megaparsec/Text/Megaparsec.hs

{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE MultiParamTypeClasses #-}
{-# LANGUAGE RankNTypes #-}
{-# LANGUAGE Safe #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE UndecidableInstances #-}

-- |
-- Module      :  Text.Megaparsec
-- Copyright   :  © 2015–present Megaparsec contributors
--                © 2007 Paolo Martini
--                © 1999–2001 Daan Leijen
-- License     :  FreeBSD
--
-- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
-- Stability   :  experimental
-- Portability :  portable
--
-- This module includes everything you need to get started writing a parser.
-- If you are new to Megaparsec and don't know where to begin, take a look
-- at the tutorial <https://markkarpov.com/tutorial/megaparsec.html>.
--
-- In addition to the "Text.Megaparsec" module, which exports and re-exports
-- almost everything that you may need, we advise to import
-- "Text.Megaparsec.Char" if you plan to work with a stream of 'Char' tokens
-- or "Text.Megaparsec.Byte" if you intend to parse binary data.
--
-- It is common to start working with the library by defining a type synonym
-- like this:
--
-- > type Parser = Parsec Void Text
-- >                      ^    ^
-- >                      |    |
-- > Custom error component    Input stream type
--
-- Then you can write type signatures like @Parser 'Int'@—for a parser that
-- returns an 'Int' for example.
--
-- Similarly (since it's known to cause confusion), you should use
-- 'ParseErrorBundle' type parametrized like this:
--
-- > ParseErrorBundle Text Void
-- >                  ^    ^
-- >                  |    |
-- >  Input stream type    Custom error component (the same you used in Parser)
--
-- Megaparsec uses some type-level machinery to provide flexibility without
-- compromising on type safety. Thus type signatures are sometimes necessary
-- to avoid ambiguous types. If you're seeing an error message that reads
-- like “Type variable @e0@ is ambiguous …”, you need to give an explicit
-- signature to your parser to resolve the ambiguity. It's a good idea to
-- provide type signatures for all top-level definitions.
module Text.Megaparsec
  ( -- * Re-exports
    -- $reexports
    module Text.Megaparsec.Pos,
    module Text.Megaparsec.Error,
    module Text.Megaparsec.Stream,
    module Control.Monad.Combinators,

    -- * Data types
    State (..),
    PosState (..),
    Parsec,
    ParsecT,

    -- * Running parser
    parse,
    parseMaybe,
    parseTest,
    runParser,
    runParser',
    runParserT,
    runParserT',

    -- * Primitive combinators
    MonadParsec (..),

    -- * Signaling parse errors
    -- $parse-errors
    failure,
    fancyFailure,
    unexpected,
    customFailure,
    region,
    registerParseError,
    registerFailure,
    registerFancyFailure,

    -- * Derivatives of primitive combinators
    single,
    satisfy,
    anySingle,
    anySingleBut,
    oneOf,
    noneOf,
    chunk,
    (<?>),
    match,
    takeRest,
    atEnd,

    -- * Parser state combinators
    getInput,
    setInput,
    getSourcePos,
    getOffset,
    setOffset,
    setParserState,
  )
where

import Control.Monad.Combinators
import Control.Monad.Identity
import Data.List.NonEmpty (NonEmpty (..))
import qualified Data.List.NonEmpty as NE
import Data.Maybe (fromJust)
import Data.Set (Set)
import qualified Data.Set as E
import Text.Megaparsec.Class
import Text.Megaparsec.Error
import Text.Megaparsec.Internal
import Text.Megaparsec.Pos
import Text.Megaparsec.State
import Text.Megaparsec.Stream

-- $reexports
--
-- Note that we re-export monadic combinators from
-- "Control.Monad.Combinators" because these are more efficient than
-- 'Applicative'-based ones (†). Thus 'many' and 'some' may clash with the
-- functions from "Control.Applicative". You need to hide the functions like
-- this:
--
-- > import Control.Applicative hiding (many, some)
--
-- † As of Megaparsec 9.7.0 'Control.Applicative.many' and
-- 'Control.Applicative.some' are as efficient as their monadic
-- counterparts.
--
-- Also note that you can import "Control.Monad.Combinators.NonEmpty" if you
-- wish that combinators like 'some' return 'NonEmpty' lists. The module
-- lives in the @parser-combinators@ package (you need at least version
-- /0.4.0/).
--
-- This module is intended to be imported qualified:
--
-- > import qualified Control.Monad.Combinators.NonEmpty as NE
--
-- Other modules of interest are:
--
--     * "Control.Monad.Combinators.Expr" for parsing of expressions.
--     * "Control.Applicative.Permutations" for parsing of permutations
--       phrases.

----------------------------------------------------------------------------
-- Data types

-- | 'Parsec' is a non-transformer variant of the more general 'ParsecT'
-- monad transformer.
type Parsec e s = ParsecT e s Identity

----------------------------------------------------------------------------
-- Running a parser

-- | @'parse' p file input@ runs parser @p@ over 'Identity' (see
-- 'runParserT' if you're using the 'ParsecT' monad transformer; 'parse'
-- itself is just a synonym for 'runParser'). It returns either a
-- 'ParseErrorBundle' ('Left') or a value of type @a@ ('Right').
-- 'errorBundlePretty' can be used to turn 'ParseErrorBundle' into the
-- string representation of the error message. See "Text.Megaparsec.Error"
-- if you need to do more advanced error analysis.
--
-- > main = case parse numbers "" "11,2,43" of
-- >          Left bundle -> putStr (errorBundlePretty bundle)
-- >          Right xs -> print (sum xs)
-- >
-- > numbers = decimal `sepBy` char ','
--
-- 'parse' is the same as 'runParser'.
parse ::
  -- | Parser to run
  Parsec e s a ->
  -- | Name of source file
  String ->
  -- | Input for parser
  s ->
  Either (ParseErrorBundle s e) a
parse = runParser

-- | @'parseMaybe' p input@ runs the parser @p@ on @input@ and returns the
-- result inside 'Just' on success and 'Nothing' on failure. This function
-- also parses 'eof', so if the parser doesn't consume all of its input, it
-- will fail.
--
-- The function is supposed to be useful for lightweight parsing, where
-- error messages (and thus file names) are not important and entire input
-- should be consumed. For example, it can be used for parsing of a single
-- number according to a specification of its format.
parseMaybe :: (Ord e, Stream s) => Parsec e s a -> s -> Maybe a
parseMaybe p s =
  case parse (p <* eof) "" s of
    Left _ -> Nothing
    Right x -> Just x

-- | The expression @'parseTest' p input@ applies the parser @p@ on the
-- input @input@ and prints the result to stdout. Useful for testing.
parseTest ::
  ( ShowErrorComponent e,
    Show a,
    VisualStream s,
    TraversableStream s
  ) =>
  -- | Parser to run
  Parsec e s a ->
  -- | Input for parser
  s ->
  IO ()
parseTest p input =
  case parse p "" input of
    Left e -> putStr (errorBundlePretty e)
    Right x -> print x

-- | @'runParser' p file input@ runs parser @p@ on the input stream of
-- tokens @input@, obtained from source @file@. The @file@ is only used in
-- error messages and may be the empty string. Returns either a
-- 'ParseErrorBundle' ('Left') or a value of type @a@ ('Right').
--
-- > parseFromFile p file = runParser p file <$> readFile file
--
-- 'runParser' is the same as 'parse'.
runParser ::
  -- | Parser to run
  Parsec e s a ->
  -- | Name of source file
  String ->
  -- | Input for parser
  s ->
  Either (ParseErrorBundle s e) a
runParser p name s = snd $ runParser' p (initialState name s)

-- | The function is similar to 'runParser' with the difference that it
-- accepts and returns the parser state. This allows us e.g. to specify
-- arbitrary textual position at the beginning of parsing. This is the most
-- general way to run a parser over the 'Identity' monad.
--
-- @since 4.2.0
runParser' ::
  -- | Parser to run
  Parsec e s a ->
  -- | Initial state
  State s e ->
  (State s e, Either (ParseErrorBundle s e) a)
runParser' p = runIdentity . runParserT' p

-- | @'runParserT' p file input@ runs parser @p@ on the input list of tokens
-- @input@, obtained from source @file@. The @file@ is only used in error
-- messages and may be the empty string. Returns a computation in the
-- underlying monad @m@ that returns either a 'ParseErrorBundle' ('Left') or
-- a value of type @a@ ('Right').
runParserT ::
  (Monad m) =>
  -- | Parser to run
  ParsecT e s m a ->
  -- | Name of source file
  String ->
  -- | Input for parser
  s ->
  m (Either (ParseErrorBundle s e) a)
runParserT p name s = snd <$> runParserT' p (initialState name s)

-- | This function is similar to 'runParserT', but like 'runParser'' it
-- accepts and returns parser state. This is thus the most general way to
-- run a parser.
--
-- @since 4.2.0
runParserT' ::
  (Monad m) =>
  -- | Parser to run
  ParsecT e s m a ->
  -- | Initial state
  State s e ->
  m (State s e, Either (ParseErrorBundle s e) a)
runParserT' p s = do
  (Reply s' _ result) <- runParsecT p s
  let toBundle es =
        ParseErrorBundle
          { bundleErrors =
              NE.sortWith errorOffset es,
            bundlePosState = statePosState s
          }
  return $ case result of
    OK _ x ->
      case NE.nonEmpty (stateParseErrors s') of
        Nothing -> (s', Right x)
        Just de -> (s', Left (toBundle de))
    Error e ->
      (s', Left (toBundle (e :| stateParseErrors s')))

----------------------------------------------------------------------------
-- Signaling parse errors

-- $parse-errors
--
-- The most general function to fail and end parsing is 'parseError'. These
-- are built on top of it. The section also includes functions starting with
-- the @register@ prefix which allow users to register “delayed”
-- 'ParseError's.

-- | Stop parsing and report a trivial 'ParseError'.
--
-- @since 6.0.0
failure ::
  (MonadParsec e s m) =>
  -- | Unexpected item (if any)
  Maybe (ErrorItem (Token s)) ->
  -- | Expected items
  Set (ErrorItem (Token s)) ->
  m a
failure us ps = do
  o <- getOffset
  parseError (TrivialError o us ps)
{-# INLINE failure #-}

-- | Stop parsing and report a fancy 'ParseError'. To report a single custom
-- parse error, see 'Text.Megaparsec.customFailure'.
--
-- @since 6.0.0
fancyFailure ::
  (MonadParsec e s m) =>
  -- | Fancy error components
  Set (ErrorFancy e) ->
  m a
fancyFailure xs = do
  o <- getOffset
  parseError (FancyError o xs)
{-# INLINE fancyFailure #-}

-- | The parser @'unexpected' item@ fails with an error message telling
-- about unexpected item @item@ without consuming any input.
--
-- > unexpected item = failure (Just item) Set.empty
unexpected :: (MonadParsec e s m) => ErrorItem (Token s) -> m a
unexpected item = failure (Just item) E.empty
{-# INLINE unexpected #-}

-- | Report a custom parse error. For a more general version, see
-- 'fancyFailure'.
--
-- > customFailure = fancyFailure . Set.singleton . ErrorCustom
--
-- @since 6.3.0
customFailure :: (MonadParsec e s m) => e -> m a
customFailure = fancyFailure . E.singleton . ErrorCustom
{-# INLINE customFailure #-}

-- | Specify how to process 'ParseError's that happen inside of this
-- wrapper. This applies to both normal and delayed 'ParseError's.
--
-- As a side-effect of the implementation the inner computation will start
-- with an empty collection of delayed errors and they will be updated and
-- “restored” on the way out of 'region'.
--
-- @since 5.3.0
region ::
  (MonadParsec e s m) =>
  -- | How to process 'ParseError's
  (ParseError s e -> ParseError s e) ->
  -- | The “region” that the processing applies to
  m a ->
  m a
region f m = do
  deSoFar <- stateParseErrors <$> getParserState
  updateParserState $ \s ->
    s {stateParseErrors = []}
  r <- observing m
  updateParserState $ \s ->
    s {stateParseErrors = (f <$> stateParseErrors s) ++ deSoFar}
  case r of
    Left err -> parseError (f err)
    Right x -> return x
{-# INLINEABLE region #-}

-- | Register a 'ParseError' for later reporting. This action does not end
-- parsing and has no effect except for adding the given 'ParseError' to the
-- collection of “delayed” 'ParseError's which will be taken into
-- consideration at the end of parsing. Only if this collection is empty the
-- parser will succeed. This is the main way to report several parse errors
-- at once.
--
-- @since 8.0.0
registerParseError :: (MonadParsec e s m) => ParseError s e -> m ()
registerParseError e = updateParserState $ \s ->
  s {stateParseErrors = e : stateParseErrors s}
{-# INLINE registerParseError #-}

-- | Like 'failure', but for delayed 'ParseError's.
--
-- @since 8.0.0
registerFailure ::
  (MonadParsec e s m) =>
  -- | Unexpected item (if any)
  Maybe (ErrorItem (Token s)) ->
  -- | Expected items
  Set (ErrorItem (Token s)) ->
  m ()
registerFailure us ps = do
  o <- getOffset
  registerParseError (TrivialError o us ps)
{-# INLINE registerFailure #-}

-- | Like 'fancyFailure', but for delayed 'ParseError's.
--
-- @since 8.0.0
registerFancyFailure ::
  (MonadParsec e s m) =>
  -- | Fancy error components
  Set (ErrorFancy e) ->
  m ()
registerFancyFailure xs = do
  o <- getOffset
  registerParseError (FancyError o xs)
{-# INLINE registerFancyFailure #-}

----------------------------------------------------------------------------
-- Derivatives of primitive combinators

-- | @'single' t@ only matches the single token @t@.
--
-- > semicolon = single ';'
--
-- See also: 'token', 'anySingle', 'Text.Megaparsec.Byte.char',
-- 'Text.Megaparsec.Char.char'.
--
-- @since 7.0.0
single ::
  (MonadParsec e s m) =>
  -- | Token to match
  Token s ->
  m (Token s)
single t = token testToken expected
  where
    testToken x = if x == t then Just x else Nothing
    expected = E.singleton (Tokens (t :| []))
{-# INLINE single #-}

-- | The parser @'satisfy' f@ succeeds for any token for which the supplied
-- function @f@ returns 'True'.
--
-- > digitChar = satisfy isDigit <?> "digit"
-- > oneOf cs  = satisfy (`elem` cs)
--
-- __Performance note__: when you need to parse a single token, it is often
-- a good idea to use 'satisfy' with the right predicate function instead of
-- creating a complex parser using the combinators.
--
-- See also: 'anySingle', 'anySingleBut', 'oneOf', 'noneOf'.
--
-- @since 7.0.0
satisfy ::
  (MonadParsec e s m) =>
  -- | Predicate to apply
  (Token s -> Bool) ->
  m (Token s)
satisfy f = token testChar E.empty
  where
    testChar x = if f x then Just x else Nothing
{-# INLINE satisfy #-}

-- | Parse and return a single token. It's a good idea to attach a 'label'
-- to this parser.
--
-- > anySingle = satisfy (const True)
--
-- See also: 'satisfy', 'anySingleBut'.
--
-- @since 7.0.0
anySingle :: (MonadParsec e s m) => m (Token s)
anySingle = satisfy (const True)
{-# INLINE anySingle #-}

-- | Match any token but the given one. It's a good idea to attach a 'label'
-- to this parser.
--
-- > anySingleBut t = satisfy (/= t)
--
-- See also: 'single', 'anySingle', 'satisfy'.
--
-- @since 7.0.0
anySingleBut ::
  (MonadParsec e s m) =>
  -- | Token we should not match
  Token s ->
  m (Token s)
anySingleBut t = satisfy (/= t)
{-# INLINE anySingleBut #-}

-- | @'oneOf' ts@ succeeds if the current token is in the supplied
-- collection of tokens @ts@. Returns the parsed token. Note that this
-- parser cannot automatically generate the “expected” component of error
-- message, so usually you should label it manually with 'label' or ('<?>').
--
-- > oneOf cs = satisfy (`elem` cs)
--
-- See also: 'satisfy'.
--
-- > digit = oneOf ['0'..'9'] <?> "digit"
--
-- __Performance note__: prefer 'satisfy' when you can because it's faster
-- when you have only a couple of tokens to compare to:
--
-- > quoteFast = satisfy (\x -> x == '\'' || x == '\"')
-- > quoteSlow = oneOf "'\""
--
-- @since 7.0.0
oneOf ::
  (Foldable f, MonadParsec e s m) =>
  -- | Collection of matching tokens
  f (Token s) ->
  m (Token s)
oneOf cs = satisfy (\x -> elem x cs)
{-# INLINE oneOf #-}

-- | As the dual of 'oneOf', @'noneOf' ts@ succeeds if the current token
-- /not/ in the supplied list of tokens @ts@. Returns the parsed character.
-- Note that this parser cannot automatically generate the “expected”
-- component of error message, so usually you should label it manually with
-- 'label' or ('<?>').
--
-- > noneOf cs = satisfy (`notElem` cs)
--
-- See also: 'satisfy'.
--
-- __Performance note__: prefer 'satisfy' and 'anySingleBut' when you can
-- because it's faster.
--
-- @since 7.0.0
noneOf ::
  (Foldable f, MonadParsec e s m) =>
  -- | Collection of taken we should not match
  f (Token s) ->
  m (Token s)
noneOf cs = satisfy (\x -> notElem x cs)
{-# INLINE noneOf #-}

-- | @'chunk' chk@ only matches the chunk @chk@.
--
-- > divOrMod = chunk "div" <|> chunk "mod"
--
-- See also: 'tokens', 'Text.Megaparsec.Char.string',
-- 'Text.Megaparsec.Byte.string'.
--
-- @since 7.0.0
chunk ::
  (MonadParsec e s m) =>
  -- | Chunk to match
  Tokens s ->
  m (Tokens s)
chunk = tokens (==)
{-# INLINE chunk #-}

-- | A synonym for 'label' in the form of an operator.
infix 0 <?>

(<?>) :: (MonadParsec e s m) => m a -> String -> m a
(<?>) = flip label
{-# INLINE (<?>) #-}

-- | Return both the result of a parse and a chunk of input that was
-- consumed during parsing. This relies on the change of the 'stateOffset'
-- value to evaluate how many tokens were consumed. If you mess with it
-- manually in the argument parser, prepare for troubles.
--
-- @since 5.3.0
match :: (MonadParsec e s m) => m a -> m (Tokens s, a)
match p = do
  o <- getOffset
  s <- getInput
  r <- p
  o' <- getOffset
  -- NOTE The 'fromJust' call here should never fail because if the stream
  -- is empty before 'p' (the only case when 'takeN_' can return 'Nothing'
  -- as per its invariants), (tp' - tp) won't be greater than 0, and in that
  -- case 'Just' is guaranteed to be returned as per another invariant of
  -- 'takeN_'.
  return ((fst . fromJust) (takeN_ (o' - o) s), r)
{-# INLINEABLE match #-}

-- | Consume the rest of the input and return it as a chunk. This parser
-- never fails, but may return the empty chunk.
--
-- > takeRest = takeWhileP Nothing (const True)
--
-- @since 6.0.0
takeRest :: (MonadParsec e s m) => m (Tokens s)
takeRest = takeWhileP Nothing (const True)
{-# INLINE takeRest #-}

-- | Return 'True' when end of input has been reached.
--
-- > atEnd = option False (True <$ hidden eof)
--
-- @since 6.0.0
atEnd :: (MonadParsec e s m) => m Bool
atEnd = option False (True <$ hidden eof)
{-# INLINE atEnd #-}

----------------------------------------------------------------------------
-- Parser state combinators

-- | Return the current input.
getInput :: (MonadParsec e s m) => m s
getInput = stateInput <$> getParserState
{-# INLINE getInput #-}

-- | @'setInput' input@ continues parsing with @input@.
setInput :: (MonadParsec e s m) => s -> m ()
setInput s = updateParserState (\(State _ o pst de) -> State s o pst de)
{-# INLINE setInput #-}

-- | Return the current source position. This function /is not cheap/, do
-- not call it e.g. on matching of every token, that's a bad idea. Still you
-- can use it to get 'SourcePos' to attach to things that you parse.
--
-- The function works under the assumption that we move in the input stream
-- only forwards and never backwards, which is always true unless the user
-- abuses the library.
--
-- @since 7.0.0
getSourcePos :: (TraversableStream s, MonadParsec e s m) => m SourcePos
getSourcePos = do
  st <- getParserState
  let pst = reachOffsetNoLine (stateOffset st) (statePosState st)
  setParserState st {statePosState = pst}
  return (pstateSourcePos pst)
{-# INLINE getSourcePos #-}

-- | Get the number of tokens processed so far.
--
-- See also: 'setOffset'.
--
-- @since 7.0.0
getOffset :: (MonadParsec e s m) => m Int
getOffset = stateOffset <$> getParserState
{-# INLINE getOffset #-}

-- | Set the number of tokens processed so far.
--
-- See also: 'getOffset'.
--
-- @since 7.0.0
setOffset :: (MonadParsec e s m) => Int -> m ()
setOffset o = updateParserState $ \(State s _ pst de) ->
  State s o pst de
{-# INLINE setOffset #-}

-- | @'setParserState' st@ sets the parser state to @st@.
--
-- See also: 'getParserState', 'updateParserState'.
setParserState :: (MonadParsec e s m) => State s e -> m ()
setParserState st = updateParserState (const st)
{-# INLINE setParserState #-}