NAME
Math::Calc::Parser - Parse and evaluate mathematical expressions
SYNOPSIS
use Math::Calc::Parser 'calc';
use utf8; # for π in source code
my $result = calc '2 + 2'; # 4
my $result = calc 'int rand 5'; # Random integer between 0 and 4
my $result = calc 'sqrt -1'; # i
my $result = calc '0xff << 2'; # 1020
my $result = calc '1/0'; # Division by 0 exception
# Class methods
my $result = Math::Calc::Parser->evaluate('2 + 2'); # 4
my $result = Math::Calc::Parser->evaluate('3π^2'); # 29.608813203268
my $result = Math::Calc::Parser->evaluate('0.7(ln 4)'); # 0.970406052783923
# With more advanced error handling
my $result = Math::Calc::Parser->try_evaluate('rand(abs'); # undef (Mismatched parentheses)
if (defined $result) {
print "Result: $result\n";
} else {
print "Error: ".Math::Calc::Parser->error."\n";
}
# Or as an object for more control
my $parser = Math::Calc::Parser->new;
$parser->add_functions(triple => { args => 1, code => sub { $_[0]*3 } });
$parser->add_functions(pow => { args => 2, code => sub { $_[0] ** $_[1] });
$parser->add_functions(one => sub { 1 }, two => sub { 2 }, three => sub { 3 });
my $result = $parser->evaluate('2(triple one)'); # 2*(1*3) = 6
my $result = $parser->evaluate('pow(triple two, three)'); # (2*3)^3 = 216
my $result = $parser->try_evaluate('triple triple'); # undef (Malformed expression)
die $parser->error unless defined $result;
$parser->remove_functions('π', 'e');
$parser->evaluate('3π'); # Invalid function exception
# Arbitrary precision calculations - use only in a controlled environment
$parser->bignum(1);
my $result = $parser->evaluate('30!'); # 265252859812191058636308480000000
my $result = $parser->evaluate('atan pi'); # 1.262627255678911683444322083605698343509
# Rational number calculations - use only in a controlled environment
$parser->bigrat(1);
my $result = $parser->evaluate('3 / 9'); # 1/3
my $result = $parser->evaluate('3 >> 2'); # 3/4
DESCRIPTION
Math::Calc::Parser is a simplified mathematical expression evaluator
with support for complex and trigonometric operations, implicit
multiplication, and perlish "parentheses optional" functions, while
being safe for arbitrary user input. It parses input strings into a
structure based on Reverse Polish notation
evaluates the result. The list of recognized functions may be
customized using "add_functions" and "remove_functions".
FUNCTIONS
calc
use Math::Calc::Parser 'calc';
my $result = calc '2+2';
$ perl -MMath::Calc::Parser=calc -E 'say calc "2+2"'
$ perl -Math -e '2+2'
Compact exportable function wrapping "evaluate" for string expressions.
Throws an exception on error. See ath for easy compact one-liners.
ATTRIBUTES
These attributes can only be set on instantiated objects.
bignum
my $bool = $parser->bignum;
$parser = $parser->bignum($bool);
Enable support for arbitrary precision numbers using Math::BigInt and
Math::BigFloat. This will avoid losing precision when working with
floats or large integers, but see "BIGNUM CAVEATS".
bigrat
my $bool = $parser->bigrat;
$parser = $parser->bigrat($bool);
Enable support for precise rational numbers using Math::BigRat. This
will avoid losing precision when working with integer divison and
similar operations, and will result in output like 3/7 where possible,
but see "BIGNUM CAVEATS".
METHODS
Aside from add_functions and remove_functions, all methods can be
called as class methods, and will act on a singleton object with the
default functions available.
new
my $parser = Math::Calc::Parser->new;
my $parser = Math::Calc::Parser->new(bignum => 1);
Creates a new Math::Calc::Parser object.
parse
my $parsed = Math::Calc::Parser->parse('5 / e^(i*pi)');
my $parsed = $parser->parse('3pi');
Parses a mathematical expression. On success, returns an array
reference representation of the expression in RPN notation which can be
passed to "evaluate". Throws an exception on failure.
evaluate
my $result = Math::Calc::Parser->evaluate($parsed);
my $result = Math::Calc::Parser->evaluate('log rand 7');
my $result = $parser->evaluate('round 13/3');
Evaluates a mathematical expression. The argument can be either an
arrayref from "parse" or a string expression which will be passed to
"parse". Returns the result of the expression on success or throws an
exception on failure.
try_evaluate
if (defined (my $result = Math::Calc::Parser->try_evaluate('floor 2.5'))) {
print "Result: $result\n";
} else {
print "Error: ".Math::Calc::Parser->error."\n";
}
if (defined (my $result = $parser->try_evaluate('log(5'))) {
print "Result: $result\n";
} else {
print "Error: ".$parser->error."\n";
}
Same as "evaluate" but instead of throwing an exception on failure,
returns undef. The "error" method can then be used to retrieve the
error message. The error message for the most recent "try_evaluate"
call can also be retrieved from the package variable
$Math::Calc::Parser::ERROR.
error
my $result = Math::Calc::Parser->try_evaluate('(i');
die Math::Calc::Parser->error unless defined $result;
my $result = $parser->try_evaluate('2//');
die $parser->error unless defined $result;
Returns the error message after a failed "try_evaluate".
add_functions
$parser->add_functions(
my_function => { args => 5, code => sub { return grep { $_ > 0 } @_; } },
other_function => sub { 20 },
bignum_function => { args => 1, code => sub { 2 ** $_[0] }, bignum_code => sub { Math::BigInt->new(2)->bpow($_[0]) } },
);
Adds functions to be recognized by the parser object. Keys are function
names which must start with an alphabetic character and consist only of
word characters
are either a hashref containing args and code keys, or a coderef that
is assumed to be a 0-argument function. args must be an integer greater
than or equal to 0. code or the passed coderef will be called with the
numeric operands passed as parameters, and must either return a numeric
result or throw an exception. Non-numeric results will be cast to
numbers in the usual perl fashion, and undefined results will throw an
evaluation error.
Alternate implementations to be used when "bignum" or "bigrat" is
enabled can be passed as bignum_code and bigrat_code respectively.
bignum_code will also be used for "bigrat" calculations if bigrat_code
is not separately defined; it is not common that these will need
separate implementations.
remove_functions
$parser->remove_functions('rand','nonexistent');
Removes functions from the parser object if they exist. Can be used to
remove default functions as well as functions previously added with
"add_functions".
OPERATORS
Math::Calc::Parser recognizes the following operators with their usual
mathematical definitions.
+, -, *, /, %, ^, !, <<, >>
Note: + and - can represent both binary addition/subtraction and unary
negation.
DEFAULT FUNCTIONS
Math::Calc::Parser parses several functions by default, which can be
customized using "add_functions" or "remove_functions" on an object
instance.
abs
Absolute value.
acos
asin
atan
Inverse sine, cosine, and tangent.
atan2
Two-argument inverse tangent of first argument divided by second
argument.
ceil
Round up to nearest integer.
cos
Cosine.
e
Euler's number.
floor
Round down to nearest integer.
i
Imaginary unit.
int
Cast (truncate) to integer.
ln
Natural log.
log
Log base 10.
logn
Log with arbitrary base given as second argument.
pi
π
π
π (this must be the decoded Unicode character)
rand
Random value between 0 and 1 (exclusive of 1). Uses
Math::Random::Secure if installed.
round
Round to nearest integer, with halfway cases rounded away from zero.
Due to bugs in Math::BigRat, precision may be lost with "bigrat"
enabled.
sin
Sine.
sqrt
Square root.
tan
Tangent.
CAVEATS
While parentheses are optional for functions with 0 or 1 argument, they
are required when a comma is used to separate multiple arguments.
Due to the nature of handling complex numbers, the evaluated result may
be a Math::Complex object. These objects can be directly printed or
used in numeric operations but may be more difficult to use in
comparisons.
Operators that are not defined to operate on complex numbers will
return the result of the operation on the real components of their
operands. This includes the operators <<, >>, %, and !.
BIGNUM CAVEATS
The Math::BigInt, Math::BigFloat, and Math::BigRat packages are useful
for working with numbers without losing precision, and can be used by
this module by setting the "bignum" or "bigrat" attributes, but care
should be taken. They will perform significantly slower than native
Perl numbers, and can result in an operation that does not terminate or
one that uses up all your memory.
Additionally, similar to when using the bignum or bigrat pragmas, the
auto-upgrading and downgrading behavior of these modules can only be
set globally, so enabling these options will affect all other uses of
these modules in your program. For the same reason, it is not
recommended to enable both "bignum" and "bigrat" in the same program.
The evaluated result may be a Math::BigInt, Math::BigFloat,
Math::BigRat, or other similar type of object. These objects can be
printed and behave normally as numbers.
Math::BigFloat defaults to rounding values at 40 digits in division.
This can be controlled by setting the global "ACCURACY AND PRECISION"
in Math::BigFloat, but may have a large impact on performance and
memory usage.
Complex math is incompatible with "bignum" and "bigrat" and will likely
result in NaN.
BUGS
Report any issues on the public bugtracker.
AUTHOR
Dan Book, dbook@cpan.org
COPYRIGHT AND LICENSE
Copyright 2015, Dan Book.
This library is free software; you may redistribute it and/or modify it
under the terms of the Artistic License version 2.0.
SEE ALSO
Math::Complex