Implement Lexer (with temporary error handling and basic tests)
This commit is contained in:
parent
61d40330be
commit
da91a5df33
@ -0,0 +1,138 @@
|
||||
package it.cavallium.warppi.math.rules.dsl.frontend;
|
||||
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.function.Function;
|
||||
import java.util.function.Predicate;
|
||||
import java.util.stream.Collectors;
|
||||
import java.util.stream.Stream;
|
||||
|
||||
import static it.cavallium.warppi.math.rules.dsl.frontend.TokenType.*;
|
||||
|
||||
/**
|
||||
* Converts the source string to a list of tokens.
|
||||
*/
|
||||
public class Lexer {
|
||||
private static final Map<String, TokenType> keywords = Stream.of(
|
||||
REDUCTION, EXPANSION, CALCULATION, EXISTENCE,
|
||||
ARCCOS, ARCSIN, ARCTAN, COS, SIN, TAN, ROOT, SQRT, LOG,
|
||||
UNDEFINED, PI, E
|
||||
).collect(Collectors.toMap(
|
||||
tokenType -> tokenType.name().toLowerCase(),
|
||||
Function.identity()
|
||||
));
|
||||
|
||||
private final String source;
|
||||
private final List<Token> tokens = new ArrayList<>();
|
||||
private int startOfLexeme = 0;
|
||||
private int curPosition = 0;
|
||||
|
||||
public Lexer(final String source) {
|
||||
this.source = source;
|
||||
}
|
||||
|
||||
public List<Token> lex() {
|
||||
while (!atEnd()) {
|
||||
startOfLexeme = curPosition;
|
||||
lexToken();
|
||||
}
|
||||
tokens.add(new Token(EOF, "", source.length()));
|
||||
return tokens;
|
||||
}
|
||||
|
||||
private void lexToken() {
|
||||
char current = popChar();
|
||||
switch (current) {
|
||||
case ':': emitToken(COLON); break;
|
||||
case ',': emitToken(COMMA); break;
|
||||
case '(': emitToken(LEFT_PAREN); break;
|
||||
case ')': emitToken(RIGHT_PAREN); break;
|
||||
case '[': emitToken(LEFT_BRACKET); break;
|
||||
case ']': emitToken(RIGHT_BRACKET); break;
|
||||
case '=': emitToken(EQUALS); break;
|
||||
case '*': emitToken(TIMES); break;
|
||||
case '/': emitToken(DIVIDE); break;
|
||||
case '^': emitToken(POWER); break;
|
||||
|
||||
case '+':
|
||||
if (matchChar('-')) {
|
||||
emitToken(PLUS_MINUS);
|
||||
} else {
|
||||
emitToken(PLUS);
|
||||
}
|
||||
break;
|
||||
|
||||
case '-':
|
||||
if (matchChar('>')) {
|
||||
emitToken(ARROW);
|
||||
} else {
|
||||
emitToken(MINUS);
|
||||
}
|
||||
break;
|
||||
|
||||
default:
|
||||
if (isAsciiDigit(current)) {
|
||||
number();
|
||||
} else if (Character.isJavaIdentifierStart(current)) {
|
||||
keywordOrIdentifier();
|
||||
} else if (!Character.isWhitespace(current)) {
|
||||
throw new RuntimeException("Unexpected character " + current);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private void number() {
|
||||
matchWhile(Lexer::isAsciiDigit);
|
||||
if (matchChar('.') && matchWhile(Lexer::isAsciiDigit) == 0) {
|
||||
throw new RuntimeException("Expected digits after decimal separator");
|
||||
}
|
||||
emitToken(NUMBER);
|
||||
}
|
||||
|
||||
private void keywordOrIdentifier() {
|
||||
matchWhile(Character::isJavaIdentifierPart);
|
||||
TokenType type = keywords.getOrDefault(currentLexeme(), IDENTIFIER);
|
||||
emitToken(type);
|
||||
}
|
||||
|
||||
private char popChar() {
|
||||
char current = source.charAt(curPosition);
|
||||
curPosition++;
|
||||
return current;
|
||||
}
|
||||
|
||||
private boolean matchChar(char expected) {
|
||||
if (atEnd() || source.charAt(curPosition) != expected) {
|
||||
return false;
|
||||
}
|
||||
curPosition++;
|
||||
return true;
|
||||
}
|
||||
|
||||
private int matchWhile(Predicate<Character> predicate) {
|
||||
int matched = 0;
|
||||
while (!atEnd() && predicate.test(source.charAt(curPosition))) {
|
||||
curPosition++;
|
||||
matched++;
|
||||
}
|
||||
return matched;
|
||||
}
|
||||
|
||||
private void emitToken(TokenType type) {
|
||||
tokens.add(new Token(type, currentLexeme(), startOfLexeme));
|
||||
}
|
||||
|
||||
private String currentLexeme() {
|
||||
return source.substring(startOfLexeme, curPosition);
|
||||
}
|
||||
|
||||
private boolean atEnd() {
|
||||
return curPosition >= source.length();
|
||||
}
|
||||
|
||||
// Character.isDigit also allows various Unicode digits
|
||||
private static boolean isAsciiDigit(char c) {
|
||||
return '0' <= c && c <= '9';
|
||||
}
|
||||
}
|
||||
@ -0,0 +1,82 @@
|
||||
package it.cavallium.warppi.math.rules.dsl.frontend;
|
||||
|
||||
import org.junit.Test;
|
||||
|
||||
import java.util.Arrays;
|
||||
import java.util.List;
|
||||
|
||||
import static it.cavallium.warppi.math.rules.dsl.frontend.TokenType.*;
|
||||
import static org.junit.Assert.*;
|
||||
|
||||
public class LexerTest {
|
||||
@Test
|
||||
public void validRule() {
|
||||
final Lexer lexer = new Lexer(
|
||||
"reduction TestRule_123:\n" +
|
||||
" x + y * z = -(a_123 +- 3 / 2.2) -> [\n" +
|
||||
" x^a_123 = cos(pi) - log(e, e),\n" +
|
||||
" undefined,\n" +
|
||||
"]\n"
|
||||
);
|
||||
final List<Token> expected = Arrays.asList(
|
||||
new Token(REDUCTION, "reduction", 0),
|
||||
new Token(IDENTIFIER, "TestRule_123", 10),
|
||||
new Token(COLON, ":", 22),
|
||||
new Token(IDENTIFIER, "x", 26),
|
||||
new Token(PLUS, "+", 28),
|
||||
new Token(IDENTIFIER, "y", 30),
|
||||
new Token(TIMES, "*", 32),
|
||||
new Token(IDENTIFIER, "z", 34),
|
||||
new Token(EQUALS, "=", 36),
|
||||
new Token(MINUS, "-", 38),
|
||||
new Token(LEFT_PAREN, "(", 39),
|
||||
new Token(IDENTIFIER, "a_123", 40),
|
||||
new Token(PLUS_MINUS, "+-", 46),
|
||||
new Token(NUMBER, "3", 49),
|
||||
new Token(DIVIDE, "/", 51),
|
||||
new Token(NUMBER, "2.2", 53),
|
||||
new Token(RIGHT_PAREN, ")", 56),
|
||||
new Token(ARROW, "->", 58),
|
||||
new Token(LEFT_BRACKET, "[", 61),
|
||||
new Token(IDENTIFIER, "x", 67),
|
||||
new Token(POWER, "^", 68),
|
||||
new Token(IDENTIFIER, "a_123", 69),
|
||||
new Token(EQUALS, "=", 75),
|
||||
new Token(COS, "cos", 77),
|
||||
new Token(LEFT_PAREN, "(", 80),
|
||||
new Token(PI, "pi", 81),
|
||||
new Token(RIGHT_PAREN, ")", 83),
|
||||
new Token(MINUS, "-", 85),
|
||||
new Token(LOG, "log", 87),
|
||||
new Token(LEFT_PAREN, "(", 90),
|
||||
new Token(E, "e", 91),
|
||||
new Token(COMMA, ",", 92),
|
||||
new Token(E, "e", 94),
|
||||
new Token(RIGHT_PAREN, ")", 95),
|
||||
new Token(COMMA, ",", 96),
|
||||
new Token(UNDEFINED, "undefined", 102),
|
||||
new Token(COMMA, ",", 111),
|
||||
new Token(RIGHT_BRACKET, "]", 113),
|
||||
new Token(EOF, "", 115)
|
||||
);
|
||||
assertEquals(expected, lexer.lex());
|
||||
}
|
||||
|
||||
@Test(expected = RuntimeException.class)
|
||||
public void incompleteNumberOtherChar() {
|
||||
final Lexer lexer = new Lexer("2. 5");
|
||||
lexer.lex();
|
||||
}
|
||||
|
||||
@Test(expected = RuntimeException.class)
|
||||
public void incompleteNumberEof() {
|
||||
final Lexer lexer = new Lexer("2.");
|
||||
lexer.lex();
|
||||
}
|
||||
|
||||
@Test(expected = RuntimeException.class)
|
||||
public void meaninglessCharacter() {
|
||||
final Lexer lexer = new Lexer("@");
|
||||
lexer.lex();
|
||||
}
|
||||
}
|
||||
Loading…
x
Reference in New Issue
Block a user