The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

Date::RangeParser::EN - Parser for plain English date/time range strings

SYNOPSIS

    use Date::RangeParser::EN;

    my $parser = Date::RangeParser::EN->new;
    my ($begin, $end) = $parser->parse_range("this week");

DESCRIPTION

Parses plain-English strings representing date/time ranges

METHODS

new

Returns a new instance of Date::RangeParser::EN.

Takes an optional hash of parameters:

  • datetime_class

    By default, Date::RangeParser::EN returns two DateTime objects representing the beginning and end of the range. If you use a subclass of DateTime (or another module that implements the DateTime API), you may pass the name of this class to use it instead.

    At the very least, this given class must implement a new method that accepts a hash of arguments, where the following keys will be set:

      year
      month
      day
      hour
      minute
      second

    This gives you the freedom to set your time zones and such however you need to.

  • now_callback

    By default, Date::RangeParser::EN uses DateTime->now to determine the current date/time for calculations. If you need to work with a different time (for instance, if you need to adjust for time zones), you may pass a callback (code reference) which returns a DateTime object.

parse_range

Accepts a string representing a plain-English date range, for instance:

  • today

  • this week

  • the past 2 months

  • next Tuesday

  • two weeks ago

  • the next 3 hours

  • the 3rd of next month

  • the end of this month

More formally, this will parse the following kinds of date strings:

  NUMBER : ordinary number (1)
  PERIOD : one of: hour, day, week, month, quarter, or year (or the plural of these)
  WEEKDAY : one of: Monday, Tuesday, Wedensday, Thursday, Friday, Saturday, or Sunday
  CARDINAL : a cardinal number (21st) or the word for that number (twenty-first) or end
  MONTH : a month name: January, Feburary, March, April, May, June, July August, 
          September, October, November, or Decmeber or any 3-letter abbreviation
  YEAR : a 4-digit year (2-digits will not work)
  RANGE : any date range that can be parsed by parse_range

  today                             : today, midnight to midnight

  this PERIOD                       : the current period, start to end
  this month

  current PERIOD                    : the current period, start to end
  current year

  this WEEKDAY                      : the WEEKDAY that is in the current week, midnight to midnight
  this Monday

  NUMBER PERIOD ago                 : past date relative to now until now
  3 days ago

  past NUMBER PERIOD                : past date relative to now until now
  past 2 weeks

  last NUMBER PERIOD                : past date relative to now until now
  last 6 hours

  past NUMBER WEEKDAY               : the weekday a number of weeks before now until now
  past 4 Saturdays

  NUMBER WEEKDAY ago                : the weekday a number of weeks before now until now
  3 Fridays ago

  yesterday                         : yesterday, midnight to midnight

  last WEEKDAY                      : the WEEKDAY that is in the week prior to this, midnight to midnight
  last Wednesday

  previous WEEKDAY                  : the WEEKDAY that is in the week prior to this, midnight to midnight
  previous Friday

  past WEEKDAY                      : the WEEKDAY that is in the week prior to this, midnight to midnight
  past Tuesday

  this past WEEKDAY                 : the WEEKDAY that is in the week prior to this, midnight to midnight
  this past Saturday

  coming WEEKDAY                    : the WEEKDAY that is in the week after this, midnight to midnight
  coming Monday

  this coming WEEKDAY               : the WEEKDAY that is in the week after this, midnight to midnight
  this coming Thursday

  NUMBER PERIOD hence               : now to a future date relative to now
  4 months hence

  NUMBER PERIOD from now            : now to a future date relative to now
  6 days from now

  next NUMBER PERIOD                : now to a future date relative to now
  next 7 years

  tomorrow                          : tomorrow, midnight to midnight

  next NUMBER WEEKDAY               : the WEEKDAY that is in a number of weeks after this, midnight to midnight
  next 4 Sundays

  CARDINAL of this month            : the specified day of the current month, midnight to midnight
  14th of this month

  CARDINAL of last month            : the specified day of the previous month, midnight to midnight
  31st of last month

  CARDINAL of next month            : the specified day of the month following this, midnight to midnight
  3rd of next month

  CARDINAL of NUMBER months ago     : the specified day of a previous month, midnight to midnight
  12th of 2 months ago

  CARDINAL of NUMBER months from now : the specified day of a following month, midnight to midnight
  7th of 22 months from now

  CARDINAL of NUMBER months hence   : the specified day of a following month, midnight to midnight
  22nd of 6 months hence

  MONTH                             : the named month of the current year, 1st to last day
  August

  this MONTH                        : the named month of the current year, 1st to last day
  this Sep

  last MONTH                        : the named month of the previous year, 1st to last day
  last January

  next MONTH                        : the named month of the next year, 1st to last day
  next Dec

  MONTH YEAR                        : the named month of the named year, 1st to last day
  June 1969

  RANGE to RANGE                    : the very start of the first range to the very end of the second
  Tuesday to Next Saturday

  RANGE thru RANGE                  : the very start of the first range to the very end of the second
  2 hours ago thru the next 6 hours

  RANGE through RANGE               : the very start of the first range to the very end of the second
  August through December

  RANGE - RANGE                     : the very start of the first range to the very end of the second
  9-1-2012 - 9-30-2012

  RANGE-RANGE                       : the very start of the first range to the very end of the second
  10/10-10/20                         (ranges must not contain hyphens, "-")

Anything else is parsed by Date::Manip. If Date::Manip is unable to parse the date given either, then the dates returned will be undefined.

Also, when parsing:

  • The word "the" will always be ignored and can appear anywhere.

  • Cardinal numbers may be spelled out as words, i.e. "September first" instead of "September 1st".

  • Any plural or singular period shown above can be used with the opposite.

  • All dates are parsed relative to the parser's notion of now. You can control this by setting the now_callback option on the constructor.

Returns two DateTime objects, reprensenting the beginning and end of the range.

TO DO

There's a lot more that this module could handle. A few items that come to mind:

  • More testing to make sure certain date configurations are handled, like start of week.

  • Handle Unicode in places where such handling makes sense (like hyphen detection)

  • Allow full words instead of digits ("two weeks ago" vs "2 weeks ago")

  • Allow "between" for ranges ("between last February and this Friday") in addition to "to/through" ranges

  • This module is US English-centric (hence the "EN") and might do some things wrong for other variants of English and a generic Date::RangeParser interface could be made to allow for other languages to be parsed this way.

  • Depends on Date::Manip. This may or may not be a good thing.

DEPENDENCIES

DateTime, Date::Manip

AUTHORS

This module was authored by Grant Street Group (http://grantstreet.com), which was kind enough to give it back to the Perl community.

The CPAN distribution is maintained by Michael Aquilina (aquilina@cpan.org).

COPYRIGHT AND LICENSE

Copyright (C) 2012 Grant Street Group.

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.