mirror of https://github.com/mrkkrp/megaparsec.git synced 2024-11-23 10:23:39 +03:00

[skip ci]

2018-09-03 22:15:54 +07:00

16 KiB

Raw Blame History

Megaparsec

Features
Documentation
Tutorials
Performance
Comparison with other solutions
Related packages
Prominent projects that use Megaparsec
Links to announcements and blog posts
Authors
Contribution
License

This is an industrial-strength monadic parser combinator library. Megaparsec is a feature-rich package that strikes a nice balance between speed, flexibility, and quality of parse errors.

Features

The project provides flexible solutions to satisfy common parsing needs. The section describes them shortly. If you're looking for comprehensive documentation, see the section about documentation.

Core features

The package is built around MonadParsec, an MTL-style monad transformer. All tools and features work with all instances of MonadParsec. You can achieve various effects combining monad transformers, i.e. building a monadic stack. Since the common monad transformers like WriterT, StateT, ReaderT and others are instances of the MonadParsec type class, you can wrap ParsecT in these monads, achieving, for example, backtracking state.

On the other hand ParsecT is an instance of many type classes as well. The most useful ones are Monad, Applicative, Alternative, and MonadParsec.

Megaparsec includes all functionality that is available in Parsec plus features some combinators that are missing in other parsing libraries:

failure allows to fail reporting a parse error with unexpected and expected items.
fancyFailure allows to fail reporting custom error messages.
withRecovery allows to recover from parse errors “on-the-fly” and continue parsing. Once parsing is finished, several parse errors may be reported or ignored altogether.
observing allows to “observe” parse errors without ending parsing (they are returned in Left, while normal results are wrapped in Right).

In addition to that, Megaparsec features high-performance combinators similar to those found in Attoparsec:

tokens makes it easy to parse several tokens in a row (string and string' are built on top of this primitive). This is about 100 times faster than matching a string token by token. tokens returns “chunk” of original input, meaning that if you parse Text, it'll return Text without any repacking.
takeWhile and takeWhile1 are about 150 times faster than approaches involving many, manyTill and other similar combinators.
takeP allows to grab n tokens from the stream and returns them as a “chunk” of the stream.

Megaparsec is about as fast as Attoparsec if you write your parser carefully (see also the section about performance).

Megaparsec can currently work with the following types of input stream out-of-the-box:

String = [Char]
ByteString (strict and lazy)
Text (strict and lazy)

It's also simple to make it work with custom token streams, and Megaparsec users have done so many times.

Error messages

Megaparsec has well-typed error messages and the ability to use custom data types to adjust the library to specific domain of interest. No need to use a shapeless bunch of strings.

Megaparsec 7 introduced the ParseErrorBundle data type that helps to manage multi-error messages and pretty-print them easily and efficiently. That version of the library also made the practice of displaying offending line the default.

Alex support

Megaparsec works well with streams of tokens produced by tools like Alex. The design of the Stream type class has been changed significantly in versions 6 and 7, but user can still work with custom streams of tokens without problems.

Character and binary parsing

Megaparsec has decent support for Unicode-aware character parsing. Functions for character parsing live in the Text.Megaparsec.Char module. Similarly, there is Text.Megaparsec.Byte module for parsing streams of bytes.

Lexer

Text.Megaparsec.Char.Lexer is a module that should help you write your lexer. If you have used Parsec in the past, this module “fixes” its particularly inflexible Text.Parsec.Token.

Text.Megaparsec.Char.Lexer is intended to be imported using a qualified import, it's not included in Text.Megaparsec. The module doesn't impose how you should write your parser, but certain approaches may be more elegant than others. An especially important theme is parsing of white space, comments, and indentation.

The design of the module allows you quickly solve simple tasks and doesn't get in your way when you want to implement something less standard.

Since Megaparsec 5, all tools for indentation-sensitive parsing are available in Text.Megaparsec.Char.Lexer module—no third party packages required.

Text.Megaparsec.Byte.Lexer is also available for users who wish to parse binary data.

Documentation

Megaparsec is well-documented. See the current version of Megaparsec documentation on Hackage.

Tutorials

You can find Megaparsec tutorials here. They should provide sufficient guidance to help you to start with your parsing tasks. The site also has instructions and tips for Parsec users who decide to migrate to Megaparsec.

Performance

Despite being flexible, Megaparsec is also fast. Here is how Megaparsec 7.0.0 compares to Attoparsec 0.13.2.2 (the fastest widely used parsing library in the Haskell ecosystem):

Test case	Execution time	Allocated	Max residency
CSV (Attoparsec)	76.50 μs	397,784	10,544
CSV (Megaparsec)	64.69 μs	352,408	9,104
Log (Attoparsec)	302.8 μs	1,150,032	10,912
Log (Megaparsec)	337.8 μs	1,246,496	10,912
JSON (Attoparsec)	18.20 μs	128,368	9,032
JSON (Megaparsec)	25.45 μs	203,824	9,176

The benchmarks were created to guide development of Megaparsec 6 and can be found here.

If you think your Megaparsec parser is not efficient enough, take a look at these instructions.

Comparison with other solutions

There are quite a few libraries that can be used for parsing in Haskell, let's compare Megaparsec with some of them.

Megaparsec vs Attoparsec

Attoparsec is another prominent Haskell library for parsing. Although the both libraries deal with parsing, it's usually easy to decide which you will need in particular project:

Attoparsec is sometimes faster but not that feature-rich. It should be used when you want to process large amounts of data where performance matters more than quality of error messages.
Megaparsec is good for parsing of source code or other human-readable texts. It has better error messages and it's implemented as monad transformer.

So, if you work with something human-readable where size of input data is usually not huge, just go with Megaparsec, otherwise Attoparsec may be a better choice.

Since version 6, Megaparsec features the same fast primitives that Attoparsec has, so in many cases the difference in speed is not that big. Megaparsec now aims to be “one size fits all” ultimate solution to parsing, so it can be used even to parse low-level binary formats.

Megaparsec vs Parsec

Since Megaparsec is a fork of Parsec, we are bound to list the main differences between the two libraries:

Better error messages. We test our error messages using numerous QuickCheck (generative) tests. Good error messages are just as important for us as correct return values of our parsers. Megaparsec will be especially useful if you write a compiler or an interpreter for some language.
Megaparsec 6 can show the line on which parse error happened as part of parse error. This makes it a lot easier to figure out where the error happened.
Some quirks and “buggy features” (as well as plain bugs) of original Parsec are fixed. There is no undocumented surprising stuff in Megaparsec.
Better support for Unicode parsing in Text.Megaparsec.Char.
Megaparsec has more powerful combinators and can parse languages where indentation matters out-of-the-box.
Comprehensive test suite covering nearly 100% of our code. Compare that to absence
We have benchmarks to detect performance regressions.
Better documentation, with 100% of functions covered, without typos and obsolete information, with working examples. Megaparsec's documentation is well-structured and doesn't contain things useless to end users.
Megaparsec's code is clearer and doesn't contain “magic” found in original Parsec.
Megaparsec has well-typed error messages and custom error messages.
Megaparsec can recover from parse errors “on the fly” and continue parsing.
Megaparsec allows to conditionally process parse errors inside your parser before parsing is finished. In particular, it's possible to define regions in which parse errors, should they happen, will get a “context tag”, e.g. we could build a context stack like “in function definition foo”, “in expression x”, etc. This is not possible with Parsec.
Megaparsec is faster and supports efficient operations on top of tokens, takeWhileP, takeWhile1P, takeP like Attoparsec.

If you want to see a detailed change log, CHANGELOG.md may be helpful. Also see this original announcement for another comparison.

Megaparsec vs Trifecta

Trifecta is another Haskell library featuring good error messages. It's probably good, but also under-documented, and has unfixed bugs and flaws. Other reasons one may question choice of Trifecta is his/her parsing library:

Complicated, doesn't have any tutorials available, and documentation doesn't help at all.
Trifecta can parse String and ByteString natively, but not Text.
Trifecta's error messages may be different with their own features, but certainly not as flexible as Megaparsec's error messages in the latest versions.
Depends on lens. This means you'll pull in half of Hackage as transitive dependencies. Also if you're not into lens and would like to keep your code “vanilla”, you may not like the API.

Idris has recently switched from Trifecta to Megaparsec which allowed it to have better error messages and fewer dependencies.

Megaparsec vs Earley

Earley is a newer library that allows to safely (it your code compiles, then it probably works) parse context-free grammars (CFG). Megaparsec is a lower-level library compared to Earley, but there are still enough reasons to choose it over Earley:

Megaparsec is faster.
Your grammar may be not context-free or you may want introduce some sort of state to the parsing process. Almost all non-trivial parsers require something of this sort. Even if your grammar is context-free, state may allow to add some additional niceties. Earley does not support that.
Megaparsec's error messages are more flexible allowing to include arbitrary data in them, return multiple error messages, mark regions that affect any error that happens in those regions, etc.
The approach Earley uses differs from the conventional monadic parsing. If you work not alone, people you work with, especially beginners, will be much more productive with libraries taking more traditional path to parsing like Megaparsec.

IOW, Megaparsec is less safe but also more powerful.

The following packages are designed to be used with Megaparsec (open a PR if you want to add something to the list):

hspec-megaparsec—utilities for testing Megaparsec parsers with with Hspec.
cassava-megaparsec—Megaparsec parser of CSV files that plays nicely with Cassava.
tagsoup-megaparsec—a library for easily using TagSoup as a token type in Megaparsec.

Prominent projects that use Megaparsec

The following are some prominent projects that use Megaparsec:

Idris—a general-purpose functional programming language with dependent types
Hledger—an accounting tool
MMark—strict markdown processor for writers
Stache—Mustache templates for Haskell
Language Puppet—library for manipulating Puppet manifests

Links to announcements and blog posts

Here are some blog posts mainly announcing new features of the project and describing what sort of things are now possible:

Authors

The project was started and is currently maintained by Mark Karpov. You can find the complete list of contributors in the AUTHORS.md file in the official repository of the project. Thanks to all the people who propose features and ideas, although they are not in AUTHORS.md, without them Megaparsec would not be that good.

Contribution

Issues (bugs, feature requests or otherwise feedback) may be reported in the GitHub issue tracker for this project.

Pull requests are also welcome (and yes, they will get attention and will be merged quickly if they are good).

License

Distributed under FreeBSD license.

16 KiB Raw Blame History Unescape Escape