Getting Started

For this introduction we assume that you have already installed java8 (jre) on MacOS X or other Unix-based operating system.

Installation

To use the nez command, download the nez.jar file:

curl -O http://nez-peg.github.io/download/nez.jar
sudo cp nez.jar /usr/local/lib/nez.jar
alias nez='java -jar /usr/local/lib/nez.jar'
alias inez='java -jar /usr/local/lib/nez.jar shell'

Also, it is good idea to put the alias setting in your .bash_profile or something like it.

To test the installation, enter the nez command and check the version:

$ nez
inez
Nez-1.0-892 (beta) on Java JVM-1.8.0_05
Copyright (c) 2014-2016, Kimio Kuramitsu and Nez project authors

Nez Grammar at a Glance

Nez grammar describes both syntactic constructs for the input and data structures for the output.

The syntactic part is based on parsing expression grammars.

Expression = Sum
Sum = Product (( '+' / '-' ) Product )*
Product = Value (( '*' / '/' ) Value)*
Value = [0-9]+ / '(' Expression ')'

The data structure part can be specified by the Nez original notations such as { } and $( ).

Expression = Sum
Sum = Product {$left ( '+' #Add / '-' #Sub ) $right(Product) }*
Product = Value {$left ( '*' #Mul / '/' #Div ) $right(Value) }*
Value = { [0-9]+ #Integer } / '(' Expression ')'

Sample files

Nez grammar is specified with Nez language and stored in a .nez file. Before you learn how to specify a grammar with Nez, why don’t you try sample grammars that Nez project has provided:

math.nez - a mathematical operator
json.nez - a syntax of JSON format
xml.nez - a syntax of XML1.0 format
js.nez - JavaScript grammar
others - avaliable on Grammar repository.

Generating a Parser

Nez generates a parser in three ways: source generation, grammar translation, and dynamic parsers.

Parser source generation

Since the Yacc invented in the early 1970s, the parser source code generation is a standard and practical approach to modern parser development. As with the Yacc, you can generate a parser from your grammar.

Use the nez parser command:

nez parser -g math.nez

The above command writes a parser written in C into a file with the name math.c.

Thanks to Nez’s declarative notations, your grammar does not depend on any programming languages. You can switch other parser languages by specifying different formats.

nez parser -g math.nez –format java

The supported parser languages are also extensible. For further information, check here.

Grammar translation

Nez can translate your grammar into another PEG-dialects, such as PEGjs (JavaScript), PEGTL (C++), LPeg (Lua), etc. To translate a grammar, use the nez peg command:

nez peg -g pegjs --format pegjs

The above command writes the translated grammar into a file with the name math.pegjs. As with in the parser source code generation, you can switch other grammar languages by specifying different formats.

The translated grammar is available to generate a parser with third-parties rules.

Dynamic parsing

Nez itself can parse the input by loading your grammar at runtime. This is a new style of parsing, which requires no source code generation.

To parse an input, use the nez parse command with input files or --input specified inline text.

nez parse -g math.nez --input '1+2*3'