Recurisve descent parsing library for Python based on functional combinators
Python Makefile
Latest commit 0b68992 Feb 24, 2016 @vlasovskikh Merge pull request #41 from ljos/patch-1
Fix spelling mistake in README.md

README.md

funcparserlib

Recurisve descent parsing library for Python based on functional combinators.

Description

Parser combinators are just higher-order functions that take parsers as their arguments and return them as result values. Parser combinators are:

  • First-class values
  • Extremely composable
  • Tend to make the code quite compact
  • Resemble the readable notation of xBNF grammars

Parsers made with funcparserlib are pure-Python LL(*) parsers. It means that it's very easy to write them without thinking about look-aheads and all that hardcore parsing stuff. But the recursive descent parsing is a rather slow method compared to LL(k) or LR(k) algorithms.

So the primary domain for funcparserlib is parsing little languages or external DSLs (domain specific languages).

The library itself is very small. Its source code is only 0.5 KLOC, with lots of comments included. It features the longest parsed prefix error reporting, as well as a tiny lexer generator for token position tracking.

Show Me the Code

This is an excerpt from a JSON parser (RFC 4627) written using funcparserlib. This full example as well as others can be found here.

def parse(seq):
    'Sequence(Token) -> object'
    ...
    n = lambda s: a(Token('Name', s)) >> tokval
    def make_array(n):
        if n is None:
            return []
        else:
            return [n[0]] + n[1]
    ...
    null = n('null') >> const(None)
    true = n('true') >> const(True)
    false = n('false') >> const(False)
    number = toktype('Number') >> make_number
    string = toktype('String') >> make_string
    value = forward_decl()
    member = string + op_(':') + value >> tuple
    object = (
        op_('{') +
        maybe(member + many(op_(',') + member)) +
        op_('}')
        >> make_object)
    array = (
        op_('[') +
        maybe(value + many(op_(',') + value)) +
        op_(']')
        >> make_array)
    value.define(
          null
        | true
        | false
        | object
        | array
        | number
        | string)
    json_text = object | array
    json_file = json_text + skip(finished)

    return json_file.parse(seq)

Installation

You can install the funcparserlib library from PyPI via pip:

$ pip install funcparserlib

There are no dependencies on other libraries.

Documentation

A short intro to funcparserlib can be found in the Nested Brackets Mini-HOWTO.

The comprehensive funcparserlib Tutorial is also available.

See also comments inside the modules funcparserlib.parser and funcparserlib.lexer or generate the API docs from the modules using pydoc.

There a couple of examples available in the funcparserlib/tests directory:

See also the changelog and FAQ.

Performance and Code Size

Despite being an LL(*) parser, funcparserlib has a reasonable performance. For example, a JSON parser written using funcparserlib is 3 times faster than a parser using the popular pyparsing library and only 5 times slower than the specialized JSON library simplejson that uses ad hoc parsing. Here are some stats1:

File Size cjson simplejson funcparserlib json-ply pyparsing
6 KB 0 ms 45 ms 228 ms n/a 802 ms
11 KB 0 ms 80 ms 395 ms 367 ms 1355 ms
100 KB 4 ms 148 ms 855 ms 1071 ms 2611 ms
134 KB 11 ms 957 ms 4775 ms n/a 16534 ms
1009 KB 87 ms 6904 ms 36826 ms n/a 116510 ms
User Code 0.9 KLOC 0.8 KLOC 0.1 KLOC 0.5 KLOC 0.1 KLOC
Library Code 0 KLOC 0 KLOC 0.5 KLOC 5.3 KLOC 3.7 KLOC

funcparserlib and pyparsing both have the smallest user code size (that is a common feature of parsing libraries compared to ad hoc parsers). The library code of funcparserlib is 7 times smaller (and much more cleaner) than pyparsing. The json-ply uses a LALR parser ply (similar to Yacc) and performs like funcparserlib. cjson is a C library, hence the incredible performance :)

Similar Projects

  • LEPL. A recursive descent parsing library that uses two-way generators for backtracking. Its source code is rather large: 17 KLOC
  • pyparsing. A recursive descent parsing library. Probably the most popular Python parsing library. Nevertheless its source code is quite dirty (though 4 KLOC only)
  • Monadic Parsing in Python. A series of blog entries on monadic parsing
  • Pysec (aka Parsec in Python). A blog entry on monadic parsing, with nice syntax for Python

1 Testing hardware: Pentium III, 1 GHz, 512 MB. JSON files were taken from a real project, in a normalized encoding, i. e. they contained no extra separators. The version 0.3.2 of the library was used.