Building Blocks
Predicates
The simplest building block is a parser that only considers the first character of an input. If the character satisfies some condition, we consume it from the input. We could write that with some if
statements and substr
calls, but Parsica provides abstractions for that.
isEqual('a')
is a predicate. If you call it with another argument, you get a boolean: isEqual('a')('b') == false
.
satisfy($predicate)
is a function returns a Parser
object. You can think of it as a parser constructor. This object will do the heavy lifting of taking the first character of $input
, and testing it with the predicate.
Parsica comes with some useful predicates, including boolean and/or/not combinators:
Character parsers
In practice, you may not need to use predicates and satisfy
very often. The characters API provides commonly used parsers for single characters instead:
char($x)
is defined as satisfy(isEqual($x))
so the code above is equivalent to the first example. charI()
is the case-insensitive version of char()
. It preserves the case as is:
Parsica provides various parsers for groups of characters, like alphaNumChar
, upperChar
, punctuationChar
, newline
, and digitChar
. You can find them all listed in the API Reference.
Note that even though we parsed a digitChar
, the output is a string, not an int. That's because at this point, we're parsing characters. We'll talk about outputting other types than string later.
Strings
For longer sequences of characters, you can use string
and stringI
. Keep in mind that stringI
is not just case-insensitive, but also case-preserving.
If you want the output to be consistent, you can use map
to convert it.
Other parsers
Parsica comes with a growing library of other useful parsers, such as numeric types, and spaces. Always make sure to check the API documentation to know what the type of a parser is (aka the tpye of the output that the parser will produce.) For example, parsers like space
, tab
, and newline
all output strings containing the characters they matched. On the other hand, skipSpace
will output null
, no matter if it consumed spaces or not. This makes sense because the point is to ignore them, not use them.
skipSpace
consumes all kinds of space, whereas skipHSpace
will stop consuming at newlines and carriage returns. They also come with two friends, skipSpace1
and skipHSpace1
, which expect at least on space to present.