From 90b80ae5299502da749dfacf5d543bcfff65c1aa Mon Sep 17 00:00:00 2001
From: s3lph <account-gitlab-ideynizv@kernelpanic.lol>
Date: Sat, 19 Feb 2022 03:29:27 +0100
Subject: [PATCH] v0.5: Exponential backoff for retrys

---
 CHANGELOG.md                            | 12 ++++++++++
 icalendar_timeseries_server/__init__.py |  2 +-
 icalendar_timeseries_server/cal.py      | 30 ++++++++++++++-----------
 3 files changed, 30 insertions(+), 14 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 202c38b..73d9cd9 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,6 +1,18 @@
 # iCalendar Timeseries Server Changelog
 
 
+<!-- BEGIN RELEASE v0.5 -->
+## Version 0.5
+
+### Changes
+
+<!-- BEGIN CHANGES 0.5 -->
+- Retry calendar scraping with exponential backoff.
+<!-- END CHANGES 0.5 -->
+
+<!-- END RELEASE v0.5 -->
+
+
 <!-- BEGIN RELEASE v0.4.1 -->
 ## Version 0.4.1
 
diff --git a/icalendar_timeseries_server/__init__.py b/icalendar_timeseries_server/__init__.py
index 489c7e1..27f4493 100644
--- a/icalendar_timeseries_server/__init__.py
+++ b/icalendar_timeseries_server/__init__.py
@@ -1,2 +1,2 @@
 
-__version__ = '0.4.1'
+__version__ = '0.5'
diff --git a/icalendar_timeseries_server/cal.py b/icalendar_timeseries_server/cal.py
index 8a6e11c..925fa09 100644
--- a/icalendar_timeseries_server/cal.py
+++ b/icalendar_timeseries_server/cal.py
@@ -60,12 +60,8 @@ def _scrape_calendar(name: str, config: CalendarConfig, start: datetime, end: da
     todos = []
 
     opener: urllib.request.OpenerDirector = config.get_url_opener()
-    try:
-        with opener.open(config.url) as response:
-            data = response.read().decode('utf-8')
-    except BaseException:
-        logging.exception(f'An error occurred while scraping the calendar endpoint "{name}" ({config.url})')
-        return
+    with opener.open(config.url) as response:
+        data = response.read().decode('utf-8')
     calendar = cal.Calendar.from_ical(data)
 
     for element in calendar.walk():
@@ -105,26 +101,34 @@ def _scrape_calendar(name: str, config: CalendarConfig, start: datetime, end: da
         _TODO_SCRAPE_CACHE[name] = todos
 
 
-def scrape_calendar(name: str, config: CalendarConfig):
+def scrape_calendar(name: str, config: CalendarConfig, retry: int):
     # Get current time in configured timezone
     tz = get_config().tz
     now: datetime = datetime.now(tz)
-    # Reschedule calendar scraping
-    cron = Timer(config.interval.totimedelta(start=now).total_seconds(),
-                 lambda: scrape_calendar(name, config))
-    cron.start()
+    # Only scrape at most once a minute
+    interval = max(int(config.interval.totimedelta(start=now).total_seconds()), 60)
     # Compute interval for which to return events
     start_delta: Duration = get_config().start_delta
     end_delta: Duration = get_config().end_delta
     start: datetime = now + start_delta
     end: datetime = now + end_delta
     # Scrape and parse the calendar
-    _scrape_calendar(name, config, start, end)
+    try:
+        _scrape_calendar(name, config, start, end)
+        # Reschedule calendar scraping
+        cron = Timer(interval, lambda: scrape_calendar(name, config, 0))
+    except BaseException:
+        # reschedule with exponential backoff, but no more than the regular scrape interval
+        backoff_seconds = min(60 * 2**retry, interval)
+        logging.exception(f'An error occurred while scraping the calendar endpoint "{name}" '
+                          f'({config.url}), retrying in {backoff_seconds}s.')
+        cron = Timer(backoff_seconds, lambda: scrape_calendar(name, config, retry+1))
+    cron.start()
 
 
 def start_scrape_calendar(name: str, config: CalendarConfig):
     # Schedule first calendar scraping
-    cron = Timer(0, lambda: scrape_calendar(name, config))
+    cron = Timer(0, lambda: scrape_calendar(name, config, retry=0))
     cron.start()