module BuildModel where import Control.Applicative ( (<|>) ) import Data.Char ( isAsciiLower , isDigit , toLower ) import Data.Maybe ( fromMaybe , mapMaybe ) import Data.List ( intercalate ) import System.FilePath ( () , (<.>) , splitDirectories ) import Text.ParserCombinators.ReadP ( ReadP , char , eof , many , many1 , option , readP_to_S , satisfy , skipSpaces , string ) data LineContents = ProgramDeclaration | ModuleDeclaration String | ModuleUsed String | ModuleSubprogramDeclaration | SubmoduleDeclaration | Other data RawSource = RawSource { rawSourceFilename :: FilePath , rawSourceContents :: String } data Source = Program { programSourceFileName :: FilePath , programObjectFileName :: FilePath -> FilePath , programModulesUsed :: [String] } | Module { moduleSourceFileName :: FilePath , moduleObjectFileName :: FilePath -> FilePath , moduleModulesUsed :: [String] , moduleName :: String , moduleProducesSmod :: Bool } | Submodule { submoduleSourceFileName :: FilePath } processRawSource :: RawSource -> Source processRawSource rawSource = let sourceFileName = rawSourceFilename rawSource parsedContents = parseContents rawSource objectFileName = \bd -> bd (pathSeparatorsToUnderscores sourceFileName) <.> "o" modulesUsed = getModulesUsed parsedContents in if hasProgramDeclaration parsedContents then Program { programSourceFileName = sourceFileName , programObjectFileName = objectFileName , programModulesUsed = modulesUsed } else if hasModuleDeclaration parsedContents then Module { moduleSourceFileName = sourceFileName , moduleObjectFileName = objectFileName , moduleModulesUsed = modulesUsed , moduleName = getModuleName parsedContents , moduleProducesSmod = hasModuleSubprogramDeclaration parsedContents } else if hasSubmoduleDeclaration parsedContents then Submodule { submoduleSourceFileName = sourceFileName } else undefined pathSeparatorsToUnderscores :: FilePath -> FilePath pathSeparatorsToUnderscores fileName = intercalate "_" (splitDirectories fileName) parseContents :: RawSource -> [LineContents] parseContents rawSource = let fileLines = lines $ rawSourceContents rawSource in map parseFortranLine fileLines hasProgramDeclaration :: [LineContents] -> Bool hasProgramDeclaration parsedContents = case filter f parsedContents of x : _ -> True _ -> False where f lc = case lc of ProgramDeclaration -> True _ -> False hasModuleDeclaration :: [LineContents] -> Bool hasModuleDeclaration parsedContents = case filter f parsedContents of x : _ -> True _ -> False where f lc = case lc of ModuleDeclaration{} -> True _ -> False hasSubmoduleDeclaration :: [LineContents] -> Bool hasSubmoduleDeclaration parsedContents = case filter f parsedContents of x : _ -> True _ -> False where f lc = case lc of SubmoduleDeclaration -> True _ -> False hasModuleSubprogramDeclaration :: [LineContents] -> Bool hasModuleSubprogramDeclaration parsedContents = case filter f parsedContents of x : _ -> True _ -> False where f lc = case lc of ModuleSubprogramDeclaration -> True _ -> False getModulesUsed :: [LineContents] -> [String] getModulesUsed = mapMaybe contentToMaybeModuleName where contentToMaybeModuleName content = case content of ModuleUsed moduleName -> Just moduleName _ -> Nothing getModuleName :: [LineContents] -> String getModuleName pc = head $ mapMaybe contentToMaybeModuleName pc where contentToMaybeModuleName content = case content of ModuleDeclaration moduleName -> Just moduleName _ -> Nothing readFileLinesIO :: FilePath -> IO [String] readFileLinesIO file = do contents <- readFile file return $ lines contents parseFortranLine :: String -> LineContents parseFortranLine line = let line' = map toLower line result = readP_to_S doFortranLineParse line' in getResult result where getResult (_ : (contents, _) : _) = contents getResult [(contents, _) ] = contents getResult [] = Other doFortranLineParse :: ReadP LineContents doFortranLineParse = option Other fortranUsefulContents fortranUsefulContents :: ReadP LineContents fortranUsefulContents = programDeclaration <|> moduleSubprogramDeclaration <|> moduleDeclaration <|> submoduleDeclaration <|> useStatement programDeclaration :: ReadP LineContents programDeclaration = do skipSpaces _ <- string "program" skipAtLeastOneWhiteSpace _ <- validIdentifier return ProgramDeclaration moduleDeclaration :: ReadP LineContents moduleDeclaration = do skipSpaces _ <- string "module" skipAtLeastOneWhiteSpace moduleName <- validIdentifier skipSpaceCommentOrEnd return $ ModuleDeclaration moduleName submoduleDeclaration :: ReadP LineContents submoduleDeclaration = do skipSpaces _ <- string "submodule" return $ SubmoduleDeclaration useStatement :: ReadP LineContents useStatement = do skipSpaces _ <- string "use" skipAtLeastOneWhiteSpace modName <- validIdentifier skipSpaceCommaOrEnd return $ ModuleUsed modName moduleSubprogramDeclaration :: ReadP LineContents moduleSubprogramDeclaration = do skipAnything _ <- string "module" skipAtLeastOneWhiteSpace skipAnything _ <- string "function" <|> string "subroutine" skipAtLeastOneWhiteSpace return $ ModuleSubprogramDeclaration skipAtLeastOneWhiteSpace :: ReadP () skipAtLeastOneWhiteSpace = do _ <- many1 whiteSpace return () skipSpaceOrEnd :: ReadP () skipSpaceOrEnd = eof <|> skipAtLeastOneWhiteSpace skipSpaceCommaOrEnd :: ReadP () skipSpaceCommaOrEnd = eof <|> skipComma <|> skipAtLeastOneWhiteSpace skipSpaceCommentOrEnd :: ReadP () skipSpaceCommentOrEnd = eof <|> skipComment <|> skipAtLeastOneWhiteSpace skipComma :: ReadP () skipComma = do _ <- char ',' return () skipComment :: ReadP () skipComment = do _ <- char '!' return () skipAnything :: ReadP () skipAnything = do _ <- many (satisfy (const True)) return () whiteSpace :: ReadP Char whiteSpace = satisfy (`elem` " \t") validIdentifier :: ReadP String validIdentifier = do first <- validFirstCharacter rest <- many validIdentifierCharacter return $ first : rest validFirstCharacter :: ReadP Char validFirstCharacter = alphabet validIdentifierCharacter :: ReadP Char validIdentifierCharacter = alphabet <|> digit <|> underscore alphabet :: ReadP Char alphabet = satisfy isAsciiLower digit :: ReadP Char digit = satisfy isDigit underscore :: ReadP Char underscore = char '_'