diff --git a/CHANGES.md b/CHANGES.md index 9fd28a6e..c8b48cb8 100644 --- a/CHANGES.md +++ b/CHANGES.md @@ -42,6 +42,7 @@ * Update Six compatibility library 1.5.2 to 1.9.0 (8a545f4) * Update SimpleJSON library 2.0.9 to 3.7.3 (0bcdf20) * Update xmltodict library 0.9.0 to 0.9.2 (579a005) +* Update dateutil library 2.2 to 2.4.2 (a6b8925) [develop changelog] * Update Requests library 2.7.0 (ab1f493) to 2.7.0 (8b5e457) diff --git a/lib/dateutil/__init__.py b/lib/dateutil/__init__.py index 1020e729..743669c7 100644 --- a/lib/dateutil/__init__.py +++ b/lib/dateutil/__init__.py @@ -1,10 +1,2 @@ # -*- coding: utf-8 -*- -""" -Copyright (c) 2003-2010 Gustavo Niemeyer - -This module offers extensions to the standard Python -datetime module. -""" -__author__ = "Tomi Pieviläinen " -__license__ = "Simplified BSD" -__version__ = "2.2" +__version__ = "2.4.2" diff --git a/lib/dateutil/easter.py b/lib/dateutil/easter.py index d8a38844..8d30c4eb 100644 --- a/lib/dateutil/easter.py +++ b/lib/dateutil/easter.py @@ -1,18 +1,17 @@ +# -*- coding: utf-8 -*- """ -Copyright (c) 2003-2007 Gustavo Niemeyer - -This module offers extensions to the standard Python -datetime module. +This module offers a generic easter computing method for any given year, using +Western, Orthodox or Julian algorithms. """ -__license__ = "Simplified BSD" import datetime __all__ = ["easter", "EASTER_JULIAN", "EASTER_ORTHODOX", "EASTER_WESTERN"] -EASTER_JULIAN = 1 +EASTER_JULIAN = 1 EASTER_ORTHODOX = 2 -EASTER_WESTERN = 3 +EASTER_WESTERN = 3 + def easter(year, method=EASTER_WESTERN): """ @@ -24,7 +23,7 @@ def easter(year, method=EASTER_WESTERN): This algorithm implements three different easter calculation methods: - + 1 - Original calculation in Julian calendar, valid in dates after 326 AD 2 - Original method, with date converted to Gregorian @@ -39,7 +38,7 @@ def easter(year, method=EASTER_WESTERN): EASTER_WESTERN = 3 The default method is method 3. - + More about the algorithm may be found at: http://users.chariot.net.au/~gmarts/eastalg.htm @@ -68,24 +67,23 @@ def easter(year, method=EASTER_WESTERN): e = 0 if method < 3: # Old method - i = (19*g+15)%30 - j = (y+y//4+i)%7 + i = (19*g + 15) % 30 + j = (y + y//4 + i) % 7 if method == 2: # Extra dates to convert Julian to Gregorian date e = 10 if y > 1600: - e = e+y//100-16-(y//100-16)//4 + e = e + y//100 - 16 - (y//100 - 16)//4 else: # New method c = y//100 - h = (c-c//4-(8*c+13)//25+19*g+15)%30 - i = h-(h//28)*(1-(h//28)*(29//(h+1))*((21-g)//11)) - j = (y+y//4+i+2-c+c//4)%7 + h = (c - c//4 - (8*c + 13)//25 + 19*g + 15) % 30 + i = h - (h//28)*(1 - (h//28)*(29//(h + 1))*((21 - g)//11)) + j = (y + y//4 + i + 2 - c + c//4) % 7 # p can be from -6 to 56 corresponding to dates 22 March to 23 May # (later dates apply to method 2, although 23 May never actually occurs) - p = i-j+e - d = 1+(p+27+(p+6)//40)%31 - m = 3+(p+26)//30 + p = i - j + e + d = 1 + (p + 27 + (p + 6)//40) % 31 + m = 3 + (p + 26)//30 return datetime.date(int(y), int(m), int(d)) - diff --git a/lib/dateutil/parser.py b/lib/dateutil/parser.py index aef83623..590ac109 100644 --- a/lib/dateutil/parser.py +++ b/lib/dateutil/parser.py @@ -1,50 +1,70 @@ # -*- coding:iso-8859-1 -*- """ -Copyright (c) 2003-2007 Gustavo Niemeyer +This module offers a generic date/time string parser which is able to parse +most known formats to represent a date and/or time. -This module offers extensions to the standard Python -datetime module. +This module attempts to be forgiving with regards to unlikely input formats, +returning a datetime object even for dates which are ambiguous. If an element of +a date/time stamp is omitted, the following rules are applied: +- If AM or PM is left unspecified, a 24-hour clock is assumed, however, an hour + on a 12-hour clock (`0 <= hour <= 12`) *must* be specified if AM or PM is + specified. +- If a time zone is omitted, it is assumed to be UTC. + +If any other elements are missing, they are taken from the `datetime.datetime` +object passed to the parameter `default`. If this results in a day number +exceeding the valid number of days per month, one can fall back to the last +day of the month by setting `fallback_on_invalid_day` parameter to `True`. + +Also provided is the `smart_defaults` option, which attempts to fill in the +missing elements from context. If specified, the logic is: +- If the omitted element is smaller than the largest specified element, select + the *earliest* time matching the specified conditions; so `"June 2010"` is + interpreted as `June 1, 2010 0:00:00`) and the (somewhat strange) + `"Feb 1997 3:15 PM"` is interpreted as `February 1, 1997 15:15:00`. +- If the element is larger than the largest specified element, select the + *most recent* time matching the specified conditions (e.g parsing `"May"` + in June 2015 returns the date May 1st, 2015, whereas parsing it in April 2015 + returns May 1st 2014). If using the `date_in_future` flag, this logic is + inverted, and instead the *next* time matching the specified conditions is + returned. + +Additional resources about date/time string formats can be found below: + +- `A summary of the international standard date and time notation + `_ +- `W3C Date and Time Formats `_ +- `Time Formats (Planetary Rings Node) `_ +- `CPAN ParseDate module + `_ +- `Java SimpleDateFormat Class + `_ """ from __future__ import unicode_literals -__license__ = "Simplified BSD" - import datetime import string import time -import sys -import os import collections - -try: - from io import StringIO -except ImportError: - from io import StringIO +from io import StringIO +from calendar import monthrange, isleap from six import text_type, binary_type, integer_types from . import relativedelta from . import tz - __all__ = ["parse", "parserinfo"] -# Some pointers: -# -# http://www.cl.cam.ac.uk/~mgk25/iso-time.html -# http://www.iso.ch/iso/en/prods-services/popstds/datesandtime.html -# http://www.w3.org/TR/NOTE-datetime -# http://ringmaster.arc.nasa.gov/tools/time_formats.html -# http://search.cpan.org/author/MUIR/Time-modules-2003.0211/lib/Time/ParseDate.pm -# http://stein.cshl.org/jade/distrib/docs/java.text.SimpleDateFormat.html - - class _timelex(object): - def __init__(self, instream): + if isinstance(instream, binary_type): + instream = instream.decode() + if isinstance(instream, text_type): instream = StringIO(instream) + self.instream = instream self.wordchars = ('abcdfeghijklmnopqrstuvwxyz' 'ABCDEFGHIJKLMNOPQRSTUVWXYZ_' @@ -57,25 +77,47 @@ class _timelex(object): self.eof = False def get_token(self): + """ + This function breaks the time string into lexical units (tokens), which + can be parsed by the parser. Lexical units are demarcated by changes in + the character set, so any continuous string of letters is considered one + unit, any continuous string of numbers is considered one unit. + + The main complication arises from the fact that dots ('.') can be used + both as separators (e.g. "Sep.20.2009") or decimal points (e.g. + "4:30:21.447"). As such, it is necessary to read the full context of + any dot-separated strings before breaking it into tokens; as such, this + function maintains a "token stack", for when the ambiguous context + demands that multiple tokens be parsed at once. + """ if self.tokenstack: return self.tokenstack.pop(0) + seenletters = False token = None state = None wordchars = self.wordchars numchars = self.numchars whitespace = self.whitespace + while not self.eof: + # We only realize that we've reached the end of a token when we find + # a character that's not part of the current token - since that + # character may be part of the next token, it's stored in the + # charstack. if self.charstack: nextchar = self.charstack.pop(0) else: nextchar = self.instream.read(1) while nextchar == '\x00': nextchar = self.instream.read(1) + if not nextchar: self.eof = True break elif not state: + # First character of the token - determines if we're starting + # to parse a word, a number or something else. token = nextchar if nextchar in wordchars: state = 'a' @@ -83,10 +125,12 @@ class _timelex(object): state = '0' elif nextchar in whitespace: token = ' ' - break # emit token + break # emit token else: - break # emit token + break # emit token elif state == 'a': + # If we've already started reading a word, we keep reading + # letters until we find something that's not part of a word. seenletters = True if nextchar in wordchars: token += nextchar @@ -95,8 +139,10 @@ class _timelex(object): state = 'a.' else: self.charstack.append(nextchar) - break # emit token + break # emit token elif state == '0': + # If we've already started reading a number, we keep reading + # numbers until we find something that doesn't fit. if nextchar in numchars: token += nextchar elif nextchar == '.': @@ -104,8 +150,10 @@ class _timelex(object): state = '0.' else: self.charstack.append(nextchar) - break # emit token + break # emit token elif state == 'a.': + # If we've seen some letters and a dot separator, continue + # parsing, and the tokens will be broken up later. seenletters = True if nextchar == '.' or nextchar in wordchars: token += nextchar @@ -114,8 +162,10 @@ class _timelex(object): state = '0.' else: self.charstack.append(nextchar) - break # emit token + break # emit token elif state == '0.': + # If we've seen at least one dot separator, keep going, we'll + # break up the tokens later. if nextchar == '.' or nextchar in numchars: token += nextchar elif nextchar in wordchars and token[-1] == '.': @@ -123,15 +173,17 @@ class _timelex(object): state = 'a.' else: self.charstack.append(nextchar) - break # emit token - if (state in ('a.', '0.') and - (seenletters or token.count('.') > 1 or token[-1] == '.')): + break # emit token + + if (state in ('a.', '0.') and (seenletters or token.count('.') > 1 or + token[-1] == '.')): l = token.split('.') token = l[0] for tok in l[1:]: self.tokenstack.append('.') if tok: self.tokenstack.append(tok) + return token def __iter__(self): @@ -141,6 +193,7 @@ class _timelex(object): token = self.get_token() if token is None: raise StopIteration + return token def next(self): @@ -170,6 +223,22 @@ class _resultbase(object): class parserinfo(object): + """ + Class which handles what inputs are accepted. Subclass this to customize the + language and acceptable values for each parameter. + + :param dayfirst: + Whether to interpret the first value in an ambiguous 3-integer date + (e.g. 01/05/09) as the day (`True`) or month (`False`). If + `yearfirst` is set to `True`, this distinguishes between YDM and + YMD. Default is `False`. + + :param yearfirst: + Whether to interpret the first value in an ambiguous 3-integer date + (e.g. 01/05/09) as the year. If `True`, the first number is taken to + be the year, otherwise the last number is taken to be the year. + Default is `False`. + """ # m from a.m/p.m, t from ISO T separator JUMP = [" ", ".", ",", ";", "-", "/", "'", @@ -183,18 +252,18 @@ class parserinfo(object): ("Fri", "Friday"), ("Sat", "Saturday"), ("Sun", "Sunday")] - MONTHS = [("Jan", "January"), - ("Feb", "February"), - ("Mar", "March"), - ("Apr", "April"), - ("May", "May"), - ("Jun", "June"), - ("Jul", "July"), - ("Aug", "August"), - ("Sep", "Sept", "September"), - ("Oct", "October"), - ("Nov", "November"), - ("Dec", "December")] + MONTHS = [("Jan", "January"), + ("Feb", "February"), + ("Mar", "March"), + ("Apr", "April"), + ("May", "May"), + ("Jun", "June"), + ("Jul", "July"), + ("Aug", "August"), + ("Sep", "Sept", "September"), + ("Oct", "October"), + ("Nov", "November"), + ("Dec", "December")] HMS = [("h", "hour", "hours"), ("m", "minute", "minutes"), ("s", "second", "seconds")] @@ -204,7 +273,7 @@ class parserinfo(object): PERTAIN = ["of"] TZOFFSET = {} - def __init__(self, dayfirst=False, yearfirst=False): + def __init__(self, dayfirst=False, yearfirst=False, smart_defaults=False): self._jump = self._convert(self.JUMP) self._weekdays = self._convert(self.WEEKDAYS) self._months = self._convert(self.MONTHS) @@ -215,14 +284,14 @@ class parserinfo(object): self.dayfirst = dayfirst self.yearfirst = yearfirst + self.smart_defaults = smart_defaults self._year = time.localtime().tm_year - self._century = self._year//100*100 + self._century = self._year // 100*100 def _convert(self, lst): dct = {} - for i in range(len(lst)): - v = lst[i] + for i, v in enumerate(lst): if isinstance(v, tuple): for v in v: dct[v.lower()] = i @@ -270,6 +339,7 @@ class parserinfo(object): def tzoffset(self, name): if name in self._utczone: return 0 + return self.TZOFFSET.get(name) def convertyear(self, year): @@ -286,6 +356,7 @@ class parserinfo(object): # move to info if res.year is not None: res.year = self.convertyear(res.year) + if res.tzoffset == 0 and not res.tzname or res.tzname == 'Z': res.tzname = "UTC" res.tzoffset = 0 @@ -295,37 +366,192 @@ class parserinfo(object): class parser(object): - def __init__(self, info=None): self.info = info or parserinfo() - def parse(self, timestr, default=None, - ignoretz=False, tzinfos=None, - **kwargs): - if not default: + def parse(self, timestr, default=None, ignoretz=False, tzinfos=None, + smart_defaults=None, date_in_future=False, + fallback_on_invalid_day=None, **kwargs): + """ + Parse the date/time string into a datetime object. + + :param timestr: + Any date/time string using the supported formats. + + :param default: + The default datetime object, if this is a datetime object and not + `None`, elements specified in `timestr` replace elements in the + default object, unless `smart_defaults` is set to `True`, in which + case to the extent necessary, timestamps are calculated relative to + this date. + + :param smart_defaults: + If using smart defaults, the `default` parameter is treated as the + effective parsing date/time, and the context of the datetime string + is determined relative to `default`. If `None`, this parameter is + inherited from the :class:`parserinfo` object. + + :param date_in_future: + If `smart_defaults` is `True`, the parser assumes by default that + the timestamp refers to a date in the past, and will return the + beginning of the most recent timespan which matches the time string + (e.g. if `default` is March 3rd, 2013, "Feb" parses to + "Feb 1, 2013" and "May 3" parses to May 3rd, 2012). Setting this + parameter to `True` inverts this assumption, and returns the + beginning of the *next* matching timespan. + + :param fallback_on_invalid_day: + If specified `True`, an otherwise invalid date such as "Feb 30" or + "June 32" falls back to the last day of the month. If specified as + "False", the parser is strict about parsing otherwise valid dates + that would turn up as invalid because of the fallback rules (e.g. + "Feb 2010" run with a default of January 30, 2010 and `smartparser` + set to `False` would would throw an error, rather than falling + back to the end of February). If `None` or unspecified, the date + falls back to the most recent valid date only if the invalid date + is created as a result of an unspecified day in the time string. + + :param ignoretz: + Whether or not to ignore the time zone. + + :param tzinfos: + A time zone, to be applied to the date, if `ignoretz` is `True`. + This can be either a subclass of `tzinfo`, a time zone string or an + integer offset. + + :param **kwargs: + Keyword arguments as passed to `_parse()`. + + :return: + Returns a `datetime.datetime` object or, if the `fuzzy_with_tokens` + option is `True`, returns a tuple, the first element being a + `datetime.datetime` object, the second a tuple containing the + fuzzy tokens. + + :raises ValueError: + Raised for invalid or unknown string format, if the provided + `tzinfo` is not in a valid format, or if an invalid date would + be created. + + :raises OverFlowError: + Raised if the parsed date exceeds the largest valid C integer on + your system. + """ + + if smart_defaults is None: + smart_defaults = self.info.smart_defaults + + if default is None: + effective_dt = datetime.datetime.now() default = datetime.datetime.now().replace(hour=0, minute=0, - second=0, microsecond=0) + second=0, microsecond=0) + else: + effective_dt = default - - res, skipped_tokens = self._parse(timestr, **kwargs) + if kwargs.get('fuzzy_with_tokens', False): + res, skipped_tokens = self._parse(timestr, **kwargs) + else: + res = self._parse(timestr, **kwargs) if res is None: - raise ValueError("unknown string format") + raise ValueError("Unknown string format") + repl = {} - for attr in ["year", "month", "day", "hour", - "minute", "second", "microsecond"]: + for attr in ("year", "month", "day", "hour", + "minute", "second", "microsecond"): value = getattr(res, attr) if value is not None: repl[attr] = value + + # Choose the correct fallback position if requested by the + # `smart_defaults` parameter. + if smart_defaults: + # Determine if it refers to this year, last year or next year + if res.year is None: + if res.month is not None: + # Explicitly deal with leap year problems + if res.month == 2 and (res.day is not None and + res.day == 29): + + ly_offset = 4 if date_in_future else -4 + next_year = 4 * (default.year // 4) + + if date_in_future: + next_year += ly_offset + + if not isleap(next_year): + next_year += ly_offset + + if not isleap(default.year): + default = default.replace(year=next_year) + elif date_in_future: + next_year = default.year + 1 + else: + next_year = default.year - 1 + + if ((res.month == default.month and res.day is not None and + ((res.day < default.day and date_in_future) or + (res.day > default.day and not date_in_future))) or + ((res.month < default.month and date_in_future) or + (res.month > default.month and not date_in_future))): + + default = default.replace(year=next_year) + + # Select a proper month + if res.month is None: + if res.year is not None: + default = default.replace(month=1) + + # I'm not sure if this is even possible. + if res.day is not None: + if res.day < default.day and date_in_future: + default += datetime.timedelta(months=1) + elif res.day > default.day and not date_in_future: + default -= datetime.timedelta(months=1) + + if res.day is None: + # Determine if it's today, tomorrow or yesterday. + if res.year is None and res.month is None: + t_repl = {} + for key, val in repl.iteritems(): + if key in ('hour', 'minute', 'second', 'microsecond'): + t_repl[key] = val + + stime = effective_dt.replace(**t_repl) + + if stime < effective_dt and date_in_future: + default += datetime.timedelta(days=1) + elif stime > effective_dt and not date_in_future: + default -= datetime.timedelta(days=1) + else: + # Otherwise it's the beginning of the month + default = default.replace(day=1) + + if fallback_on_invalid_day or (fallback_on_invalid_day is None and + 'day' not in repl): + # If the default day exceeds the last day of the month, fall back to + # the end of the month. + cyear = default.year if res.year is None else res.year + cmonth = default.month if res.month is None else res.month + cday = default.day if res.day is None else res.day + + if cday > monthrange(cyear, cmonth)[1]: + repl['day'] = monthrange(cyear, cmonth)[1] + ret = default.replace(**repl) + if res.weekday is not None and not res.day: ret = ret+relativedelta.relativedelta(weekday=res.weekday) + if not ignoretz: - if isinstance(tzinfos, collections.Callable) or tzinfos and res.tzname in tzinfos: + if (isinstance(tzinfos, collections.Callable) or + tzinfos and res.tzname in tzinfos): + if isinstance(tzinfos, collections.Callable): tzdata = tzinfos(res.tzname, res.tzoffset) else: tzdata = tzinfos.get(res.tzname) + if isinstance(tzdata, datetime.tzinfo): tzinfo = tzdata elif isinstance(tzdata, text_type): @@ -333,8 +559,8 @@ class parser(object): elif isinstance(tzdata, integer_types): tzinfo = tz.tzoffset(res.tzname, tzdata) else: - raise ValueError("offset must be tzinfo subclass, " \ - "tz string, or int offset") + raise ValueError("Offset must be tzinfo subclass, " + "tz string, or int offset.") ret = ret.replace(tzinfo=tzinfo) elif res.tzname and res.tzname in time.tzname: ret = ret.replace(tzinfo=tz.tzlocal()) @@ -343,28 +569,64 @@ class parser(object): elif res.tzoffset: ret = ret.replace(tzinfo=tz.tzoffset(res.tzname, res.tzoffset)) - if skipped_tokens: + if kwargs.get('fuzzy_with_tokens', False): return ret, skipped_tokens - - return ret + else: + return ret class _result(_resultbase): __slots__ = ["year", "month", "day", "weekday", "hour", "minute", "second", "microsecond", - "tzname", "tzoffset"] + "tzname", "tzoffset", "ampm"] - def _parse(self, timestr, dayfirst=None, yearfirst=None, fuzzy=False, fuzzy_with_tokens=False): + def _parse(self, timestr, dayfirst=None, yearfirst=None, fuzzy=False, + fuzzy_with_tokens=False): + """ + Private method which performs the heavy lifting of parsing, called from + `parse()`, which passes on its `kwargs` to this function. + + :param timestr: + The string to parse. + + :param dayfirst: + Whether to interpret the first value in an ambiguous 3-integer date + (e.g. 01/05/09) as the day (`True`) or month (`False`). If + `yearfirst` is set to `True`, this distinguishes between YDM and + YMD. If set to `None`, this value is retrieved from the current + `parserinfo` object (which itself defaults to `False`). + + :param yearfirst: + Whether to interpret the first value in an ambiguous 3-integer date + (e.g. 01/05/09) as the year. If `True`, the first number is taken to + be the year, otherwise the last number is taken to be the year. If + this is set to `None`, the value is retrieved from the current + `parserinfo` object (which itself defaults to `False`). + + :param fuzzy: + Whether to allow fuzzy parsing, allowing for string like "Today is + January 1, 2047 at 8:21:00AM". + + :param fuzzy_with_tokens: + If `True`, `fuzzy` is automatically set to True, and the parser will + return a tuple where the first element is the parsed + `datetime.datetime` datetimestamp and the second element is a tuple + containing the portions of the string which were ignored, e.g. + "Today is January 1, 2047 at 8:21:00AM" should return + `(datetime.datetime(2011, 1, 1, 8, 21), (u'Today is ', u' ', u'at '))` + """ if fuzzy_with_tokens: fuzzy = True info = self.info + if dayfirst is None: dayfirst = info.dayfirst + if yearfirst is None: yearfirst = info.yearfirst - res = self._result() - l = _timelex.split(timestr) + res = self._result() + l = _timelex.split(timestr) # Splits the timestr into tokens # keep up with the last token skipped so we can recombine # consecutively skipped tokens (-2 for when i begins at 0). @@ -372,7 +634,6 @@ class parser(object): skipped_tokens = list() try: - # year/month/day list ymd = [] @@ -394,17 +655,21 @@ class parser(object): # Token is a number len_li = len(l[i]) i += 1 + if (len(ymd) == 3 and len_li in (2, 4) - and (i >= len_l or (l[i] != ':' and - info.hms(l[i]) is None))): + and res.hour is None and (i >= len_l or (l[i] != ':' and + info.hms(l[i]) is None))): # 19990101T23[59] s = l[i-1] res.hour = int(s[:2]) + if len_li == 4: res.minute = int(s[2:]) + elif len_li == 6 or (len_li > 6 and l[i-1].find('.') == 6): # YYMMDD or HHMMSS[.ss] s = l[i-1] + if not ymd and l[i-1].find('.') == -1: ymd.append(info.convertyear(int(s[:2]))) ymd.append(int(s[2:4])) @@ -414,12 +679,14 @@ class parser(object): res.hour = int(s[:2]) res.minute = int(s[2:4]) res.second, res.microsecond = _parsems(s[4:]) + elif len_li == 8: # YYYYMMDD s = l[i-1] ymd.append(int(s[:4])) ymd.append(int(s[4:6])) ymd.append(int(s[6:])) + elif len_li in (12, 14): # YYYYMMDDhhmm[ss] s = l[i-1] @@ -428,30 +695,42 @@ class parser(object): ymd.append(int(s[6:8])) res.hour = int(s[8:10]) res.minute = int(s[10:12]) + if len_li == 14: res.second = int(s[12:]) + elif ((i < len_l and info.hms(l[i]) is not None) or (i+1 < len_l and l[i] == ' ' and info.hms(l[i+1]) is not None)): + # HH[ ]h or MM[ ]m or SS[.ss][ ]s if l[i] == ' ': i += 1 + idx = info.hms(l[i]) + while True: if idx == 0: res.hour = int(value) - if value%1: - res.minute = int(60*(value%1)) + + if value % 1: + res.minute = int(60*(value % 1)) + elif idx == 1: res.minute = int(value) - if value%1: - res.second = int(60*(value%1)) + + if value % 1: + res.second = int(60*(value % 1)) + elif idx == 2: res.second, res.microsecond = \ _parsems(value_repr) + i += 1 + if i >= len_l or idx == 2: break + # 12h00 try: value_repr = l[i] @@ -461,37 +740,49 @@ class parser(object): else: i += 1 idx += 1 + if i < len_l: newidx = info.hms(l[i]) + if newidx is not None: idx = newidx - elif i == len_l and l[i-2] == ' ' and info.hms(l[i-3]) is not None: + + elif (i == len_l and l[i-2] == ' ' and + info.hms(l[i-3]) is not None): # X h MM or X m SS idx = info.hms(l[i-3]) + 1 + if idx == 1: res.minute = int(value) - if value%1: - res.second = int(60*(value%1)) + + if value % 1: + res.second = int(60*(value % 1)) elif idx == 2: res.second, res.microsecond = \ - _parsems(value_repr) + _parsems(value_repr) i += 1 + elif i+1 < len_l and l[i] == ':': # HH:MM[:SS[.ss]] res.hour = int(value) i += 1 value = float(l[i]) res.minute = int(value) - if value%1: - res.second = int(60*(value%1)) + + if value % 1: + res.second = int(60*(value % 1)) + i += 1 + if i < len_l and l[i] == ':': res.second, res.microsecond = _parsems(l[i+1]) i += 2 + elif i < len_l and l[i] in ('-', '/', '.'): sep = l[i] ymd.append(int(value)) i += 1 + if i < len_l and not info.jump(l[i]): try: # 01-01[-01] @@ -499,45 +790,55 @@ class parser(object): except ValueError: # 01-Jan[-01] value = info.month(l[i]) + if value is not None: ymd.append(value) assert mstridx == -1 mstridx = len(ymd)-1 else: return None + i += 1 + if i < len_l and l[i] == sep: # We have three members i += 1 value = info.month(l[i]) + if value is not None: ymd.append(value) mstridx = len(ymd)-1 assert mstridx == -1 else: ymd.append(int(l[i])) + i += 1 elif i >= len_l or info.jump(l[i]): if i+1 < len_l and info.ampm(l[i+1]) is not None: # 12 am res.hour = int(value) + if res.hour < 12 and info.ampm(l[i+1]) == 1: res.hour += 12 elif res.hour == 12 and info.ampm(l[i+1]) == 0: res.hour = 0 + i += 1 else: # Year, month or day ymd.append(int(value)) i += 1 elif info.ampm(l[i]) is not None: + # 12am res.hour = int(value) + if res.hour < 12 and info.ampm(l[i]) == 1: res.hour += 12 elif res.hour == 12 and info.ampm(l[i]) == 0: res.hour = 0 i += 1 + elif not fuzzy: return None else: @@ -557,6 +858,7 @@ class parser(object): ymd.append(value) assert mstridx == -1 mstridx = len(ymd)-1 + i += 1 if i < len_l: if l[i] in ('-', '/'): @@ -565,11 +867,13 @@ class parser(object): i += 1 ymd.append(int(l[i])) i += 1 + if i < len_l and l[i] == sep: # Jan-01-99 i += 1 ymd.append(int(l[i])) i += 1 + elif (i+3 < len_l and l[i] == l[i+2] == ' ' and info.pertain(l[i+1])): # Jan of 01 @@ -588,17 +892,47 @@ class parser(object): # Check am/pm value = info.ampm(l[i]) if value is not None: - if value == 1 and res.hour < 12: - res.hour += 12 - elif value == 0 and res.hour == 12: - res.hour = 0 + # For fuzzy parsing, 'a' or 'am' (both valid English words) + # may erroneously trigger the AM/PM flag. Deal with that + # here. + val_is_ampm = True + + # If there's already an AM/PM flag, this one isn't one. + if fuzzy and res.ampm is not None: + val_is_ampm = False + + # If AM/PM is found and hour is not, raise a ValueError + if res.hour is None: + if fuzzy: + val_is_ampm = False + else: + raise ValueError('No hour specified with ' + + 'AM or PM flag.') + elif not 0 <= res.hour <= 12: + # If AM/PM is found, it's a 12 hour clock, so raise + # an error for invalid range + if fuzzy: + val_is_ampm = False + else: + raise ValueError('Invalid hour specified for ' + + '12-hour clock.') + + if val_is_ampm: + if value == 1 and res.hour < 12: + res.hour += 12 + elif value == 0 and res.hour == 12: + res.hour = 0 + + res.ampm = value + i += 1 continue # Check for a timezone name if (res.hour is not None and len(l[i]) <= 5 and - res.tzname is None and res.tzoffset is None and - not [x for x in l[i] if x not in string.ascii_uppercase]): + res.tzname is None and res.tzoffset is None and + not [x for x in l[i] if x not in + string.ascii_uppercase]): res.tzname = l[i] res.tzoffset = info.tzoffset(res.tzname) i += 1 @@ -623,6 +957,7 @@ class parser(object): signal = (-1, 1)[l[i] == '+'] i += 1 len_li = len(l[i]) + if len_li == 4: # -0300 res.tzoffset = int(l[i][:2])*3600+int(l[i][2:])*60 @@ -636,6 +971,7 @@ class parser(object): else: return None i += 1 + res.tzoffset *= signal # Look for a timezone name between parenthesis @@ -643,7 +979,7 @@ class parser(object): info.jump(l[i]) and l[i+1] == '(' and l[i+3] == ')' and 3 <= len(l[i+2]) <= 5 and not [x for x in l[i+2] - if x not in string.ascii_uppercase]): + if x not in string.ascii_uppercase]): # -0300 (BRST) res.tzname = l[i+2] i += 4 @@ -672,11 +1008,13 @@ class parser(object): if mstridx != -1: res.month = ymd[mstridx] del ymd[mstridx] + if len_ymd > 1 or mstridx == -1: if ymd[0] > 31: res.year = ymd[0] else: res.day = ymd[0] + elif len_ymd == 2: # Two members with numbers if ymd[0] > 31: @@ -691,7 +1029,8 @@ class parser(object): else: # 01-13 res.month, res.day = ymd - if len_ymd == 3: + + elif len_ymd == 3: # Three members if mstridx == 0: res.month, res.day, res.year = ymd @@ -704,6 +1043,7 @@ class parser(object): # Give precendence to day-first, since # two-digit years is usually hand-written. res.day, res.month, res.year = ymd + elif mstridx == 2: # WTF!? if ymd[1] > 31: @@ -712,6 +1052,7 @@ class parser(object): else: # 99-01-Jan res.year, res.day, res.month = ymd + else: if ymd[0] > 31 or \ (yearfirst and ymd[1] <= 12 and ymd[2] <= 31): @@ -732,16 +1073,66 @@ class parser(object): if fuzzy_with_tokens: return res, tuple(skipped_tokens) - - return res, None + else: + return res DEFAULTPARSER = parser() + + def parse(timestr, parserinfo=None, **kwargs): - # Python 2.x support: datetimes return their string presentation as - # bytes in 2.x and unicode in 3.x, so it's reasonable to expect that - # the parser will get both kinds. Internally we use unicode only. - if isinstance(timestr, binary_type): - timestr = timestr.decode() + """ + Parse a string in one of the supported formats, using the `parserinfo` + parameters. + + :param timestr: + A string containing a date/time stamp. + + :param parserinfo: + A :class:`parserinfo` object containing parameters for the parser. + If `None`, the default arguments to the `parserinfo` constructor are + used. + + The `**kwargs` parameter takes the following keyword arguments: + + :param default: + The default datetime object, if this is a datetime object and not + `None`, elements specified in `timestr` replace elements in the + default object. + + :param ignoretz: + Whether or not to ignore the time zone (boolean). + + :param tzinfos: + A time zone, to be applied to the date, if `ignoretz` is `True`. + This can be either a subclass of `tzinfo`, a time zone string or an + integer offset. + + :param dayfirst: + Whether to interpret the first value in an ambiguous 3-integer date + (e.g. 01/05/09) as the day (`True`) or month (`False`). If + `yearfirst` is set to `True`, this distinguishes between YDM and + YMD. If set to `None`, this value is retrieved from the current + :class:`parserinfo` object (which itself defaults to `False`). + + :param yearfirst: + Whether to interpret the first value in an ambiguous 3-integer date + (e.g. 01/05/09) as the year. If `True`, the first number is taken to + be the year, otherwise the last number is taken to be the year. If + this is set to `None`, the value is retrieved from the current + :class:`parserinfo` object (which itself defaults to `False`). + + :param fuzzy: + Whether to allow fuzzy parsing, allowing for string like "Today is + January 1, 2047 at 8:21:00AM". + + :param fuzzy_with_tokens: + If `True`, `fuzzy` is automatically set to True, and the parser will + return a tuple where the first element is the parsed + `datetime.datetime` datetimestamp and the second element is a tuple + containing the portions of the string which were ignored, e.g. + "Today is January 1, 2047 at 8:21:00AM" should return + `(datetime.datetime(2011, 1, 1, 8, 21), (u'Today is ', u' ', u'at '))` + """ if parserinfo: return parser(parserinfo).parse(timestr, **kwargs) else: @@ -779,7 +1170,7 @@ class _tzparser(object): # BRST+3[BRDT[+2]] j = i while j < len_l and not [x for x in l[j] - if x in "0123456789:,-+"]: + if x in "0123456789:,-+"]: j += 1 if j != i: if not res.stdabbr: @@ -789,8 +1180,8 @@ class _tzparser(object): offattr = "dstoffset" res.dstabbr = "".join(l[i:j]) i = j - if (i < len_l and - (l[i] in ('+', '-') or l[i][0] in "0123456789")): + if (i < len_l and (l[i] in ('+', '-') or l[i][0] in + "0123456789")): if l[i] in ('+', '-'): # Yes, that's right. See the TZ variable # documentation. @@ -801,8 +1192,8 @@ class _tzparser(object): len_li = len(l[i]) if len_li == 4: # -0300 - setattr(res, offattr, - (int(l[i][:2])*3600+int(l[i][2:])*60)*signal) + setattr(res, offattr, (int(l[i][:2])*3600 + + int(l[i][2:])*60)*signal) elif i+1 < len_l and l[i+1] == ':': # -03:00 setattr(res, offattr, @@ -822,7 +1213,8 @@ class _tzparser(object): if i < len_l: for j in range(i, len_l): - if l[j] == ';': l[j] = ',' + if l[j] == ';': + l[j] = ',' assert l[i] == ',' @@ -831,7 +1223,7 @@ class _tzparser(object): if i >= len_l: pass elif (8 <= l.count(',') <= 9 and - not [y for x in l[i:] if x != ',' + not [y for x in l[i:] if x != ',' for y in x if y not in "0123456789"]): # GMT0BST,3,0,30,3600,10,0,26,7200[,3600] for x in (res.start, res.end): @@ -845,7 +1237,7 @@ class _tzparser(object): i += 2 if value: x.week = value - x.weekday = (int(l[i])-1)%7 + x.weekday = (int(l[i])-1) % 7 else: x.day = int(l[i]) i += 2 @@ -861,7 +1253,7 @@ class _tzparser(object): elif (l.count(',') == 2 and l[i:].count('/') <= 2 and not [y for x in l[i:] if x not in (',', '/', 'J', 'M', '.', '-', ':') - for y in x if y not in "0123456789"]): + for y in x if y not in "0123456789"]): for x in (res.start, res.end): if l[i] == 'J': # non-leap year day (1 based) @@ -880,7 +1272,7 @@ class _tzparser(object): i += 1 assert l[i] in ('-', '.') i += 1 - x.weekday = (int(l[i])-1)%7 + x.weekday = (int(l[i])-1) % 7 else: # year day (zero based) x.yday = int(l[i])+1 @@ -921,6 +1313,8 @@ class _tzparser(object): DEFAULTTZPARSER = _tzparser() + + def _parsetz(tzstr): return DEFAULTTZPARSER.parse(tzstr) diff --git a/lib/dateutil/relativedelta.py b/lib/dateutil/relativedelta.py index 4393bcbc..557d5210 100644 --- a/lib/dateutil/relativedelta.py +++ b/lib/dateutil/relativedelta.py @@ -1,11 +1,4 @@ -""" -Copyright (c) 2003-2010 Gustavo Niemeyer - -This module offers extensions to the standard Python -datetime module. -""" -__license__ = "Simplified BSD" - +# -*- coding: utf-8 -*- import datetime import calendar @@ -13,6 +6,7 @@ from six import integer_types __all__ = ["relativedelta", "MO", "TU", "WE", "TH", "FR", "SA", "SU"] + class weekday(object): __slots__ = ["weekday", "n"] @@ -43,25 +37,35 @@ class weekday(object): MO, TU, WE, TH, FR, SA, SU = weekdays = tuple([weekday(x) for x in range(7)]) + class relativedelta(object): """ -The relativedelta type is based on the specification of the excelent -work done by M.-A. Lemburg in his mx.DateTime extension. However, -notice that this type does *NOT* implement the same algorithm as +The relativedelta type is based on the specification of the excellent +work done by M.-A. Lemburg in his +`mx.DateTime `_ extension. +However, notice that this type does *NOT* implement the same algorithm as his work. Do *NOT* expect it to behave like mx.DateTime's counterpart. -There's two different ways to build a relativedelta instance. The -first one is passing it two date/datetime classes: +There are two different ways to build a relativedelta instance. The +first one is passing it two date/datetime classes:: relativedelta(datetime1, datetime2) -And the other way is to use the following keyword arguments: +The second one is passing it any number of the following keyword arguments:: + + relativedelta(arg1=x,arg2=y,arg3=z...) year, month, day, hour, minute, second, microsecond: - Absolute information. + Absolute information (argument is singular); adding or subtracting a + relativedelta with absolute information does not perform an aritmetic + operation, but rather REPLACES the corresponding value in the + original datetime with the value(s) in relativedelta. years, months, weeks, days, hours, minutes, seconds, microseconds: - Relative information, may be negative. + Relative information, may be negative (argument is plural); adding + or subtracting a relativedelta with relative information performs + the corresponding aritmetic operation on the original datetime value + with the information in the relativedelta. weekday: One of the weekday instances (MO, TU, etc). These instances may @@ -80,26 +84,26 @@ And the other way is to use the following keyword arguments: Here is the behavior of operations with relativedelta: -1) Calculate the absolute year, using the 'year' argument, or the +1. Calculate the absolute year, using the 'year' argument, or the original datetime year, if the argument is not present. -2) Add the relative 'years' argument to the absolute year. +2. Add the relative 'years' argument to the absolute year. -3) Do steps 1 and 2 for month/months. +3. Do steps 1 and 2 for month/months. -4) Calculate the absolute day, using the 'day' argument, or the +4. Calculate the absolute day, using the 'day' argument, or the original datetime day, if the argument is not present. Then, subtract from the day until it fits in the year and month found after their operations. -5) Add the relative 'days' argument to the absolute day. Notice +5. Add the relative 'days' argument to the absolute day. Notice that the 'weeks' argument is multiplied by 7 and added to 'days'. -6) Do steps 1 and 2 for hour/hours, minute/minutes, second/seconds, +6. Do steps 1 and 2 for hour/hours, minute/minutes, second/seconds, microsecond/microseconds. -7) If the 'weekday' argument is present, calculate the weekday, +7. If the 'weekday' argument is present, calculate the weekday, with the given (wday, nth) tuple. wday is the index of the weekday (0-6, 0=Mon), and nth is the number of weeks to add forward or backward, depending on its signal. Notice that if @@ -114,9 +118,14 @@ Here is the behavior of operations with relativedelta: yearday=None, nlyearday=None, hour=None, minute=None, second=None, microsecond=None): if dt1 and dt2: - if (not isinstance(dt1, datetime.date)) or (not isinstance(dt2, datetime.date)): + # datetime is a subclass of date. So both must be date + if not (isinstance(dt1, datetime.date) and + isinstance(dt2, datetime.date)): raise TypeError("relativedelta only diffs datetime/date") - if not type(dt1) == type(dt2): #isinstance(dt1, type(dt2)): + # We allow two dates, or two datetimes, so we coerce them to be + # of the same type + if (isinstance(dt1, datetime.datetime) != + isinstance(dt2, datetime.datetime)): if not isinstance(dt1, datetime.datetime): dt1 = datetime.datetime.fromordinal(dt1.toordinal()) elif not isinstance(dt2, datetime.datetime): @@ -158,7 +167,7 @@ Here is the behavior of operations with relativedelta: else: self.years = years self.months = months - self.days = days+weeks*7 + self.days = days + weeks * 7 self.leapdays = leapdays self.hours = hours self.minutes = minutes @@ -185,7 +194,8 @@ Here is the behavior of operations with relativedelta: if yearday > 59: self.leapdays = -1 if yday: - ydayidx = [31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 366] + ydayidx = [31, 59, 90, 120, 151, 181, 212, + 243, 273, 304, 334, 366] for idx, ydays in enumerate(ydayidx): if yday <= ydays: self.month = idx+1 @@ -225,13 +235,20 @@ Here is the behavior of operations with relativedelta: div, mod = divmod(self.months*s, 12) self.months = mod*s self.years += div*s - if (self.hours or self.minutes or self.seconds or self.microseconds or - self.hour is not None or self.minute is not None or - self.second is not None or self.microsecond is not None): + if (self.hours or self.minutes or self.seconds or self.microseconds + or self.hour is not None or self.minute is not None or + self.second is not None or self.microsecond is not None): self._has_time = 1 else: self._has_time = 0 + @property + def weeks(self): + return self.days // 7 + @weeks.setter + def weeks(self, value): + self.days = self.days - (self.weeks * 7) + value*7 + def _set_months(self, months): self.months = months if abs(self.months) > 11: @@ -244,22 +261,24 @@ Here is the behavior of operations with relativedelta: def __add__(self, other): if isinstance(other, relativedelta): - return relativedelta(years=other.years+self.years, - months=other.months+self.months, - days=other.days+self.days, - hours=other.hours+self.hours, - minutes=other.minutes+self.minutes, - seconds=other.seconds+self.seconds, - microseconds=other.microseconds+self.microseconds, - leapdays=other.leapdays or self.leapdays, - year=other.year or self.year, - month=other.month or self.month, - day=other.day or self.day, - weekday=other.weekday or self.weekday, - hour=other.hour or self.hour, - minute=other.minute or self.minute, - second=other.second or self.second, - microsecond=other.microsecond or self.microsecond) + return self.__class__(years=other.years+self.years, + months=other.months+self.months, + days=other.days+self.days, + hours=other.hours+self.hours, + minutes=other.minutes+self.minutes, + seconds=other.seconds+self.seconds, + microseconds=(other.microseconds + + self.microseconds), + leapdays=other.leapdays or self.leapdays, + year=other.year or self.year, + month=other.month or self.month, + day=other.day or self.day, + weekday=other.weekday or self.weekday, + hour=other.hour or self.hour, + minute=other.minute or self.minute, + second=other.second or self.second, + microsecond=(other.microsecond or + self.microsecond)) if not isinstance(other, datetime.date): raise TypeError("unsupported type for add operation") elif self._has_time and not isinstance(other, datetime.datetime): @@ -295,9 +314,9 @@ Here is the behavior of operations with relativedelta: weekday, nth = self.weekday.weekday, self.weekday.n or 1 jumpdays = (abs(nth)-1)*7 if nth > 0: - jumpdays += (7-ret.weekday()+weekday)%7 + jumpdays += (7-ret.weekday()+weekday) % 7 else: - jumpdays += (ret.weekday()-weekday)%7 + jumpdays += (ret.weekday()-weekday) % 7 jumpdays *= -1 ret += datetime.timedelta(days=jumpdays) return ret @@ -311,7 +330,7 @@ Here is the behavior of operations with relativedelta: def __sub__(self, other): if not isinstance(other, relativedelta): raise TypeError("unsupported type for sub operation") - return relativedelta(years=self.years-other.years, + return self.__class__(years=self.years-other.years, months=self.months-other.months, days=self.days-other.days, hours=self.hours-other.hours, @@ -329,7 +348,7 @@ Here is the behavior of operations with relativedelta: microsecond=self.microsecond or other.microsecond) def __neg__(self): - return relativedelta(years=-self.years, + return self.__class__(years=-self.years, months=-self.months, days=-self.days, hours=-self.hours, @@ -363,10 +382,12 @@ Here is the behavior of operations with relativedelta: self.minute is None and self.second is None and self.microsecond is None) + # Compatibility with Python 2.x + __nonzero__ = __bool__ def __mul__(self, other): f = float(other) - return relativedelta(years=int(self.years*f), + return self.__class__(years=int(self.years*f), months=int(self.months*f), days=int(self.days*f), hours=int(self.hours*f), diff --git a/lib/dateutil/rrule.py b/lib/dateutil/rrule.py index ad4d3ba7..57b034a4 100644 --- a/lib/dateutil/rrule.py +++ b/lib/dateutil/rrule.py @@ -1,21 +1,19 @@ +# -*- coding: utf-8 -*- """ -Copyright (c) 2003-2010 Gustavo Niemeyer - -This module offers extensions to the standard Python -datetime module. +The rrule module offers a small, complete, and very fast, implementation of +the recurrence rules documented in the +`iCalendar RFC `_, +including support for caching of results. """ -__license__ = "Simplified BSD" - import itertools import datetime import calendar -try: - import _thread -except ImportError: - import thread as _thread import sys +from fractions import gcd + from six import advance_iterator, integer_types +from six.moves import _thread __all__ = ["rrule", "rruleset", "rrulestr", "YEARLY", "MONTHLY", "WEEKLY", "DAILY", @@ -23,7 +21,7 @@ __all__ = ["rrule", "rruleset", "rrulestr", "MO", "TU", "WE", "TH", "FR", "SA", "SU"] # Every mask is 7 days longer to handle cross-year weekly periods. -M366MASK = tuple([1]*31+[2]*29+[3]*31+[4]*30+[5]*31+[6]*30+ +M366MASK = tuple([1]*31+[2]*29+[3]*31+[4]*30+[5]*31+[6]*30 + [7]*31+[8]*31+[9]*30+[10]*31+[11]*30+[12]*31+[1]*7) M365MASK = list(M366MASK) M29, M30, M31 = list(range(1, 30)), list(range(1, 31)), list(range(1, 32)) @@ -39,6 +37,8 @@ del M29, M30, M31, M365MASK[59], MDAY365MASK[59], NMDAY365MASK[31] MDAY365MASK = tuple(MDAY365MASK) M365MASK = tuple(M365MASK) +FREQNAMES = ['YEARLY','MONTHLY','WEEKLY','DAILY','HOURLY','MINUTELY','SECONDLY'] + (YEARLY, MONTHLY, WEEKLY, @@ -51,6 +51,7 @@ M365MASK = tuple(M365MASK) easter = None parser = None + class weekday(object): __slots__ = ["weekday", "n"] @@ -83,12 +84,13 @@ class weekday(object): MO, TU, WE, TH, FR, SA, SU = weekdays = tuple([weekday(x) for x in range(7)]) + class rrulebase(object): def __init__(self, cache=False): if cache: self._cache = [] self._cache_lock = _thread.allocate_lock() - self._cache_gen = self._iter() + self._cache_gen = self._iter() self._cache_complete = False else: self._cache = None @@ -163,11 +165,17 @@ class rrulebase(object): # __len__() introduces a large performance penality. def count(self): + """ Returns the number of recurrences in this set. It will have go + trough the whole recurrence, if this hasn't been done before. """ if self._len is None: - for x in self: pass + for x in self: + pass return self._len def before(self, dt, inc=False): + """ Returns the last recurrence before the given datetime instance. The + inc keyword defines what happens if dt is an occurrence. With + inc=True, if dt itself is an occurrence, it will be returned. """ if self._cache_complete: gen = self._cache else: @@ -186,6 +194,9 @@ class rrulebase(object): return last def after(self, dt, inc=False): + """ Returns the first recurrence after the given datetime instance. The + inc keyword defines what happens if dt is an occurrence. With + inc=True, if dt itself is an occurrence, it will be returned. """ if self._cache_complete: gen = self._cache else: @@ -200,7 +211,52 @@ class rrulebase(object): return i return None - def between(self, after, before, inc=False): + def xafter(self, dt, count=None, inc=False): + """ + Generator which yields up to `count` recurrences after the given + datetime instance, equivalent to `after`. + + :param dt: + The datetime at which to start generating recurrences. + + :param count: + The maximum number of recurrences to generate. If `None` (default), + dates are generated until the recurrence rule is exhausted. + + :param inc: + If `dt` is an instance of the rule and `inc` is `True`, it is + included in the output. + + :yields: Yields a sequence of `datetime` objects. + """ + + if self._cache_complete: + gen = self._cache + else: + gen = self + + # Select the comparison function + if inc: + comp = lambda dc, dtc: dc >= dtc + else: + comp = lambda dc, dtc: dc > dtc + + # Generate dates + n = 0 + for d in gen: + if comp(d, dt): + yield d + + if count is not None: + n += 1 + if n >= count: + break + + def between(self, after, before, inc=False, count=1): + """ Returns all the occurrences of the rrule between after and before. + The inc keyword defines what happens if after and/or before are + themselves occurrences. With inc=True, they will be included in the + list, if they are found in the recurrence set. """ if self._cache_complete: gen = self._cache else: @@ -229,7 +285,93 @@ class rrulebase(object): l.append(i) return l + class rrule(rrulebase): + """ + That's the base of the rrule operation. It accepts all the keywords + defined in the RFC as its constructor parameters (except byday, + which was renamed to byweekday) and more. The constructor prototype is:: + + rrule(freq) + + Where freq must be one of YEARLY, MONTHLY, WEEKLY, DAILY, HOURLY, MINUTELY, + or SECONDLY. + + Additionally, it supports the following keyword arguments: + + :param cache: + If given, it must be a boolean value specifying to enable or disable + caching of results. If you will use the same rrule instance multiple + times, enabling caching will improve the performance considerably. + :param dtstart: + The recurrence start. Besides being the base for the recurrence, + missing parameters in the final recurrence instances will also be + extracted from this date. If not given, datetime.now() will be used + instead. + :param interval: + The interval between each freq iteration. For example, when using + YEARLY, an interval of 2 means once every two years, but with HOURLY, + it means once every two hours. The default interval is 1. + :param wkst: + The week start day. Must be one of the MO, TU, WE constants, or an + integer, specifying the first day of the week. This will affect + recurrences based on weekly periods. The default week start is got + from calendar.firstweekday(), and may be modified by + calendar.setfirstweekday(). + :param count: + How many occurrences will be generated. + :param until: + If given, this must be a datetime instance, that will specify the + limit of the recurrence. If a recurrence instance happens to be the + same as the datetime instance given in the until keyword, this will + be the last occurrence. + :param bysetpos: + If given, it must be either an integer, or a sequence of integers, + positive or negative. Each given integer will specify an occurrence + number, corresponding to the nth occurrence of the rule inside the + frequency period. For example, a bysetpos of -1 if combined with a + MONTHLY frequency, and a byweekday of (MO, TU, WE, TH, FR), will + result in the last work day of every month. + :param bymonth: + If given, it must be either an integer, or a sequence of integers, + meaning the months to apply the recurrence to. + :param bymonthday: + If given, it must be either an integer, or a sequence of integers, + meaning the month days to apply the recurrence to. + :param byyearday: + If given, it must be either an integer, or a sequence of integers, + meaning the year days to apply the recurrence to. + :param byweekno: + If given, it must be either an integer, or a sequence of integers, + meaning the week numbers to apply the recurrence to. Week numbers + have the meaning described in ISO8601, that is, the first week of + the year is that containing at least four days of the new year. + :param byweekday: + If given, it must be either an integer (0 == MO), a sequence of + integers, one of the weekday constants (MO, TU, etc), or a sequence + of these constants. When given, these variables will define the + weekdays where the recurrence will be applied. It's also possible to + use an argument n for the weekday instances, which will mean the nth + occurrence of this weekday in the period. For example, with MONTHLY, + or with YEARLY and BYMONTH, using FR(+1) in byweekday will specify the + first friday of the month where the recurrence happens. Notice that in + the RFC documentation, this is specified as BYDAY, but was renamed to + avoid the ambiguity of that keyword. + :param byhour: + If given, it must be either an integer, or a sequence of integers, + meaning the hours to apply the recurrence to. + :param byminute: + If given, it must be either an integer, or a sequence of integers, + meaning the minutes to apply the recurrence to. + :param bysecond: + If given, it must be either an integer, or a sequence of integers, + meaning the seconds to apply the recurrence to. + :param byeaster: + If given, it must be either an integer, or a sequence of integers, + positive or negative. Each integer will define an offset from the + Easter Sunday. Passing the offset 0 to byeaster will yield the Easter + Sunday itself. This is an extension to the RFC specification. + """ def __init__(self, freq, dtstart=None, interval=1, wkst=None, count=None, until=None, bysetpos=None, bymonth=None, bymonthday=None, byyearday=None, byeaster=None, @@ -249,15 +391,24 @@ class rrule(rrulebase): self._freq = freq self._interval = interval self._count = count + + # Cache the original byxxx rules, if they are provided, as the _byxxx + # attributes do not necessarily map to the inputs, and this can be + # a problem in generating the strings. Only store things if they've + # been supplied (the string retrieval will just use .get()) + self._original_rule = {} + if until and not isinstance(until, datetime.datetime): until = datetime.datetime.fromordinal(until.toordinal()) self._until = until + if wkst is None: self._wkst = calendar.firstweekday() elif isinstance(wkst, integer_types): self._wkst = wkst else: self._wkst = wkst.weekday + if bysetpos is None: self._bysetpos = None elif isinstance(bysetpos, integer_types): @@ -271,30 +422,47 @@ class rrule(rrulebase): if pos == 0 or not (-366 <= pos <= 366): raise ValueError("bysetpos must be between 1 and 366, " "or between -366 and -1") - if not (byweekno or byyearday or bymonthday or - byweekday is not None or byeaster is not None): + + if self._bysetpos: + self._original_rule['bysetpos'] = self._bysetpos + + if (byweekno is None and byyearday is None and bymonthday is None and + byweekday is None and byeaster is None): if freq == YEARLY: - if not bymonth: + if bymonth is None: bymonth = dtstart.month + self._original_rule['bymonth'] = None bymonthday = dtstart.day + self._original_rule['bymonthday'] = None elif freq == MONTHLY: bymonthday = dtstart.day + self._original_rule['bymonthday'] = None elif freq == WEEKLY: byweekday = dtstart.weekday() + self._original_rule['byweekday'] = None + # bymonth - if not bymonth: + if bymonth is None: self._bymonth = None - elif isinstance(bymonth, integer_types): - self._bymonth = (bymonth,) else: - self._bymonth = tuple(bymonth) + if isinstance(bymonth, integer_types): + bymonth = (bymonth,) + + self._bymonth = tuple(sorted(set(bymonth))) + + if 'bymonth' not in self._original_rule: + self._original_rule['bymonth'] = self._bymonth + # byyearday - if not byyearday: + if byyearday is None: self._byyearday = None - elif isinstance(byyearday, integer_types): - self._byyearday = (byyearday,) else: - self._byyearday = tuple(byyearday) + if isinstance(byyearday, integer_types): + byyearday = (byyearday,) + + self._byyearday = tuple(sorted(set(byyearday))) + self._original_rule['byyearday'] = self._byyearday + # byeaster if byeaster is not None: if not easter: @@ -302,90 +470,144 @@ class rrule(rrulebase): if isinstance(byeaster, integer_types): self._byeaster = (byeaster,) else: - self._byeaster = tuple(byeaster) + self._byeaster = tuple(sorted(byeaster)) + + self._original_rule['byeaster'] = self._byeaster else: self._byeaster = None - # bymonthay - if not bymonthday: + + # bymonthday + if bymonthday is None: self._bymonthday = () self._bynmonthday = () - elif isinstance(bymonthday, integer_types): - if bymonthday < 0: - self._bynmonthday = (bymonthday,) - self._bymonthday = () - else: - self._bymonthday = (bymonthday,) - self._bynmonthday = () else: - self._bymonthday = tuple([x for x in bymonthday if x > 0]) - self._bynmonthday = tuple([x for x in bymonthday if x < 0]) + if isinstance(bymonthday, integer_types): + bymonthday = (bymonthday,) + + bymonthday = set(bymonthday) # Ensure it's unique + + self._bymonthday = tuple(sorted([x for x in bymonthday if x > 0])) + self._bynmonthday = tuple(sorted([x for x in bymonthday if x < 0])) + + # Storing positive numbers first, then negative numbers + if 'bymonthday' not in self._original_rule: + self._original_rule['bymonthday'] = tuple( + itertools.chain(self._bymonthday, self._bynmonthday)) + # byweekno if byweekno is None: self._byweekno = None - elif isinstance(byweekno, integer_types): - self._byweekno = (byweekno,) else: - self._byweekno = tuple(byweekno) + if isinstance(byweekno, integer_types): + byweekno = (byweekno,) + + self._byweekno = tuple(sorted(set(byweekno))) + + self._original_rule['byweekno'] = self._byweekno + # byweekday / bynweekday if byweekday is None: self._byweekday = None self._bynweekday = None - elif isinstance(byweekday, integer_types): - self._byweekday = (byweekday,) - self._bynweekday = None - elif hasattr(byweekday, "n"): - if not byweekday.n or freq > MONTHLY: - self._byweekday = (byweekday.weekday,) - self._bynweekday = None - else: - self._bynweekday = ((byweekday.weekday, byweekday.n),) - self._byweekday = None else: - self._byweekday = [] - self._bynweekday = [] + # If it's one of the valid non-sequence types, convert to a + # single-element sequence before the iterator that builds the + # byweekday set. + if isinstance(byweekday, integer_types) or hasattr(byweekday, "n"): + byweekday = (byweekday,) + + self._byweekday = set() + self._bynweekday = set() for wday in byweekday: if isinstance(wday, integer_types): - self._byweekday.append(wday) + self._byweekday.add(wday) elif not wday.n or freq > MONTHLY: - self._byweekday.append(wday.weekday) + self._byweekday.add(wday.weekday) else: - self._bynweekday.append((wday.weekday, wday.n)) - self._byweekday = tuple(self._byweekday) - self._bynweekday = tuple(self._bynweekday) + self._bynweekday.add((wday.weekday, wday.n)) + if not self._byweekday: self._byweekday = None elif not self._bynweekday: self._bynweekday = None + + if self._byweekday is not None: + self._byweekday = tuple(sorted(self._byweekday)) + orig_byweekday = [weekday(x) for x in self._byweekday] + else: + orig_byweekday = tuple() + + if self._bynweekday is not None: + self._bynweekday = tuple(sorted(self._bynweekday)) + orig_bynweekday = [weekday(*x) for x in self._bynweekday] + else: + orig_bynweekday = tuple() + + if 'byweekday' not in self._original_rule: + self._original_rule['byweekday'] = tuple(itertools.chain( + orig_byweekday, orig_bynweekday)) + # byhour if byhour is None: if freq < HOURLY: - self._byhour = (dtstart.hour,) + self._byhour = set((dtstart.hour,)) else: self._byhour = None - elif isinstance(byhour, integer_types): - self._byhour = (byhour,) else: - self._byhour = tuple(byhour) + if isinstance(byhour, integer_types): + byhour = (byhour,) + + if freq == HOURLY: + self._byhour = self.__construct_byset(start=dtstart.hour, + byxxx=byhour, + base=24) + else: + self._byhour = set(byhour) + + self._byhour = tuple(sorted(self._byhour)) + self._original_rule['byhour'] = self._byhour + # byminute if byminute is None: if freq < MINUTELY: - self._byminute = (dtstart.minute,) + self._byminute = set((dtstart.minute,)) else: self._byminute = None - elif isinstance(byminute, integer_types): - self._byminute = (byminute,) else: - self._byminute = tuple(byminute) + if isinstance(byminute, integer_types): + byminute = (byminute,) + + if freq == MINUTELY: + self._byminute = self.__construct_byset(start=dtstart.minute, + byxxx=byminute, + base=60) + else: + self._byminute = set(byminute) + + self._byminute = tuple(sorted(self._byminute)) + self._original_rule['byminute'] = self._byminute + # bysecond if bysecond is None: if freq < SECONDLY: - self._bysecond = (dtstart.second,) + self._bysecond = ((dtstart.second,)) else: self._bysecond = None - elif isinstance(bysecond, integer_types): - self._bysecond = (bysecond,) else: - self._bysecond = tuple(bysecond) + if isinstance(bysecond, integer_types): + bysecond = (bysecond,) + + self._bysecond = set(bysecond) + + if freq == SECONDLY: + self._bysecond = self.__construct_byset(start=dtstart.second, + byxxx=bysecond, + base=60) + else: + self._bysecond = set(bysecond) + + self._bysecond = tuple(sorted(self._bysecond)) + self._original_rule['bysecond'] = self._bysecond if self._freq >= HOURLY: self._timeset = None @@ -395,11 +617,70 @@ class rrule(rrulebase): for minute in self._byminute: for second in self._bysecond: self._timeset.append( - datetime.time(hour, minute, second, - tzinfo=self._tzinfo)) + datetime.time(hour, minute, second, + tzinfo=self._tzinfo)) self._timeset.sort() self._timeset = tuple(self._timeset) + def __str__(self): + """ + Output a string that would generate this RRULE if passed to rrulestr. + This is mostly compatible with RFC2445, except for the + dateutil-specific extension BYEASTER. + """ + + output = [] + h, m, s = [None] * 3 + if self._dtstart: + output.append(self._dtstart.strftime('DTSTART:%Y%m%dT%H%M%S')) + h, m, s = self._dtstart.timetuple()[3:6] + + parts = ['FREQ=' + FREQNAMES[self._freq]] + if self._interval != 1: + parts.append('INTERVAL=' + str(self._interval)) + + if self._wkst: + parts.append('WKST=' + str(self._wkst)) + + if self._count: + parts.append('COUNT=' + str(self._count)) + + if self._original_rule.get('byweekday') is not None: + # The str() method on weekday objects doesn't generate + # RFC2445-compliant strings, so we should modify that. + original_rule = dict(self._original_rule) + wday_strings = [] + for wday in original_rule['byweekday']: + if wday.n: + wday_strings.append('{n:+d}{wday}'.format( + n=wday.n, + wday=repr(wday)[0:2])) + else: + wday_strings.append(repr(wday)) + + original_rule['byweekday'] = wday_strings + else: + original_rule = self._original_rule + + partfmt = '{name}={vals}' + for name, key in [('BYSETPOS', 'bysetpos'), + ('BYMONTH', 'bymonth'), + ('BYMONTHDAY', 'bymonthday'), + ('BYYEARDAY', 'byyearday'), + ('BYWEEKNO', 'byweekno'), + ('BYDAY', 'byweekday'), + ('BYHOUR', 'byhour'), + ('BYMINUTE', 'byminute'), + ('BYSECOND', 'bysecond'), + ('BYEASTER', 'byeaster')]: + value = original_rule.get(key) + if value: + parts.append(partfmt.format(name=name, vals=(','.join(str(v) + for v in value)))) + + output.append(';'.join(parts)) + return '\n'.join(output) + def _iter(self): year, month, day, hour, minute, second, weekday, yearday, _ = \ self._dtstart.timetuple() @@ -424,20 +705,20 @@ class rrule(rrulebase): ii = _iterinfo(self) ii.rebuild(year, month) - getdayset = {YEARLY:ii.ydayset, - MONTHLY:ii.mdayset, - WEEKLY:ii.wdayset, - DAILY:ii.ddayset, - HOURLY:ii.ddayset, - MINUTELY:ii.ddayset, - SECONDLY:ii.ddayset}[freq] - + getdayset = {YEARLY: ii.ydayset, + MONTHLY: ii.mdayset, + WEEKLY: ii.wdayset, + DAILY: ii.ddayset, + HOURLY: ii.ddayset, + MINUTELY: ii.ddayset, + SECONDLY: ii.ddayset}[freq] + if freq < HOURLY: timeset = self._timeset else: - gettimeset = {HOURLY:ii.htimeset, - MINUTELY:ii.mtimeset, - SECONDLY:ii.stimeset}[freq] + gettimeset = {HOURLY: ii.htimeset, + MINUTELY: ii.mtimeset, + SECONDLY: ii.stimeset}[freq] if ((freq >= HOURLY and self._byhour and hour not in self._byhour) or (freq >= MINUTELY and @@ -466,11 +747,10 @@ class rrule(rrulebase): ii.mdaymask[i] not in bymonthday and ii.nmdaymask[i] not in bynmonthday) or (byyearday and - ((i < ii.yearlen and i+1 not in byyearday - and -ii.yearlen+i not in byyearday) or - (i >= ii.yearlen and i+1-ii.yearlen not in byyearday - and -ii.nextyearlen+i-ii.yearlen - not in byyearday)))): + ((i < ii.yearlen and i+1 not in byyearday and + -ii.yearlen+i not in byyearday) or + (i >= ii.yearlen and i+1-ii.yearlen not in byyearday and + -ii.nextyearlen+i-ii.yearlen not in byyearday)))): dayset[i] = None filtered = True @@ -484,7 +764,7 @@ class rrule(rrulebase): daypos, timepos = divmod(pos-1, len(timeset)) try: i = [x for x in dayset[start:end] - if x is not None][daypos] + if x is not None][daypos] time = timeset[timepos] except IndexError: pass @@ -509,7 +789,7 @@ class rrule(rrulebase): else: for i in dayset[start:end]: if i is not None: - date = datetime.date.fromordinal(ii.yearordinal+i) + date = datetime.date.fromordinal(ii.yearordinal + i) for time in timeset: res = datetime.datetime.combine(date, time) if until and res > until: @@ -559,60 +839,86 @@ class rrule(rrulebase): if filtered: # Jump to one iteration before next day hour += ((23-hour)//interval)*interval - while True: - hour += interval - div, mod = divmod(hour, 24) - if div: - hour = mod - day += div - fixday = True - if not byhour or hour in byhour: - break + + if byhour: + ndays, hour = self.__mod_distance(value=hour, + byxxx=self._byhour, + base=24) + else: + ndays, hour = divmod(hour+interval, 24) + + if ndays: + day += ndays + fixday = True + timeset = gettimeset(hour, minute, second) elif freq == MINUTELY: if filtered: # Jump to one iteration before next day minute += ((1439-(hour*60+minute))//interval)*interval - while True: - minute += interval - div, mod = divmod(minute, 60) + + valid = False + rep_rate = (24*60) + for j in range(rep_rate // gcd(interval, rep_rate)): + if byminute: + nhours, minute = \ + self.__mod_distance(value=minute, + byxxx=self._byminute, + base=60) + else: + nhours, minute = divmod(minute+interval, 60) + + div, hour = divmod(hour+nhours, 24) if div: - minute = mod - hour += div - div, mod = divmod(hour, 24) - if div: - hour = mod - day += div - fixday = True - filtered = False - if ((not byhour or hour in byhour) and - (not byminute or minute in byminute)): + day += div + fixday = True + filtered = False + + if not byhour or hour in byhour: + valid = True break + + if not valid: + raise ValueError('Invalid combination of interval and ' + + 'byhour resulting in empty rule.') + timeset = gettimeset(hour, minute, second) elif freq == SECONDLY: if filtered: # Jump to one iteration before next day - second += (((86399-(hour*3600+minute*60+second)) - //interval)*interval) - while True: - second += self._interval - div, mod = divmod(second, 60) + second += (((86399 - (hour * 3600 + minute * 60 + second)) + // interval) * interval) + + rep_rate = (24 * 3600) + valid = False + for j in range(0, rep_rate // gcd(interval, rep_rate)): + if bysecond: + nminutes, second = \ + self.__mod_distance(value=second, + byxxx=self._bysecond, + base=60) + else: + nminutes, second = divmod(second+interval, 60) + + div, minute = divmod(minute+nminutes, 60) if div: - second = mod - minute += div - div, mod = divmod(minute, 60) + hour += div + div, hour = divmod(hour, 24) if div: - minute = mod - hour += div - div, mod = divmod(hour, 24) - if div: - hour = mod - day += div - fixday = True + day += div + fixday = True + if ((not byhour or hour in byhour) and - (not byminute or minute in byminute) and - (not bysecond or second in bysecond)): + (not byminute or minute in byminute) and + (not bysecond or second in bysecond)): + valid = True break + + if not valid: + raise ValueError('Invalid combination of interval, ' + + 'byhour and byminute resulting in empty' + + ' rule.') + timeset = gettimeset(hour, minute, second) if fixday and day > 28: @@ -630,6 +936,86 @@ class rrule(rrulebase): daysinmonth = calendar.monthrange(year, month)[1] ii.rebuild(year, month) + def __construct_byset(self, start, byxxx, base): + """ + If a `BYXXX` sequence is passed to the constructor at the same level as + `FREQ` (e.g. `FREQ=HOURLY,BYHOUR={2,4,7},INTERVAL=3`), there are some + specifications which cannot be reached given some starting conditions. + + This occurs whenever the interval is not coprime with the base of a + given unit and the difference between the starting position and the + ending position is not coprime with the greatest common denominator + between the interval and the base. For example, with a FREQ of hourly + starting at 17:00 and an interval of 4, the only valid values for + BYHOUR would be {21, 1, 5, 9, 13, 17}, because 4 and 24 are not + coprime. + + :param start: + Specifies the starting position. + :param byxxx: + An iterable containing the list of allowed values. + :param base: + The largest allowable value for the specified frequency (e.g. + 24 hours, 60 minutes). + + This does not preserve the type of the iterable, returning a set, since + the values should be unique and the order is irrelevant, this will + speed up later lookups. + + In the event of an empty set, raises a :exception:`ValueError`, as this + results in an empty rrule. + """ + + cset = set() + + # Support a single byxxx value. + if isinstance(byxxx, integer_types): + byxxx = (byxxx, ) + + for num in byxxx: + i_gcd = gcd(self._interval, base) + # Use divmod rather than % because we need to wrap negative nums. + if i_gcd == 1 or divmod(num - start, i_gcd)[1] == 0: + cset.add(num) + + if len(cset) == 0: + raise ValueError("Invalid rrule byxxx generates an empty set.") + + return cset + + def __mod_distance(self, value, byxxx, base): + """ + Calculates the next value in a sequence where the `FREQ` parameter is + specified along with a `BYXXX` parameter at the same "level" + (e.g. `HOURLY` specified with `BYHOUR`). + + :param value: + The old value of the component. + :param byxxx: + The `BYXXX` set, which should have been generated by + `rrule._construct_byset`, or something else which checks that a + valid rule is present. + :param base: + The largest allowable value for the specified frequency (e.g. + 24 hours, 60 minutes). + + If a valid value is not found after `base` iterations (the maximum + number before the sequence would start to repeat), this raises a + :exception:`ValueError`, as no valid values were found. + + This returns a tuple of `divmod(n*interval, base)`, where `n` is the + smallest number of `interval` repetitions until the next specified + value in `byxxx` is found. + """ + accumulator = 0 + for ii in range(1, base + 1): + # Using divmod() over % to account for negative intervals + div, value = divmod(value + self._interval, base) + accumulator += div + if value in byxxx: + return (accumulator, value) + + class _iterinfo(object): __slots__ = ["rrule", "lastyear", "lastmonth", "yearlen", "nextyearlen", "yearordinal", "yearweekday", @@ -645,8 +1031,8 @@ class _iterinfo(object): # Every mask is 7 days longer to handle cross-year weekly periods. rr = self.rrule if year != self.lastyear: - self.yearlen = 365+calendar.isleap(year) - self.nextyearlen = 365+calendar.isleap(year+1) + self.yearlen = 365 + calendar.isleap(year) + self.nextyearlen = 365 + calendar.isleap(year + 1) firstyday = datetime.date(year, 1, 1) self.yearordinal = firstyday.toordinal() self.yearweekday = firstyday.weekday() @@ -669,13 +1055,13 @@ class _iterinfo(object): self.wnomask = None else: self.wnomask = [0]*(self.yearlen+7) - #no1wkst = firstwkst = self.wdaymask.index(rr._wkst) - no1wkst = firstwkst = (7-self.yearweekday+rr._wkst)%7 + # no1wkst = firstwkst = self.wdaymask.index(rr._wkst) + no1wkst = firstwkst = (7-self.yearweekday+rr._wkst) % 7 if no1wkst >= 4: no1wkst = 0 # Number of days in the year, plus the days we got # from last year. - wyearlen = self.yearlen+(self.yearweekday-rr._wkst)%7 + wyearlen = self.yearlen+(self.yearweekday-rr._wkst) % 7 else: # Number of days in the year, minus the days we # left in last year. @@ -721,22 +1107,22 @@ class _iterinfo(object): # this year. if -1 not in rr._byweekno: lyearweekday = datetime.date(year-1, 1, 1).weekday() - lno1wkst = (7-lyearweekday+rr._wkst)%7 + lno1wkst = (7-lyearweekday+rr._wkst) % 7 lyearlen = 365+calendar.isleap(year-1) if lno1wkst >= 4: lno1wkst = 0 - lnumweeks = 52+(lyearlen+ - (lyearweekday-rr._wkst)%7)%7//4 + lnumweeks = 52+(lyearlen + + (lyearweekday-rr._wkst) % 7) % 7//4 else: - lnumweeks = 52+(self.yearlen-no1wkst)%7//4 + lnumweeks = 52+(self.yearlen-no1wkst) % 7//4 else: lnumweeks = -1 if lnumweeks in rr._byweekno: for i in range(no1wkst): self.wnomask[i] = 1 - if (rr._bynweekday and - (month != self.lastmonth or year != self.lastyear)): + if (rr._bynweekday and (month != self.lastmonth or + year != self.lastyear)): ranges = [] if rr._freq == YEARLY: if rr._bymonth: @@ -755,10 +1141,10 @@ class _iterinfo(object): for wday, n in rr._bynweekday: if n < 0: i = last+(n+1)*7 - i -= (self.wdaymask[i]-wday)%7 + i -= (self.wdaymask[i]-wday) % 7 else: i = first+(n-1)*7 - i += (7-self.wdaymask[i]+wday)%7 + i += (7-self.wdaymask[i]+wday) % 7 if first <= i <= last: self.nwdaymask[i] = 1 @@ -775,50 +1161,50 @@ class _iterinfo(object): return list(range(self.yearlen)), 0, self.yearlen def mdayset(self, year, month, day): - set = [None]*self.yearlen + dset = [None]*self.yearlen start, end = self.mrange[month-1:month+1] for i in range(start, end): - set[i] = i - return set, start, end + dset[i] = i + return dset, start, end def wdayset(self, year, month, day): # We need to handle cross-year weeks here. - set = [None]*(self.yearlen+7) + dset = [None]*(self.yearlen+7) i = datetime.date(year, month, day).toordinal()-self.yearordinal start = i for j in range(7): - set[i] = i + dset[i] = i i += 1 - #if (not (0 <= i < self.yearlen) or + # if (not (0 <= i < self.yearlen) or # self.wdaymask[i] == self.rrule._wkst): # This will cross the year boundary, if necessary. if self.wdaymask[i] == self.rrule._wkst: break - return set, start, i + return dset, start, i def ddayset(self, year, month, day): - set = [None]*self.yearlen - i = datetime.date(year, month, day).toordinal()-self.yearordinal - set[i] = i - return set, i, i+1 + dset = [None] * self.yearlen + i = datetime.date(year, month, day).toordinal() - self.yearordinal + dset[i] = i + return dset, i, i + 1 def htimeset(self, hour, minute, second): - set = [] + tset = [] rr = self.rrule for minute in rr._byminute: for second in rr._bysecond: - set.append(datetime.time(hour, minute, second, - tzinfo=rr._tzinfo)) - set.sort() - return set + tset.append(datetime.time(hour, minute, second, + tzinfo=rr._tzinfo)) + tset.sort() + return tset def mtimeset(self, hour, minute, second): - set = [] + tset = [] rr = self.rrule for second in rr._bysecond: - set.append(datetime.time(hour, minute, second, tzinfo=rr._tzinfo)) - set.sort() - return set + tset.append(datetime.time(hour, minute, second, tzinfo=rr._tzinfo)) + tset.sort() + return tset def stimeset(self, hour, minute, second): return (datetime.time(hour, minute, second, @@ -826,6 +1212,12 @@ class _iterinfo(object): class rruleset(rrulebase): + """ The rruleset type allows more complex recurrence setups, mixing + multiple rules, dates, exclusion rules, and exclusion dates. The type + constructor takes the following keyword arguments: + + :param cache: If True, caching of results will be enabled, improving + performance of multiple queries considerably. """ class _genitem(object): def __init__(self, genlist, gen): @@ -865,15 +1257,26 @@ class rruleset(rrulebase): self._exdate = [] def rrule(self, rrule): + """ Include the given :py:class:`rrule` instance in the recurrence set + generation. """ self._rrule.append(rrule) def rdate(self, rdate): + """ Include the given :py:class:`datetime` instance in the recurrence + set generation. """ self._rdate.append(rdate) def exrule(self, exrule): + """ Include the given rrule instance in the recurrence set exclusion + list. Dates which are part of the given recurrence rules will not + be generated, even if some inclusive rrule or rdate matches them. + """ self._exrule.append(exrule) def exdate(self, exdate): + """ Include the given datetime instance in the recurrence set + exclusion list. Dates included that way will not be generated, + even if some inclusive rrule or rdate matches them. """ self._exdate.append(exdate) def _iter(self): @@ -905,6 +1308,7 @@ class rruleset(rrulebase): rlist.sort() self._len = total + class _rrulestr(object): _freq_map = {"YEARLY": YEARLY, @@ -915,7 +1319,8 @@ class _rrulestr(object): "MINUTELY": MINUTELY, "SECONDLY": SECONDLY} - _weekday_map = {"MO":0,"TU":1,"WE":2,"TH":3,"FR":4,"SA":5,"SU":6} + _weekday_map = {"MO": 0, "TU": 1, "WE": 2, "TH": 3, + "FR": 4, "SA": 5, "SU": 6} def _handle_int(self, rrkwargs, name, value, **kwargs): rrkwargs[name.lower()] = int(value) @@ -923,17 +1328,17 @@ class _rrulestr(object): def _handle_int_list(self, rrkwargs, name, value, **kwargs): rrkwargs[name.lower()] = [int(x) for x in value.split(',')] - _handle_INTERVAL = _handle_int - _handle_COUNT = _handle_int - _handle_BYSETPOS = _handle_int_list - _handle_BYMONTH = _handle_int_list + _handle_INTERVAL = _handle_int + _handle_COUNT = _handle_int + _handle_BYSETPOS = _handle_int_list + _handle_BYMONTH = _handle_int_list _handle_BYMONTHDAY = _handle_int_list - _handle_BYYEARDAY = _handle_int_list - _handle_BYEASTER = _handle_int_list - _handle_BYWEEKNO = _handle_int_list - _handle_BYHOUR = _handle_int_list - _handle_BYMINUTE = _handle_int_list - _handle_BYSECOND = _handle_int_list + _handle_BYYEARDAY = _handle_int_list + _handle_BYEASTER = _handle_int_list + _handle_BYWEEKNO = _handle_int_list + _handle_BYHOUR = _handle_int_list + _handle_BYMINUTE = _handle_int_list + _handle_BYSECOND = _handle_int_list def _handle_FREQ(self, rrkwargs, name, value, **kwargs): rrkwargs["freq"] = self._freq_map[value] @@ -944,23 +1349,34 @@ class _rrulestr(object): from dateutil import parser try: rrkwargs["until"] = parser.parse(value, - ignoretz=kwargs.get("ignoretz"), - tzinfos=kwargs.get("tzinfos")) + ignoretz=kwargs.get("ignoretz"), + tzinfos=kwargs.get("tzinfos")) except ValueError: raise ValueError("invalid until date") def _handle_WKST(self, rrkwargs, name, value, **kwargs): rrkwargs["wkst"] = self._weekday_map[value] - def _handle_BYWEEKDAY(self, rrkwargs, name, value, **kwarsg): + def _handle_BYWEEKDAY(self, rrkwargs, name, value, **kwargs): + """ + Two ways to specify this: +1MO or MO(+1) + """ l = [] for wday in value.split(','): - for i in range(len(wday)): - if wday[i] not in '+-0123456789': - break - n = wday[:i] or None - w = wday[i:] - if n: n = int(n) + if '(' in wday: + # If it's of the form TH(+1), etc. + splt = wday.split('(') + w = splt[0] + n = int(splt[1][:-1]) + else: + # If it's of the form +1MO + for i in range(len(wday)): + if wday[i] not in '+-0123456789': + break + n = wday[:i] or None + w = wday[i:] + if n: + n = int(n) l.append(weekdays[self._weekday_map[w]](n)) rrkwargs["byweekday"] = l @@ -1021,8 +1437,8 @@ class _rrulestr(object): i += 1 else: lines = s.split() - if (not forceset and len(lines) == 1 and - (s.find(':') == -1 or s.startswith('RRULE:'))): + if (not forceset and len(lines) == 1 and (s.find(':') == -1 or + s.startswith('RRULE:'))): return self._parse_rfc_rrule(lines[0], cache=cache, dtstart=dtstart, ignoretz=ignoretz, tzinfos=tzinfos) @@ -1071,32 +1487,32 @@ class _rrulestr(object): tzinfos=tzinfos) else: raise ValueError("unsupported property: "+name) - if (forceset or len(rrulevals) > 1 or - rdatevals or exrulevals or exdatevals): + if (forceset or len(rrulevals) > 1 or rdatevals + or exrulevals or exdatevals): if not parser and (rdatevals or exdatevals): from dateutil import parser - set = rruleset(cache=cache) + rset = rruleset(cache=cache) for value in rrulevals: - set.rrule(self._parse_rfc_rrule(value, dtstart=dtstart, - ignoretz=ignoretz, - tzinfos=tzinfos)) - for value in rdatevals: - for datestr in value.split(','): - set.rdate(parser.parse(datestr, - ignoretz=ignoretz, - tzinfos=tzinfos)) - for value in exrulevals: - set.exrule(self._parse_rfc_rrule(value, dtstart=dtstart, + rset.rrule(self._parse_rfc_rrule(value, dtstart=dtstart, ignoretz=ignoretz, tzinfos=tzinfos)) - for value in exdatevals: + for value in rdatevals: for datestr in value.split(','): - set.exdate(parser.parse(datestr, + rset.rdate(parser.parse(datestr, ignoretz=ignoretz, tzinfos=tzinfos)) + for value in exrulevals: + rset.exrule(self._parse_rfc_rrule(value, dtstart=dtstart, + ignoretz=ignoretz, + tzinfos=tzinfos)) + for value in exdatevals: + for datestr in value.split(','): + rset.exdate(parser.parse(datestr, + ignoretz=ignoretz, + tzinfos=tzinfos)) if compatible and dtstart: - set.rdate(dtstart) - return set + rset.rdate(dtstart) + return rset else: return self._parse_rfc_rrule(rrulevals[0], dtstart=dtstart, diff --git a/lib/dateutil/tz.py b/lib/dateutil/tz.py index e849fc24..444775e7 100644 --- a/lib/dateutil/tz.py +++ b/lib/dateutil/tz.py @@ -1,19 +1,25 @@ +# -*- coding: utf-8 -*- """ -Copyright (c) 2003-2007 Gustavo Niemeyer - -This module offers extensions to the standard Python -datetime module. +This module offers timezone implementations subclassing the abstract +:py:`datetime.tzinfo` type. There are classes to handle tzfile format files +(usually are in :file:`/etc/localtime`, :file:`/usr/share/zoneinfo`, etc), TZ +environment string (in all known formats), given ranges (with help from +relative deltas), local machine timezone, fixed offset timezone, and UTC +timezone. """ -__license__ = "Simplified BSD" - -from six import string_types, PY3 - import datetime import struct import time import sys import os +from six import string_types, PY3 + +try: + from dateutil.tzwin import tzwin, tzwinlocal +except ImportError: + tzwin = tzwinlocal = None + relativedelta = None parser = None rrule = None @@ -21,32 +27,31 @@ rrule = None __all__ = ["tzutc", "tzoffset", "tzlocal", "tzfile", "tzrange", "tzstr", "tzical", "tzwin", "tzwinlocal", "gettz"] -try: - from dateutil.tzwin import tzwin, tzwinlocal -except (ImportError, OSError): - tzwin, tzwinlocal = None, None -def tzname_in_python2(myfunc): +def tzname_in_python2(namefunc): """Change unicode output into bytestrings in Python 2 tzname() API changed in Python 3. It used to return bytes, but was changed to unicode strings """ - def inner_func(*args, **kwargs): - if PY3: - return myfunc(*args, **kwargs) - else: - return myfunc(*args, **kwargs).encode() - return inner_func + def adjust_encoding(*args, **kwargs): + name = namefunc(*args, **kwargs) + if name is not None and not PY3: + name = name.encode() + + return name + + return adjust_encoding ZERO = datetime.timedelta(0) EPOCHORDINAL = datetime.datetime.utcfromtimestamp(0).toordinal() + class tzutc(datetime.tzinfo): def utcoffset(self, dt): return ZERO - + def dst(self, dt): return ZERO @@ -66,6 +71,7 @@ class tzutc(datetime.tzinfo): __reduce__ = object.__reduce__ + class tzoffset(datetime.tzinfo): def __init__(self, name, offset): @@ -96,6 +102,7 @@ class tzoffset(datetime.tzinfo): __reduce__ = object.__reduce__ + class tzlocal(datetime.tzinfo): _std_offset = datetime.timedelta(seconds=-time.timezone) @@ -123,25 +130,25 @@ class tzlocal(datetime.tzinfo): def _isdst(self, dt): # We can't use mktime here. It is unstable when deciding if # the hour near to a change is DST or not. - # + # # timestamp = time.mktime((dt.year, dt.month, dt.day, dt.hour, # dt.minute, dt.second, dt.weekday(), 0, -1)) # return time.localtime(timestamp).tm_isdst # # The code above yields the following result: # - #>>> import tz, datetime - #>>> t = tz.tzlocal() - #>>> datetime.datetime(2003,2,15,23,tzinfo=t).tzname() - #'BRDT' - #>>> datetime.datetime(2003,2,16,0,tzinfo=t).tzname() - #'BRST' - #>>> datetime.datetime(2003,2,15,23,tzinfo=t).tzname() - #'BRST' - #>>> datetime.datetime(2003,2,15,22,tzinfo=t).tzname() - #'BRDT' - #>>> datetime.datetime(2003,2,15,23,tzinfo=t).tzname() - #'BRDT' + # >>> import tz, datetime + # >>> t = tz.tzlocal() + # >>> datetime.datetime(2003,2,15,23,tzinfo=t).tzname() + # 'BRDT' + # >>> datetime.datetime(2003,2,16,0,tzinfo=t).tzname() + # 'BRST' + # >>> datetime.datetime(2003,2,15,23,tzinfo=t).tzname() + # 'BRST' + # >>> datetime.datetime(2003,2,15,22,tzinfo=t).tzname() + # 'BRDT' + # >>> datetime.datetime(2003,2,15,23,tzinfo=t).tzname() + # 'BRDT' # # Here is a more stable implementation: # @@ -166,6 +173,7 @@ class tzlocal(datetime.tzinfo): __reduce__ = object.__reduce__ + class _ttinfo(object): __slots__ = ["offset", "delta", "isdst", "abbr", "isstd", "isgmt"] @@ -205,15 +213,20 @@ class _ttinfo(object): if name in state: setattr(self, name, state[name]) + class tzfile(datetime.tzinfo): # http://www.twinsun.com/tz/tz-link.htm # ftp://ftp.iana.org/tz/tz*.tar.gz - - def __init__(self, fileobj): + + def __init__(self, fileobj, filename=None): + file_opened_here = False if isinstance(fileobj, string_types): self._filename = fileobj fileobj = open(fileobj, 'rb') + file_opened_here = True + elif filename is not None: + self._filename = filename elif hasattr(fileobj, "name"): self._filename = fileobj.name else: @@ -228,125 +241,128 @@ class tzfile(datetime.tzinfo): # six four-byte values of type long, written in a # ``standard'' byte order (the high-order byte # of the value is written first). + try: + if fileobj.read(4).decode() != "TZif": + raise ValueError("magic not found") - if fileobj.read(4).decode() != "TZif": - raise ValueError("magic not found") + fileobj.read(16) - fileobj.read(16) + ( + # The number of UTC/local indicators stored in the file. + ttisgmtcnt, - ( - # The number of UTC/local indicators stored in the file. - ttisgmtcnt, + # The number of standard/wall indicators stored in the file. + ttisstdcnt, - # The number of standard/wall indicators stored in the file. - ttisstdcnt, - - # The number of leap seconds for which data is - # stored in the file. - leapcnt, + # The number of leap seconds for which data is + # stored in the file. + leapcnt, - # The number of "transition times" for which data - # is stored in the file. - timecnt, + # The number of "transition times" for which data + # is stored in the file. + timecnt, - # The number of "local time types" for which data - # is stored in the file (must not be zero). - typecnt, + # The number of "local time types" for which data + # is stored in the file (must not be zero). + typecnt, - # The number of characters of "time zone - # abbreviation strings" stored in the file. - charcnt, + # The number of characters of "time zone + # abbreviation strings" stored in the file. + charcnt, - ) = struct.unpack(">6l", fileobj.read(24)) + ) = struct.unpack(">6l", fileobj.read(24)) - # The above header is followed by tzh_timecnt four-byte - # values of type long, sorted in ascending order. - # These values are written in ``standard'' byte order. - # Each is used as a transition time (as returned by - # time(2)) at which the rules for computing local time - # change. + # The above header is followed by tzh_timecnt four-byte + # values of type long, sorted in ascending order. + # These values are written in ``standard'' byte order. + # Each is used as a transition time (as returned by + # time(2)) at which the rules for computing local time + # change. - if timecnt: - self._trans_list = struct.unpack(">%dl" % timecnt, - fileobj.read(timecnt*4)) - else: - self._trans_list = [] + if timecnt: + self._trans_list = struct.unpack(">%dl" % timecnt, + fileobj.read(timecnt*4)) + else: + self._trans_list = [] - # Next come tzh_timecnt one-byte values of type unsigned - # char; each one tells which of the different types of - # ``local time'' types described in the file is associated - # with the same-indexed transition time. These values - # serve as indices into an array of ttinfo structures that - # appears next in the file. - - if timecnt: - self._trans_idx = struct.unpack(">%dB" % timecnt, - fileobj.read(timecnt)) - else: - self._trans_idx = [] - - # Each ttinfo structure is written as a four-byte value - # for tt_gmtoff of type long, in a standard byte - # order, followed by a one-byte value for tt_isdst - # and a one-byte value for tt_abbrind. In each - # structure, tt_gmtoff gives the number of - # seconds to be added to UTC, tt_isdst tells whether - # tm_isdst should be set by localtime(3), and - # tt_abbrind serves as an index into the array of - # time zone abbreviation characters that follow the - # ttinfo structure(s) in the file. + # Next come tzh_timecnt one-byte values of type unsigned + # char; each one tells which of the different types of + # ``local time'' types described in the file is associated + # with the same-indexed transition time. These values + # serve as indices into an array of ttinfo structures that + # appears next in the file. - ttinfo = [] + if timecnt: + self._trans_idx = struct.unpack(">%dB" % timecnt, + fileobj.read(timecnt)) + else: + self._trans_idx = [] - for i in range(typecnt): - ttinfo.append(struct.unpack(">lbb", fileobj.read(6))) + # Each ttinfo structure is written as a four-byte value + # for tt_gmtoff of type long, in a standard byte + # order, followed by a one-byte value for tt_isdst + # and a one-byte value for tt_abbrind. In each + # structure, tt_gmtoff gives the number of + # seconds to be added to UTC, tt_isdst tells whether + # tm_isdst should be set by localtime(3), and + # tt_abbrind serves as an index into the array of + # time zone abbreviation characters that follow the + # ttinfo structure(s) in the file. - abbr = fileobj.read(charcnt).decode() + ttinfo = [] - # Then there are tzh_leapcnt pairs of four-byte - # values, written in standard byte order; the - # first value of each pair gives the time (as - # returned by time(2)) at which a leap second - # occurs; the second gives the total number of - # leap seconds to be applied after the given time. - # The pairs of values are sorted in ascending order - # by time. + for i in range(typecnt): + ttinfo.append(struct.unpack(">lbb", fileobj.read(6))) - # Not used, for now - if leapcnt: - leap = struct.unpack(">%dl" % (leapcnt*2), - fileobj.read(leapcnt*8)) + abbr = fileobj.read(charcnt).decode() - # Then there are tzh_ttisstdcnt standard/wall - # indicators, each stored as a one-byte value; - # they tell whether the transition times associated - # with local time types were specified as standard - # time or wall clock time, and are used when - # a time zone file is used in handling POSIX-style - # time zone environment variables. + # Then there are tzh_leapcnt pairs of four-byte + # values, written in standard byte order; the + # first value of each pair gives the time (as + # returned by time(2)) at which a leap second + # occurs; the second gives the total number of + # leap seconds to be applied after the given time. + # The pairs of values are sorted in ascending order + # by time. - if ttisstdcnt: - isstd = struct.unpack(">%db" % ttisstdcnt, - fileobj.read(ttisstdcnt)) + # Not used, for now + # if leapcnt: + # leap = struct.unpack(">%dl" % (leapcnt*2), + # fileobj.read(leapcnt*8)) - # Finally, there are tzh_ttisgmtcnt UTC/local - # indicators, each stored as a one-byte value; - # they tell whether the transition times associated - # with local time types were specified as UTC or - # local time, and are used when a time zone file - # is used in handling POSIX-style time zone envi- - # ronment variables. + # Then there are tzh_ttisstdcnt standard/wall + # indicators, each stored as a one-byte value; + # they tell whether the transition times associated + # with local time types were specified as standard + # time or wall clock time, and are used when + # a time zone file is used in handling POSIX-style + # time zone environment variables. - if ttisgmtcnt: - isgmt = struct.unpack(">%db" % ttisgmtcnt, - fileobj.read(ttisgmtcnt)) + if ttisstdcnt: + isstd = struct.unpack(">%db" % ttisstdcnt, + fileobj.read(ttisstdcnt)) - # ** Everything has been read ** + # Finally, there are tzh_ttisgmtcnt UTC/local + # indicators, each stored as a one-byte value; + # they tell whether the transition times associated + # with local time types were specified as UTC or + # local time, and are used when a time zone file + # is used in handling POSIX-style time zone envi- + # ronment variables. + + if ttisgmtcnt: + isgmt = struct.unpack(">%db" % ttisgmtcnt, + fileobj.read(ttisgmtcnt)) + + # ** Everything has been read ** + finally: + if file_opened_here: + fileobj.close() # Build ttinfo list self._ttinfo_list = [] for i in range(typecnt): - gmtoff, isdst, abbrind = ttinfo[i] + gmtoff, isdst, abbrind = ttinfo[i] # Round to full-minutes if that's not the case. Python's # datetime doesn't accept sub-minute timezones. Check # http://python.org/sf/1447945 for some information. @@ -464,7 +480,7 @@ class tzfile(datetime.tzinfo): # However, this class stores historical changes in the # dst offset, so I belive that this wouldn't be the right # way to implement this. - + @tzname_in_python2 def tzname(self, dt): if not self._ttinfo_std: @@ -481,7 +497,6 @@ class tzfile(datetime.tzinfo): def __ne__(self, other): return not self.__eq__(other) - def __repr__(self): return "%s(%s)" % (self.__class__.__name__, repr(self._filename)) @@ -490,8 +505,8 @@ class tzfile(datetime.tzinfo): raise ValueError("Unpickable %s class" % self.__class__.__name__) return (self.__class__, (self._filename,)) -class tzrange(datetime.tzinfo): +class tzrange(datetime.tzinfo): def __init__(self, stdabbr, stdoffset=None, dstabbr=None, dstoffset=None, start=None, end=None): @@ -512,12 +527,12 @@ class tzrange(datetime.tzinfo): self._dst_offset = ZERO if dstabbr and start is None: self._start_delta = relativedelta.relativedelta( - hours=+2, month=4, day=1, weekday=relativedelta.SU(+1)) + hours=+2, month=4, day=1, weekday=relativedelta.SU(+1)) else: self._start_delta = start if dstabbr and end is None: self._end_delta = relativedelta.relativedelta( - hours=+1, month=10, day=31, weekday=relativedelta.SU(-1)) + hours=+1, month=10, day=31, weekday=relativedelta.SU(-1)) else: self._end_delta = end @@ -570,8 +585,9 @@ class tzrange(datetime.tzinfo): __reduce__ = object.__reduce__ + class tzstr(tzrange): - + def __init__(self, s): global parser if not parser: @@ -645,9 +661,10 @@ class tzstr(tzrange): def __repr__(self): return "%s(%s)" % (self.__class__.__name__, repr(self._s)) + class _tzicalvtzcomp(object): def __init__(self, tzoffsetfrom, tzoffsetto, isdst, - tzname=None, rrule=None): + tzname=None, rrule=None): self.tzoffsetfrom = datetime.timedelta(seconds=tzoffsetfrom) self.tzoffsetto = datetime.timedelta(seconds=tzoffsetto) self.tzoffsetdiff = self.tzoffsetto-self.tzoffsetfrom @@ -655,6 +672,7 @@ class _tzicalvtzcomp(object): self.tzname = tzname self.rrule = rrule + class _tzicalvtz(datetime.tzinfo): def __init__(self, tzid, comps=[]): self._tzid = tzid @@ -718,6 +736,7 @@ class _tzicalvtz(datetime.tzinfo): __reduce__ = object.__reduce__ + class tzical(object): def __init__(self, fileobj): global rrule @@ -726,7 +745,8 @@ class tzical(object): if isinstance(fileobj, string_types): self._s = fileobj - fileobj = open(fileobj, 'r') # ical should be encoded in UTF-8 with CRLF + # ical should be encoded in UTF-8 with CRLF + fileobj = open(fileobj, 'r') elif hasattr(fileobj, "name"): self._s = fileobj.name else: @@ -754,7 +774,7 @@ class tzical(object): if not s: raise ValueError("empty offset") if s[0] in ('+', '-'): - signal = (-1, +1)[s[0]=='+'] + signal = (-1, +1)[s[0] == '+'] s = s[1:] else: signal = +1 @@ -815,7 +835,8 @@ class tzical(object): if not tzid: raise ValueError("mandatory TZID not found") if not comps: - raise ValueError("at least one component is needed") + raise ValueError( + "at least one component is needed") # Process vtimezone self._vtz[tzid] = _tzicalvtz(tzid, comps) invtz = False @@ -823,9 +844,11 @@ class tzical(object): if not founddtstart: raise ValueError("mandatory DTSTART not found") if tzoffsetfrom is None: - raise ValueError("mandatory TZOFFSETFROM not found") + raise ValueError( + "mandatory TZOFFSETFROM not found") if tzoffsetto is None: - raise ValueError("mandatory TZOFFSETFROM not found") + raise ValueError( + "mandatory TZOFFSETFROM not found") # Process component rr = None if rrulelines: @@ -848,15 +871,18 @@ class tzical(object): rrulelines.append(line) elif name == "TZOFFSETFROM": if parms: - raise ValueError("unsupported %s parm: %s "%(name, parms[0])) + raise ValueError( + "unsupported %s parm: %s " % (name, parms[0])) tzoffsetfrom = self._parse_offset(value) elif name == "TZOFFSETTO": if parms: - raise ValueError("unsupported TZOFFSETTO parm: "+parms[0]) + raise ValueError( + "unsupported TZOFFSETTO parm: "+parms[0]) tzoffsetto = self._parse_offset(value) elif name == "TZNAME": if parms: - raise ValueError("unsupported TZNAME parm: "+parms[0]) + raise ValueError( + "unsupported TZNAME parm: "+parms[0]) tzname = value elif name == "COMMENT": pass @@ -865,7 +891,8 @@ class tzical(object): else: if name == "TZID": if parms: - raise ValueError("unsupported TZID parm: "+parms[0]) + raise ValueError( + "unsupported TZID parm: "+parms[0]) tzid = value elif name in ("TZURL", "LAST-MODIFIED", "COMMENT"): pass @@ -886,6 +913,7 @@ else: TZFILES = [] TZPATHS = [] + def gettz(name=None): tz = None if not name: @@ -933,11 +961,11 @@ def gettz(name=None): pass else: tz = None - if tzwin: + if tzwin is not None: try: tz = tzwin(name) - except OSError: - pass + except WindowsError: + tz = None if not tz: from dateutil.zoneinfo import gettz tz = gettz(name) diff --git a/lib/dateutil/tzwin.py b/lib/dateutil/tzwin.py index 041c6cc3..e8a82d75 100644 --- a/lib/dateutil/tzwin.py +++ b/lib/dateutil/tzwin.py @@ -1,8 +1,8 @@ # This code was originally contributed by Jeffrey Harris. import datetime import struct -import winreg +from six.moves import winreg __all__ = ["tzwin", "tzwinlocal"] @@ -12,8 +12,8 @@ TZKEYNAMENT = r"SOFTWARE\Microsoft\Windows NT\CurrentVersion\Time Zones" TZKEYNAME9X = r"SOFTWARE\Microsoft\Windows\CurrentVersion\Time Zones" TZLOCALKEYNAME = r"SYSTEM\CurrentControlSet\Control\TimeZoneInformation" + def _settzkeyname(): - global TZKEYNAME handle = winreg.ConnectRegistry(None, winreg.HKEY_LOCAL_MACHINE) try: winreg.OpenKey(handle, TZKEYNAMENT).Close() @@ -21,8 +21,10 @@ def _settzkeyname(): except WindowsError: TZKEYNAME = TZKEYNAME9X handle.Close() + return TZKEYNAME + +TZKEYNAME = _settzkeyname() -_settzkeyname() class tzwinbase(datetime.tzinfo): """tzinfo class based on win32's timezones available in the registry.""" @@ -39,7 +41,7 @@ class tzwinbase(datetime.tzinfo): return datetime.timedelta(minutes=minutes) else: return datetime.timedelta(0) - + def tzname(self, dt): if self._isdst(dt): return self._dstname @@ -59,8 +61,11 @@ class tzwinbase(datetime.tzinfo): def display(self): return self._display - + def _isdst(self, dt): + if not self._dstmonth: + # dstmonth == 0 signals the zone has no daylight saving time + return False dston = picknthweekday(dt.year, self._dstmonth, self._dstdayofweek, self._dsthour, self._dstminute, self._dstweeknumber) @@ -78,31 +83,33 @@ class tzwin(tzwinbase): def __init__(self, name): self._name = name - handle = winreg.ConnectRegistry(None, winreg.HKEY_LOCAL_MACHINE) - tzkey = winreg.OpenKey(handle, "%s\%s" % (TZKEYNAME, name)) - keydict = valuestodict(tzkey) - tzkey.Close() - handle.Close() + # multiple contexts only possible in 2.7 and 3.1, we still support 2.6 + with winreg.ConnectRegistry(None, winreg.HKEY_LOCAL_MACHINE) as handle: + with winreg.OpenKey(handle, + "%s\%s" % (TZKEYNAME, name)) as tzkey: + keydict = valuestodict(tzkey) self._stdname = keydict["Std"].encode("iso-8859-1") self._dstname = keydict["Dlt"].encode("iso-8859-1") self._display = keydict["Display"] - + # See http://ww_winreg.jsiinc.com/SUBA/tip0300/rh0398.htm tup = struct.unpack("=3l16h", keydict["TZI"]) - self._stdoffset = -tup[0]-tup[1] # Bias + StandardBias * -1 - self._dstoffset = self._stdoffset-tup[2] # + DaylightBias * -1 - + self._stdoffset = -tup[0]-tup[1] # Bias + StandardBias * -1 + self._dstoffset = self._stdoffset-tup[2] # + DaylightBias * -1 + + # for the meaning see the win32 TIME_ZONE_INFORMATION structure docs + # http://msdn.microsoft.com/en-us/library/windows/desktop/ms725481(v=vs.85).aspx (self._stdmonth, - self._stddayofweek, # Sunday = 0 - self._stdweeknumber, # Last = 5 + self._stddayofweek, # Sunday = 0 + self._stdweeknumber, # Last = 5 self._stdhour, self._stdminute) = tup[4:9] (self._dstmonth, - self._dstdayofweek, # Sunday = 0 - self._dstweeknumber, # Last = 5 + self._dstdayofweek, # Sunday = 0 + self._dstweeknumber, # Last = 5 self._dsthour, self._dstminute) = tup[12:17] @@ -114,61 +121,59 @@ class tzwin(tzwinbase): class tzwinlocal(tzwinbase): - + def __init__(self): - handle = winreg.ConnectRegistry(None, winreg.HKEY_LOCAL_MACHINE) + with winreg.ConnectRegistry(None, winreg.HKEY_LOCAL_MACHINE) as handle: - tzlocalkey = winreg.OpenKey(handle, TZLOCALKEYNAME) - keydict = valuestodict(tzlocalkey) - tzlocalkey.Close() + with winreg.OpenKey(handle, TZLOCALKEYNAME) as tzlocalkey: + keydict = valuestodict(tzlocalkey) - self._stdname = keydict["StandardName"].encode("iso-8859-1") - self._dstname = keydict["DaylightName"].encode("iso-8859-1") + self._stdname = keydict["StandardName"].encode("iso-8859-1") + self._dstname = keydict["DaylightName"].encode("iso-8859-1") - try: - tzkey = winreg.OpenKey(handle, "%s\%s"%(TZKEYNAME, self._stdname)) - _keydict = valuestodict(tzkey) - self._display = _keydict["Display"] - tzkey.Close() - except OSError: - self._display = None + try: + with winreg.OpenKey( + handle, "%s\%s" % (TZKEYNAME, self._stdname)) as tzkey: + _keydict = valuestodict(tzkey) + self._display = _keydict["Display"] + except OSError: + self._display = None - handle.Close() - self._stdoffset = -keydict["Bias"]-keydict["StandardBias"] self._dstoffset = self._stdoffset-keydict["DaylightBias"] - # See http://ww_winreg.jsiinc.com/SUBA/tip0300/rh0398.htm tup = struct.unpack("=8h", keydict["StandardStart"]) (self._stdmonth, - self._stddayofweek, # Sunday = 0 - self._stdweeknumber, # Last = 5 + self._stddayofweek, # Sunday = 0 + self._stdweeknumber, # Last = 5 self._stdhour, self._stdminute) = tup[1:6] tup = struct.unpack("=8h", keydict["DaylightStart"]) (self._dstmonth, - self._dstdayofweek, # Sunday = 0 - self._dstweeknumber, # Last = 5 + self._dstdayofweek, # Sunday = 0 + self._dstweeknumber, # Last = 5 self._dsthour, self._dstminute) = tup[1:6] def __reduce__(self): return (self.__class__, ()) + def picknthweekday(year, month, dayofweek, hour, minute, whichweek): """dayofweek == 0 means Sunday, whichweek 5 means last instance""" first = datetime.datetime(year, month, 1, hour, minute) - weekdayone = first.replace(day=((dayofweek-first.isoweekday())%7+1)) + weekdayone = first.replace(day=((dayofweek-first.isoweekday()) % 7+1)) for n in range(whichweek): dt = weekdayone+(whichweek-n)*ONEWEEK if dt.month == month: return dt + def valuestodict(key): """Convert a registry key's values to a dictionary.""" dict = {} diff --git a/lib/dateutil/zoneinfo/.gitignore b/lib/dateutil/zoneinfo/.gitignore deleted file mode 100644 index 335ec957..00000000 --- a/lib/dateutil/zoneinfo/.gitignore +++ /dev/null @@ -1 +0,0 @@ -*.tar.gz diff --git a/lib/dateutil/zoneinfo/__init__.py b/lib/dateutil/zoneinfo/__init__.py index 81db1405..ff59639c 100644 --- a/lib/dateutil/zoneinfo/__init__.py +++ b/lib/dateutil/zoneinfo/__init__.py @@ -1,109 +1,135 @@ # -*- coding: utf-8 -*- -""" -Copyright (c) 2003-2005 Gustavo Niemeyer - -This module offers extensions to the standard Python -datetime module. -""" import logging import os -from subprocess import call +import warnings +import tempfile +import shutil +import json + +from subprocess import check_call from tarfile import TarFile +from pkgutil import get_data +from io import BytesIO +from contextlib import closing from dateutil.tz import tzfile -__author__ = "Tomi Pieviläinen " -__license__ = "Simplified BSD" +__all__ = ["gettz", "gettz_db_metadata", "rebuild"] -__all__ = ["setcachesize", "gettz", "rebuild"] +_ZONEFILENAME = "dateutil-zoneinfo.tar.gz" +_METADATA_FN = 'METADATA' + +# python2.6 compatability. Note that TarFile.__exit__ != TarFile.close, but +# it's close enough for python2.6 +_tar_open = TarFile.open +if not hasattr(TarFile, '__exit__'): + def _tar_open(*args, **kwargs): + return closing(TarFile.open(*args, **kwargs)) -CACHE = [] -CACHESIZE = 10 class tzfile(tzfile): def __reduce__(self): return (gettz, (self._filename,)) -def getzoneinfofile(): - filenames = sorted(os.listdir(os.path.join(os.path.dirname(__file__)))) - filenames.reverse() - for entry in filenames: - if entry.startswith("zoneinfo") and ".tar." in entry: - return os.path.join(os.path.dirname(__file__), entry) - return None -ZONEINFOFILE = getzoneinfofile() +def getzoneinfofile_stream(): + try: + return BytesIO(get_data(__name__, _ZONEFILENAME)) + except IOError as e: # TODO switch to FileNotFoundError? + warnings.warn("I/O error({0}): {1}".format(e.errno, e.strerror)) + return None -del getzoneinfofile -def setcachesize(size): - global CACHESIZE, CACHE - CACHESIZE = size - del CACHE[size:] +class ZoneInfoFile(object): + def __init__(self, zonefile_stream=None): + if zonefile_stream is not None: + with _tar_open(fileobj=zonefile_stream, mode='r') as tf: + # dict comprehension does not work on python2.6 + # TODO: get back to the nicer syntax when we ditch python2.6 + # self.zones = {zf.name: tzfile(tf.extractfile(zf), + # filename = zf.name) + # for zf in tf.getmembers() if zf.isfile()} + self.zones = dict((zf.name, tzfile(tf.extractfile(zf), + filename=zf.name)) + for zf in tf.getmembers() + if zf.isfile() and zf.name != _METADATA_FN) + # deal with links: They'll point to their parent object. Less + # waste of memory + # links = {zl.name: self.zones[zl.linkname] + # for zl in tf.getmembers() if zl.islnk() or zl.issym()} + links = dict((zl.name, self.zones[zl.linkname]) + for zl in tf.getmembers() if + zl.islnk() or zl.issym()) + self.zones.update(links) + try: + metadata_json = tf.extractfile(tf.getmember(_METADATA_FN)) + metadata_str = metadata_json.read().decode('UTF-8') + self.metadata = json.loads(metadata_str) + except KeyError: + # no metadata in tar file + self.metadata = None + else: + self.zones = dict() + self.metadata = None + + +# The current API has gettz as a module function, although in fact it taps into +# a stateful class. So as a workaround for now, without changing the API, we +# will create a new "global" class instance the first time a user requests a +# timezone. Ugly, but adheres to the api. +# +# TODO: deprecate this. +_CLASS_ZONE_INSTANCE = list() + def gettz(name): - tzinfo = None - if ZONEINFOFILE: - for cachedname, tzinfo in CACHE: - if cachedname == name: - break - else: - tf = TarFile.open(ZONEINFOFILE) - try: - zonefile = tf.extractfile(name) - except KeyError: - tzinfo = None - else: - tzinfo = tzfile(zonefile) - tf.close() - CACHE.insert(0, (name, tzinfo)) - del CACHE[CACHESIZE:] - return tzinfo + if len(_CLASS_ZONE_INSTANCE) == 0: + _CLASS_ZONE_INSTANCE.append(ZoneInfoFile(getzoneinfofile_stream())) + return _CLASS_ZONE_INSTANCE[0].zones.get(name) -def rebuild(filename, tag=None, format="gz"): + +def gettz_db_metadata(): + """ Get the zonefile metadata + + See `zonefile_metadata`_ + + :returns: A dictionary with the database metadata + """ + if len(_CLASS_ZONE_INSTANCE) == 0: + _CLASS_ZONE_INSTANCE.append(ZoneInfoFile(getzoneinfofile_stream())) + return _CLASS_ZONE_INSTANCE[0].metadata + + +def rebuild(filename, tag=None, format="gz", zonegroups=[], metadata=None): """Rebuild the internal timezone info in dateutil/zoneinfo/zoneinfo*tar* filename is the timezone tarball from ftp.iana.org/tz. """ - import tempfile, shutil tmpdir = tempfile.mkdtemp() zonedir = os.path.join(tmpdir, "zoneinfo") moduledir = os.path.dirname(__file__) - if tag: tag = "-"+tag - targetname = "zoneinfo%s.tar.%s" % (tag, format) try: - tf = TarFile.open(filename) - # The "backwards" zone file contains links to other files, so must be - # processed as last - for name in sorted(tf.getnames(), - key=lambda k: k != "backward" and k or "z"): - if not (name.endswith(".sh") or - name.endswith(".tab") or - name == "leapseconds"): + with _tar_open(filename) as tf: + for name in zonegroups: tf.extract(name, tmpdir) - filepath = os.path.join(tmpdir, name) - try: - # zic will return errors for nontz files in the package - # such as the Makefile or README, so check_call cannot - # be used (or at least extra checks would be needed) - call(["zic", "-d", zonedir, filepath]) - except OSError as e: - if e.errno == 2: - logging.error( - "Could not find zic. Perhaps you need to install " - "libc-bin or some other package that provides it, " - "or it's not in your PATH?") + filepaths = [os.path.join(tmpdir, n) for n in zonegroups] + try: + check_call(["zic", "-d", zonedir] + filepaths) + except OSError as e: + if e.errno == 2: + logging.error( + "Could not find zic. Perhaps you need to install " + "libc-bin or some other package that provides it, " + "or it's not in your PATH?") raise - tf.close() - target = os.path.join(moduledir, targetname) - for entry in os.listdir(moduledir): - if entry.startswith("zoneinfo") and ".tar." in entry: - os.unlink(os.path.join(moduledir, entry)) - tf = TarFile.open(target, "w:%s" % format) - for entry in os.listdir(zonedir): - entrypath = os.path.join(zonedir, entry) - tf.add(entrypath, entry) - tf.close() + # write metadata file + with open(os.path.join(zonedir, _METADATA_FN), 'w') as f: + json.dump(metadata, f, indent=4, sort_keys=True) + target = os.path.join(moduledir, _ZONEFILENAME) + with _tar_open(target, "w:%s" % format) as tf: + for entry in os.listdir(zonedir): + entrypath = os.path.join(zonedir, entry) + tf.add(entrypath, entry) finally: shutil.rmtree(tmpdir)