Description

A generator of random texts from EBNF-like grammars.

Install

  • Install Ruby 1.9.3 or higher.
  • gem install gen-text

Example

Simple grammar:

expr =
  expr ('+'|'-') expr |
  expr ('*'|'/') expr |
  '(' expr ')' |
  0...1000 ;

Sample output:

(914-548*155+131*799)*827

See more examples in "sample" directory.

Usage

Run it with the command:

gen-text file.g

Here file.g is a file containing the grammar description which consists of the rule definitions:

nonterminal1 = expr1 ;

nonterminal2 = expr2 ;

nonterminal3 = expr3 ;

Nonterminals start from a letter or "_" and may contain alphanumeric characters, "_", "-" and ":".

You may also use backquoted nonterminals:

`nonterminal1` = expr1 ;

`Nonterminal with arbitrary characters: [:|:]\/!` = expr2 ;

Trailing ";" can be omitted:

nonterminal1 = expr1

nonterminal2 = expr2

nonterminal3 = expr3

Also you may omit the left part in the first rule:

expr1

nonterminal2 = expr2

nonterminal3 = expr3

You may use C++-like comments:

// A comment

/* Another comment */

Expressions

You may use the following expressions in the rules' right part:

Expression Meaning
"str"
'str' 		Generate a string.

Following escape sequences are allowed: "\n", "\t", "\e" and "\." where "." is an arbitrary character.
U+HHHH Generate an UTF-8 character sequence corresponding to the Unicode code. E. g.: U+000A is equivalent to "\n".
n (a number) Generate a number
m...n Generate a random number between m and n (inclusive).
nonterm —
Combinators
expr expr Sequence.
expr | expr Random choice. 
| expr
| expr
 		Random choice (another form). 
| [m%] expr
| [n%] expr
| expr
 		Random choice with specific probabilities.

If probability is unspecified then it is calculated automatically. 
| [0.1] expr
| [0.3] expr
| expr
 		The same as above. Probabilities may be specified as floating point numbers between 0.0 and 1.0.
expr*
expr+
expr?
expr*[n]
expr*[m...n]
Repeat expr many times:
  • 0 or more times
  • 1 or more times
  • 0 or 1 time
  • exactly n times
  • between m and n times

Note: you may use inf ("infinity") instead of m or n.
n:expr Capture a string generated by expr into variable n. The variable may then be used in Ruby code insertions.

You may also use names like $n and @n.

Note: presence of this expression turns on backtracking and output buffering and may result in enormous memory usage.
n := expr The same as above but the output generated by expr is discarded.
Ruby code insertions
{ code } Execute the code. Generate nothing.

Note: all code insertions inside a rule share the same scope.
{= code } Generate a string returned by the code.
{? code } Condition. A code which must evaluate to true.

Note: presence of this expression turns on backtracking and output buffering and may result in enormous memory usage.

Alternate syntax

You may use "<-" instead of "=":

nonterm1 <- expr1 ;

nonterm2 <- expr2 ;

and "/" instead of "|":

`simple choice` <- "a" / "b" ;