NAME
   DateTime::Format::Alami - Parse human date/time expression (base class)

VERSION
   This document describes version 0.16 of DateTime::Format::Alami (from
   Perl distribution DateTime-Format-Alami), released on 2017-07-10.

SYNOPSIS
   For English:

    use DateTime::Format::Alami::EN;
    my $parser = DateTime::Format::Alami::EN->new();
    my $dt = $parser->parse_datetime("2 hours 13 minutes from now");

   Or you can also call as class method:

    my $dt = DateTime::Format::Alami::EN->parse_datetime("yesterday");

   To parse duration:

    my $dtdur = DateTime::Format::Alami::EN->parse_datetime_duration("2h"); # 2 hours

   For Indonesian:

    use DateTime::Format::Alami::ID;
    my $parser = DateTime::Format::Alami::ID->new();
    my $dt = $parser->parse_datetime("5 jam lagi");

   Or you can also call as class method:

    my $dt = DateTime::Format::Alami::ID->parse_datetime("hari ini");

   To parse duration:

    my $dtdur = DateTime::Format::Alami::ID->parse_datetime_duration("2h"); # 2 days

DESCRIPTION
   This class parses human/natural date/time/duration string and returns
   DateTime (or DateTime::Duration) object. Currently it supports English
   and Indonesian. The goal of this module is to make it easier to add
   support for other human languages.

   To actually use this class, you must use one of its subclasses for each
   human language that you want to parse.

   There are already some other DateTime human language parsers on CPAN and
   elsewhere, see "SEE ALSO".

HOW IT WORKS
   DateTime::Format::Alami is base class. Each human language is
   implemented in a separate "DateTime::Format::Alami::<ISO_CODE>" module
   (e.g. DateTime::Format::Alami::EN and DateTime::Format::Alami::EN) which
   is a subclass.

   Parsing is done using a single recursive regex (i.e. containing
   "(?&NAME)" and "(?(DEFINE))" patterns, see perlre). This regex is
   composed from pieces of pattern strings in the "p_*" and "o_*" methods,
   to make it easier to override in an OO-fashion.

   A pattern string that is returned by the "p_*" method is a normal regex
   pattern string that will be compiled using the /x and /i regex modifier.
   The pattern string can also refer to pattern in other "o_*" or "p_*"
   method using syntax "<o_foo>" or "<p_foo>". Example, "o_today" for
   English might be something like:

    sub p_today { "(?: today | this \s+ day )" }

   Other examples:

    sub p_yesterday { "(?: yesterday )" }

    sub p_dateymd { join(
        "",
       '(?: <o_dayint> \\s* ?<o_monthname> | <o_monthname> \\s* <o_dayint>\\b|<o_monthint>[ /-]<o_dayint>\\b )',
       '(?: \\s*[,/-]?\\s* <o_yearint>)?'
    )}

    sub o_date { "(?: <p_today>|<p_yesterday>|<p_dateymd>)" }

    sub p_time { "(?: <o_hour>:<o_minute>(?:<o_second>)? \s* <o_ampm> )" }

    sub p_date_time { "(?: <o_date> (?:\s+ at)? <o_time> )" }

   When a pattern from "p_*" matches, a corresponding action method "a_*"
   will be invoked. Usually the method will set or modify a DateTime object
   in "$self->{_dt}". For example, this is code for "a_today":

    sub a_today {
        my $self = shift;
        $self->{_dt} = DateTime->today;
    }

   The patterns from all "p_*" methods will be combined in an alternation
   to form the final pattern.

   An "o_*" pattern is just like "p_*", but they will not be combined into
   the final pattern and matching it won't execute a corresponding "a_*"
   method.

   And there are also "w_*" methods which return array of strings.

   Parsing duration is similar, except the method names are "pdur_*",
   "odur_*" and "adur_*".

ADDING A NEW HUMAN LANGUAGE
   See an example in existing "DateTime::Format::Alami::*" module.
   Basically you just need to supply the necessary patterns in the "p_*"
   methods. If you want to introduce new "p_*" method, don't forget to
   supply the action too in the "a_*" method.

METHODS
 new => obj
   Constructor. You actually must instantiate subclass instead.

 parse_datetime($str[ , \%opts ]) => obj
   Parse/extract date/time expression in $str. Die if expression cannot be
   parsed. Otherwise return DateTime object (or string/number if "format"
   option is "verbatim"/"epoch", or hash if "format" option is "combined")
   or array of objects/strings/numbers (if "returns" option is
   "all"/"all_cron").

   Known options:

   *   time_zone => str

       Will be passed to DateTime constructor.

   *   format => str (DateTime|verbatim|epoch|combined)

       The default is "DateTime", which will return DateTime object. Other
       choices include "verbatim" (returns the original text), "epoch"
       (returns Unix timestamp), "combined" (returns a hash containing keys
       like "DateTime", "verbatim", "epoch", and other extra information:
       "pos" [position of pattern in the string], "pattern" [pattern name],
       "m" [raw named capture groups], "uses_time" [whether the date
       involves time of day]).

       You might think that choosing "epoch" or "verbatim" could avoid the
       overhead of DateTime, but actually you can't since DateTime is used
       as the primary format during parsing. The epoch is retrieved from
       the DateTime object using the "epoch" method.

   *   prefers => str (nearest|future|past)

       NOT YET IMPLEMENTED.

       This option decides what happens when an ambiguous date appears in
       the input. For example, "Friday" may refer to any number of Fridays.
       Possible choices are: "nearest" (prefer the nearest date, the
       default), "future" (prefer the closest future date), "past" (prefer
       the closest past date).

   *   returns => str (first|last|earliest|latest|all|all_cron)

       If the text has multiple possible dates, then this argument
       determines which date will be returned. Possible choices are:
       "first" (return the first date found in the string, the default),
       "last" (return the final date found in the string), "earliest"
       (return the date found in the string that chronologically precedes
       any other date in the string), "latest" (return the date found in
       the string that chronologically follows any other date in the
       string), "all" (return all dates found in the string, in the order
       they were found in the string), "all_cron" (return all dates found
       in the string, in chronological order).

       When "all" or "all_cron" is chosen, function will return array(ref)
       of results instead of a single result, even if there is only a
       single actual result.

 parse_datetime_duration($str[ , \%opts ]) => obj
   Parse/extract duration expression in $str. Die if expression cannot be
   parsed. Otherwise return DateTime::Duration object (or string/number if
   "format" option is "verbatim"/"seconds", or hash if "format" option is
   "combined") or array of objects/strings/numbers (if "returns" option is
   "all"/"all_sorted").

   Known options:

   *   format => str (Duration|verbatim|seconds|combined)

       The default is "Duration", which will return DateTime::Duration
       object. Other choices include "verbatim" (returns the original
       text), "seconds" (returns number of seconds, approximated),
       "combined" (returns a hash containing keys like "Duration",
       "verbatim", "seconds", and other extra information: "pos" [position
       of pattern in the string], "pattern" [pattern name], "m" [raw named
       capture groups]).

       You might think that choosing "seconds" or "verbatim" could avoid
       the overhead of DateTime::Duration, but actually you can't since
       DateTime::Duration is used as the primary format during parsing. The
       number of seconds is calculated from the DateTime::Duration object
       *using an approximation* (for example, "1 month" does not convert
       exactly to seconds).

   *   returns => str (first|last|smallest|largest|all|all_sorted)

       If the text has multiple possible durations, then this argument
       determines which date will be returned. Possible choices are:
       "first" (return the first duration found in the string, the
       default), "last" (return the final duration found in the string),
       "smallest" (return the smallest duration), "largest" (return the
       largest duration), "all" (return all durations found in the string,
       in the order they were found in the string), "all_sorted" (return
       all durations found in the string, in smallest-to-largest order).

       When "all" or "all_sorted" is chosen, function will return
       array(ref) of results instead of a single result, even if there is
       only a single actual result.

FAQ
 What does "alami" mean?
   It is an Indonesian word, meaning "natural".

 How does it compare to similar modules?
   DateTime::Format::Natural (DF:Natural) is a more established module
   (first released on 2006) and can understand a bit more English
   expression like 'last day of Sep'. Aside from English, it does not yet
   support other languages.

   DFA:EN's "parse_datetime_duration()" produces a DateTime::Duration
   object while DF:Natural's "parse_datetime_duration()" returns two
   DateTime objects instead. In other words, DF:Natural can parse "from 23
   Jun to 29 Jun" in addition to "for 2 weeks".

   DF:Natural in general is slightly more strict about the formats it
   accepts, e.g. it rejects "Jun 23st" (the error message even gives hints
   that the suffix must be 'rd'). DF:Natural can give a detailed error
   message on why parsing has failed (see its "error()" method).

   DateTime::Format::Flexible (DF:Flexible) is another established module
   (first released in 2007) that, aside from parsing human expression (like
   'tomorrow', 'sep 1st') can also parse date/time in several other formats
   like RFC 822, making it a convenient module to use as a 'one-stop'
   solution to parse date. Compared to DF:Natural, it has better support
   for timezone but cannot parse some English expressions. Aside from
   English, it currently supports German and Spanish. It does not support
   parsing duration expression.

   This module itself: DateTime::Format::Alami (DF:Alami) is yet another
   implementation. Internally, it uses recursive regex to make parsing
   simpler and adding more languages easier. It requires perl 5.14.0 or
   newer due to the use of "(?{ ... })" code blocks inside regular
   expression (while DF:Natural and DF:Flexible can run on perl 5.8+). It
   currently supports English and Indonesian. It supports parsing duration
   expression and returns DateTime::Duration object. It has the smallest
   startup time (see see Bencher::Scenario::DateTimeFormatAlami::Startup).

   Performance-wise, all the modules are within the same order of magnitude
   (see Bencher::Scenario::DateTimeFormatAlami::Parsing).

HOMEPAGE
   Please visit the project's homepage at
   <https://metacpan.org/release/DateTime-Format-Alami>.

SOURCE
   Source repository is at
   <https://github.com/perlancar/perl-DateTime-Format-Alami>.

BUGS
   Please report any bugs or feature requests on the bugtracker website
   <https://rt.cpan.org/Public/Dist/Display.html?Name=DateTime-Format-Alami
   >

   When submitting a bug or request, please include a test-file or a patch
   to an existing test-file that illustrates the bug or desired feature.

SEE ALSO
 Similar modules on CPAN
   Date::Extract. DateTime::Format::Alami has some features of
   Date::Extract so it can be used to replace Date::Extract.

   DateTime::Format::Flexible. See "FAQ".

   For Indonesian: DateTime::Format::Indonesian, Date::Extract::ID
   (currently this module uses DateTime::Format::Alami::ID as its backend).

   For English: DateTime::Format::Natural. See "FAQ".

 Other modules on CPAN
   DateTime::Format::Human deals with formatting and not parsing.

 Similar non-Perl libraries
   Natt Java library, which the last time I tried sometimes gives weird
   answer, e.g. "32 Oct" becomes 1 Oct in the far future.
   http://natty.joestelmach.com/

   Duckling Clojure library, which can parse date/time as well as numbers
   with some other units like temperature.
   https://github.com/wit-ai/duckling

AUTHOR
   perlancar <[email protected]>

COPYRIGHT AND LICENSE
   This software is copyright (c) 2017, 2016, 2014 by [email protected].

   This is free software; you can redistribute it and/or modify it under
   the same terms as the Perl 5 programming language system itself.