| 1 | n/a | """Strptime-related classes and functions. |
|---|
| 2 | n/a | |
|---|
| 3 | n/a | CLASSES: |
|---|
| 4 | n/a | LocaleTime -- Discovers and stores locale-specific time information |
|---|
| 5 | n/a | TimeRE -- Creates regexes for pattern matching a string of text containing |
|---|
| 6 | n/a | time information |
|---|
| 7 | n/a | |
|---|
| 8 | n/a | FUNCTIONS: |
|---|
| 9 | n/a | _getlang -- Figure out what language is being used for the locale |
|---|
| 10 | n/a | strptime -- Calculates the time struct represented by the passed-in string |
|---|
| 11 | n/a | |
|---|
| 12 | n/a | """ |
|---|
| 13 | n/a | import time |
|---|
| 14 | n/a | import locale |
|---|
| 15 | n/a | import calendar |
|---|
| 16 | n/a | from re import compile as re_compile |
|---|
| 17 | n/a | from re import IGNORECASE |
|---|
| 18 | n/a | from re import escape as re_escape |
|---|
| 19 | n/a | from datetime import (date as datetime_date, |
|---|
| 20 | n/a | timedelta as datetime_timedelta, |
|---|
| 21 | n/a | timezone as datetime_timezone) |
|---|
| 22 | n/a | try: |
|---|
| 23 | n/a | from _thread import allocate_lock as _thread_allocate_lock |
|---|
| 24 | n/a | except ImportError: |
|---|
| 25 | n/a | from _dummy_thread import allocate_lock as _thread_allocate_lock |
|---|
| 26 | n/a | |
|---|
| 27 | n/a | __all__ = [] |
|---|
| 28 | n/a | |
|---|
| 29 | n/a | def _getlang(): |
|---|
| 30 | n/a | # Figure out what the current language is set to. |
|---|
| 31 | n/a | return locale.getlocale(locale.LC_TIME) |
|---|
| 32 | n/a | |
|---|
| 33 | n/a | class LocaleTime(object): |
|---|
| 34 | n/a | """Stores and handles locale-specific information related to time. |
|---|
| 35 | n/a | |
|---|
| 36 | n/a | ATTRIBUTES: |
|---|
| 37 | n/a | f_weekday -- full weekday names (7-item list) |
|---|
| 38 | n/a | a_weekday -- abbreviated weekday names (7-item list) |
|---|
| 39 | n/a | f_month -- full month names (13-item list; dummy value in [0], which |
|---|
| 40 | n/a | is added by code) |
|---|
| 41 | n/a | a_month -- abbreviated month names (13-item list, dummy value in |
|---|
| 42 | n/a | [0], which is added by code) |
|---|
| 43 | n/a | am_pm -- AM/PM representation (2-item list) |
|---|
| 44 | n/a | LC_date_time -- format string for date/time representation (string) |
|---|
| 45 | n/a | LC_date -- format string for date representation (string) |
|---|
| 46 | n/a | LC_time -- format string for time representation (string) |
|---|
| 47 | n/a | timezone -- daylight- and non-daylight-savings timezone representation |
|---|
| 48 | n/a | (2-item list of sets) |
|---|
| 49 | n/a | lang -- Language used by instance (2-item tuple) |
|---|
| 50 | n/a | """ |
|---|
| 51 | n/a | |
|---|
| 52 | n/a | def __init__(self): |
|---|
| 53 | n/a | """Set all attributes. |
|---|
| 54 | n/a | |
|---|
| 55 | n/a | Order of methods called matters for dependency reasons. |
|---|
| 56 | n/a | |
|---|
| 57 | n/a | The locale language is set at the offset and then checked again before |
|---|
| 58 | n/a | exiting. This is to make sure that the attributes were not set with a |
|---|
| 59 | n/a | mix of information from more than one locale. This would most likely |
|---|
| 60 | n/a | happen when using threads where one thread calls a locale-dependent |
|---|
| 61 | n/a | function while another thread changes the locale while the function in |
|---|
| 62 | n/a | the other thread is still running. Proper coding would call for |
|---|
| 63 | n/a | locks to prevent changing the locale while locale-dependent code is |
|---|
| 64 | n/a | running. The check here is done in case someone does not think about |
|---|
| 65 | n/a | doing this. |
|---|
| 66 | n/a | |
|---|
| 67 | n/a | Only other possible issue is if someone changed the timezone and did |
|---|
| 68 | n/a | not call tz.tzset . That is an issue for the programmer, though, |
|---|
| 69 | n/a | since changing the timezone is worthless without that call. |
|---|
| 70 | n/a | |
|---|
| 71 | n/a | """ |
|---|
| 72 | n/a | self.lang = _getlang() |
|---|
| 73 | n/a | self.__calc_weekday() |
|---|
| 74 | n/a | self.__calc_month() |
|---|
| 75 | n/a | self.__calc_am_pm() |
|---|
| 76 | n/a | self.__calc_timezone() |
|---|
| 77 | n/a | self.__calc_date_time() |
|---|
| 78 | n/a | if _getlang() != self.lang: |
|---|
| 79 | n/a | raise ValueError("locale changed during initialization") |
|---|
| 80 | n/a | if time.tzname != self.tzname or time.daylight != self.daylight: |
|---|
| 81 | n/a | raise ValueError("timezone changed during initialization") |
|---|
| 82 | n/a | |
|---|
| 83 | n/a | def __pad(self, seq, front): |
|---|
| 84 | n/a | # Add '' to seq to either the front (is True), else the back. |
|---|
| 85 | n/a | seq = list(seq) |
|---|
| 86 | n/a | if front: |
|---|
| 87 | n/a | seq.insert(0, '') |
|---|
| 88 | n/a | else: |
|---|
| 89 | n/a | seq.append('') |
|---|
| 90 | n/a | return seq |
|---|
| 91 | n/a | |
|---|
| 92 | n/a | def __calc_weekday(self): |
|---|
| 93 | n/a | # Set self.a_weekday and self.f_weekday using the calendar |
|---|
| 94 | n/a | # module. |
|---|
| 95 | n/a | a_weekday = [calendar.day_abbr[i].lower() for i in range(7)] |
|---|
| 96 | n/a | f_weekday = [calendar.day_name[i].lower() for i in range(7)] |
|---|
| 97 | n/a | self.a_weekday = a_weekday |
|---|
| 98 | n/a | self.f_weekday = f_weekday |
|---|
| 99 | n/a | |
|---|
| 100 | n/a | def __calc_month(self): |
|---|
| 101 | n/a | # Set self.f_month and self.a_month using the calendar module. |
|---|
| 102 | n/a | a_month = [calendar.month_abbr[i].lower() for i in range(13)] |
|---|
| 103 | n/a | f_month = [calendar.month_name[i].lower() for i in range(13)] |
|---|
| 104 | n/a | self.a_month = a_month |
|---|
| 105 | n/a | self.f_month = f_month |
|---|
| 106 | n/a | |
|---|
| 107 | n/a | def __calc_am_pm(self): |
|---|
| 108 | n/a | # Set self.am_pm by using time.strftime(). |
|---|
| 109 | n/a | |
|---|
| 110 | n/a | # The magic date (1999,3,17,hour,44,55,2,76,0) is not really that |
|---|
| 111 | n/a | # magical; just happened to have used it everywhere else where a |
|---|
| 112 | n/a | # static date was needed. |
|---|
| 113 | n/a | am_pm = [] |
|---|
| 114 | n/a | for hour in (1, 22): |
|---|
| 115 | n/a | time_tuple = time.struct_time((1999,3,17,hour,44,55,2,76,0)) |
|---|
| 116 | n/a | am_pm.append(time.strftime("%p", time_tuple).lower()) |
|---|
| 117 | n/a | self.am_pm = am_pm |
|---|
| 118 | n/a | |
|---|
| 119 | n/a | def __calc_date_time(self): |
|---|
| 120 | n/a | # Set self.date_time, self.date, & self.time by using |
|---|
| 121 | n/a | # time.strftime(). |
|---|
| 122 | n/a | |
|---|
| 123 | n/a | # Use (1999,3,17,22,44,55,2,76,0) for magic date because the amount of |
|---|
| 124 | n/a | # overloaded numbers is minimized. The order in which searches for |
|---|
| 125 | n/a | # values within the format string is very important; it eliminates |
|---|
| 126 | n/a | # possible ambiguity for what something represents. |
|---|
| 127 | n/a | time_tuple = time.struct_time((1999,3,17,22,44,55,2,76,0)) |
|---|
| 128 | n/a | date_time = [None, None, None] |
|---|
| 129 | n/a | date_time[0] = time.strftime("%c", time_tuple).lower() |
|---|
| 130 | n/a | date_time[1] = time.strftime("%x", time_tuple).lower() |
|---|
| 131 | n/a | date_time[2] = time.strftime("%X", time_tuple).lower() |
|---|
| 132 | n/a | replacement_pairs = [('%', '%%'), (self.f_weekday[2], '%A'), |
|---|
| 133 | n/a | (self.f_month[3], '%B'), (self.a_weekday[2], '%a'), |
|---|
| 134 | n/a | (self.a_month[3], '%b'), (self.am_pm[1], '%p'), |
|---|
| 135 | n/a | ('1999', '%Y'), ('99', '%y'), ('22', '%H'), |
|---|
| 136 | n/a | ('44', '%M'), ('55', '%S'), ('76', '%j'), |
|---|
| 137 | n/a | ('17', '%d'), ('03', '%m'), ('3', '%m'), |
|---|
| 138 | n/a | # '3' needed for when no leading zero. |
|---|
| 139 | n/a | ('2', '%w'), ('10', '%I')] |
|---|
| 140 | n/a | replacement_pairs.extend([(tz, "%Z") for tz_values in self.timezone |
|---|
| 141 | n/a | for tz in tz_values]) |
|---|
| 142 | n/a | for offset,directive in ((0,'%c'), (1,'%x'), (2,'%X')): |
|---|
| 143 | n/a | current_format = date_time[offset] |
|---|
| 144 | n/a | for old, new in replacement_pairs: |
|---|
| 145 | n/a | # Must deal with possible lack of locale info |
|---|
| 146 | n/a | # manifesting itself as the empty string (e.g., Swedish's |
|---|
| 147 | n/a | # lack of AM/PM info) or a platform returning a tuple of empty |
|---|
| 148 | n/a | # strings (e.g., MacOS 9 having timezone as ('','')). |
|---|
| 149 | n/a | if old: |
|---|
| 150 | n/a | current_format = current_format.replace(old, new) |
|---|
| 151 | n/a | # If %W is used, then Sunday, 2005-01-03 will fall on week 0 since |
|---|
| 152 | n/a | # 2005-01-03 occurs before the first Monday of the year. Otherwise |
|---|
| 153 | n/a | # %U is used. |
|---|
| 154 | n/a | time_tuple = time.struct_time((1999,1,3,1,1,1,6,3,0)) |
|---|
| 155 | n/a | if '00' in time.strftime(directive, time_tuple): |
|---|
| 156 | n/a | U_W = '%W' |
|---|
| 157 | n/a | else: |
|---|
| 158 | n/a | U_W = '%U' |
|---|
| 159 | n/a | date_time[offset] = current_format.replace('11', U_W) |
|---|
| 160 | n/a | self.LC_date_time = date_time[0] |
|---|
| 161 | n/a | self.LC_date = date_time[1] |
|---|
| 162 | n/a | self.LC_time = date_time[2] |
|---|
| 163 | n/a | |
|---|
| 164 | n/a | def __calc_timezone(self): |
|---|
| 165 | n/a | # Set self.timezone by using time.tzname. |
|---|
| 166 | n/a | # Do not worry about possibility of time.tzname[0] == time.tzname[1] |
|---|
| 167 | n/a | # and time.daylight; handle that in strptime. |
|---|
| 168 | n/a | try: |
|---|
| 169 | n/a | time.tzset() |
|---|
| 170 | n/a | except AttributeError: |
|---|
| 171 | n/a | pass |
|---|
| 172 | n/a | self.tzname = time.tzname |
|---|
| 173 | n/a | self.daylight = time.daylight |
|---|
| 174 | n/a | no_saving = frozenset({"utc", "gmt", self.tzname[0].lower()}) |
|---|
| 175 | n/a | if self.daylight: |
|---|
| 176 | n/a | has_saving = frozenset({self.tzname[1].lower()}) |
|---|
| 177 | n/a | else: |
|---|
| 178 | n/a | has_saving = frozenset() |
|---|
| 179 | n/a | self.timezone = (no_saving, has_saving) |
|---|
| 180 | n/a | |
|---|
| 181 | n/a | |
|---|
| 182 | n/a | class TimeRE(dict): |
|---|
| 183 | n/a | """Handle conversion from format directives to regexes.""" |
|---|
| 184 | n/a | |
|---|
| 185 | n/a | def __init__(self, locale_time=None): |
|---|
| 186 | n/a | """Create keys/values. |
|---|
| 187 | n/a | |
|---|
| 188 | n/a | Order of execution is important for dependency reasons. |
|---|
| 189 | n/a | |
|---|
| 190 | n/a | """ |
|---|
| 191 | n/a | if locale_time: |
|---|
| 192 | n/a | self.locale_time = locale_time |
|---|
| 193 | n/a | else: |
|---|
| 194 | n/a | self.locale_time = LocaleTime() |
|---|
| 195 | n/a | base = super() |
|---|
| 196 | n/a | base.__init__({ |
|---|
| 197 | n/a | # The " \d" part of the regex is to make %c from ANSI C work |
|---|
| 198 | n/a | 'd': r"(?P<d>3[0-1]|[1-2]\d|0[1-9]|[1-9]| [1-9])", |
|---|
| 199 | n/a | 'f': r"(?P<f>[0-9]{1,6})", |
|---|
| 200 | n/a | 'H': r"(?P<H>2[0-3]|[0-1]\d|\d)", |
|---|
| 201 | n/a | 'I': r"(?P<I>1[0-2]|0[1-9]|[1-9])", |
|---|
| 202 | n/a | 'G': r"(?P<G>\d\d\d\d)", |
|---|
| 203 | n/a | '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])", |
|---|
| 204 | n/a | 'm': r"(?P<m>1[0-2]|0[1-9]|[1-9])", |
|---|
| 205 | n/a | 'M': r"(?P<M>[0-5]\d|\d)", |
|---|
| 206 | n/a | 'S': r"(?P<S>6[0-1]|[0-5]\d|\d)", |
|---|
| 207 | n/a | 'U': r"(?P<U>5[0-3]|[0-4]\d|\d)", |
|---|
| 208 | n/a | 'w': r"(?P<w>[0-6])", |
|---|
| 209 | n/a | 'u': r"(?P<u>[1-7])", |
|---|
| 210 | n/a | 'V': r"(?P<V>5[0-3]|0[1-9]|[1-4]\d|\d)", |
|---|
| 211 | n/a | # W is set below by using 'U' |
|---|
| 212 | n/a | 'y': r"(?P<y>\d\d)", |
|---|
| 213 | n/a | #XXX: Does 'Y' need to worry about having less or more than |
|---|
| 214 | n/a | # 4 digits? |
|---|
| 215 | n/a | 'Y': r"(?P<Y>\d\d\d\d)", |
|---|
| 216 | n/a | 'z': r"(?P<z>[+-]\d\d[0-5]\d)", |
|---|
| 217 | n/a | 'A': self.__seqToRE(self.locale_time.f_weekday, 'A'), |
|---|
| 218 | n/a | 'a': self.__seqToRE(self.locale_time.a_weekday, 'a'), |
|---|
| 219 | n/a | 'B': self.__seqToRE(self.locale_time.f_month[1:], 'B'), |
|---|
| 220 | n/a | 'b': self.__seqToRE(self.locale_time.a_month[1:], 'b'), |
|---|
| 221 | n/a | 'p': self.__seqToRE(self.locale_time.am_pm, 'p'), |
|---|
| 222 | n/a | 'Z': self.__seqToRE((tz for tz_names in self.locale_time.timezone |
|---|
| 223 | n/a | for tz in tz_names), |
|---|
| 224 | n/a | 'Z'), |
|---|
| 225 | n/a | '%': '%'}) |
|---|
| 226 | n/a | base.__setitem__('W', base.__getitem__('U').replace('U', 'W')) |
|---|
| 227 | n/a | base.__setitem__('c', self.pattern(self.locale_time.LC_date_time)) |
|---|
| 228 | n/a | base.__setitem__('x', self.pattern(self.locale_time.LC_date)) |
|---|
| 229 | n/a | base.__setitem__('X', self.pattern(self.locale_time.LC_time)) |
|---|
| 230 | n/a | |
|---|
| 231 | n/a | def __seqToRE(self, to_convert, directive): |
|---|
| 232 | n/a | """Convert a list to a regex string for matching a directive. |
|---|
| 233 | n/a | |
|---|
| 234 | n/a | Want possible matching values to be from longest to shortest. This |
|---|
| 235 | n/a | prevents the possibility of a match occurring for a value that also |
|---|
| 236 | n/a | a substring of a larger value that should have matched (e.g., 'abc' |
|---|
| 237 | n/a | matching when 'abcdef' should have been the match). |
|---|
| 238 | n/a | |
|---|
| 239 | n/a | """ |
|---|
| 240 | n/a | to_convert = sorted(to_convert, key=len, reverse=True) |
|---|
| 241 | n/a | for value in to_convert: |
|---|
| 242 | n/a | if value != '': |
|---|
| 243 | n/a | break |
|---|
| 244 | n/a | else: |
|---|
| 245 | n/a | return '' |
|---|
| 246 | n/a | regex = '|'.join(re_escape(stuff) for stuff in to_convert) |
|---|
| 247 | n/a | regex = '(?P<%s>%s' % (directive, regex) |
|---|
| 248 | n/a | return '%s)' % regex |
|---|
| 249 | n/a | |
|---|
| 250 | n/a | def pattern(self, format): |
|---|
| 251 | n/a | """Return regex pattern for the format string. |
|---|
| 252 | n/a | |
|---|
| 253 | n/a | Need to make sure that any characters that might be interpreted as |
|---|
| 254 | n/a | regex syntax are escaped. |
|---|
| 255 | n/a | |
|---|
| 256 | n/a | """ |
|---|
| 257 | n/a | processed_format = '' |
|---|
| 258 | n/a | # The sub() call escapes all characters that might be misconstrued |
|---|
| 259 | n/a | # as regex syntax. Cannot use re.escape since we have to deal with |
|---|
| 260 | n/a | # format directives (%m, etc.). |
|---|
| 261 | n/a | regex_chars = re_compile(r"([\\.^$*+?\(\){}\[\]|])") |
|---|
| 262 | n/a | format = regex_chars.sub(r"\\\1", format) |
|---|
| 263 | n/a | whitespace_replacement = re_compile(r'\s+') |
|---|
| 264 | n/a | format = whitespace_replacement.sub(r'\\s+', format) |
|---|
| 265 | n/a | while '%' in format: |
|---|
| 266 | n/a | directive_index = format.index('%')+1 |
|---|
| 267 | n/a | processed_format = "%s%s%s" % (processed_format, |
|---|
| 268 | n/a | format[:directive_index-1], |
|---|
| 269 | n/a | self[format[directive_index]]) |
|---|
| 270 | n/a | format = format[directive_index+1:] |
|---|
| 271 | n/a | return "%s%s" % (processed_format, format) |
|---|
| 272 | n/a | |
|---|
| 273 | n/a | def compile(self, format): |
|---|
| 274 | n/a | """Return a compiled re object for the format string.""" |
|---|
| 275 | n/a | return re_compile(self.pattern(format), IGNORECASE) |
|---|
| 276 | n/a | |
|---|
| 277 | n/a | _cache_lock = _thread_allocate_lock() |
|---|
| 278 | n/a | # DO NOT modify _TimeRE_cache or _regex_cache without acquiring the cache lock |
|---|
| 279 | n/a | # first! |
|---|
| 280 | n/a | _TimeRE_cache = TimeRE() |
|---|
| 281 | n/a | _CACHE_MAX_SIZE = 5 # Max number of regexes stored in _regex_cache |
|---|
| 282 | n/a | _regex_cache = {} |
|---|
| 283 | n/a | |
|---|
| 284 | n/a | def _calc_julian_from_U_or_W(year, week_of_year, day_of_week, week_starts_Mon): |
|---|
| 285 | n/a | """Calculate the Julian day based on the year, week of the year, and day of |
|---|
| 286 | n/a | the week, with week_start_day representing whether the week of the year |
|---|
| 287 | n/a | assumes the week starts on Sunday or Monday (6 or 0).""" |
|---|
| 288 | n/a | first_weekday = datetime_date(year, 1, 1).weekday() |
|---|
| 289 | n/a | # If we are dealing with the %U directive (week starts on Sunday), it's |
|---|
| 290 | n/a | # easier to just shift the view to Sunday being the first day of the |
|---|
| 291 | n/a | # week. |
|---|
| 292 | n/a | if not week_starts_Mon: |
|---|
| 293 | n/a | first_weekday = (first_weekday + 1) % 7 |
|---|
| 294 | n/a | day_of_week = (day_of_week + 1) % 7 |
|---|
| 295 | n/a | # Need to watch out for a week 0 (when the first day of the year is not |
|---|
| 296 | n/a | # the same as that specified by %U or %W). |
|---|
| 297 | n/a | week_0_length = (7 - first_weekday) % 7 |
|---|
| 298 | n/a | if week_of_year == 0: |
|---|
| 299 | n/a | return 1 + day_of_week - first_weekday |
|---|
| 300 | n/a | else: |
|---|
| 301 | n/a | days_to_week = week_0_length + (7 * (week_of_year - 1)) |
|---|
| 302 | n/a | return 1 + days_to_week + day_of_week |
|---|
| 303 | n/a | |
|---|
| 304 | n/a | |
|---|
| 305 | n/a | def _calc_julian_from_V(iso_year, iso_week, iso_weekday): |
|---|
| 306 | n/a | """Calculate the Julian day based on the ISO 8601 year, week, and weekday. |
|---|
| 307 | n/a | ISO weeks start on Mondays, with week 01 being the week containing 4 Jan. |
|---|
| 308 | n/a | ISO week days range from 1 (Monday) to 7 (Sunday). |
|---|
| 309 | n/a | """ |
|---|
| 310 | n/a | correction = datetime_date(iso_year, 1, 4).isoweekday() + 3 |
|---|
| 311 | n/a | ordinal = (iso_week * 7) + iso_weekday - correction |
|---|
| 312 | n/a | # ordinal may be negative or 0 now, which means the date is in the previous |
|---|
| 313 | n/a | # calendar year |
|---|
| 314 | n/a | if ordinal < 1: |
|---|
| 315 | n/a | ordinal += datetime_date(iso_year, 1, 1).toordinal() |
|---|
| 316 | n/a | iso_year -= 1 |
|---|
| 317 | n/a | ordinal -= datetime_date(iso_year, 1, 1).toordinal() |
|---|
| 318 | n/a | return iso_year, ordinal |
|---|
| 319 | n/a | |
|---|
| 320 | n/a | |
|---|
| 321 | n/a | def _strptime(data_string, format="%a %b %d %H:%M:%S %Y"): |
|---|
| 322 | n/a | """Return a 2-tuple consisting of a time struct and an int containing |
|---|
| 323 | n/a | the number of microseconds based on the input string and the |
|---|
| 324 | n/a | format string.""" |
|---|
| 325 | n/a | |
|---|
| 326 | n/a | for index, arg in enumerate([data_string, format]): |
|---|
| 327 | n/a | if not isinstance(arg, str): |
|---|
| 328 | n/a | msg = "strptime() argument {} must be str, not {}" |
|---|
| 329 | n/a | raise TypeError(msg.format(index, type(arg))) |
|---|
| 330 | n/a | |
|---|
| 331 | n/a | global _TimeRE_cache, _regex_cache |
|---|
| 332 | n/a | with _cache_lock: |
|---|
| 333 | n/a | locale_time = _TimeRE_cache.locale_time |
|---|
| 334 | n/a | if (_getlang() != locale_time.lang or |
|---|
| 335 | n/a | time.tzname != locale_time.tzname or |
|---|
| 336 | n/a | time.daylight != locale_time.daylight): |
|---|
| 337 | n/a | _TimeRE_cache = TimeRE() |
|---|
| 338 | n/a | _regex_cache.clear() |
|---|
| 339 | n/a | locale_time = _TimeRE_cache.locale_time |
|---|
| 340 | n/a | if len(_regex_cache) > _CACHE_MAX_SIZE: |
|---|
| 341 | n/a | _regex_cache.clear() |
|---|
| 342 | n/a | format_regex = _regex_cache.get(format) |
|---|
| 343 | n/a | if not format_regex: |
|---|
| 344 | n/a | try: |
|---|
| 345 | n/a | format_regex = _TimeRE_cache.compile(format) |
|---|
| 346 | n/a | # KeyError raised when a bad format is found; can be specified as |
|---|
| 347 | n/a | # \\, in which case it was a stray % but with a space after it |
|---|
| 348 | n/a | except KeyError as err: |
|---|
| 349 | n/a | bad_directive = err.args[0] |
|---|
| 350 | n/a | if bad_directive == "\\": |
|---|
| 351 | n/a | bad_directive = "%" |
|---|
| 352 | n/a | del err |
|---|
| 353 | n/a | raise ValueError("'%s' is a bad directive in format '%s'" % |
|---|
| 354 | n/a | (bad_directive, format)) from None |
|---|
| 355 | n/a | # IndexError only occurs when the format string is "%" |
|---|
| 356 | n/a | except IndexError: |
|---|
| 357 | n/a | raise ValueError("stray %% in format '%s'" % format) from None |
|---|
| 358 | n/a | _regex_cache[format] = format_regex |
|---|
| 359 | n/a | found = format_regex.match(data_string) |
|---|
| 360 | n/a | if not found: |
|---|
| 361 | n/a | raise ValueError("time data %r does not match format %r" % |
|---|
| 362 | n/a | (data_string, format)) |
|---|
| 363 | n/a | if len(data_string) != found.end(): |
|---|
| 364 | n/a | raise ValueError("unconverted data remains: %s" % |
|---|
| 365 | n/a | data_string[found.end():]) |
|---|
| 366 | n/a | |
|---|
| 367 | n/a | iso_year = year = None |
|---|
| 368 | n/a | month = day = 1 |
|---|
| 369 | n/a | hour = minute = second = fraction = 0 |
|---|
| 370 | n/a | tz = -1 |
|---|
| 371 | n/a | tzoffset = None |
|---|
| 372 | n/a | # Default to -1 to signify that values not known; not critical to have, |
|---|
| 373 | n/a | # though |
|---|
| 374 | n/a | iso_week = week_of_year = None |
|---|
| 375 | n/a | week_of_year_start = None |
|---|
| 376 | n/a | # weekday and julian defaulted to None so as to signal need to calculate |
|---|
| 377 | n/a | # values |
|---|
| 378 | n/a | weekday = julian = None |
|---|
| 379 | n/a | found_dict = found.groupdict() |
|---|
| 380 | n/a | for group_key in found_dict.keys(): |
|---|
| 381 | n/a | # Directives not explicitly handled below: |
|---|
| 382 | n/a | # c, x, X |
|---|
| 383 | n/a | # handled by making out of other directives |
|---|
| 384 | n/a | # U, W |
|---|
| 385 | n/a | # worthless without day of the week |
|---|
| 386 | n/a | if group_key == 'y': |
|---|
| 387 | n/a | year = int(found_dict['y']) |
|---|
| 388 | n/a | # Open Group specification for strptime() states that a %y |
|---|
| 389 | n/a | #value in the range of [00, 68] is in the century 2000, while |
|---|
| 390 | n/a | #[69,99] is in the century 1900 |
|---|
| 391 | n/a | if year <= 68: |
|---|
| 392 | n/a | year += 2000 |
|---|
| 393 | n/a | else: |
|---|
| 394 | n/a | year += 1900 |
|---|
| 395 | n/a | elif group_key == 'Y': |
|---|
| 396 | n/a | year = int(found_dict['Y']) |
|---|
| 397 | n/a | elif group_key == 'G': |
|---|
| 398 | n/a | iso_year = int(found_dict['G']) |
|---|
| 399 | n/a | elif group_key == 'm': |
|---|
| 400 | n/a | month = int(found_dict['m']) |
|---|
| 401 | n/a | elif group_key == 'B': |
|---|
| 402 | n/a | month = locale_time.f_month.index(found_dict['B'].lower()) |
|---|
| 403 | n/a | elif group_key == 'b': |
|---|
| 404 | n/a | month = locale_time.a_month.index(found_dict['b'].lower()) |
|---|
| 405 | n/a | elif group_key == 'd': |
|---|
| 406 | n/a | day = int(found_dict['d']) |
|---|
| 407 | n/a | elif group_key == 'H': |
|---|
| 408 | n/a | hour = int(found_dict['H']) |
|---|
| 409 | n/a | elif group_key == 'I': |
|---|
| 410 | n/a | hour = int(found_dict['I']) |
|---|
| 411 | n/a | ampm = found_dict.get('p', '').lower() |
|---|
| 412 | n/a | # If there was no AM/PM indicator, we'll treat this like AM |
|---|
| 413 | n/a | if ampm in ('', locale_time.am_pm[0]): |
|---|
| 414 | n/a | # We're in AM so the hour is correct unless we're |
|---|
| 415 | n/a | # looking at 12 midnight. |
|---|
| 416 | n/a | # 12 midnight == 12 AM == hour 0 |
|---|
| 417 | n/a | if hour == 12: |
|---|
| 418 | n/a | hour = 0 |
|---|
| 419 | n/a | elif ampm == locale_time.am_pm[1]: |
|---|
| 420 | n/a | # We're in PM so we need to add 12 to the hour unless |
|---|
| 421 | n/a | # we're looking at 12 noon. |
|---|
| 422 | n/a | # 12 noon == 12 PM == hour 12 |
|---|
| 423 | n/a | if hour != 12: |
|---|
| 424 | n/a | hour += 12 |
|---|
| 425 | n/a | elif group_key == 'M': |
|---|
| 426 | n/a | minute = int(found_dict['M']) |
|---|
| 427 | n/a | elif group_key == 'S': |
|---|
| 428 | n/a | second = int(found_dict['S']) |
|---|
| 429 | n/a | elif group_key == 'f': |
|---|
| 430 | n/a | s = found_dict['f'] |
|---|
| 431 | n/a | # Pad to always return microseconds. |
|---|
| 432 | n/a | s += "0" * (6 - len(s)) |
|---|
| 433 | n/a | fraction = int(s) |
|---|
| 434 | n/a | elif group_key == 'A': |
|---|
| 435 | n/a | weekday = locale_time.f_weekday.index(found_dict['A'].lower()) |
|---|
| 436 | n/a | elif group_key == 'a': |
|---|
| 437 | n/a | weekday = locale_time.a_weekday.index(found_dict['a'].lower()) |
|---|
| 438 | n/a | elif group_key == 'w': |
|---|
| 439 | n/a | weekday = int(found_dict['w']) |
|---|
| 440 | n/a | if weekday == 0: |
|---|
| 441 | n/a | weekday = 6 |
|---|
| 442 | n/a | else: |
|---|
| 443 | n/a | weekday -= 1 |
|---|
| 444 | n/a | elif group_key == 'u': |
|---|
| 445 | n/a | weekday = int(found_dict['u']) |
|---|
| 446 | n/a | weekday -= 1 |
|---|
| 447 | n/a | elif group_key == 'j': |
|---|
| 448 | n/a | julian = int(found_dict['j']) |
|---|
| 449 | n/a | elif group_key in ('U', 'W'): |
|---|
| 450 | n/a | week_of_year = int(found_dict[group_key]) |
|---|
| 451 | n/a | if group_key == 'U': |
|---|
| 452 | n/a | # U starts week on Sunday. |
|---|
| 453 | n/a | week_of_year_start = 6 |
|---|
| 454 | n/a | else: |
|---|
| 455 | n/a | # W starts week on Monday. |
|---|
| 456 | n/a | week_of_year_start = 0 |
|---|
| 457 | n/a | elif group_key == 'V': |
|---|
| 458 | n/a | iso_week = int(found_dict['V']) |
|---|
| 459 | n/a | elif group_key == 'z': |
|---|
| 460 | n/a | z = found_dict['z'] |
|---|
| 461 | n/a | tzoffset = int(z[1:3]) * 60 + int(z[3:5]) |
|---|
| 462 | n/a | if z.startswith("-"): |
|---|
| 463 | n/a | tzoffset = -tzoffset |
|---|
| 464 | n/a | elif group_key == 'Z': |
|---|
| 465 | n/a | # Since -1 is default value only need to worry about setting tz if |
|---|
| 466 | n/a | # it can be something other than -1. |
|---|
| 467 | n/a | found_zone = found_dict['Z'].lower() |
|---|
| 468 | n/a | for value, tz_values in enumerate(locale_time.timezone): |
|---|
| 469 | n/a | if found_zone in tz_values: |
|---|
| 470 | n/a | # Deal with bad locale setup where timezone names are the |
|---|
| 471 | n/a | # same and yet time.daylight is true; too ambiguous to |
|---|
| 472 | n/a | # be able to tell what timezone has daylight savings |
|---|
| 473 | n/a | if (time.tzname[0] == time.tzname[1] and |
|---|
| 474 | n/a | time.daylight and found_zone not in ("utc", "gmt")): |
|---|
| 475 | n/a | break |
|---|
| 476 | n/a | else: |
|---|
| 477 | n/a | tz = value |
|---|
| 478 | n/a | break |
|---|
| 479 | n/a | # Deal with the cases where ambiguities arize |
|---|
| 480 | n/a | # don't assume default values for ISO week/year |
|---|
| 481 | n/a | if year is None and iso_year is not None: |
|---|
| 482 | n/a | if iso_week is None or weekday is None: |
|---|
| 483 | n/a | raise ValueError("ISO year directive '%G' must be used with " |
|---|
| 484 | n/a | "the ISO week directive '%V' and a weekday " |
|---|
| 485 | n/a | "directive ('%A', '%a', '%w', or '%u').") |
|---|
| 486 | n/a | if julian is not None: |
|---|
| 487 | n/a | raise ValueError("Day of the year directive '%j' is not " |
|---|
| 488 | n/a | "compatible with ISO year directive '%G'. " |
|---|
| 489 | n/a | "Use '%Y' instead.") |
|---|
| 490 | n/a | elif week_of_year is None and iso_week is not None: |
|---|
| 491 | n/a | if weekday is None: |
|---|
| 492 | n/a | raise ValueError("ISO week directive '%V' must be used with " |
|---|
| 493 | n/a | "the ISO year directive '%G' and a weekday " |
|---|
| 494 | n/a | "directive ('%A', '%a', '%w', or '%u').") |
|---|
| 495 | n/a | else: |
|---|
| 496 | n/a | raise ValueError("ISO week directive '%V' is incompatible with " |
|---|
| 497 | n/a | "the year directive '%Y'. Use the ISO year '%G' " |
|---|
| 498 | n/a | "instead.") |
|---|
| 499 | n/a | |
|---|
| 500 | n/a | leap_year_fix = False |
|---|
| 501 | n/a | if year is None and month == 2 and day == 29: |
|---|
| 502 | n/a | year = 1904 # 1904 is first leap year of 20th century |
|---|
| 503 | n/a | leap_year_fix = True |
|---|
| 504 | n/a | elif year is None: |
|---|
| 505 | n/a | year = 1900 |
|---|
| 506 | n/a | |
|---|
| 507 | n/a | |
|---|
| 508 | n/a | # If we know the week of the year and what day of that week, we can figure |
|---|
| 509 | n/a | # out the Julian day of the year. |
|---|
| 510 | n/a | if julian is None and weekday is not None: |
|---|
| 511 | n/a | if week_of_year is not None: |
|---|
| 512 | n/a | week_starts_Mon = True if week_of_year_start == 0 else False |
|---|
| 513 | n/a | julian = _calc_julian_from_U_or_W(year, week_of_year, weekday, |
|---|
| 514 | n/a | week_starts_Mon) |
|---|
| 515 | n/a | elif iso_year is not None and iso_week is not None: |
|---|
| 516 | n/a | year, julian = _calc_julian_from_V(iso_year, iso_week, weekday + 1) |
|---|
| 517 | n/a | if julian is not None and julian <= 0: |
|---|
| 518 | n/a | year -= 1 |
|---|
| 519 | n/a | yday = 366 if calendar.isleap(year) else 365 |
|---|
| 520 | n/a | julian += yday |
|---|
| 521 | n/a | |
|---|
| 522 | n/a | if julian is None: |
|---|
| 523 | n/a | # Cannot pre-calculate datetime_date() since can change in Julian |
|---|
| 524 | n/a | # calculation and thus could have different value for the day of |
|---|
| 525 | n/a | # the week calculation. |
|---|
| 526 | n/a | # Need to add 1 to result since first day of the year is 1, not 0. |
|---|
| 527 | n/a | julian = datetime_date(year, month, day).toordinal() - \ |
|---|
| 528 | n/a | datetime_date(year, 1, 1).toordinal() + 1 |
|---|
| 529 | n/a | else: # Assume that if they bothered to include Julian day (or if it was |
|---|
| 530 | n/a | # calculated above with year/week/weekday) it will be accurate. |
|---|
| 531 | n/a | datetime_result = datetime_date.fromordinal( |
|---|
| 532 | n/a | (julian - 1) + |
|---|
| 533 | n/a | datetime_date(year, 1, 1).toordinal()) |
|---|
| 534 | n/a | year = datetime_result.year |
|---|
| 535 | n/a | month = datetime_result.month |
|---|
| 536 | n/a | day = datetime_result.day |
|---|
| 537 | n/a | if weekday is None: |
|---|
| 538 | n/a | weekday = datetime_date(year, month, day).weekday() |
|---|
| 539 | n/a | # Add timezone info |
|---|
| 540 | n/a | tzname = found_dict.get("Z") |
|---|
| 541 | n/a | if tzoffset is not None: |
|---|
| 542 | n/a | gmtoff = tzoffset * 60 |
|---|
| 543 | n/a | else: |
|---|
| 544 | n/a | gmtoff = None |
|---|
| 545 | n/a | |
|---|
| 546 | n/a | if leap_year_fix: |
|---|
| 547 | n/a | # the caller didn't supply a year but asked for Feb 29th. We couldn't |
|---|
| 548 | n/a | # use the default of 1900 for computations. We set it back to ensure |
|---|
| 549 | n/a | # that February 29th is smaller than March 1st. |
|---|
| 550 | n/a | year = 1900 |
|---|
| 551 | n/a | |
|---|
| 552 | n/a | return (year, month, day, |
|---|
| 553 | n/a | hour, minute, second, |
|---|
| 554 | n/a | weekday, julian, tz, tzname, gmtoff), fraction |
|---|
| 555 | n/a | |
|---|
| 556 | n/a | def _strptime_time(data_string, format="%a %b %d %H:%M:%S %Y"): |
|---|
| 557 | n/a | """Return a time struct based on the input string and the |
|---|
| 558 | n/a | format string.""" |
|---|
| 559 | n/a | tt = _strptime(data_string, format)[0] |
|---|
| 560 | n/a | return time.struct_time(tt[:time._STRUCT_TM_ITEMS]) |
|---|
| 561 | n/a | |
|---|
| 562 | n/a | def _strptime_datetime(cls, data_string, format="%a %b %d %H:%M:%S %Y"): |
|---|
| 563 | n/a | """Return a class cls instance based on the input string and the |
|---|
| 564 | n/a | format string.""" |
|---|
| 565 | n/a | tt, fraction = _strptime(data_string, format) |
|---|
| 566 | n/a | tzname, gmtoff = tt[-2:] |
|---|
| 567 | n/a | args = tt[:6] + (fraction,) |
|---|
| 568 | n/a | if gmtoff is not None: |
|---|
| 569 | n/a | tzdelta = datetime_timedelta(seconds=gmtoff) |
|---|
| 570 | n/a | if tzname: |
|---|
| 571 | n/a | tz = datetime_timezone(tzdelta, tzname) |
|---|
| 572 | n/a | else: |
|---|
| 573 | n/a | tz = datetime_timezone(tzdelta) |
|---|
| 574 | n/a | args += (tz,) |
|---|
| 575 | n/a | |
|---|
| 576 | n/a | return cls(*args) |
|---|