blob: 896c7986c4e1911d7e4e4555b18d15d9d73cfa12 [file] [log] [blame]
Guido van Rossum00efe7e2002-07-19 17:04:46 +00001"""Strptime-related classes and functions.
2
3CLASSES:
Brett Cannon474335c2003-08-05 04:02:49 +00004 LocaleTime -- Discovers and stores locale-specific time information
Barry Warsaw4d895fa2002-09-23 22:46:49 +00005 TimeRE -- Creates regexes for pattern matching a string of text containing
Brett Cannon474335c2003-08-05 04:02:49 +00006 time information
Guido van Rossum00efe7e2002-07-19 17:04:46 +00007
8FUNCTIONS:
Raymond Hettinger1fdb6332003-03-09 07:44:42 +00009 _getlang -- Figure out what language is being used for the locale
Guido van Rossum00efe7e2002-07-19 17:04:46 +000010 strptime -- Calculates the time struct represented by the passed-in string
11
Guido van Rossum00efe7e2002-07-19 17:04:46 +000012"""
13import time
14import locale
15import calendar
16from re import compile as re_compile
Antoine Pitroufd036452008-08-19 17:56:33 +000017from re import IGNORECASE, ASCII
Brett Cannon4f35c712004-10-06 02:11:37 +000018from re import escape as re_escape
Raymond Hettinger1fdb6332003-03-09 07:44:42 +000019from datetime import date as datetime_date
Brett Cannon474335c2003-08-05 04:02:49 +000020try:
Georg Brandl2067bfd2008-05-25 13:05:15 +000021 from _thread import allocate_lock as _thread_allocate_lock
Brett Cannon474335c2003-08-05 04:02:49 +000022except:
Georg Brandl2067bfd2008-05-25 13:05:15 +000023 from _dummy_thread import allocate_lock as _thread_allocate_lock
Guido van Rossum00efe7e2002-07-19 17:04:46 +000024
Christian Heimesdd15f6c2008-03-16 00:07:10 +000025__all__ = []
Guido van Rossum00efe7e2002-07-19 17:04:46 +000026
Tim Peters80cebc12003-01-19 04:40:44 +000027def _getlang():
28 # Figure out what the current language is set to.
Brett Cannon175ddb52003-07-24 06:27:17 +000029 return locale.getlocale(locale.LC_TIME)
Barry Warsaw35816e62002-08-29 16:24:50 +000030
Guido van Rossum00efe7e2002-07-19 17:04:46 +000031class LocaleTime(object):
32 """Stores and handles locale-specific information related to time.
33
Brett Cannon474335c2003-08-05 04:02:49 +000034 ATTRIBUTES:
Guido van Rossum00efe7e2002-07-19 17:04:46 +000035 f_weekday -- full weekday names (7-item list)
36 a_weekday -- abbreviated weekday names (7-item list)
Brett Cannonf5c96fb2003-08-08 01:53:05 +000037 f_month -- full month names (13-item list; dummy value in [0], which
Guido van Rossum00efe7e2002-07-19 17:04:46 +000038 is added by code)
Brett Cannonf5c96fb2003-08-08 01:53:05 +000039 a_month -- abbreviated month names (13-item list, dummy value in
Guido van Rossum00efe7e2002-07-19 17:04:46 +000040 [0], which is added by code)
41 am_pm -- AM/PM representation (2-item list)
42 LC_date_time -- format string for date/time representation (string)
43 LC_date -- format string for date representation (string)
44 LC_time -- format string for time representation (string)
Tim Peters469cdad2002-08-08 20:19:19 +000045 timezone -- daylight- and non-daylight-savings timezone representation
Brett Cannon474335c2003-08-05 04:02:49 +000046 (2-item list of sets)
47 lang -- Language used by instance (2-item tuple)
Guido van Rossum00efe7e2002-07-19 17:04:46 +000048 """
49
Brett Cannon474335c2003-08-05 04:02:49 +000050 def __init__(self):
51 """Set all attributes.
Raymond Hettinger6b59f5f2003-10-16 05:53:16 +000052
Brett Cannon474335c2003-08-05 04:02:49 +000053 Order of methods called matters for dependency reasons.
54
55 The locale language is set at the offset and then checked again before
56 exiting. This is to make sure that the attributes were not set with a
57 mix of information from more than one locale. This would most likely
58 happen when using threads where one thread calls a locale-dependent
59 function while another thread changes the locale while the function in
60 the other thread is still running. Proper coding would call for
61 locks to prevent changing the locale while locale-dependent code is
62 running. The check here is done in case someone does not think about
63 doing this.
Brett Cannon5187a3b2003-08-11 07:24:05 +000064
65 Only other possible issue is if someone changed the timezone and did
66 not call tz.tzset . That is an issue for the programmer, though,
67 since changing the timezone is worthless without that call.
Raymond Hettinger6b59f5f2003-10-16 05:53:16 +000068
Brett Cannon474335c2003-08-05 04:02:49 +000069 """
70 self.lang = _getlang()
71 self.__calc_weekday()
72 self.__calc_month()
73 self.__calc_am_pm()
74 self.__calc_timezone()
75 self.__calc_date_time()
76 if _getlang() != self.lang:
77 raise ValueError("locale changed during initialization")
Guido van Rossum00efe7e2002-07-19 17:04:46 +000078
79 def __pad(self, seq, front):
Brett Cannon474335c2003-08-05 04:02:49 +000080 # Add '' to seq to either the front (is True), else the back.
Guido van Rossum00efe7e2002-07-19 17:04:46 +000081 seq = list(seq)
Barry Warsaw35816e62002-08-29 16:24:50 +000082 if front:
83 seq.insert(0, '')
84 else:
85 seq.append('')
Guido van Rossum00efe7e2002-07-19 17:04:46 +000086 return seq
87
Guido van Rossum00efe7e2002-07-19 17:04:46 +000088 def __calc_weekday(self):
Brett Cannon474335c2003-08-05 04:02:49 +000089 # Set self.a_weekday and self.f_weekday using the calendar
Barry Warsaw35816e62002-08-29 16:24:50 +000090 # module.
Brett Cannon474335c2003-08-05 04:02:49 +000091 a_weekday = [calendar.day_abbr[i].lower() for i in range(7)]
92 f_weekday = [calendar.day_name[i].lower() for i in range(7)]
93 self.a_weekday = a_weekday
94 self.f_weekday = f_weekday
Tim Peters469cdad2002-08-08 20:19:19 +000095
Guido van Rossum00efe7e2002-07-19 17:04:46 +000096 def __calc_month(self):
Brett Cannon474335c2003-08-05 04:02:49 +000097 # Set self.f_month and self.a_month using the calendar module.
98 a_month = [calendar.month_abbr[i].lower() for i in range(13)]
99 f_month = [calendar.month_name[i].lower() for i in range(13)]
100 self.a_month = a_month
101 self.f_month = f_month
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000102
103 def __calc_am_pm(self):
Brett Cannon474335c2003-08-05 04:02:49 +0000104 # Set self.am_pm by using time.strftime().
Tim Peters469cdad2002-08-08 20:19:19 +0000105
Barry Warsaw35816e62002-08-29 16:24:50 +0000106 # The magic date (1999,3,17,hour,44,55,2,76,0) is not really that
107 # magical; just happened to have used it everywhere else where a
108 # static date was needed.
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000109 am_pm = []
Guido van Rossumcd16bf62007-06-13 18:07:49 +0000110 for hour in (1, 22):
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000111 time_tuple = time.struct_time((1999,3,17,hour,44,55,2,76,0))
Brett Cannon474335c2003-08-05 04:02:49 +0000112 am_pm.append(time.strftime("%p", time_tuple).lower())
113 self.am_pm = am_pm
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000114
115 def __calc_date_time(self):
Brett Cannon474335c2003-08-05 04:02:49 +0000116 # Set self.date_time, self.date, & self.time by using
Barry Warsaw35816e62002-08-29 16:24:50 +0000117 # time.strftime().
Tim Peters469cdad2002-08-08 20:19:19 +0000118
Barry Warsaw35816e62002-08-29 16:24:50 +0000119 # Use (1999,3,17,22,44,55,2,76,0) for magic date because the amount of
120 # overloaded numbers is minimized. The order in which searches for
121 # values within the format string is very important; it eliminates
122 # possible ambiguity for what something represents.
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000123 time_tuple = time.struct_time((1999,3,17,22,44,55,2,76,0))
124 date_time = [None, None, None]
Brett Cannon474335c2003-08-05 04:02:49 +0000125 date_time[0] = time.strftime("%c", time_tuple).lower()
126 date_time[1] = time.strftime("%x", time_tuple).lower()
127 date_time[2] = time.strftime("%X", time_tuple).lower()
128 replacement_pairs = [('%', '%%'), (self.f_weekday[2], '%A'),
Barry Warsaw4d895fa2002-09-23 22:46:49 +0000129 (self.f_month[3], '%B'), (self.a_weekday[2], '%a'),
130 (self.a_month[3], '%b'), (self.am_pm[1], '%p'),
Barry Warsaw4d895fa2002-09-23 22:46:49 +0000131 ('1999', '%Y'), ('99', '%y'), ('22', '%H'),
132 ('44', '%M'), ('55', '%S'), ('76', '%j'),
133 ('17', '%d'), ('03', '%m'), ('3', '%m'),
134 # '3' needed for when no leading zero.
Brett Cannon474335c2003-08-05 04:02:49 +0000135 ('2', '%w'), ('10', '%I')]
136 replacement_pairs.extend([(tz, "%Z") for tz_values in self.timezone
137 for tz in tz_values])
138 for offset,directive in ((0,'%c'), (1,'%x'), (2,'%X')):
139 current_format = date_time[offset]
140 for old, new in replacement_pairs:
Jack Jansen62fe7552003-01-15 22:59:39 +0000141 # Must deal with possible lack of locale info
142 # manifesting itself as the empty string (e.g., Swedish's
143 # lack of AM/PM info) or a platform returning a tuple of empty
144 # strings (e.g., MacOS 9 having timezone as ('','')).
145 if old:
Barry Warsaw4d895fa2002-09-23 22:46:49 +0000146 current_format = current_format.replace(old, new)
Brett Cannonf1b2ba62005-08-29 18:25:55 +0000147 # If %W is used, then Sunday, 2005-01-03 will fall on week 0 since
148 # 2005-01-03 occurs before the first Monday of the year. Otherwise
149 # %U is used.
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000150 time_tuple = time.struct_time((1999,1,3,1,1,1,6,3,0))
Brett Cannon6e372d12005-08-27 19:25:59 +0000151 if '00' in time.strftime(directive, time_tuple):
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000152 U_W = '%W'
Brett Cannonf1b2ba62005-08-29 18:25:55 +0000153 else:
154 U_W = '%U'
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000155 date_time[offset] = current_format.replace('11', U_W)
Brett Cannon474335c2003-08-05 04:02:49 +0000156 self.LC_date_time = date_time[0]
157 self.LC_date = date_time[1]
Raymond Hettinger6b59f5f2003-10-16 05:53:16 +0000158 self.LC_time = date_time[2]
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000159
160 def __calc_timezone(self):
Brett Cannon474335c2003-08-05 04:02:49 +0000161 # Set self.timezone by using time.tzname.
Brett Cannon5187a3b2003-08-11 07:24:05 +0000162 # Do not worry about possibility of time.tzname[0] == timetzname[1]
163 # and time.daylight; handle that in strptime .
Brett Cannonabe8eb02003-05-13 20:28:15 +0000164 try:
165 time.tzset()
166 except AttributeError:
167 pass
Raymond Hettingera690a992003-11-16 16:17:49 +0000168 no_saving = frozenset(["utc", "gmt", time.tzname[0].lower()])
Brett Cannon172d9ef2003-05-11 06:23:36 +0000169 if time.daylight:
Raymond Hettingera690a992003-11-16 16:17:49 +0000170 has_saving = frozenset([time.tzname[1].lower()])
Brett Cannon172d9ef2003-05-11 06:23:36 +0000171 else:
Raymond Hettingera690a992003-11-16 16:17:49 +0000172 has_saving = frozenset()
Brett Cannon474335c2003-08-05 04:02:49 +0000173 self.timezone = (no_saving, has_saving)
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000174
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000175
176class TimeRE(dict):
177 """Handle conversion from format directives to regexes."""
178
Brett Cannon2c24d422003-07-24 20:02:28 +0000179 def __init__(self, locale_time=None):
Brett Cannon474335c2003-08-05 04:02:49 +0000180 """Create keys/values.
Raymond Hettinger6b59f5f2003-10-16 05:53:16 +0000181
Brett Cannon474335c2003-08-05 04:02:49 +0000182 Order of execution is important for dependency reasons.
Raymond Hettinger6b59f5f2003-10-16 05:53:16 +0000183
Brett Cannon474335c2003-08-05 04:02:49 +0000184 """
185 if locale_time:
186 self.locale_time = locale_time
187 else:
188 self.locale_time = LocaleTime()
Guido van Rossumcd16bf62007-06-13 18:07:49 +0000189 base = super()
Neal Norwitz5efc50d2002-12-30 22:23:12 +0000190 base.__init__({
Brett Cannon474335c2003-08-05 04:02:49 +0000191 # The " \d" part of the regex is to make %c from ANSI C work
Neal Norwitz5efc50d2002-12-30 22:23:12 +0000192 'd': r"(?P<d>3[0-1]|[1-2]\d|0[1-9]|[1-9]| [1-9])",
Christian Heimesdd15f6c2008-03-16 00:07:10 +0000193 'f': r"(?P<f>[0-9]{1,6})",
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000194 'H': r"(?P<H>2[0-3]|[0-1]\d|\d)",
Neal Norwitz5efc50d2002-12-30 22:23:12 +0000195 'I': r"(?P<I>1[0-2]|0[1-9]|[1-9])",
196 'j': r"(?P<j>36[0-6]|3[0-5]\d|[1-2]\d\d|0[1-9]\d|00[1-9]|[1-9]\d|0[1-9]|[1-9])",
197 'm': r"(?P<m>1[0-2]|0[1-9]|[1-9])",
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000198 'M': r"(?P<M>[0-5]\d|\d)",
199 'S': r"(?P<S>6[0-1]|[0-5]\d|\d)",
200 'U': r"(?P<U>5[0-3]|[0-4]\d|\d)",
201 'w': r"(?P<w>[0-6])",
Neal Norwitz5efc50d2002-12-30 22:23:12 +0000202 # W is set below by using 'U'
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000203 'y': r"(?P<y>\d\d)",
Brett Cannon474335c2003-08-05 04:02:49 +0000204 #XXX: Does 'Y' need to worry about having less or more than
205 # 4 digits?
206 'Y': r"(?P<Y>\d\d\d\d)",
207 'A': self.__seqToRE(self.locale_time.f_weekday, 'A'),
208 'a': self.__seqToRE(self.locale_time.a_weekday, 'a'),
209 'B': self.__seqToRE(self.locale_time.f_month[1:], 'B'),
210 'b': self.__seqToRE(self.locale_time.a_month[1:], 'b'),
211 'p': self.__seqToRE(self.locale_time.am_pm, 'p'),
Brett Cannonf7948c22004-10-06 02:23:14 +0000212 'Z': self.__seqToRE((tz for tz_names in self.locale_time.timezone
213 for tz in tz_names),
Brett Cannon474335c2003-08-05 04:02:49 +0000214 'Z'),
215 '%': '%'})
Brett Cannon8abcc5d2004-10-18 01:37:57 +0000216 base.__setitem__('W', base.__getitem__('U').replace('U', 'W'))
Brett Cannon474335c2003-08-05 04:02:49 +0000217 base.__setitem__('c', self.pattern(self.locale_time.LC_date_time))
218 base.__setitem__('x', self.pattern(self.locale_time.LC_date))
219 base.__setitem__('X', self.pattern(self.locale_time.LC_time))
Tim Peters469cdad2002-08-08 20:19:19 +0000220
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000221 def __seqToRE(self, to_convert, directive):
Brett Cannon474335c2003-08-05 04:02:49 +0000222 """Convert a list to a regex string for matching a directive.
Raymond Hettinger6b59f5f2003-10-16 05:53:16 +0000223
Brett Cannon474335c2003-08-05 04:02:49 +0000224 Want possible matching values to be from longest to shortest. This
225 prevents the possibility of a match occuring for a value that also
226 a substring of a larger value that should have matched (e.g., 'abc'
227 matching when 'abcdef' should have been the match).
Raymond Hettinger6b59f5f2003-10-16 05:53:16 +0000228
Brett Cannon474335c2003-08-05 04:02:49 +0000229 """
Brett Cannonffa5cf92004-10-06 22:48:58 +0000230 to_convert = sorted(to_convert, key=len, reverse=True)
Jack Jansen62fe7552003-01-15 22:59:39 +0000231 for value in to_convert:
232 if value != '':
233 break
234 else:
235 return ''
Brett Cannon4f35c712004-10-06 02:11:37 +0000236 regex = '|'.join(re_escape(stuff) for stuff in to_convert)
Barry Warsaw4d895fa2002-09-23 22:46:49 +0000237 regex = '(?P<%s>%s' % (directive, regex)
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000238 return '%s)' % regex
239
240 def pattern(self, format):
Brett Cannon474335c2003-08-05 04:02:49 +0000241 """Return regex pattern for the format string.
Tim Peters0eadaac2003-04-24 16:02:54 +0000242
Brett Cannon1e91d8e2003-04-19 04:00:56 +0000243 Need to make sure that any characters that might be interpreted as
Brett Cannon5187a3b2003-08-11 07:24:05 +0000244 regex syntax are escaped.
Tim Peters0eadaac2003-04-24 16:02:54 +0000245
Brett Cannon1e91d8e2003-04-19 04:00:56 +0000246 """
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000247 processed_format = ''
Brett Cannon1e91d8e2003-04-19 04:00:56 +0000248 # The sub() call escapes all characters that might be misconstrued
Brett Cannon4f35c712004-10-06 02:11:37 +0000249 # as regex syntax. Cannot use re.escape since we have to deal with
250 # format directives (%m, etc.).
Brett Cannon953c6f52003-08-29 02:28:54 +0000251 regex_chars = re_compile(r"([\\.^$*+?\(\){}\[\]|])")
Brett Cannon1e91d8e2003-04-19 04:00:56 +0000252 format = regex_chars.sub(r"\\\1", format)
Tim Peters80cebc12003-01-19 04:40:44 +0000253 whitespace_replacement = re_compile('\s+')
Guido van Rossumcd16bf62007-06-13 18:07:49 +0000254 format = whitespace_replacement.sub('\s+', format)
Raymond Hettingerbac788a2004-05-04 09:21:43 +0000255 while '%' in format:
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000256 directive_index = format.index('%')+1
Tim Peters469cdad2002-08-08 20:19:19 +0000257 processed_format = "%s%s%s" % (processed_format,
Barry Warsaw35816e62002-08-29 16:24:50 +0000258 format[:directive_index-1],
259 self[format[directive_index]])
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000260 format = format[directive_index+1:]
261 return "%s%s" % (processed_format, format)
262
263 def compile(self, format):
264 """Return a compiled re object for the format string."""
Antoine Pitroufd036452008-08-19 17:56:33 +0000265 return re_compile(self.pattern(format), IGNORECASE | ASCII)
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000266
Brett Cannon474335c2003-08-05 04:02:49 +0000267_cache_lock = _thread_allocate_lock()
268# DO NOT modify _TimeRE_cache or _regex_cache without acquiring the cache lock
269# first!
270_TimeRE_cache = TimeRE()
Brett Cannon5187a3b2003-08-11 07:24:05 +0000271_CACHE_MAX_SIZE = 5 # Max number of regexes stored in _regex_cache
Brett Cannon474335c2003-08-05 04:02:49 +0000272_regex_cache = {}
Guido van Rossum00efe7e2002-07-19 17:04:46 +0000273
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000274def _calc_julian_from_U_or_W(year, week_of_year, day_of_week, week_starts_Mon):
275 """Calculate the Julian day based on the year, week of the year, and day of
276 the week, with week_start_day representing whether the week of the year
277 assumes the week starts on Sunday or Monday (6 or 0)."""
278 first_weekday = datetime_date(year, 1, 1).weekday()
279 # If we are dealing with the %U directive (week starts on Sunday), it's
280 # easier to just shift the view to Sunday being the first day of the
281 # week.
282 if not week_starts_Mon:
283 first_weekday = (first_weekday + 1) % 7
284 day_of_week = (day_of_week + 1) % 7
285 # Need to watch out for a week 0 (when the first day of the year is not
286 # the same as that specified by %U or %W).
287 week_0_length = (7 - first_weekday) % 7
288 if week_of_year == 0:
289 return 1 + day_of_week - first_weekday
290 else:
291 days_to_week = week_0_length + (7 * (week_of_year - 1))
292 return 1 + days_to_week + day_of_week
293
294
Christian Heimesdd15f6c2008-03-16 00:07:10 +0000295def _strptime(data_string, format="%a %b %d %H:%M:%S %Y"):
Brett Cannon5187a3b2003-08-11 07:24:05 +0000296 """Return a time struct based on the input string and the format string."""
Brett Cannona783d062005-09-15 02:34:56 +0000297 global _TimeRE_cache, _regex_cache
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000298 with _cache_lock:
Guido van Rossumd8faa362007-04-27 19:54:29 +0000299 if _getlang() != _TimeRE_cache.locale_time.lang:
Brett Cannon474335c2003-08-05 04:02:49 +0000300 _TimeRE_cache = TimeRE()
Guido van Rossumd8faa362007-04-27 19:54:29 +0000301 _regex_cache.clear()
Brett Cannon474335c2003-08-05 04:02:49 +0000302 if len(_regex_cache) > _CACHE_MAX_SIZE:
303 _regex_cache.clear()
Guido van Rossumd8faa362007-04-27 19:54:29 +0000304 locale_time = _TimeRE_cache.locale_time
Brett Cannon474335c2003-08-05 04:02:49 +0000305 format_regex = _regex_cache.get(format)
306 if not format_regex:
Brett Cannon5d0bf942005-11-02 23:04:26 +0000307 try:
Guido van Rossumd8faa362007-04-27 19:54:29 +0000308 format_regex = _TimeRE_cache.compile(format)
Brett Cannon5d0bf942005-11-02 23:04:26 +0000309 # KeyError raised when a bad format is found; can be specified as
310 # \\, in which case it was a stray % but with a space after it
Guido van Rossumb940e112007-01-10 16:19:56 +0000311 except KeyError as err:
Brett Cannon5d0bf942005-11-02 23:04:26 +0000312 bad_directive = err.args[0]
313 if bad_directive == "\\":
314 bad_directive = "%"
315 del err
316 raise ValueError("'%s' is a bad directive in format '%s'" %
317 (bad_directive, format))
318 # IndexError only occurs when the format string is "%"
319 except IndexError:
320 raise ValueError("stray %% in format '%s'" % format)
Brett Cannon474335c2003-08-05 04:02:49 +0000321 _regex_cache[format] = format_regex
Tim Peters80cebc12003-01-19 04:40:44 +0000322 found = format_regex.match(data_string)
Tim Peters08e54272003-01-18 03:53:49 +0000323 if not found:
Thomas Wouters89f507f2006-12-13 04:49:30 +0000324 raise ValueError("time data %r does not match format %r" %
Raymond Hettinger4a6302b2003-07-13 01:31:38 +0000325 (data_string, format))
Brett Cannon2b6dfec2003-04-28 21:30:13 +0000326 if len(data_string) != found.end():
327 raise ValueError("unconverted data remains: %s" %
328 data_string[found.end():])
Tim Peters08e54272003-01-18 03:53:49 +0000329 year = 1900
330 month = day = 1
Christian Heimesdd15f6c2008-03-16 00:07:10 +0000331 hour = minute = second = fraction = 0
Tim Peters08e54272003-01-18 03:53:49 +0000332 tz = -1
Brett Cannon8dc25ad2004-10-18 01:47:46 +0000333 # Default to -1 to signify that values not known; not critical to have,
334 # though
Brett Cannon8abcc5d2004-10-18 01:37:57 +0000335 week_of_year = -1
336 week_of_year_start = -1
Brett Cannon8dc25ad2004-10-18 01:47:46 +0000337 # weekday and julian defaulted to -1 so as to signal need to calculate
338 # values
Tim Peters08e54272003-01-18 03:53:49 +0000339 weekday = julian = -1
340 found_dict = found.groupdict()
Guido van Rossumcc2b0162007-02-11 06:12:03 +0000341 for group_key in found_dict.keys():
Brett Cannon8abcc5d2004-10-18 01:37:57 +0000342 # Directives not explicitly handled below:
343 # c, x, X
344 # handled by making out of other directives
345 # U, W
346 # worthless without day of the week
Tim Peters08e54272003-01-18 03:53:49 +0000347 if group_key == 'y':
348 year = int(found_dict['y'])
349 # Open Group specification for strptime() states that a %y
350 #value in the range of [00, 68] is in the century 2000, while
351 #[69,99] is in the century 1900
352 if year <= 68:
353 year += 2000
354 else:
355 year += 1900
356 elif group_key == 'Y':
357 year = int(found_dict['Y'])
358 elif group_key == 'm':
359 month = int(found_dict['m'])
360 elif group_key == 'B':
Brett Cannon474335c2003-08-05 04:02:49 +0000361 month = locale_time.f_month.index(found_dict['B'].lower())
Tim Peters08e54272003-01-18 03:53:49 +0000362 elif group_key == 'b':
Brett Cannon474335c2003-08-05 04:02:49 +0000363 month = locale_time.a_month.index(found_dict['b'].lower())
Tim Peters08e54272003-01-18 03:53:49 +0000364 elif group_key == 'd':
365 day = int(found_dict['d'])
Neal Norwitz77290f22003-06-29 04:16:49 +0000366 elif group_key == 'H':
Tim Peters08e54272003-01-18 03:53:49 +0000367 hour = int(found_dict['H'])
368 elif group_key == 'I':
369 hour = int(found_dict['I'])
370 ampm = found_dict.get('p', '').lower()
371 # If there was no AM/PM indicator, we'll treat this like AM
Brett Cannon474335c2003-08-05 04:02:49 +0000372 if ampm in ('', locale_time.am_pm[0]):
Tim Peters08e54272003-01-18 03:53:49 +0000373 # We're in AM so the hour is correct unless we're
374 # looking at 12 midnight.
375 # 12 midnight == 12 AM == hour 0
376 if hour == 12:
377 hour = 0
Brett Cannon474335c2003-08-05 04:02:49 +0000378 elif ampm == locale_time.am_pm[1]:
Tim Peters08e54272003-01-18 03:53:49 +0000379 # We're in PM so we need to add 12 to the hour unless
380 # we're looking at 12 noon.
381 # 12 noon == 12 PM == hour 12
382 if hour != 12:
383 hour += 12
384 elif group_key == 'M':
385 minute = int(found_dict['M'])
386 elif group_key == 'S':
387 second = int(found_dict['S'])
Christian Heimesdd15f6c2008-03-16 00:07:10 +0000388 elif group_key == 'f':
389 s = found_dict['f']
390 # Pad to always return microseconds.
391 s += "0" * (6 - len(s))
392 fraction = int(s)
Tim Peters08e54272003-01-18 03:53:49 +0000393 elif group_key == 'A':
Brett Cannon474335c2003-08-05 04:02:49 +0000394 weekday = locale_time.f_weekday.index(found_dict['A'].lower())
Tim Peters08e54272003-01-18 03:53:49 +0000395 elif group_key == 'a':
Brett Cannon474335c2003-08-05 04:02:49 +0000396 weekday = locale_time.a_weekday.index(found_dict['a'].lower())
Tim Peters08e54272003-01-18 03:53:49 +0000397 elif group_key == 'w':
398 weekday = int(found_dict['w'])
399 if weekday == 0:
400 weekday = 6
401 else:
402 weekday -= 1
403 elif group_key == 'j':
404 julian = int(found_dict['j'])
Brett Cannon8abcc5d2004-10-18 01:37:57 +0000405 elif group_key in ('U', 'W'):
406 week_of_year = int(found_dict[group_key])
407 if group_key == 'U':
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000408 # U starts week on Sunday.
Brett Cannon8abcc5d2004-10-18 01:37:57 +0000409 week_of_year_start = 6
410 else:
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000411 # W starts week on Monday.
Brett Cannon8abcc5d2004-10-18 01:37:57 +0000412 week_of_year_start = 0
Tim Peters08e54272003-01-18 03:53:49 +0000413 elif group_key == 'Z':
Brett Cannon172d9ef2003-05-11 06:23:36 +0000414 # Since -1 is default value only need to worry about setting tz if
415 # it can be something other than -1.
Tim Peters08e54272003-01-18 03:53:49 +0000416 found_zone = found_dict['Z'].lower()
Brett Cannon5187a3b2003-08-11 07:24:05 +0000417 for value, tz_values in enumerate(locale_time.timezone):
418 if found_zone in tz_values:
419 # Deal with bad locale setup where timezone names are the
420 # same and yet time.daylight is true; too ambiguous to
421 # be able to tell what timezone has daylight savings
Brett Cannon8172ac32004-03-07 23:16:27 +0000422 if (time.tzname[0] == time.tzname[1] and
423 time.daylight and found_zone not in ("utc", "gmt")):
Tim Peters58eb11c2004-01-18 20:29:55 +0000424 break
Brett Cannon5187a3b2003-08-11 07:24:05 +0000425 else:
Brett Cannon474335c2003-08-05 04:02:49 +0000426 tz = value
Brett Cannon5187a3b2003-08-11 07:24:05 +0000427 break
Brett Cannon8abcc5d2004-10-18 01:37:57 +0000428 # If we know the week of the year and what day of that week, we can figure
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000429 # out the Julian day of the year.
Brett Cannon14adbe72004-10-28 04:49:21 +0000430 if julian == -1 and week_of_year != -1 and weekday != -1:
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000431 week_starts_Mon = True if week_of_year_start == 0 else False
432 julian = _calc_julian_from_U_or_W(year, week_of_year, weekday,
433 week_starts_Mon)
Raymond Hettinger1fdb6332003-03-09 07:44:42 +0000434 # Cannot pre-calculate datetime_date() since can change in Julian
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000435 # calculation and thus could have different value for the day of the week
436 # calculation.
Tim Peters08e54272003-01-18 03:53:49 +0000437 if julian == -1:
Raymond Hettinger1fdb6332003-03-09 07:44:42 +0000438 # Need to add 1 to result since first day of the year is 1, not 0.
439 julian = datetime_date(year, month, day).toordinal() - \
440 datetime_date(year, 1, 1).toordinal() + 1
441 else: # Assume that if they bothered to include Julian day it will
Thomas Wouters9fe394c2007-02-05 01:24:16 +0000442 # be accurate.
Raymond Hettinger1fdb6332003-03-09 07:44:42 +0000443 datetime_result = datetime_date.fromordinal((julian - 1) + datetime_date(year, 1, 1).toordinal())
444 year = datetime_result.year
445 month = datetime_result.month
446 day = datetime_result.day
Tim Peters08e54272003-01-18 03:53:49 +0000447 if weekday == -1:
Raymond Hettinger1fdb6332003-03-09 07:44:42 +0000448 weekday = datetime_date(year, month, day).weekday()
Christian Heimesdd15f6c2008-03-16 00:07:10 +0000449 return (time.struct_time((year, month, day,
450 hour, minute, second,
451 weekday, julian, tz)), fraction)
452
453def _strptime_time(data_string, format="%a %b %d %H:%M:%S %Y"):
454 return _strptime(data_string, format)[0]