transcendental-lisp/src/main/kotlin/scanner/LispScanner.java

package scanner;

import error.LineColumnException;
import file.FilePosition;
import file.FilePositionTracker;
import token.Token;
import token.TokenFactory;
import token.TokenFactoryImpl;
import util.Characters;

import java.io.InputStream;
import java.util.function.Function;

import static java.lang.Character.isDigit;
import static java.lang.Character.isWhitespace;

/**
 * Converts a stream of bytes into a stream of Lisp tokens.
 */
public class LispScanner {

    private LispInputStream inputStream;
    private FilePositionTracker positionTracker;
    private TokenFactory tokenFactory;

    public LispScanner(InputStream inputStream, String fileName) {
        this.inputStream = new LispCommentRemovingInputStream(inputStream);
        this.positionTracker = new FilePositionTracker(fileName);
        this.tokenFactory = new TokenFactoryImpl();
    }

    public Token getNextToken() {
        for (int c = inputStream.read(); c != Characters.EOF; c = inputStream.read()) {
            char currentCharacter = (char) c;
            positionTracker.incrementColumn();

            if (!isWhitespace(currentCharacter))
                return createTokenFromCharacter(currentCharacter);
            else if (currentCharacter == Characters.NEWLINE)
                positionTracker.incrementLine();
        }

        return tokenFactory.createEofToken(positionTracker.currentPosition());
    }

    private Token createTokenFromCharacter(char c) {
        FilePosition currentPosition = positionTracker.currentPosition();
        String tokenText = retrieveTokenText(c);

        return tokenFactory.createToken(tokenText, currentPosition);
    }

    private String retrieveTokenText(char firstCharacter) {
        String tokenText = "" + firstCharacter;

        if (firstCharacter == Characters.DOUBLE_QUOTE)
            tokenText = retrieveStringTokenText(firstCharacter);
        else if (Characters.INSTANCE.isNumberPrefix(firstCharacter))
            tokenText = retrieveNumberOrIdentifierTokenText(firstCharacter);
        else if (isDigit(firstCharacter))
            tokenText = retrieveNumberTokenText(firstCharacter);
        else if (Characters.INSTANCE.isLegalIdentifierCharacter(firstCharacter))
            tokenText = retrieveIdentifierTokenText(firstCharacter);

        return tokenText;
    }

    private String retrieveStringTokenText(char firstDoubleQuote) {
        ComplexTokenTextRetriever retriever = new ComplexTokenTextRetriever(firstDoubleQuote,
                                                                            Characters.INSTANCE::isLegalStringCharacter);

        return retriever.retrieveToken();
    }

    private String retrieveNumberOrIdentifierTokenText(char firstCharacter) {
        char nextCharacter = (char) inputStream.read();
        inputStream.unreadLastCharacter();

        if (isDigit(nextCharacter))
            return retrieveNumberTokenText(firstCharacter);

        return retrieveIdentifierTokenText(firstCharacter);
    }

    private String retrieveNumberTokenText(char firstCharacter) {
        ComplexTokenTextRetriever retriever = new ComplexTokenTextRetriever(firstCharacter, Character::isDigit);

        return retriever.retrieveToken();
    }

    private String retrieveIdentifierTokenText(char firstCharacter) {
        ComplexTokenTextRetriever retriever = new ComplexTokenTextRetriever(firstCharacter,
                                                                            Characters.INSTANCE::isLegalIdentifierCharacter);

        return retriever.retrieveToken();
    }

    private class ComplexTokenTextRetriever {

        Function<Character, Boolean> isPartOfToken;
        StringBuilder text;
        FilePosition position;
        char firstCharacter;
        char currentCharacter;
        char previousCharacter;

        public ComplexTokenTextRetriever(char firstCharacter, Function<Character, Boolean> isPartOfToken) {
            this.isPartOfToken = isPartOfToken;
            this.text = new StringBuilder();
            this.position = positionTracker.currentPosition();
            this.firstCharacter = firstCharacter;
            this.currentCharacter = firstCharacter;
            this.previousCharacter = firstCharacter;
        }

        public String retrieveToken() {
            text.append(firstCharacter);

            for (int c = inputStream.read(); c != Characters.EOF; c = inputStream.read()) {
                currentCharacter = (char) c;

                if (!isPartOfToken.apply(currentCharacter)) {
                    inputStream.unreadLastCharacter();

                    return text.toString();
                }

                addCharacterToToken();

                if (isTerminatingCharacter())
                    return text.toString();

                previousCharacter = currentCharacter;
            }

            return terminateTokenWithEof();
        }

        private void addCharacterToToken() {
            text.append(currentCharacter);
            positionTracker.incrementColumn();

            if (currentCharacter == Characters.NEWLINE)
                positionTracker.incrementLine();
        }

        private boolean isTerminatingCharacter() {
            return isStringToken() && isTerminatingDoubleQuote();
        }

        private boolean isStringToken() {
            return firstCharacter == Characters.DOUBLE_QUOTE;
        }

        private boolean isTerminatingDoubleQuote() {
            return (currentCharacter == Characters.DOUBLE_QUOTE) &&
                   (previousCharacter != Characters.BACKSLASH);
        }

        private String terminateTokenWithEof() {
            if (isStringToken())
                throw new UnterminatedStringException(position);

            return text.toString();
        }
    }

    public static class UnterminatedStringException extends LineColumnException {

        private static final long serialVersionUID = 1L;

        public UnterminatedStringException(FilePosition position) {
            super(position);
        }

        @Override
        public String getMessagePrefix() {
            return "unterminated quoted string";
        }
    }
}