megaparsy

Work In Progress - ignore for now ;)

Initial motivation for this project was to build an indentation-sensitive parser. This is problematic for any parser library which is limited to expressing context-free grammars (which is most of them), unless you use a separate pre-processing step (like e.g. the parser for the Python language does).

Googling around this topic I came upon https://markkarpov.com/tutorial/megaparsec.html#indentationsensitive-parsing

So in Haskell-land we can find ‘Monadic Parser Combinators’, which do not have this restriciton to CFGs. And the Megaparsec library defines many useful combinators, including some which designed specifically to help with building indentation-sensitive parsers.

But… it was difficult for me to learn to make a parser for the first time and learn Haskell simultaneously. So I looked around for a Python option and found the brilliant Parsy library. It provides an implementation of ‘Monadic Parser Combinators’ in Python.

Nicely and simply explained here:

So in Megaparsy I have ported a bunch of the Megaparsec combinators over to Parsy. It is surprising how smoothly it went and how similar they come out - this is testament to the genius of Parsy, which is where all the clever part is.

(I haven’t done all of them yet, but the indentation-sensitive ones are here)

I’ve also roughly translated across the property-based tests from their Haskell QuickCheck originals into Python Hypothesis ones.