else 0
+    if (unixtime() - mtime) < ttl:  # use cache
+        return open(filename).read()
+    r = requests.get(url)
+    open(filename, "w").write(r.text)
+    return r.text
+
+
+def find_gwr_advance_ticket_date() -> date | None:
+    """Get GWR advance tickets date with cache."""
+    html = get_gwr_advance_tickets_page_html()
+    return extract_weekday_date(html)
+
+
 def next_uk_mothers_day(input_date: date) -> date:
     """Calculate the date of the next UK Mother's Day from the current date."""
     current_year = input_date.year
@@ -420,7 +458,7 @@ def waste_collection_events() -> list[Event]:
     return events
 
 
-def get_data(now: datetime) -> dict[str, str | object]:
+def get_data(now: datetime) -> typing.Mapping[str, str | object]:
     """Get data to display on agenda dashboard."""
     rocket_dir = os.path.join(data_dir, "thespacedevs")
     today = now.date()
@@ -439,6 +477,7 @@ def get_data(now: datetime) -> dict[str, str | object]:
         "fathers_day": next_uk_fathers_day(today),
         "uk_financial_year_end": uk_financial_year_end(today),
         "xmas_last_posting_dates": xmas_last_posting_dates,
+        "gwr_advance_tickets": find_gwr_advance_ticket_date(),
         "rockets": thespacedevs.get_launches(rocket_dir, limit=40),
     }
 
diff --git a/templates/index.html b/templates/index.html
index 8b4d99b..b942d06 100644
--- a/templates/index.html
+++ b/templates/index.html
@@ -24,6 +24,7 @@
   "xmas_day": "Christmas day",
   "next_up_series": "Next Up documentary",
   "waste_schedule": "Waste schedule",
+  "gwr_advance_tickets": "GWR advance tickets",
 }
 %}
 
@@ -46,6 +47,13 @@
     <ul>
     <li>Today is {{now.strftime("%A, %-d %b %Y")}}</li>
     <li>GBPUSD: {{"{:,.3f}".format(gbpusd)}}</li>
+    <li>GWR advance ticket furthest date:
+      {% if gwr_advance_tickets %}
+        {{ gwr_advance_tickets.strftime("%A, %-d %b %Y") }}
+      {% else %}
+        unknown
+      {% endif %}
+    </li>
     </ul>
 
   <table class="table table-hover w-auto">