package carpet.script;

import carpet.script.exception.ExpressionException;
import carpet.script.exception.InternalExpressionException;

import java.util.ArrayList;
import java.util.Collections;
import java.util.Iterator;
import java.util.List;
import java.util.stream.Collectors;
import java.util.stream.StreamSupport;

/**
 * Expression tokenizer that allows to iterate over a {@link String}
 * expression token by token. Blank characters will be skipped.
 */
public class Tokenizer implements Iterator<Tokenizer.Token>
{
    /** What character to use for decimal separators. */
    private static final char decimalSeparator = '.';
    /** What character to use for minus sign (negative values). */
    private static final char minusSign = '-';
    /** Actual position in expression string. */
    private int pos = 0;
    private int lineno = 0;
    private int linepos = 0;
    private boolean comments;
    private boolean newLinesMarkers;
    /** The original input expression. */
    private String input;
    /** The previous token or <code>null</code> if none. */
    private Token previousToken;

    private Expression expression;
    private Context context;

    Tokenizer(Context c, Expression expr, String input, boolean allowComments, boolean allowNewLineMakers)
    {
        this.input = input;
        this.expression = expr;
        this.context = c;
        this.comments = allowComments;
        this.newLinesMarkers = allowNewLineMakers;
    }

    public List<Token> postProcess()
    {
        Iterable<Token> iterable = () -> this;
        List<Token> originalTokens = StreamSupport.stream(iterable.spliterator(), false).collect(Collectors.toList());
        List<Token> cleanedTokens = new ArrayList<>();
        Token last = null;
        while (originalTokens.size() > 0)
        {
            Token current = originalTokens.remove(originalTokens.size()-1);
            if (current.type == Token.TokenType.MARKER && current.surface.startsWith("//"))
                continue;
                // skipping comments
            if (!isSemicolon(current)
                    || (last != null && last.type != Token.TokenType.CLOSE_PAREN && last.type != Token.TokenType.COMMA && !isSemicolon(last)))
            {
                if (isSemicolon(current))
                {
                    current.surface = ";";
                    current.type = Token.TokenType.OPERATOR;
                }
                if (current.type == Token.TokenType.MARKER)
                {
                    // dealing with tokens in reversed order
                    if ("{".equals(current.surface))
                    {
                        cleanedTokens.add(current.morphedInto(Token.TokenType.OPEN_PAREN, "("));
                        current.morph(Token.TokenType.FUNCTION, "m");
                    }
                    else if ("[".equals(current.surface))
                    {
                        cleanedTokens.add(current.morphedInto(Token.TokenType.OPEN_PAREN, "("));
                        current.morph(Token.TokenType.FUNCTION, "l");
                    }
                    else if ("}".equals(current.surface) || "]".equals(current.surface))
                    {
                        current.morph(Token.TokenType.CLOSE_PAREN, ")");
                    }
                }
                cleanedTokens.add(current);
            }
            if (!(current.type == Token.TokenType.MARKER && current.surface.equals("$")))
                last = current;
        }
        Collections.reverse(cleanedTokens);
        return cleanedTokens;
    }

    @Override
    public boolean hasNext()
    {
        return (pos < input.length());
    }

    /**
     * Peek at the next character, without advancing the iterator.
     *
     * @return The next character or character 0, if at end of string.
     */
    private char peekNextChar()
    {
        return (pos < (input.length() - 1)) ? input.charAt(pos + 1) : 0;
    }

    private boolean isHexDigit(char ch)
    {
        return ch == 'x' || ch == 'X' || (ch >= '0' && ch <= '9') || (ch >= 'a' && ch <= 'f')
                || (ch >= 'A' && ch <= 'F');
    }

    private static boolean isSemicolon(Token tok)
    {
        return (    tok.type == Token.TokenType.OPERATOR && tok.surface.equals(";") )
                || (tok.type == Token.TokenType.UNARY_OPERATOR && tok.surface.equals(";u") );
    }

    public static List<Token> simplepass(String input)
    {
        Tokenizer tok = new Tokenizer(null, null, input, false, false);
        List<Token> res = new ArrayList<>();
        while (tok.hasNext()) res.add(tok.next());
        return res;
    }

    @Override
    public Token next()
    {
        Token token = new Token();

        if (pos >= input.length())
        {
            return previousToken = null;
        }
        char ch = input.charAt(pos);
        while (Character.isWhitespace(ch) && pos < input.length())
        {
            linepos++;
            if (ch=='\n')
            {
                lineno++;
                linepos = 0;
            }
            ch = input.charAt(++pos);
        }
        token.pos = pos;
        token.lineno = lineno;
        token.linepos = linepos;

        boolean isHex = false;

        if (Character.isDigit(ch)) // || (ch == decimalSeparator && Character.isDigit(peekNextChar())))
                                   // decided to no support this notation to favour element access via . operator
        {
            if (ch == '0' && (peekNextChar() == 'x' || peekNextChar() == 'X'))
                isHex = true;
            while ((isHex
                    && isHexDigit(
                    ch))
                    || (Character.isDigit(ch) || ch == decimalSeparator || ch == 'e' || ch == 'E'
                    || (ch == minusSign && token.length() > 0
                    && ('e' == token.charAt(token.length() - 1)
                    || 'E' == token.charAt(token.length() - 1)))
                    || (ch == '+' && token.length() > 0
                    && ('e' == token.charAt(token.length() - 1)
                    || 'E' == token.charAt(token.length() - 1))))
                    && (pos < input.length()))
            {
                token.append(input.charAt(pos++));
                linepos++;
                ch = pos == input.length() ? 0 : input.charAt(pos);
            }
            token.type = isHex ? Token.TokenType.HEX_LITERAL : Token.TokenType.LITERAL;
        }
        else if (ch == '\'')
        {
            pos++;
            linepos++;
            token.type = Token.TokenType.STRINGPARAM;
            if (pos == input.length() && expression != null && context != null)
                throw new ExpressionException(context, this.expression, token, "Program truncated");
            ch = input.charAt(pos);
            while (ch != '\'')
            {
                if (ch == '\\')
                {
                    char nextChar = peekNextChar();
                    if (nextChar == 'n')
                    {
                        token.append('\n');
                    }
                    else if (nextChar == 't')
                    {
                        //throw new ExpressionException(context, this.expression, token,
                        //        "Tab character is not supported");
                        token.append('\t');
                    }
                    else if (nextChar == 'r')
                    {
                        throw new ExpressionException(context, this.expression, token,
                                "Carriage return character is not supported");
                        //token.append('\r');
                    }
                    else if (nextChar == '\\' || nextChar == '\'')
                    {
                        token.append(nextChar);
                    }
                    else
                    {
                        pos--;
                        linepos--;
                    }
                    pos+=2;
                    linepos+=2;
                    if (pos == input.length() && expression != null && context != null)
                        throw new ExpressionException(context, this.expression, token, "Program truncated");
                }
                else
                {
                    token.append(input.charAt(pos++));
                    linepos++;
                    if (ch=='\n')
                    {
                        lineno++;
                        linepos = 0;
                    }
                    if (pos == input.length() && expression != null && context != null)
                        throw new ExpressionException(context, this.expression, token, "Program truncated");
                }
                ch = input.charAt(pos);
            }
            pos++;
            linepos++;

        }
        else if (Character.isLetter(ch) || "_".indexOf(ch) >= 0)
        {
            while ((Character.isLetter(ch) || Character.isDigit(ch) || "_".indexOf(ch) >= 0
                    || token.length() == 0 && "_".indexOf(ch) >= 0) && (pos < input.length()))
            {
                token.append(input.charAt(pos++));
                linepos++;
                ch = pos == input.length() ? 0 : input.charAt(pos);
            }
            // Remove optional white spaces after function or variable name
            if (Character.isWhitespace(ch))
            {
                while (Character.isWhitespace(ch) && pos < input.length())
                {
                    ch = input.charAt(pos++);
                    linepos++;
                    if (ch=='\n')
                    {
                        lineno++;
                        linepos = 0;
                    }
                }
                pos--;
                linepos--;
            }
            token.type = ch == '(' ? Token.TokenType.FUNCTION : Token.TokenType.VARIABLE;
        }
        else if (ch == '(' || ch == ')' || ch == ',' ||
                ch == '{' || ch == '}' || ch == '[' || ch == ']')
        {
            if (ch == '(')
            {
                token.type = Token.TokenType.OPEN_PAREN;
            }
            else if (ch == ')')
            {
                token.type = Token.TokenType.CLOSE_PAREN;
            }
            else if (ch == ',')
            {
                token.type = Token.TokenType.COMMA;
            }
            else
            {
                token.type = Token.TokenType.MARKER;
            }
            token.append(ch);
            pos++;
            linepos++;

            if (expression != null && context != null && previousToken != null &&
                    previousToken.type == Token.TokenType.OPERATOR &&
                    (ch == ')' || ch == ',' || ch == ']' || ch == '}') &&
                    !previousToken.surface.equalsIgnoreCase(";")
            )
                throw new ExpressionException(context, this.expression, previousToken,
                        "Can't have operator " + previousToken.surface + " at the end of a subexpression");
        }
        else
        {
            String greedyMatch = "";
            int initialPos = pos;
            int initialLinePos = linepos;
            ch = input.charAt(pos);
            int validOperatorSeenUntil = -1;
            while (!Character.isLetter(ch) && !Character.isDigit(ch) && "_".indexOf(ch) < 0
                    && !Character.isWhitespace(ch) && ch != '(' && ch != ')' && ch != ','
                    && (pos < input.length()))
            {
                greedyMatch += ch;
                if (comments && "//".equals(greedyMatch))
                {

                    while ( ch != '\n' && pos < input.length())
                    {
                        ch = input.charAt(pos++);
                        linepos++;
                        greedyMatch += ch;
                    }
                    if (ch=='\n')
                    {
                        lineno++;
                        linepos = 0;
                    }
                    token.append(greedyMatch);
                    token.type = Token.TokenType.MARKER;
                    return token; // skipping setting previous
                }
                pos++;
                linepos++;
                if (Expression.none.isAnOperator(greedyMatch))
                {
                    validOperatorSeenUntil = pos;
                }
                ch = pos == input.length() ? 0 : input.charAt(pos);
            }
            if (newLinesMarkers && "$".equals(greedyMatch))
            {
                lineno++;
                linepos = 0;
                token.type = Token.TokenType.MARKER;
                token.append('$');
                return token; // skipping previous token lookback
            }
            if (validOperatorSeenUntil != -1)
            {
                token.append(input.substring(initialPos, validOperatorSeenUntil));
                pos = validOperatorSeenUntil;
                linepos = initialLinePos+validOperatorSeenUntil-initialPos;
            }
            else
            {
                token.append(greedyMatch);
            }

            if (previousToken == null || previousToken.type == Token.TokenType.OPERATOR
                    || previousToken.type == Token.TokenType.OPEN_PAREN || previousToken.type == Token.TokenType.COMMA
                    || (previousToken.type == Token.TokenType.MARKER && ( previousToken.surface.equals("{") || previousToken.surface.equals("[") ) )
            )
            {
                token.surface += "u";
                token.type = Token.TokenType.UNARY_OPERATOR;
            }
            else
            {
                token.type = Token.TokenType.OPERATOR;
            }
        }
        if (expression != null && context != null && previousToken != null &&
            (
                token.type == Token.TokenType.LITERAL ||
                token.type == Token.TokenType.HEX_LITERAL ||
                token.type == Token.TokenType.VARIABLE ||
                token.type == Token.TokenType.STRINGPARAM ||
              ( token.type == Token.TokenType.MARKER && ( previousToken.surface.equalsIgnoreCase("{") || previousToken.surface.equalsIgnoreCase("["))) ||
                token.type == Token.TokenType.FUNCTION
            ) &&(
                previousToken.type == Token.TokenType.VARIABLE ||
                previousToken.type == Token.TokenType.FUNCTION ||
                previousToken.type == Token.TokenType.LITERAL ||
                previousToken.type == Token.TokenType.CLOSE_PAREN ||
              ( previousToken.type == Token.TokenType.MARKER && ( previousToken.surface.equalsIgnoreCase("}") || previousToken.surface.equalsIgnoreCase("]"))) ||
                previousToken.type == Token.TokenType.HEX_LITERAL ||
                previousToken.type == Token.TokenType.STRINGPARAM
            )
        )
        {
            throw new ExpressionException(context, this.expression, previousToken, "'"+token.surface +"' is not allowed after '"+previousToken.surface+"'");
        }
        return previousToken = token;
    }

    @Override
    public void remove()
    {
        throw new InternalExpressionException("remove() not supported");
    }

    public static class Token
    {
        enum TokenType
        {
            FUNCTION(true, false), OPERATOR(true, false), UNARY_OPERATOR(true, false),
            VARIABLE(false, false), CONSTANT(false, true),
            LITERAL(false, true), HEX_LITERAL(false, true), STRINGPARAM(false, true),
            OPEN_PAREN(false, true), COMMA(false, true), CLOSE_PAREN(false, true), MARKER(false, true);

            boolean functional;
            boolean constant;

            TokenType(boolean functional, boolean constant)
            {
                this.functional = functional;
                this.constant = constant;
            }

            public boolean isFunctional()
            {
                return functional;
            }

            public boolean isConstant() {
                return constant;
            }
        }
        public String surface = "";
        public TokenType type;
        public int pos;
        public int linepos;
        public int lineno;
        public static final Token NONE = new Token();
        public Token morphedInto(TokenType newType, String newSurface)
        {
            Token created = new Token();
            created.surface = newSurface;
            created.type = newType;
            created.pos = pos;
            created.linepos = linepos;
            created.lineno= lineno;
            return created;
        }

        public void morph(TokenType type, String s)
        {
            this.type = type;
            this.surface = s;
        }

        public void append(char c)
        {
            surface += c;
        }

        public void append(String s)
        {
            surface += s;
        }

        public char charAt(int pos)
        {
            return surface.charAt(pos);
        }

        public int length()
        {
            return surface.length();
        }

        @Override
        public String toString()
        {
            return surface;
        }
    }
}