Rewrote the parse_date utility function to allow arbitrary specificat…

…ion of time formats. Added timestamp to the attribute map with the twitter api time format specified. This change should be backwards compatible because the parse_date function defaults to the previously hardcoded timeestamp format.

pyposterous/models.py

@@ -135,5 +135,6 @@ class User(PosterousData):
     ('views', 'filesize', 'height', 'width', 'commentscount', 'num_posts', 'size', ):int,
     ('private', 'commentsenabled', 'primary'):lambda x: x.upper() == 'TRUE',
     ('body',):lambda x: x.strip(), # Hopefully whitespace will not be significant. 
+    ('timestamp',):lambda x: parse_date(x, '%a %b %d %H:%M:%S %Y'),
     ('date',):lambda x: parse_date(x),} 
 

pyposterous/utils.py

@@ -30,23 +30,40 @@ def docstring_trim(docstring):
         trimmed.pop(0)
     # Return a single string:
     return '\n'.join(trimmed)
+
+def parse_date(time_string, time_format="%a, %d %b %Y %H:%M:%S"):
+    """Converts a string representation of a time stamp to a UTC datetime object.
+    A utcoffset of -HHMM is expected in the time_string, but should not be
+    represented in the time_format string by %z. Just leave it out. This
+    function will find it, and do the necessary calculation to convert the
+    output to UTC.
+    
+    time_string -- a string to be converted to a UTC datetime object. Must contain a utcoffset string +HHMM or -HHMM
+    time_format -- a string representing the format of time_string. See strptime documentation for an example. Leave out %z.    
+    """
+    from re import findall
+    from datetime import datetime
 
-def parse_date(time_string):
-    """Returns a UTC datetime object based on a string representation of time
-    in this format: %a, %d %b %Y %H:%M:%S %z"""
-    from datetime import datetime, timedelta
+    utc_offset_str = findall(r'\+[0-9]{4}|-[0-9]{4}', time_string)[0]
+    time_string = time_string.replace(utc_offset_str, "", 1)
 
-    utc_offset_str = time_string[-6:].strip()
-    sign = 1
-    # UTC offset is negative. Trim that -!
-    if utc_offset_str[0] == '-':
-        sign = -1
-        utc_offset_str = utc_offset_str[1:5]
+    # the substitution above may have resulted in unnecessary whitespace
+    time_string = time_string.strip().replace('  ', ' ')
+
+    return convert_to_utc(datetime.strptime(time_string, time_format), utc_offset_str)
 
-    utcoffset = sign * timedelta(hours=int(utc_offset_str[0:2]), minutes=int(utc_offset_str[2:4]))
+def convert_to_utc(date, offset):
+    from datetime import timedelta
+
+    sign = offset[0]
+    hour = int(offset[1:3])
+    minute = int(offset[3:5])
+
+    if sign == '-': sign = 1
+    if sign == '+': sign = -1
+
+    return date + sign * timedelta(hours=hour, minutes=minute)
 
-    return datetime.strptime(time_string[:-6], '%a, %d %b %Y %H:%M:%S') - utcoffset
-
 def try_parse_int(number):
     try:
         return int(number)