NAME DateTime::Util::Calc - DateTime Calculation Utilities SYNOPSIS use DateTime::Util::Calc qw(polynomial); my @coeffs = qw(2 3 -2); my \$x = 5; my \$rv = polynomial(\$x, @coeffs); DESCRIPTION This module contains some common calculation utilities that are required to perform datetime calculations, specifically from "Calendrical Calculations" -- they are NOT meant to be general purpose. Nothing is exported by default. You must either explicitly export them, or use as fully qualified function names. FUNCTIONS max(\$a, \$b) min(\$a, \$b) max() returns the bigger of \$a and \$b. min() returns the smaller of \$a and \$b. polynomial(\$x, @coefs) Calculates the value of a polynomial equation, based on Horner's Rule. c + b * x + a * (x ** 2) x = 5 is expressed as: polynomial(5, c, b, a); moment(\$dt) dt_from_moment(\$moment) moment() converts a DateTime object to moment, which is RD days + the time of day as fraction of the total seconds in a day. dt_from_moment() converts a moment to DateTime object. rata_die() Returns a new DateTime object that is set to Rata Die, 0001-01-01 00:00:00 UTC bigfloat(\$v) bigint(\$v) If the value \$v is not a Math::BigFloat object, returns the value converted to Math::BigFloat. Otherwise returns the value itself. bigint() does the same for Math::BigInt. bf_downgrade(\$v) bi_downgrade(\$v) If the value \$v is a Math::BigFloat object, returns the "downgraded", regular Perl scalar version of \$v. This is sometimes required for functions or objects that do not accept Math::BigFloat. If \$v is not Math::BigFloat object, just returns the value itself. bi_downgrade() does the same for Math::BigInt. truncate_to_midday(\$dt) Truncates the DateTime object to 12:00 noon. sin_deg(\$degrees) cos_deg(\$degrees) tan_deg(\$degrees) asin_deg(\$degrees) acos_deg(\$degrees) Each of these functions calculates their respective values based on degrees, not radians (as Perl's version of sin() and cos() would do). Since Math::BigFloat does not have corresponding trigonemetric functions, values passed to these funtions will be automatically downgraded via bf_downgrade() mod(\$v,\$mod) Calculates the modulus of \$v over \$mod. Perl's built-in modulus operator (%) for some reason rounds numbers UP when a fractional number's modulus is taken. Many of the calculations also needed the fractional part of the calculation, so this function takes care of both. Example: mod(12.234, 5) = 2.234 amod(\$v,\$mod) This function is almost identical to mod(), but when the regular modulus value is 0, returns \$mod instead of 0. Example: amod(11, 5) = 1 amod(10, 5) = 5 amod(9, 5) = 4 amod(8, 5) = 3 binary_search(\$hi, \$lo, \$mu, \$phi) This is a special version of binary search, where the terminating condition is determined by the result of coderefs \$mu and \$phi. \$mu is passed the value of \$hi and \$lo. If it returns true upon execution, then the search terminates. \$phi is passed the next median value. If it returns true upon execution, then the search terminates. If the above two fails, then \$hi and \$lo are re-computed for the next iteration. search_next(%opts) Performs a "linear" search until some condition is met. This is a generalized version of the formula defined in [1] p.22. The basic idea is : x = base while (! check(x) ) { x = next(x); } return x %opts can contain the following parameters: base The initial value to use to start the search process. The value can be anything, but you must provide "check" and "next" parameters that are capable of handling the type of thing you specified. check (coderef) Code to be executed to determine the end of the search. The function receives the current value of "x", and should return a true value if the condition to end the loop has been reached next (coderef, optional) Code to be executed to determine the next value of "x". The function receives the current value of "x", and should return the value to be used for the next iteration. If unspecified, it will use a function that blindly adds 1 to whatever x is. (so if you specified a number for "base", it should work -- but if you passed an object like DateTime, it will probably be an error) So for example, to iterate through 1 through 9, you could do something like this my \$x = search_next( base => 1, check => sub { \$_[0] == 9 } ); And \$x will be set to 9. For a more interesting example, we could look for a DateTime object \$dt matching a certain condition "foo()": my \$dt = search_next( base => \$base_date, check => \&foo, next => sub { \$_[0] + DateTime::Duration->new(days => 1) } ); revolution(\$angle_in_degrees) Reduces any angle to within the first revolution by sbtracting or adding even multiples of 360.0. rev180(\$angle_in_degrees) Reduces input to within +180..+180 degrees angle(\$h, \$m, \$s) CAVEATS For performance reasons, there is absolutely no parameter validation via Params::Validate in this module! AUTHOR (c) Copyright 2004-2005 Daisuke Maki . All rights reserved.