| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 1 | """Strptime-related classes and functions. | 
|  | 2 |  | 
|  | 3 | CLASSES: | 
|  | 4 | LocaleTime -- Discovers and/or stores locale-specific time information | 
| Barry Warsaw | 4d895fa | 2002-09-23 22:46:49 +0000 | [diff] [blame] | 5 | TimeRE -- Creates regexes for pattern matching a string of text containing | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 6 | time information as is returned by time.strftime() | 
|  | 7 |  | 
|  | 8 | FUNCTIONS: | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 9 | _getlang -- Figure out what language is being used for the locale | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 10 | strptime -- Calculates the time struct represented by the passed-in string | 
|  | 11 |  | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 12 | Requires Python 2.2.1 or higher (mainly because of the use of property()). | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 13 | Can be used in Python 2.2 if the following line is added: | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 14 | True = 1; False = 0 | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 15 | """ | 
|  | 16 | import time | 
|  | 17 | import locale | 
|  | 18 | import calendar | 
|  | 19 | from re import compile as re_compile | 
|  | 20 | from re import IGNORECASE | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 21 | from datetime import date as datetime_date | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 22 |  | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 23 | __author__ = "Brett Cannon" | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 24 | __email__ = "brett@python.org" | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 25 |  | 
|  | 26 | __all__ = ['strptime'] | 
|  | 27 |  | 
| Tim Peters | 80cebc1 | 2003-01-19 04:40:44 +0000 | [diff] [blame] | 28 | def _getlang(): | 
|  | 29 | # Figure out what the current language is set to. | 
|  | 30 | current_lang = locale.getlocale(locale.LC_TIME)[0] | 
|  | 31 | if current_lang: | 
|  | 32 | return current_lang | 
|  | 33 | else: | 
|  | 34 | current_lang = locale.getdefaultlocale()[0] | 
|  | 35 | if current_lang: | 
|  | 36 | return current_lang | 
|  | 37 | else: | 
|  | 38 | return '' | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 39 |  | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 40 | class LocaleTime(object): | 
|  | 41 | """Stores and handles locale-specific information related to time. | 
|  | 42 |  | 
|  | 43 | ATTRIBUTES (all read-only after instance creation! Instance variables that | 
|  | 44 | store the values have mangled names): | 
|  | 45 | f_weekday -- full weekday names (7-item list) | 
|  | 46 | a_weekday -- abbreviated weekday names (7-item list) | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 47 | f_month -- full weekday names (14-item list; dummy value in [0], which | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 48 | is added by code) | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 49 | a_month -- abbreviated weekday names (13-item list, dummy value in | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 50 | [0], which is added by code) | 
|  | 51 | am_pm -- AM/PM representation (2-item list) | 
|  | 52 | LC_date_time -- format string for date/time representation (string) | 
|  | 53 | LC_date -- format string for date representation (string) | 
|  | 54 | LC_time -- format string for time representation (string) | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 55 | timezone -- daylight- and non-daylight-savings timezone representation | 
|  | 56 | (3-item list; code tacks on blank item at end for | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 57 | possible lack of timezone such as UTC) | 
|  | 58 | lang -- Language used by instance (string) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 59 | """ | 
|  | 60 |  | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 61 | def __init__(self, f_weekday=None, a_weekday=None, f_month=None, | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 62 | a_month=None, am_pm=None, LC_date_time=None, LC_time=None, | 
|  | 63 | LC_date=None, timezone=None, lang=None): | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 64 | """Optionally set attributes with passed-in values.""" | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 65 | if f_weekday is None: | 
|  | 66 | self.__f_weekday = None | 
|  | 67 | elif len(f_weekday) == 7: | 
|  | 68 | self.__f_weekday = list(f_weekday) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 69 | else: | 
|  | 70 | raise TypeError("full weekday names must be a 7-item sequence") | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 71 | if a_weekday is None: | 
|  | 72 | self.__a_weekday = None | 
|  | 73 | elif len(a_weekday) == 7: | 
|  | 74 | self.__a_weekday = list(a_weekday) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 75 | else: | 
|  | 76 | raise TypeError( | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 77 | "abbreviated weekday names must be a 7-item  sequence") | 
|  | 78 | if f_month is None: | 
|  | 79 | self.__f_month = None | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 80 | elif len(f_month) == 12: | 
|  | 81 | self.__f_month = self.__pad(f_month, True) | 
|  | 82 | else: | 
|  | 83 | raise TypeError("full month names must be a 12-item sequence") | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 84 | if a_month is None: | 
|  | 85 | self.__a_month = None | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 86 | elif len(a_month) == 12: | 
|  | 87 | self.__a_month = self.__pad(a_month, True) | 
|  | 88 | else: | 
|  | 89 | raise TypeError( | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 90 | "abbreviated month names must be a 12-item sequence") | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 91 | if am_pm is None: | 
|  | 92 | self.__am_pm = None | 
|  | 93 | elif len(am_pm) == 2: | 
|  | 94 | self.__am_pm = am_pm | 
|  | 95 | else: | 
|  | 96 | raise TypeError("AM/PM representation must be a 2-item sequence") | 
|  | 97 | self.__LC_date_time = LC_date_time | 
|  | 98 | self.__LC_time = LC_time | 
|  | 99 | self.__LC_date = LC_date | 
|  | 100 | self.__timezone = timezone | 
|  | 101 | if timezone: | 
|  | 102 | if len(timezone) != 2: | 
|  | 103 | raise TypeError("timezone names must contain 2 items") | 
|  | 104 | else: | 
|  | 105 | self.__timezone = self.__pad(timezone, False) | 
|  | 106 | self.__lang = lang | 
|  | 107 |  | 
|  | 108 | def __pad(self, seq, front): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 109 | # Add '' to seq to either front (is True), else the back. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 110 | seq = list(seq) | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 111 | if front: | 
|  | 112 | seq.insert(0, '') | 
|  | 113 | else: | 
|  | 114 | seq.append('') | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 115 | return seq | 
|  | 116 |  | 
|  | 117 | def __set_nothing(self, stuff): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 118 | # Raise TypeError when trying to set an attribute. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 119 | raise TypeError("attribute does not support assignment") | 
|  | 120 |  | 
|  | 121 | def __get_f_weekday(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 122 | # Fetch self.f_weekday. | 
|  | 123 | if not self.__f_weekday: | 
|  | 124 | self.__calc_weekday() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 125 | return self.__f_weekday | 
|  | 126 |  | 
|  | 127 | def __get_a_weekday(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 128 | # Fetch self.a_weekday. | 
|  | 129 | if not self.__a_weekday: | 
|  | 130 | self.__calc_weekday() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 131 | return self.__a_weekday | 
|  | 132 |  | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 133 | f_weekday = property(__get_f_weekday, __set_nothing, | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 134 | doc="Full weekday names") | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 135 | a_weekday = property(__get_a_weekday, __set_nothing, | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 136 | doc="Abbreviated weekday names") | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 137 |  | 
|  | 138 | def __get_f_month(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 139 | # Fetch self.f_month. | 
|  | 140 | if not self.__f_month: | 
|  | 141 | self.__calc_month() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 142 | return self.__f_month | 
|  | 143 |  | 
|  | 144 | def __get_a_month(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 145 | # Fetch self.a_month. | 
|  | 146 | if not self.__a_month: | 
|  | 147 | self.__calc_month() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 148 | return self.__a_month | 
|  | 149 |  | 
|  | 150 | f_month = property(__get_f_month, __set_nothing, | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 151 | doc="Full month names (dummy value at index 0)") | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 152 | a_month = property(__get_a_month, __set_nothing, | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 153 | doc="Abbreviated month names (dummy value at index 0)") | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 154 |  | 
|  | 155 | def __get_am_pm(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 156 | # Fetch self.am_pm. | 
|  | 157 | if not self.__am_pm: | 
|  | 158 | self.__calc_am_pm() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 159 | return self.__am_pm | 
|  | 160 |  | 
|  | 161 | am_pm = property(__get_am_pm, __set_nothing, doc="AM/PM representation") | 
|  | 162 |  | 
|  | 163 | def __get_timezone(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 164 | # Fetch self.timezone. | 
|  | 165 | if not self.__timezone: | 
|  | 166 | self.__calc_timezone() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 167 | return self.__timezone | 
|  | 168 |  | 
|  | 169 | timezone = property(__get_timezone, __set_nothing, | 
|  | 170 | doc="Timezone representation (dummy value at index 2)") | 
|  | 171 |  | 
|  | 172 | def __get_LC_date_time(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 173 | # Fetch self.LC_date_time. | 
|  | 174 | if not self.__LC_date_time: | 
|  | 175 | self.__calc_date_time() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 176 | return self.__LC_date_time | 
|  | 177 |  | 
|  | 178 | def __get_LC_date(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 179 | # Fetch self.LC_date. | 
|  | 180 | if not self.__LC_date: | 
|  | 181 | self.__calc_date_time() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 182 | return self.__LC_date | 
|  | 183 |  | 
|  | 184 | def __get_LC_time(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 185 | # Fetch self.LC_time. | 
|  | 186 | if not self.__LC_time: | 
|  | 187 | self.__calc_date_time() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 188 | return self.__LC_time | 
|  | 189 |  | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 190 | LC_date_time = property( | 
|  | 191 | __get_LC_date_time, __set_nothing, | 
|  | 192 | doc= | 
|  | 193 | "Format string for locale's date/time representation ('%c' format)") | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 194 | LC_date = property(__get_LC_date, __set_nothing, | 
|  | 195 | doc="Format string for locale's date representation ('%x' format)") | 
|  | 196 | LC_time = property(__get_LC_time, __set_nothing, | 
|  | 197 | doc="Format string for locale's time representation ('%X' format)") | 
|  | 198 |  | 
|  | 199 | def __get_lang(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 200 | # Fetch self.lang. | 
|  | 201 | if not self.__lang: | 
|  | 202 | self.__calc_lang() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 203 | return self.__lang | 
|  | 204 |  | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 205 | lang = property(__get_lang, __set_nothing, | 
|  | 206 | doc="Language used for instance") | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 207 |  | 
|  | 208 | def __calc_weekday(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 209 | # Set self.__a_weekday and self.__f_weekday using the calendar | 
|  | 210 | # module. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 211 | a_weekday = [calendar.day_abbr[i] for i in range(7)] | 
|  | 212 | f_weekday = [calendar.day_name[i] for i in range(7)] | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 213 | if not self.__a_weekday: | 
|  | 214 | self.__a_weekday = a_weekday | 
|  | 215 | if not self.__f_weekday: | 
|  | 216 | self.__f_weekday = f_weekday | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 217 |  | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 218 | def __calc_month(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 219 | # Set self.__f_month and self.__a_month using the calendar module. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 220 | a_month = [calendar.month_abbr[i] for i in range(13)] | 
|  | 221 | f_month = [calendar.month_name[i] for i in range(13)] | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 222 | if not self.__a_month: | 
|  | 223 | self.__a_month = a_month | 
|  | 224 | if not self.__f_month: | 
|  | 225 | self.__f_month = f_month | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 226 |  | 
|  | 227 | def __calc_am_pm(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 228 | # Set self.__am_pm by using time.strftime(). | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 229 |  | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 230 | # The magic date (1999,3,17,hour,44,55,2,76,0) is not really that | 
|  | 231 | # magical; just happened to have used it everywhere else where a | 
|  | 232 | # static date was needed. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 233 | am_pm = [] | 
|  | 234 | for hour in (01,22): | 
|  | 235 | time_tuple = time.struct_time((1999,3,17,hour,44,55,2,76,0)) | 
|  | 236 | am_pm.append(time.strftime("%p", time_tuple)) | 
|  | 237 | self.__am_pm = am_pm | 
|  | 238 |  | 
|  | 239 | def __calc_date_time(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 240 | # Set self.__date_time, self.__date, & self.__time by using | 
|  | 241 | # time.strftime(). | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 242 |  | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 243 | # Use (1999,3,17,22,44,55,2,76,0) for magic date because the amount of | 
|  | 244 | # overloaded numbers is minimized.  The order in which searches for | 
|  | 245 | # values within the format string is very important; it eliminates | 
|  | 246 | # possible ambiguity for what something represents. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 247 | time_tuple = time.struct_time((1999,3,17,22,44,55,2,76,0)) | 
|  | 248 | date_time = [None, None, None] | 
|  | 249 | date_time[0] = time.strftime("%c", time_tuple) | 
|  | 250 | date_time[1] = time.strftime("%x", time_tuple) | 
|  | 251 | date_time[2] = time.strftime("%X", time_tuple) | 
|  | 252 | for offset,directive in ((0,'%c'), (1,'%x'), (2,'%X')): | 
|  | 253 | current_format = date_time[offset] | 
| Barry Warsaw | 4d895fa | 2002-09-23 22:46:49 +0000 | [diff] [blame] | 254 | for old, new in ( | 
|  | 255 | ('%', '%%'), (self.f_weekday[2], '%A'), | 
|  | 256 | (self.f_month[3], '%B'), (self.a_weekday[2], '%a'), | 
|  | 257 | (self.a_month[3], '%b'), (self.am_pm[1], '%p'), | 
|  | 258 | (self.timezone[0], '%Z'), (self.timezone[1], '%Z'), | 
|  | 259 | ('1999', '%Y'), ('99', '%y'), ('22', '%H'), | 
|  | 260 | ('44', '%M'), ('55', '%S'), ('76', '%j'), | 
|  | 261 | ('17', '%d'), ('03', '%m'), ('3', '%m'), | 
|  | 262 | # '3' needed for when no leading zero. | 
|  | 263 | ('2', '%w'), ('10', '%I')): | 
| Jack Jansen | 62fe755 | 2003-01-15 22:59:39 +0000 | [diff] [blame] | 264 | # Must deal with possible lack of locale info | 
|  | 265 | # manifesting itself as the empty string (e.g., Swedish's | 
|  | 266 | # lack of AM/PM info) or a platform returning a tuple of empty | 
|  | 267 | # strings (e.g., MacOS 9 having timezone as ('','')). | 
|  | 268 | if old: | 
| Barry Warsaw | 4d895fa | 2002-09-23 22:46:49 +0000 | [diff] [blame] | 269 | current_format = current_format.replace(old, new) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 270 | time_tuple = time.struct_time((1999,1,3,1,1,1,6,3,0)) | 
|  | 271 | if time.strftime(directive, time_tuple).find('00'): | 
|  | 272 | U_W = '%U' | 
|  | 273 | else: | 
|  | 274 | U_W = '%W' | 
|  | 275 | date_time[offset] = current_format.replace('11', U_W) | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 276 | if not self.__LC_date_time: | 
|  | 277 | self.__LC_date_time = date_time[0] | 
|  | 278 | if not self.__LC_date: | 
|  | 279 | self.__LC_date = date_time[1] | 
|  | 280 | if not self.__LC_time: | 
|  | 281 | self.__LC_time = date_time[2] | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 282 |  | 
|  | 283 | def __calc_timezone(self): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 284 | # Set self.__timezone by using time.tzname. | 
|  | 285 | # | 
|  | 286 | # Empty string used for matching when timezone is not used/needed such | 
|  | 287 | # as with UTC. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 288 | self.__timezone = self.__pad(time.tzname, 0) | 
|  | 289 |  | 
|  | 290 | def __calc_lang(self): | 
| Tim Peters | 80cebc1 | 2003-01-19 04:40:44 +0000 | [diff] [blame] | 291 | # Set self.__lang by using __getlang(). | 
|  | 292 | self.__lang = _getlang() | 
|  | 293 |  | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 294 |  | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 295 |  | 
|  | 296 | class TimeRE(dict): | 
|  | 297 | """Handle conversion from format directives to regexes.""" | 
|  | 298 |  | 
|  | 299 | def __init__(self, locale_time=LocaleTime()): | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 300 | """Init inst with non-locale regexes and store LocaleTime object.""" | 
| Neal Norwitz | 5efc50d | 2002-12-30 22:23:12 +0000 | [diff] [blame] | 301 | #XXX: Does 'Y' need to worry about having less or more than 4 digits? | 
|  | 302 | base = super(TimeRE, self) | 
|  | 303 | base.__init__({ | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 304 | # The " \d" option is to make %c from ANSI C work | 
| Neal Norwitz | 5efc50d | 2002-12-30 22:23:12 +0000 | [diff] [blame] | 305 | 'd': r"(?P<d>3[0-1]|[1-2]\d|0[1-9]|[1-9]| [1-9])", | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 306 | 'H': r"(?P<H>2[0-3]|[0-1]\d|\d)", | 
| Neal Norwitz | 5efc50d | 2002-12-30 22:23:12 +0000 | [diff] [blame] | 307 | 'I': r"(?P<I>1[0-2]|0[1-9]|[1-9])", | 
|  | 308 | '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])", | 
|  | 309 | 'm': r"(?P<m>1[0-2]|0[1-9]|[1-9])", | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 310 | 'M': r"(?P<M>[0-5]\d|\d)", | 
|  | 311 | 'S': r"(?P<S>6[0-1]|[0-5]\d|\d)", | 
|  | 312 | 'U': r"(?P<U>5[0-3]|[0-4]\d|\d)", | 
|  | 313 | 'w': r"(?P<w>[0-6])", | 
| Neal Norwitz | 5efc50d | 2002-12-30 22:23:12 +0000 | [diff] [blame] | 314 | # W is set below by using 'U' | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 315 | 'y': r"(?P<y>\d\d)", | 
|  | 316 | 'Y': r"(?P<Y>\d\d\d\d)"}) | 
| Neal Norwitz | 5efc50d | 2002-12-30 22:23:12 +0000 | [diff] [blame] | 317 | base.__setitem__('W', base.__getitem__('U')) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 318 | self.locale_time = locale_time | 
|  | 319 |  | 
|  | 320 | def __getitem__(self, fetch): | 
|  | 321 | """Try to fetch regex; if it does not exist, construct it.""" | 
|  | 322 | try: | 
| Barry Warsaw | 4d895fa | 2002-09-23 22:46:49 +0000 | [diff] [blame] | 323 | return super(TimeRE, self).__getitem__(fetch) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 324 | except KeyError: | 
| Barry Warsaw | 4d895fa | 2002-09-23 22:46:49 +0000 | [diff] [blame] | 325 | constructors = { | 
|  | 326 | 'A': lambda: self.__seqToRE(self.locale_time.f_weekday, fetch), | 
|  | 327 | 'a': lambda: self.__seqToRE(self.locale_time.a_weekday, fetch), | 
|  | 328 | 'B': lambda: self.__seqToRE(self.locale_time.f_month[1:], | 
|  | 329 | fetch), | 
|  | 330 | 'b': lambda: self.__seqToRE(self.locale_time.a_month[1:], | 
|  | 331 | fetch), | 
|  | 332 | 'c': lambda: self.pattern(self.locale_time.LC_date_time), | 
|  | 333 | 'p': lambda: self.__seqToRE(self.locale_time.am_pm, fetch), | 
|  | 334 | 'x': lambda: self.pattern(self.locale_time.LC_date), | 
|  | 335 | 'X': lambda: self.pattern(self.locale_time.LC_time), | 
|  | 336 | 'Z': lambda: self.__seqToRE(self.locale_time.timezone, fetch), | 
|  | 337 | '%': lambda: '%', | 
|  | 338 | } | 
|  | 339 | if fetch in constructors: | 
|  | 340 | self[fetch] = constructors[fetch]() | 
|  | 341 | return self[fetch] | 
|  | 342 | else: | 
|  | 343 | raise | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 344 |  | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 345 | def __seqToRE(self, to_convert, directive): | 
| Jack Jansen | 62fe755 | 2003-01-15 22:59:39 +0000 | [diff] [blame] | 346 | """Convert a list to a regex string for matching a directive.""" | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 347 | def sorter(a, b): | 
|  | 348 | """Sort based on length. | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 349 |  | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 350 | Done in case for some strange reason that names in the locale only | 
|  | 351 | differ by a suffix and thus want the name with the suffix to match | 
|  | 352 | first. | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 353 | """ | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 354 | try: | 
|  | 355 | a_length = len(a) | 
|  | 356 | except TypeError: | 
|  | 357 | a_length = 0 | 
|  | 358 | try: | 
|  | 359 | b_length = len(b) | 
|  | 360 | except TypeError: | 
|  | 361 | b_length = 0 | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 362 | return cmp(b_length, a_length) | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 363 |  | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 364 | to_convert = to_convert[:]  # Don't want to change value in-place. | 
| Jack Jansen | 62fe755 | 2003-01-15 22:59:39 +0000 | [diff] [blame] | 365 | for value in to_convert: | 
|  | 366 | if value != '': | 
|  | 367 | break | 
|  | 368 | else: | 
|  | 369 | return '' | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 370 | to_convert.sort(sorter) | 
| Barry Warsaw | 4d895fa | 2002-09-23 22:46:49 +0000 | [diff] [blame] | 371 | regex = '|'.join(to_convert) | 
|  | 372 | regex = '(?P<%s>%s' % (directive, regex) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 373 | return '%s)' % regex | 
|  | 374 |  | 
|  | 375 | def pattern(self, format): | 
| Brett Cannon | 1e91d8e | 2003-04-19 04:00:56 +0000 | [diff] [blame] | 376 | """Return re pattern for the format string. | 
| Tim Peters | 0eadaac | 2003-04-24 16:02:54 +0000 | [diff] [blame] | 377 |  | 
| Brett Cannon | 1e91d8e | 2003-04-19 04:00:56 +0000 | [diff] [blame] | 378 | Need to make sure that any characters that might be interpreted as | 
|  | 379 | regex syntax is escaped. | 
| Tim Peters | 0eadaac | 2003-04-24 16:02:54 +0000 | [diff] [blame] | 380 |  | 
| Brett Cannon | 1e91d8e | 2003-04-19 04:00:56 +0000 | [diff] [blame] | 381 | """ | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 382 | processed_format = '' | 
| Brett Cannon | 1e91d8e | 2003-04-19 04:00:56 +0000 | [diff] [blame] | 383 | # The sub() call escapes all characters that might be misconstrued | 
|  | 384 | # as regex syntax. | 
|  | 385 | regex_chars = re_compile(r"([\\.^$*+?{}\[\]|])") | 
|  | 386 | format = regex_chars.sub(r"\\\1", format) | 
| Tim Peters | 80cebc1 | 2003-01-19 04:40:44 +0000 | [diff] [blame] | 387 | whitespace_replacement = re_compile('\s+') | 
|  | 388 | format = whitespace_replacement.sub('\s*', format) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 389 | while format.find('%') != -1: | 
|  | 390 | directive_index = format.index('%')+1 | 
| Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 391 | processed_format = "%s%s%s" % (processed_format, | 
| Barry Warsaw | 35816e6 | 2002-08-29 16:24:50 +0000 | [diff] [blame] | 392 | format[:directive_index-1], | 
|  | 393 | self[format[directive_index]]) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 394 | format = format[directive_index+1:] | 
|  | 395 | return "%s%s" % (processed_format, format) | 
|  | 396 |  | 
|  | 397 | def compile(self, format): | 
|  | 398 | """Return a compiled re object for the format string.""" | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 399 | return re_compile(self.pattern(format), IGNORECASE) | 
|  | 400 |  | 
| Tim Peters | 80cebc1 | 2003-01-19 04:40:44 +0000 | [diff] [blame] | 401 | # Cached TimeRE; probably only need one instance ever so cache it for performance | 
|  | 402 | _locale_cache = TimeRE() | 
|  | 403 | # Cached regex objects; same reason as for TimeRE cache | 
|  | 404 | _regex_cache = dict() | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 405 |  | 
|  | 406 | def strptime(data_string, format="%a %b %d %H:%M:%S %Y"): | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 407 | """Return a time struct based on the input data and the format string.""" | 
| Tim Peters | 80cebc1 | 2003-01-19 04:40:44 +0000 | [diff] [blame] | 408 | global _locale_cache | 
|  | 409 | global _regex_cache | 
|  | 410 | locale_time = _locale_cache.locale_time | 
|  | 411 | # If the language changes, caches are invalidated, so clear them | 
|  | 412 | if locale_time.lang != _getlang(): | 
|  | 413 | _locale_cache = TimeRE() | 
|  | 414 | _regex_cache.clear() | 
|  | 415 | format_regex = _regex_cache.get(format) | 
|  | 416 | if not format_regex: | 
|  | 417 | # Limit regex cache size to prevent major bloating of the module; | 
|  | 418 | # The value 5 is arbitrary | 
|  | 419 | if len(_regex_cache) > 5: | 
|  | 420 | _regex_cache.clear() | 
|  | 421 | format_regex = _locale_cache.compile(format) | 
|  | 422 | _regex_cache[format] = format_regex | 
|  | 423 | found = format_regex.match(data_string) | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 424 | if not found: | 
|  | 425 | raise ValueError("time data did not match format") | 
| Brett Cannon | 2b6dfec | 2003-04-28 21:30:13 +0000 | [diff] [blame] | 426 | if len(data_string) != found.end(): | 
|  | 427 | raise ValueError("unconverted data remains: %s" % | 
|  | 428 | data_string[found.end():]) | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 429 | year = 1900 | 
|  | 430 | month = day = 1 | 
|  | 431 | hour = minute = second = 0 | 
|  | 432 | tz = -1 | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 433 | # weekday and julian defaulted to -1 so as to signal need to calculate values | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 434 | weekday = julian = -1 | 
|  | 435 | found_dict = found.groupdict() | 
|  | 436 | for group_key in found_dict.iterkeys(): | 
|  | 437 | if group_key == 'y': | 
|  | 438 | year = int(found_dict['y']) | 
|  | 439 | # Open Group specification for strptime() states that a %y | 
|  | 440 | #value in the range of [00, 68] is in the century 2000, while | 
|  | 441 | #[69,99] is in the century 1900 | 
|  | 442 | if year <= 68: | 
|  | 443 | year += 2000 | 
|  | 444 | else: | 
|  | 445 | year += 1900 | 
|  | 446 | elif group_key == 'Y': | 
|  | 447 | year = int(found_dict['Y']) | 
|  | 448 | elif group_key == 'm': | 
|  | 449 | month = int(found_dict['m']) | 
|  | 450 | elif group_key == 'B': | 
|  | 451 | month = _insensitiveindex(locale_time.f_month, found_dict['B']) | 
|  | 452 | elif group_key == 'b': | 
|  | 453 | month = _insensitiveindex(locale_time.a_month, found_dict['b']) | 
|  | 454 | elif group_key == 'd': | 
|  | 455 | day = int(found_dict['d']) | 
|  | 456 | elif group_key is 'H': | 
|  | 457 | hour = int(found_dict['H']) | 
|  | 458 | elif group_key == 'I': | 
|  | 459 | hour = int(found_dict['I']) | 
|  | 460 | ampm = found_dict.get('p', '').lower() | 
|  | 461 | # If there was no AM/PM indicator, we'll treat this like AM | 
|  | 462 | if ampm in ('', locale_time.am_pm[0].lower()): | 
|  | 463 | # We're in AM so the hour is correct unless we're | 
|  | 464 | # looking at 12 midnight. | 
|  | 465 | # 12 midnight == 12 AM == hour 0 | 
|  | 466 | if hour == 12: | 
|  | 467 | hour = 0 | 
|  | 468 | elif ampm == locale_time.am_pm[1].lower(): | 
|  | 469 | # We're in PM so we need to add 12 to the hour unless | 
|  | 470 | # we're looking at 12 noon. | 
|  | 471 | # 12 noon == 12 PM == hour 12 | 
|  | 472 | if hour != 12: | 
|  | 473 | hour += 12 | 
|  | 474 | elif group_key == 'M': | 
|  | 475 | minute = int(found_dict['M']) | 
|  | 476 | elif group_key == 'S': | 
|  | 477 | second = int(found_dict['S']) | 
|  | 478 | elif group_key == 'A': | 
|  | 479 | weekday = _insensitiveindex(locale_time.f_weekday, | 
|  | 480 | found_dict['A']) | 
|  | 481 | elif group_key == 'a': | 
|  | 482 | weekday = _insensitiveindex(locale_time.a_weekday, | 
|  | 483 | found_dict['a']) | 
|  | 484 | elif group_key == 'w': | 
|  | 485 | weekday = int(found_dict['w']) | 
|  | 486 | if weekday == 0: | 
|  | 487 | weekday = 6 | 
|  | 488 | else: | 
|  | 489 | weekday -= 1 | 
|  | 490 | elif group_key == 'j': | 
|  | 491 | julian = int(found_dict['j']) | 
|  | 492 | elif group_key == 'Z': | 
|  | 493 | found_zone = found_dict['Z'].lower() | 
|  | 494 | if locale_time.timezone[0] == locale_time.timezone[1]: | 
|  | 495 | pass #Deals with bad locale setup where timezone info is | 
|  | 496 | # the same; first found on FreeBSD 4.4. | 
|  | 497 | elif locale_time.timezone[0].lower() == found_zone: | 
|  | 498 | tz = 0 | 
|  | 499 | elif locale_time.timezone[1].lower() == found_zone: | 
|  | 500 | tz = 1 | 
|  | 501 | elif locale_time.timezone[2].lower() == found_zone: | 
|  | 502 | tz = -1 | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 503 | # Cannot pre-calculate datetime_date() since can change in Julian | 
|  | 504 | #calculation and thus could have different value for the day of the week | 
|  | 505 | #calculation | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 506 | if julian == -1: | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 507 | # Need to add 1 to result since first day of the year is 1, not 0. | 
|  | 508 | julian = datetime_date(year, month, day).toordinal() - \ | 
|  | 509 | datetime_date(year, 1, 1).toordinal() + 1 | 
|  | 510 | else:  # Assume that if they bothered to include Julian day it will | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 511 | #be accurate | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 512 | datetime_result = datetime_date.fromordinal((julian - 1) + datetime_date(year, 1, 1).toordinal()) | 
|  | 513 | year = datetime_result.year | 
|  | 514 | month = datetime_result.month | 
|  | 515 | day = datetime_result.day | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 516 | if weekday == -1: | 
| Raymond Hettinger | 1fdb633 | 2003-03-09 07:44:42 +0000 | [diff] [blame] | 517 | weekday = datetime_date(year, month, day).weekday() | 
| Tim Peters | 08e5427 | 2003-01-18 03:53:49 +0000 | [diff] [blame] | 518 | return time.struct_time((year, month, day, | 
|  | 519 | hour, minute, second, | 
|  | 520 | weekday, julian, tz)) | 
| Guido van Rossum | 00efe7e | 2002-07-19 17:04:46 +0000 | [diff] [blame] | 521 |  | 
| Barry Warsaw | 4d895fa | 2002-09-23 22:46:49 +0000 | [diff] [blame] | 522 | def _insensitiveindex(lst, findme): | 
|  | 523 | # Perform a case-insensitive index search. | 
|  | 524 |  | 
|  | 525 | #XXX <bc>: If LocaleTime is not exposed, then consider removing this and | 
|  | 526 | #          just lowercase when LocaleTime sets its vars and lowercasing | 
|  | 527 | #          search values. | 
|  | 528 | findme = findme.lower() | 
|  | 529 | for key,item in enumerate(lst): | 
|  | 530 | if item.lower() == findme: | 
|  | 531 | return key | 
|  | 532 | else: | 
|  | 533 | raise ValueError("value not in list") |