A linear-time C++17 PEG parser generator supporting memoization, left-recursion and context-dependent grammars.

The following defines a simple calculator program. It is able to parse and evaluate the basic operations +, -, *, / while obeying operator and bracket precedence and ignoring whitespace characters between tokens.

#include <peg_parser/generator.h>
#include <iostream>

void example() {
  peg_parser::ParserGenerator<float> g;

  // Define grammar and evaluation rules
  g.setSeparator(g["Whitespace"] << "[\t ]");
  g["Sum"     ] << "Add | Subtract | Product";
  g["Product" ] << "Multiply | Divide | Atomic";
  g["Atomic"  ] << "Number | '(' Sum ')'";
  g["Add"     ] << "Sum '+' Product"    >> [](auto e){ return e[0].evaluate() + e[1].evaluate(); };
  g["Subtract"] << "Sum '-' Product"    >> [](auto e){ return e[0].evaluate() - e[1].evaluate(); };
  g["Multiply"] << "Product '*' Atomic" >> [](auto e){ return e[0].evaluate() * e[1].evaluate(); };
  g["Divide"  ] << "Product '/' Atomic" >> [](auto e){ return e[0].evaluate() / e[1].evaluate(); };
  g["Number"  ] << "'-'? [0-9]+ ('.' [0-9]+)?" >> [](auto e){ return stof(e.string()); };
  g.setStart(g["Sum"]);

  // Execute a string
  auto input = "1 + 2 * (3+4)/2 - 3";
  float result = g.run(input); // -> 5
  std::cout << input << " = " << result << std::endl;
}

PEGParser requires at least cmake 3.14 and the ability to compile C++17 code. The following shows how to compile and run the calculator example.

cmake -Hexample -Bbuild/example
cmake --build build/example -j8
./build/examples/calculator

You should familiarize yourself with the syntax of parsing expression grammars. The included examples should help you to get started.

PEGParser can be easily added to your project through CPM.cmake.

CPMAddPackage(
  NAME PEGParser
  VERSION 2.0
  GIT_REPOSITORY https://github.com/TheLartians/PEGParser.git
)

target_link_libraries(myProject PEGParser)

PEGParser is designed for ease-of-use and rapid prototyping of grammars with arbitrary complexity, and builds its parsers at run time. So far no work has been invested on optimizing the library, however it runs fast enough to be used in several production projects.

PEGParser uses memoization, resulting in linear time complexity (as a function of input string length) for grammars without left-recursion. Left-recursive grammars have squared time complexity in worst case. Memoization can also be disabled on a per-rule basis, reducing the memory footprint and allowing context-dependent rules.