Approaching Text Parsing in Scala

Question

I'm making an application that will parse commands in Scala. An example of a command would be:

todo get milk for friday

So the plan is to have a pretty smart parser break the line apart and recognize the command part and the fact that there is a reference to time in the string.

In general I need to make a tokenizer in Scala. So I'm wondering what my options are for this. I'm familiar with regular expressions but I plan on making an SQL like search feature also:

search todo for today with tags shopping

And I feel that regular expressions will be inflexible implementing commands with a lot of variation. This leads me to think of implementing some sort of grammar.

What are my options in this regard in Scala?

Answer 1

You want to search for "parser combinators". I have a blog post using this approach (http://cleverlytitled.blogspot.com/2009/04/shunting-yard-algorithm.html), but I think the best reference is this series of posts by Stefan Zieger (http://szeiger.de/blog/2008/07/27/formal-language-processing-in-scala-part-1/)

Answer 2

Here are slides from a presentation I did in Sept. 2009 on Scala parser combinators. (http://sites.google.com/site/compulsiontocode/files/lambdalounge/ImplementingExternalDSLsUsingScalaParserCombinators.ppt) An implementation of a simple Logo-like language is demonstrated. It might provide some insights.

Answer 3

Scala has a parser library (scala.util.parsing.combinator) which enables one to write a parser directly from its EBNF specification. If you have an EBNF for your language, it should be easy to write the Scala parser. If not, you'd better first try to define your language formally.