I'm following an introductory course on functional programming, where we use Haskell.
Part of an excercise is to write a parser for the input string.
However I can't solve the following error, or get what is actually happening.
Parser.hs:29:71:
Couldn't match expected type `String' with actual type `Char'
In the first argument of `readPoint', namely `start'
In the expression: readPoint start
In the expression:
(readLines track, readPoint start, readLine finish)
The error originates from this line:
readTrack str = parseTrack (lines str) where
parseTrack (start : finish : track) = (readLines track, readPoint start, readLine finish)
What I expected to happen is that the input string got split into a list of lines, which get passed to parseTrack.
parseTrack would then use pattern matching to name the top two strings(lines) from the list and the rest.
However what I believe is happening is that finish is the top element from the list, and start gets assigned the top char from that string.
I would really like to know how to solve this problem and what is actually happening.
Thanks a lot!
Parser.hs
module Parser where
import Types
readFloat :: String -> Float
readFloat str = case reads str of
[] -> error "not a floating point number"
(p,_):_ -> p
readInt :: String -> Int
readInt str = case reads str of
[] -> error "not an integer"
(p,_):_ -> p
readPoint :: String -> Point
readPoint str = parsePoint (words str) where
parsePoint (x : y : _) = (readInt x, readInt y)
readLine :: String -> Line
readLine str = parseLine (words str) where
parseLine (x1 : y1 : x2 : y2 : _) = ((readInt x1, readInt y1), (readInt x2, readInt y2))
readLines :: String -> [Line]
readLines str = parseLines (lines str) where
parseLines (line : rest) = readLine line : parseLines rest
readTrack :: String -> Track
readTrack str = parseTrack (lines str) where
parseTrack (start : finish : track) = (readLines track, readPoint start, readLine finish)
Types.hs
module Types where
type Vector2D = (Int, Int)
type Point = Vector2D
type Line = (Point, Point)
type Velocity = Vector2D
type CarState = (Position, Velocity)
type Position = Vector2D
type Trace = [Position]
type Track = ([Line], Point, Line)
Your variable track was actually a list of single lines, not a string with '\n's in it. Since you've already split it into lines, you can just map readLine over it, giving:
readTrack str = parseTrack (lines str) where
parseTrack (start:finish:tracks)
= (map readLine tracks, readPoint start, readLine finish)
Here tracks :: [String], which is why you can map readLine on them all - you don't need to use readLines to split it into lines first. (You can tell it's a list because it's the final thing to the right hand side of a :.)
You say
However what I believe is happening is that finish is the top element from the list, and start gets assigned the top char from that string.
Well what was happening was: because you asked for readLines track as the first output, Haskell started there, and since you declared
readLines :: String -> [Line]
that meant that track had to be a String - that's the only thing readLines can deal with.
First, you need to remember that : has an element on its left and a list on its right, so in
3:4:stuff
stuff has to be [Integer] because it's on the right of some Integer elements. Similarly,
c:"a string"
means c has to be a Char because String = [Char].
In your code, we've worked out that track is a String, so that means that when you write
(start : finish : track)
both start and finish have to be elements you can put at the front of a String, so both start and finish have to be Char.
Haskell then looks at your code readPoint start, but because it's worked out that start has type Char, but
readPoint :: String -> Point
it complains that Char and String don't match.
I think you made the mistake because you forgot that readLines takes a single string, but it felt (from the name) like it should happily take a list of strings. Your parseLines looks like it does a similar thing, but it takes a list of strings, so copes, whereas readLines takes a single string with newline characters, so can't cope with a list.
UPD. Oh, sorry, I didn't get that track means multiple tracks and has to be of type [String]. So the answer from AndrewC fits more.
Since in haskell the pattern (x:xs) means that if x has type a then xs has to be of type [a] your pattern in parseTrack means in types smth like (a : a : [a]).
Compiler wanna evaluate type a and first what it see on the right is readLines track. Func readLines has type String -> [Line] so compiler has track as String which means [a] is of type String. Also in haskell String is [Char] so a is Char.
But you need a as String. So you just need to take first three strings and throw out the rest tail of [String]. In types it'll mean smth like (String : String : String : [String]). For it you can rewrite matching pattern in parseTrack to:
parseTrack (start : finish : track : _)
Related
Im trying to replace all strings which contain a substring by itself, in a list.
I've tried it by using the map function:
cleanUpChars = map(\w -> if isInfixOf "**" w then map(\c -> if c == "*" then ""; else c); else w)
To me this reads as: map elements in a list, such that if a character of a word contains * replace it with nothing
To Haskell: "Couldnt match expected type [[Char]] -> [[Char]] with actual type [Char] in the expression: w" (and the last w is underlined)
Any help is appreciated
To answer the revised question (when isInfixOf has been imported correctly):
cleanUpChars = map(\w -> if isInfixOf "**" w then map(\c -> if c == "*" then ""; else c); else w)
The most obvious thing wrong here is that c in the inner parentheses will be a Char (since it's the input to a function which is mapped over a String) - and characters use single quotes, not double quotes. This isn't just a case of a typo or wrong syntax, however - "" works fine as an empty string (and is equivalent to [] since Strings are just lists), but there is no such thing as an "empty character".
If, as it seems, your aim is to remove all *s from each string in the list that contains **, then the right tool is filter rather than map:
Prelude Data.List> cleanUpChars = map(\w -> if isInfixOf "**" w then filter (/= '*') w; else w)
Prelude Data.List> cleanUpChars ["th**is", "is", "a*", "t**es*t"]
["this","is","a*","test"]
(Note that in the example I made up, it removes all asterisks from t**es*t, even the single one. This may not be what you actually wanted, but it's what your logic in the faulty version implied - you'll have to be a little more sophisticated to only remove pairs of consecutive *'s.)
PS I would certainly never write the function like that, with the semicolon - it really doesn't gain you anything. I would also use the infix form of isInfixOf, which makes it much clearer which string you are looking for inside the other:
cleanUpChars :: [String] -> [String]
cleanUpChars = map (\w -> if "**" `isInfixOf` w then filter (/= '*') w else w)
I'm still not particularly happy with that for readability - there's probably some nice way to tidy it up that I'm overlooking for now. But even if not, it helps readability imo to give the function a local name (hopefully you can come up with a more concise name than my version!):
cleanUpChars :: [String] -> [String]
cleanUpChars = map possiblyRemoveAsterisks
where possiblyRemoveAsterisks w = if "**" `isInfixOf` w then filter (/= '*') w else w
So my problem is to take a string in haskell and to modify it so that if there are certain characters, they are changed to other characters, and I have created a helper function to do this, however there is one case where if the character is '!' then it become '!!!111oneone', so i figure to do this you would need to concatenate the current string with '!!111oneone', the trouble is that my function was working with chars however to do this we would need to work with the string, how would you combine this, ie a helper to modify the chars if necessary and implementing the conversion if there is a '!'.
Here is what i have so far
convert :: String -> String
convert [] = []
convert (x:xs) =
| x == '!' = !helper
| otherwise = converthelper x
Assuming your helper is something like
helper :: Char -> String
helper '!' = "!!!111oneone"
helper c = [c]
then you can use concatMap to map helper over each character in your string, and then concatenate the results into a single string.
convert :: String -> String
convert = concatMap helper
-- convert msg = concatMap helper msg
The trick is that your helper promotes every character to a list of characters; most characters just become the corresponding one-character string, but ! becomes something more.
(Note that concatMap forms the basis of the Monad instance for lists. You could also write convert msg = msg >>= helper.)
I want to make string "t1", "t2","t3", ...so on.
so I did like this
let i =0;...
(something) -> let z = "t" ^ string_of_int (i+1)
my intention is every time that program enters (something), i increases.
because I have to make new string when program enters (something).
but it had syntax error
What should I do?
If you want to embed an integer as part of a string, use Printf.sprintf:
let make_string i =
Printf.sprintf "t%d" i
You'll have to take care of tht part that generates increasing numbers with something else, like a for loop.
input file is txt :
000011S\n
0001110\n
001G111\n
0001000\n
Result is:
[["0","0","0","0","1","1","S"], ["0","0","0","1","1","1","0"] [...]]
Read a text file with
file <- openFile nameFile ReadMode
and the final output
[["a","1","0","b"],["d","o","t","2"]]
is a map with list of char
try to:
convert x = map (map read . words) $ lines x
but return [[string ]]
As it could do to return the output I want? [[Char]],
is there any equivalent for word but for char?
one solution
convert :: String -> [[String]]
convert = map (map return) . lines
should do the trick
remark
the return here is a neat trick to write \c -> [c] - wrapping a Char into a singleton list as lists are a monad
how it works
Let me try to explain this:
lines will split the input into lines: [String] which each element in this list being one line
the outer map (...) . lines will then apply the function in (...) to each of this lines
the function inside: map return will again map each character of a line (remember: a String is just a list of Char) and will so apply return to each of this characters
now return here will just take a character and put it into a singleton list: 'a' -> [a] = "a" which is exactly what you wanted
your example
Prelude> convert "000011S\n0001110\n001G111\n0001000\n"
[["0","0","0","0","1","1","S"]
,["0","0","0","1","1","1","0"]
,["0","0","1","G","1","1","1"]
,["0","0","0","1","0","0","0"]]
concerning your comment
if you expect convert :: String -> [[Char]] (which is just String -> [String] then all you need is convert = lines!
[[Char]] == [String]
Prelude> map (map head) [["a","1","0","b"],["d","o","t","2"]]
["a10b","dot2"]
will fail for empty Strings though.
or map concat [[...]]
Im trying to create a list of strings using some recursion.
Basically i want to take a part of a string up to a certain point. Create a list from that and then process the rest of the string through recursion.
type DocName = FilePath
type Line = (Int,String)
type Document = [Line]
splitLines :: String -> Document
splitLines [] = []
splitLines str | length str == 0 = []
| otherwise = zip [0..(length listStr)] listStr
where
listStr = [getLine] ++ splitLines getRest
getLine = (takeWhile (/='\n') str)
getRest = (dropWhile (=='\n') (dropWhile (/='\n') str))
Thats what i got. But it just concats the strings back together since they are list of characters themselves. But i want to create a list of strings.
["test","123"] if the input was "test\n123\n"
Thanks
If you try to compile your code, you'll get an error message telling you that in the line
listStr = [getLine] ++ splitLines getRest
splitLines getRest has type Document, but it should have type [String]. This is easy enough to understand, since [getLine] is a list of strings (well a list of one string) and so it can only be concatenated with another list of strings, not a list of int-string-tuples.
So to fix this we can use map to replace each int-string-tuple in the Document with only the string to get a list of strings, i.e.:
listStr = [getLine] ++ map snd (splitLines getRest)
After changing the line to the above your code will compile and run just fine.
But it just concats the strings back together since they are list of characters themselves.
I'm not sure why you think that.
The reason your code did not compile was because of the type of splitLines as I explained above. Once you fix that error, the code behaves exactly as you want it to, returning a list of integer-string-tuples. At no point are strings concatenated.
Well, if you wrote this just to practice recursion then it is fine once you fix error mentioned by sepp2k. But in real code, I would prefer -
splitLines str = zip [0..] (lines str)
Or even
splitLines = zip [0..] . lines