Sprout

Warning: This crate is still in the early stages of development, and is not guaranteed to be stable. Also, I built this for fun to use with my personal projects without having any idea about how stuff like this usually works, so until I eventually look into this topic more and find time for optimiziations, be warned that this implementation is likely not optimal.

sprout is a rust crate that parses text into ASTs (Abstract Syntax Trees), given definitions for tokens and grammars built from those tokens.

Usage

You can start by importing the sprout prelude. For most projects, this will be sufficient.

use sprout::prelude::*;

Example

First, define an enum for your tokens. It should derive the necessary traits shown below.

#[derive(Debug, PartialEq, Eq, Clone, Copy)]
pub enum Token {
   Number,
   Word,
   Space
}

Implement std::fmt::Display for your token enum. This will be used to generate human-readable error messages.

use std::fmt;

impl fmt::Display for Token {
   fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
      match self {
         Self::Number => write!(f, "number"),
         Self::Word => write!(f, "word"),
         Self::Space => write!(f, "space")
      }
   }
}

Now, you can use the alphabet macro to provide definitions for your tokens using a subset of regular expressions syntax that includes parentheses (()), square brackets ([]), ranges ([a-z]), as well as the operators *, + and ?.

use Token::*;

let alphabet = alphabet! {
   Number => "[0-9]+";
   Word => "[a-z]+";
   Space => " "
};

Next, you define your grammar, in terms of "procedures", or "procs". They are the abstract parts of your language that will end up forming the nodes of your syntax tree.

First of all, you again create an enum:

#[derive(Debug, PartialEq, Eq, Clone, Copy)]
pub enum Proc {
   TwoOrThreeWords,
   WordOrNumber,
   Sequence
}

And, again, implement std::fmt::Display:

impl fmt::Display for Proc {
   fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
      match self {
         Self::TwoOrThreeWords => write!(f, "two or three words"),
         Self::WordOrNumber => write!(f, "word or number"),
         Self::Sequence => write!(f, "word/number sequence")
      }
   }
}

Finally, define your procedures using the grammar macro.

use Token::*;
use Proc::*;

let grammar = grammar! {
   #TwoOrThreeWords => Word, Space, Word, (Space, Word)?;
   #WordOrNumber => [Word; Number];
   #Sequence =>
      ([#TwoOrThreeWords; #WordOrNumber], Space)*,
      [#TwoOrThreeWords; #WordOrNumber];
};

As you can see in this example, within a grammar definition, names of procedures are always prefixed with a #. Names of tokens are simply left as-is. Sequences of tokens/procedures are comma-separated, and you can use the following special syntax for more complex patterns:

Now, finally, from your alphabet and grammar, you can construct a parser:

let parser = Parser::new(alphabet, grammar);

This parser will now simply spit out an AST (or a ParsingError) for any string you throw at it! For example this input

let tree = parser.parse(Proc::Sequence, "abc ab 123 xyz 69".to_string());

would produce an AST like this:

Sequence "abc ab 123 xyz 69"
   TwoOrThreeWords "abc ab"
   WordOrNumber "123"
   WordOrNumber "xyz"
   WordOrNumber "69"

Specifically, the output type, AST, is an alias for a trees::Tree of ASTNodes, where ASTNode has these fields:

For more information about the API of trees::Tree see the trees documentation