EBNF.cr

Library for working with Context free Grammar:

• Parse EBNF, BNF and Bison Grammar
• Convert EBNF to BNF
• Generate CNF
• Generate First/Follow sets
• Generate LR(0)/LR(1)/LALR(1)/GLR parsing Tables

Note: EBNF Grammar should follow the ISO/IEC 14977 standard as it is described here

Warning: This Project is experimental. Its APIs are not yet solidified, and are subject to change at any time.

Installation

Add this to your application's shard.yml:

dependencies:
  ebnf:
    github: jrester/EBNF.cr

Usage

• CLI
• Parse Grammar
  • EBNF Grammar
  • BNF Grammar
  • Bison/YACC Grammar
• Convert Grammar
  • Convert to BNF
  • Generate CNF
• FIRST/FOLLOW Set
• Parsing tables
  • LR(0)
  • LR(1)
  • LALR(1)
  • LL(0)
  • LL(1)
  • GLR

CLI

The library provides a simple cli to identify, convert or export grammar

$ crystal run bin/ebnf.cr -- --help
Usage: ebnf [OPTIONS] file
    --stdin                          Read from stdin
    -j, --json                       Export Grammar as json
    -c, --cnf                        Convert grammar to cnf
    -b, --bnf                        Convert grammar to bnf
    -t TYPE, --type=TYPE             Provide type of grammar. If not provided grammar will be detected automatically.
    -i, --identify                   Identify grammar
    -o FILE, --out=FILE              Output file
    -v                               --verbose
    -h, --help                       Show this help

Parsing

Grammar can be built from a string directly with #from or from a file with #from_file which will return an EBNF::Grammar. #from and #from_file raise UnknownTokenError when a token is not known and UnexpectedTokenError if the token was not expected. #from? and from_file? will return nil if an error is encountered.

require "ebnf"

EBNF::Grammar.from_file "grammar.y" #=> EBNF::Grammar

Note: This will try to recognize your Grammar and will throw an UnkownGrammarError if no grammar was recognzized. If you already know which grammar type you want to parse use EBNF::<EBNF/BNF/Bison>.from or see the examples below.

EBNF Grammar

require "ebnf"

grammar = <<-Grammar
letter = "A" | "B" | "C" | "D" | "E" | "F" | "G"
       | "H" | "I" | "J" | "K" | "L" | "M" | "N"
       | "O" | "P" | "Q" | "R" | "S" | "T" | "U"
       | "V" | "W" | "X" | "Y" | "Z" | "a" | "b"
       | "c" | "d" | "e" | "f" | "g" | "h" | "i"
       | "j" | "k" | "l" | "m" | "n" | "o" | "p"
       | "q" | "r" | "s" | "t" | "u" | "v" | "w"
       | "x" | "y" | "z" ;
digit = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9" ;
symbol = "[" | "]" | "{" | "}" | "(" | ")" | "<" | ">"
       | "'" | '"' | "=" | "|" | "." | "," | ";" ;
character = letter | digit | symbol | "_" ;

identifier = letter , { letter | digit | "_" } ;
terminal = "'" , character , { character } , "'"
         | '"' , character , { character } , '"' ;

lhs = identifier ;
rhs = identifier
     | terminal
     | "[" , rhs , "]"
     | "{" , rhs , "}"
     | "(" , rhs , ")"
     | rhs , "|" , rhs
     | rhs , "," , rhs ;

rule = lhs , "=" , rhs , ";" ;
grammar = { rule } ;
Grammar

# Parse the string directly
ebnf = EBNF::EBNF.from grammar #=> EBNF::Grammar
puts ebnf #=> letter = "A" | "B" | ...

BNF Grammar

require "ebnf"

grammar <<-BNF_Grammar
<root> ::= <foo> | <bar>
<foo> ::= "A" "B" | "B" "B"
<bar> ::= "B" "A" | "A" "B"
BNF_Grammar

bnf = EBNF::BNF.from grammar # => EBNF::Grammar

Bison/Yacc Grammar

require "ebnf"

grammar = <<-Grammar
root:
    foo             { puts "foo" }
    | bar           { puts "bar" }

foo:
    A B
    | B B

bar:
    B A
    | A B
Grammar

bison = EBNF::Bison.from grammar #=> EBNF::Grammar

Every Grammar can be exported to json with #to_json and be converted to BNF grammar using #to_bnf.

Conversions

Convert Grammar

Use Grammar#to_bnf to convert a grammar to BNF. This function transforms the grammar in place. If you want to still use your old, unconverted grammar use Grammar#to_bnf! to retrive a copy of the Grammar.

Note: This may introduce new production each of them with a unique name like 'Special_350257660880508218' To make sure each name is unique the hash value of the rules in a special segment is used.

require "ebnf"

grammar = EBNF::EBNF.from_file "grammar.y"
grammar.to_bnf.type # =>  #=> EBNF::Grammar::GrammarType::BNF

CNF

grammar.to_cnf #=> nil

This will convert grammar to CNF. The order and number of steps can be specified by an Array of EBNF::CNF::Step.

grammar.to_cnf [EBNF::CNF::START, EBNF::CNF::UNIT, EBNF::CNF::START]

This will run frist START, then UNIT and again START. The default order is:

• START
• TERM
• BIN
• DEL
• UNIT

Note: Every step will be run in the way you pass it, so in the above example START will be run two times even if that wasn't your intention.

FIRST/FOLLOW Set

Grammar#first_follow generates FIRST/FOLLOW sets. It returns a Tuple with two hashes each of them containing either the first or follow table indexed by each production.

The start production of the grammar will, if not other specified with Grammar#start, be the first production of the parsed grammar.

grammar.first_follow
  #=> (Hash(String, Set(Terminal)), Hash(String, Set(Terminal)))

Parsing Tables

LR(0)

ebnf = <<-Grammar
e = e "*" b | e '+' b | b;
b = '0' | '1';
Grammar

grammar = EBNF::EBNF.from ebnf
pp EBNF::LR.generate grammar # =>

0     {"e" => [{:goto, 1_u64}], "b" => [{:goto, 2_u64}]}
1     {"\"*\"" => [{:shift, 3_u64}], "\"+\"" => [{:shift, 4_u64}], "EOS" => [{:accept, 0_u64}]}
2     {"*" => [{:reduce, "e"}], "+" => [{:reduce, "e"}], "0" => [{:reduce, "e"}], "1" => [{:reduce, "e"}]}
3     {"b" => [{:goto, 5_u64}], "\"0\"" => [{:shift, 6_u64}], "\"1\"" => [{:shift, 7_u64}]}
4     {"b" => [{:goto, 8_u64}], "\"0\"" => [{:shift, 9_u64}], "\"1\"" => [{:shift, 10_u64}]}
5     {"*" => [{:reduce, "e"}], "+" => [{:reduce, "e"}], "0" => [{:reduce, "e"}], "1" => [{:reduce, "e"}]}
6     {"*" => [{:reduce, "b"}], "+" => [{:reduce, "b"}], "0" => [{:reduce, "b"}], "1" => [{:reduce, "b"}]}
7     {"*" => [{:reduce, "b"}], "+" => [{:reduce, "b"}], "0" => [{:reduce, "b"}], "1" => [{:reduce, "b"}]}
8     {"*" => [{:reduce, "e"}], "+" => [{:reduce, "e"}], "0" => [{:reduce, "e"}], "1" => [{:reduce, "e"}]}
9     {"*" => [{:reduce, "b"}], "+" => [{:reduce, "b"}], "0" => [{:reduce, "b"}], "1" => [{:reduce, "b"}]}
10     {"*" => [{:reduce, "b"}], "+" => [{:reduce, "b"}], "0" => [{:reduce, "b"}], "1" => [{:reduce, "b"}]}

Roadmap

• Parser
  • EBNF
  • BNF
  • Bison/YACC
  • JSON
  • YAML
• Conversions
  • EBNF to BNF
  • Bison to BNF
  • CNF
  • JSON
  • YAML
• FIRST/FOLLOW Set
• Parsig tables
  • LR(1)
  • LR(0)
  • LALR(1)
  • LL(0)
  • LL(1)
  • GLR

Contributing

1. Fork it (https://github.com/jrester/ebnf/fork)
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create a new Pull Request

Contributors

• jrester - creator, maintainer