When parsing datetimes, parse out the timezone, and returns this

Rory McCann · Rory McCann · commit f542b1b37c61 · 2015-03-21T10:33:31.000+01:00
see #9
diff --git a/apache_log_parser/__init__.py b/apache_log_parser/__init__.py
@@ -1,5 +1,5 @@
 import re
-from datetime import datetime
+from datetime import datetime, tzinfo, timedelta
 
 import user_agents
 
@@ -58,23 +58,81 @@ def parse_user_agent(matched_strings):
 
     return matched_strings
 
+class FixedOffset(tzinfo):
+    """Fixed offset in minutes east from UTC."""
+
+    def __init__(self, string):
+        #import pudb ; pudb.set_trace()
+        if string[0] == '-':
+            direction = -1
+            string = string[1:]
+        elif string[0] == '+':
+            direction = +1
+            string = string[1:]
+        else:
+            direction = +1
+            string = string
+
+        hr_offset = int(string[0:2], 10)
+        min_offset = int(string[2:3], 10)
+        min_offset = hr_offset * 60 + min_offset
+        min_offset = direction * min_offset
+
+        self.__offset = timedelta(minutes = min_offset)
+
+        self.__name = string
+
+    def utcoffset(self, dt):
+        return self.__offset
+
+    def tzname(self, dt):
+        return self.__name
+
+    def dst(self, dt):
+        return timedelta(0)
+
+    def __repr__(self):
+        return repr(self.__name)
 
 
 def apachetime(s):
     """
     Given a string representation of a datetime in apache format (e.g.
-    "01/Sep/2012:06:05:11 +0000"), return the python datetime for that string
+    "01/Sep/2012:06:05:11 +0000"), return the python datetime for that string, with timezone
     """
     month_map = {'Jan': 1, 'Feb': 2, 'Mar':3, 'Apr':4, 'May':5, 'Jun':6, 'Jul':7,
         'Aug':8,  'Sep': 9, 'Oct':10, 'Nov': 11, 'Dec': 12}
     s = s[1:-1]
-    return datetime(int(s[7:11]), month_map[s[3:6]], int(s[0:2]), \
-         int(s[12:14]), int(s[15:17]), int(s[18:20]))
+
+    tz_string = s[21:26]
+    tz = FixedOffset(tz_string)
+
+    obj = datetime(year=int(s[7:11]), month=month_map[s[3:6]], day=int(s[0:2]),
+                hour=int(s[12:14]), minute=int(s[15:17]), second=int(s[18:20]),
+                tzinfo=tz )
+
+    return obj
 
 def format_time(matched_strings):
+
     time_received = matched_strings['time_received']
+
+    # Parse it to a timezone string
     obj = apachetime(time_received)
-    return {'time_received':time_received, 'time_received_datetimeobj': obj, 'time_received_isoformat': obj.isoformat()}
+    
+    # For backwards compatibility, time_received_datetimeobj is a naive
+    # datetime, so we have to create a timezone less version
+    naive_obj = obj.replace(tzinfo=None)
+
+    utc = FixedOffset('0000')
+    utc_obj = obj.astimezone(utc)
+
+    return {
+        'time_received':time_received,
+        'time_received_datetimeobj': naive_obj, 'time_received_isoformat': naive_obj.isoformat(),
+        'time_received_tz_datetimeobj': obj, 'time_received_tz_isoformat': obj.isoformat(),
+        'time_received_utc_datetimeobj': utc_obj, 'time_received_utc_isoformat': utc_obj.isoformat(),
+    }
 
 
 FORMAT_STRINGS = [
diff --git a/apache_log_parser/tests.py b/apache_log_parser/tests.py
@@ -42,6 +42,10 @@ def test_pr8(self):
             'time_received_datetimeobj': datetime.datetime(2014, 11, 28, 10, 3, 40),
             'time_received_isoformat': '2014-11-28T10:03:40', 'remote_host': '127.0.0.1',
             'time_received': '[28/Nov/2014:10:03:40 +0000]',
+            'time_received_tz_datetimeobj': datetime.datetime(2014, 11, 28, 10, 3, 40, tzinfo=apache_log_parser.FixedOffset("0000")),
+            'time_received_tz_isoformat': '2014-11-28T10:03:40+00:00', 'remote_host': '127.0.0.1',
+            'time_received_utc_datetimeobj': datetime.datetime(2014, 11, 28, 10, 3, 40, tzinfo=apache_log_parser.FixedOffset("0000")),
+            'time_received_utc_isoformat': '2014-11-28T10:03:40+00:00', 'remote_host': '127.0.0.1',
             'extension_ssl_cipher': 'MY-CYPHER',
         })
 
@@ -77,6 +81,10 @@ def test_issue9(self):
             'time_received': '[08/Mar/2015:18:06:58 -0400]',
             'time_received_datetimeobj': datetime.datetime(2015, 3, 8, 18, 6, 58),
             'time_received_isoformat': '2015-03-08T18:06:58',
+            'time_received_tz_datetimeobj': datetime.datetime(2015, 3, 8, 18, 6, 58, tzinfo=apache_log_parser.FixedOffset('-0400')),
+            'time_received_tz_isoformat': '2015-03-08T18:06:58-04:00',
+            'time_received_utc_datetimeobj': datetime.datetime(2015, 3, 8, 22, 6, 58, tzinfo=apache_log_parser.FixedOffset('0000')),
+            'time_received_utc_isoformat': '2015-03-08T22:06:58+00:00',
         })
 
 if __name__ == '__main__':
