Parsing and pretty printing with prisms

YOW! Lambda Jam 2016 talk on the fresnel library for unified parsing and printing using prisms.

Transcript

1. Parsing and pretty printing with prisms Fraser Tweedale @hackuador April 29, 2016
2. Intro
3. Background ASN.1 DER is “special” Existing libraries fell short
4. Assumed knowledge Haskell syntax “Traditional” functional parser combinators
5. Prism primer ghci> :t _Right _Right :: Prism (Either c a) (Either c b) a b ghci> preview _Right (Right 42) Just 42 ghci> preview _Right (Left "nope") Nothing ghci> review _Right 42 Right 42 ghci> preview (_Left . _Right) (Left (Right 42)) Just 42
6. ASN.1 and DER BasicConstraints ::= SEQUENCE { cA BOOLEAN DEFAULT FALSE, pathLenConstraint INTEGER (0..MAX) OPTIONAL }
7. Data.X509.Ext instance Extension ExtBasicConstraints where extDecode [Start Sequence,End Sequence] = ... extDecode [Start Sequence,Boolean b,End Sequence] = ... extDecode [Start Sequence,Boolean b,IntVal v,End Sequence] = ... extDecode _ = Left "unknown sequence"
8. Parsers
9. Parser newtype Parser a = Parser { runParser :: String -> Maybe (a, String) } char :: Parser Char char = Parser $ \s -> case s of "" -> Nothing (c : s ) -> Just (c, s )
10. What if. . . input type is not String? element type is not Char?
11. Control.Lens.Cons uncons :: (Cons s s a a) => s -> Maybe (a, s) -- look familiar?
12. Control.Lens.Cons class Cons s t a b where _Cons :: Prism s t (a, s) (b, t) instance Cons ByteString ByteString Word8 Word8 instance Cons Text Text Char Char instance Cons [a] [b] a b instance Cons (Vector a) (Vector b) a b -- and many more!
13. Redefining the parser type Parser s a = Prism s s (a, s) (a, s) char :: (Cons s s a a) => Parser s a char = _Cons runParser :: Parser s a -> s -> Maybe (a, s) runParser = preview
14. But if the parser is a Prism. . . type Parser s a = Prism s s (a, s) (a, s) char :: (Cons s s a a) => Parser s a char = _Cons runParser :: Parser s a -> s -> Maybe (a, s) runParser = preview runPrinter :: Parser s a -> (a, s) -> s runPrinter = review
15. Parser ∧ printer → grammar type Grammar s a = Prism s s (a, s) (a, s) element :: (Cons s s a a) => Grammar s a element = _Cons runParser :: Grammar s a -> s -> Maybe (a, s) runParser = preview runPrinter :: Grammar s a -> (a, s) -> s runPrinter = review
16. Convenience functions parse :: Grammar s a -> s -> Maybe a parse g s = fmap fst (preview g s) print :: (Monoid s) => Grammar s a -> a -> s print g a = review g (a, mempty)
17. Grammar - map (<<$>>) :: Prism a a b b -> Grammar s a -> Grammar s b p <<$>> g = g . swapped . aside p . swapped swapped :: Iso (a, b) (c, d) (b, a) (d, c) aside :: Prism s t a b -> Prism (e, s) (e, t) (e, a) (e, b)
18. Grammar - product (<<*>>) :: Grammar s a -> Grammar s b -> Grammar s (a, b) g1 <<*>> g2 = g1 . aside g2 . productIso where productIso = iso (\(a, (b, s)) -> ((a, b), s)) (\((a, b), s) -> (a, (b, s)))
19. Grammar - sum (<<+>>) :: Grammar s a -> Grammar s b -> Grammar s (Either a b) g1 <<+>> g2 = prism (\(x, s) -> either (review g1 . (,s)) (review g2 . (,s)) x) (\s -> first Left <$> preview g1 s <|> first Right <$> preview g2 s)
20. Grammar - list many :: Grammar s a -> Grammar s [a] many g = isoList <<$>> (g <<*>> many g) <<+>> success () isoList :: Iso (Either (a, [a]) ()) [a] isoList = ... -- like pure :: Applicative f => a -> f a success :: a -> Grammar s a success a = prism snd (Just . (a,))
21. Grammar - basic grammars satisfy :: (Cons s s a a) => (a -> Bool) -> Grammar s a satisfy f = prism id (\a -> guard (f a) >> pure a) <<$>> element symbol :: (Cons s s a a, Eq a) => a -> Grammar s a symbol a = satisfy (== a) eof :: (Cons s s a a) => Grammar s () eof = prism snd (\s -> maybe (Just ((), s)) (const Nothing) (uncons s))
22. Grammar - delimiters literal :: (Cons s s a a, Eq a) => a -> Grammar s () match :: Grammar s a -> a -> Grammar s () between :: Grammar s () -> Grammar s () -> Grammar s a -> Grammar s a
23. Grammar - combinators (<<*) :: Grammar s a -> Grammar s () -> Grammar s a (*>>) :: Grammar s () -> Grammar s a -> Grammar s a many1 :: Grammar s a -> Grammar s (NonEmpty a) replicate :: Natural -> Grammar s a -> Grammar s [a]
24. Grammar - bind bind :: Grammar s a -> (a -> Grammar s b) -> (b -> a) -> Grammar s b bind p f g = prism (\(b, s) -> review p (g b, review (f (g b)) (b, s))) (preview p >=> \(a, s ) -> preview (f a) s )
25. Grammar - bind ghci> let g = bind integral (\n -> replicate n alpha)) length ghci> parse (many g) "1a2ab3abc4bad!" Just ["a","ab","abc"] ghci> print (many g) ["hello", "world"] :: String "5hello5world"
26. Deriving Isos for custom types {-# LANGUAGE TemplateHaskell #-} import Data.Fresnel.TH (makeIso) data Foo = A Int Char | B Bool makeIso Foo -- results in splice -- _Foo :: Iso (Either (Int, Char) Bool) Foo _Foo = ...
27. Putting it all together data PhoneNumber = PhoneNumber { areaCode :: String , localNumber :: String } deriving (Show) makeIso PhoneNumber phoneNumber :: Cons s s Char Char => Grammar s PhoneNumber phoneNumber = _PhoneNumber <<$>> between (literal ( ) (literal ) ) (replicate 2 digit) <<* match (many space) " " <<*>> replicate 8 (match (many space) "" *>> digit)
28. Putting it all together ghci> parse phoneNumber ("(07)3456 78 9 0" :: String) Just (PhoneNumber "07" "34567890") ghci> print phoneNumber (PhoneNumber "07" "34567890") :: String "(07) 34567890"
29. ASN.1 grammars
30. Primitive types boolean :: (Cons s s ASN1 ASN1) => Grammar s Bool integer :: (Cons s s ASN1 ASN1) => Grammar s Integer octetString :: (Cons s s ASN1 ASN1) => Grammar s B.ByteString oid :: (Cons s s ASN1 ASN1) => Grammar s OID
31. OPTIONAL, DEFAULT and SEQUENCE opt :: Grammar s a -> Grammar s (Maybe a) def :: (Eq a) => a -> Grammar s a -> Grammar s a sequence :: (Cons s s ASN1 ASN1) => Grammar s a -> Grammar s a
32. ASN.1 grammar - example data BasicConstraints = NotCA | CA (Maybe Natural) _BasicConstraints :: Iso (Bool, Maybe Natural) BasicConstraints _BasicConstraints = ... basicConstraints :: (Cons s s ASN1 ASN1) => Grammar s BasicConstraints basicConstraints = _BasicConstraints <<$>> sequence ( def False boolean <<*>> opt (natural <<$>> integer) )
33. Wrapping up
34. How did I get here?
35. s/standards/libraries/g CC-BY-NC 2.5 https://xkcd.com/927/
36. How did I get here? 1. Evaluate existing solutions
37. How did I get here? 1. Evaluate existing solutions 2. Write Cons-based parser (to parse [ASN1])
38. How did I get here? 1. Evaluate existing solutions 2. Write Cons-based parser (to parse [ASN1]) 3. Hmm. . . _Cons is a prism. . .
39. How did I get here? 1. Evaluate existing solutions 2. Write Cons-based parser (to parse [ASN1]) 3. Hmm. . . _Cons is a prism. . . 4. Type tetris
40. How did I get here? 1. Evaluate existing solutions 2. Write Cons-based parser (to parse [ASN1]) 3. Hmm. . . _Cons is a prism. . . 4. Type tetris 5. fresnel is born
41. fresnel and fresnel-asn1 Repos https://github.com/frasertweedale/hs-fresnel https://github.com/frasertweedale/hs-fresnel-asn1 AGPLv3+ I have actually used it!
42. Limitations and caveats Error reporting Performance
43. Advantages One module; many stream and element types Don’t repeat yourself Reuse existing optics ASN.1 grammar corresponds to spec
44. What’s next? Hackage Kerberos, X.509 Implement binary DER decoding?
45. Resources and related topics Invertible Syntax Descriptions Paper: www.informatik.uni-marburg.de/~rendel/unparse/ Libraries: boomerang, roundtrip, invertible-syntax Christian Marie’s YLJ2015 talk Experiment in combining Applicative and Divisible https://github.com/charleso/portmanteau lens
46. Fin © 2016 Fraser Tweedale Except where otherwise noted this work is licensed under http://creativecommons.org/licenses/by/4.0/ Slides https://github.com/ frasertweedale/talks/ Email frase@frase.id.au Twitter @hackuador