Numeric.Parse.Flexible

Flexible numeric parsers for real-world programming languages. These parsers aim to be a superset of the numeric syntaxes across the most popular programming languages.

All parsers assume any trailing whitespace has already been consumed, and places no requirement for an endOfInput at the end of a literal. Be sure to handle these in a calling context. These parsers do not use TokenParsing, and therefore may fail while consuming input, depending on if you use a parser that automatically backtracks or not. Apply try if needed.

Documentation

integer :: (CharParsing m, Monad m) => m Integer

Parse an integer in decimal, hexadecimal, octal, or binary, with optional leading sign.

Note that because the octal parser takes primacy over decimal, numbers with a leading 0 will be parsed as octal. This is unfortunate, but matches the behavior of C, Python, and Ruby.

natural :: (CharParsing m, Monad m) => m Natural

Parse a natural number in decimal, hexadecimal, octal, or binary. As with integer, a leading 0 is interpreted as octal. Leading signs are not accepted.

decimal :: (CharParsing m, Monad m) => m Integer

Parse an integer in base 10.

Accepts 0..9 and underscore separators. No leading signs are accepted.

hexadecimal :: forall a m. (Eq a, Num a, CharParsing m, Monad m) => m a

Parse a number in hexadecimal.

Requires a 0x or 0X prefix. No leading signs are accepted. Accepts A..F, a..f, 0..9 and underscore separators.

octal :: forall a m. (Num a, CharParsing m, Monad m) => m a

Parse a number in octal.

Requires a 0, 0o or 0O prefix. No leading signs are accepted. Accepts 0..7 and underscore separators.

binary :: forall a m. (Show a, Num a, CharParsing m, Monad m) => m a

Parse a number in binary.

Requires a 0b or 0B prefix. No leading signs are accepted. Accepts 0, 1, and underscore separators.

floating :: (CharParsing m, Monad m) => m Scientific

Parse an arbitrary-precision number with an optional decimal part.

Unlike scientificP or Scientific's Read instance, this handles:

• omitted whole parts, e.g. .5
• omitted decimal parts, e.g. 5.
• exponential notation, e.g. 3.14e+1
• numeric parts, in whole or decimal or exponent parts, with _ characters
• hexadecimal, octal, and binary integer literals, without a decimal part.

You may either omit the whole or the leading part, not both; this parser also rejects the empty string. It does not handle hexadecimal floating-point numbers.

signed :: forall a m. (CharParsing m, Num a) => m a -> m a

Converts a numeric parser to one that accepts an optional leading sign.

imaginary :: forall a m. (CharParsing m, Monad m, Num a) => m a -> m (Complex a)

Converts a numeric parser to one that accepts a trailing imaginary specifier i or j. This does not add facilities for two-valued literals, i.e. 1+4j, as those are generally best left to high-level expression facilities.