Haskell creating new data

问题

{- P A R T 2 : Implementation of a parser for Łukasiewicz expressions
--TODO Define the type LExpTree, using the constructors L, V, N, Q, S, K, A, O, E, I
L for Lukasiewicz literals (i.e. C, I or U)
V for Variables (string) 
N for a node representing the prefix neg
Q for a node representing the prefix pos
S for a node representing the prefix cer
K for a node representing the prefix unk
A for a node representing the infix And
O for a node representing the infix Or
E for a node representing the infix Equivalence
M for a node representing the infix Implication
It will be convenient to have the new type derive from the classes Show and Eq
-}
data LExpTree a = L a
                | V [Char]
                | N (LExpTree a)
                | Q (LExpTree a)
                | S (LExpTree a)
                | K (LExpTree a)
                | A (LExpTree a) (LExpTree a)
                | O (LExpTree a) (LExpTree a)
                | E (LExpTree a) (LExpTree a)
                | M (LExpTree a) (LExpTree a)
                deriving (Show, Eq)

This is my Haskell Project. I have been trying it for 4 days. I just wanna verify whether my LExpTree data is correct?

I don't think so though..

After that, I have to create a series of parsers as grammar

This is 3 of them.

{- 
lukExp :: lukOpd ( '<->' lukExp | '-->' lukExp| e )
lukVar :: lowercase (Alphanumeric)*
lukLit :: C | I | U
-}
lukExp :: Parser LExpTree
lukExp = do o <- lukOpd
            ( do { symbol "<->";
                   e <- lukExp;
                   return (E (o <-> e)); }
             +++ do { symbol "-->";
                      e <- lukExp;
                      return (E (o --> e)); }
             +++ return (E o) )

lukVar :: Parser LExpTree 
lukVar = ident

lukLit :: Parser LExpTree
lukLit = do { symbol "C";
              return (L C); }
         +++ do { symbol "I";
                  return (L I); }
         +++ do { symbol "U";
                  return (L U); }

And I am getting these errors.

proj.hs:131:18:
    Expecting one more argument to `LExpTree'
    The first argument of `Parser' should have kind `*',
      but `LExpTree' has kind `* -> *'
    In the type signature for `lukExp': lukExp :: Parser LExpTree

proj.hs:184:18:
    Expecting one more argument to `LExpTree'
    The first argument of `Parser' should have kind `*',
      but `LExpTree' has kind `* -> *'
    In the type signature for `lukVar': lukVar :: Parser LExpTree

proj.hs:187:18:
    Expecting one more argument to `LExpTree'
    The first argument of `Parser' should have kind `*',
      but `LExpTree' has kind `* -> *'
    In the type signature for `lukLit': lukLit :: Parser LExpTree
Failed, modules loaded: Parser.

The project is to create Łukasiewicz expressions C for certain, U for uncertain and I for impossible.

And I already have a Parser.hs loaded and it has all the Parser and parse types and its associated functions.

I understand this is a school project and supposed to try it my best. I still have 2 more parts and this is supposed to be an easy part.

I would appreciate if someone can help me.

Thanks.

回答1:

The main problem is that the LExpTree type needs an additional type parameter to become a concrete type. This is, I believe, is a correct version of lukVar:

lukVar :: Parser (LExpTree a)
lukVar = do i <- ident; return (V i)
         -- alternatively: V <$> ident

Note that this is a parser for LExpTree a for any type a.

Now in lukLit you are trying to return things like L C. I don't see where C is defined, but let's suppose that the type of C is MyLits. Then the signature of lukLit is:

lukLit :: Parser (LExpTree MyLits)

Evidently the type parameter a in LExpTree a is the type of the literals you want to use.

Note the difference between lukVar and lukLit - your code for lukVar will work with any literal type hence the variable a in the type signature. The code for lukLit returns literals from MyLits and thus returns a LExpTree MyLits.

Now you should be able to fix up lukExp. My guess is that it will have the signature

lukExp :: Parser (LExpTree a)

I bet you can just remove the type signatures and let GHC infer them. In these cases having incorrect type signatures is causing the errors.

来源：https://stackoverflow.com/questions/37100836/haskell-creating-new-data

标签

haskell

functional-programming

expression-trees