package com.creativewidgetworks.goldparser.parser;
import java.io.File;
import java.io.IOException;
import java.io.InputStream;
import java.io.Reader;
import java.io.StringReader;
import java.lang.annotation.Annotation;
import java.lang.reflect.Constructor;
import java.lang.reflect.InvocationTargetException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.Stack;
import java.util.TreeMap;
import com.creativewidgetworks.goldparser.engine.Parser;
import com.creativewidgetworks.goldparser.engine.ParserException;
import com.creativewidgetworks.goldparser.engine.Position;
import com.creativewidgetworks.goldparser.engine.Production;
import com.creativewidgetworks.goldparser.engine.Reduction;
import com.creativewidgetworks.goldparser.engine.Token;
import com.creativewidgetworks.goldparser.engine.enums.ParseMessage;
import com.creativewidgetworks.goldparser.util.FormatHelper;
import com.creativewidgetworks.goldparser.util.ResourceHelper;
/**
* General purpose wrapper class used to process source input and generate
* a parse tree.
*
* Dependencies:
* @Parser
* @Reduction
* @Scope
* @FormatHelper
* @ResourceHelper
*
* @author Ralph Iden (http://www.creativewidgetworks.com)
* @version 5.0.0
*/
public class GOLDParser extends Parser {
// Virtual terminal constants
public static final String VT_INDENT_DECREASE = "IndentDecrease";
public static final String VT_INDENT_INCREASE = "IndentIncrease";
private Map<String, Class> ruleHandlers = new TreeMap<String, Class>();
private boolean ignoreCase;
private boolean generateTree;
private Reduction root;
private List<String> messages = new ArrayList<String>();
private Scope currentScope;
private Map<String, Scope> scopes;
// Indentation-sensitive support
private boolean useIndentVirtualTerminals;
private int ivtLine;
private Stack<Token> ivtTokens = new Stack<Token>();
private Stack<Position> ivtIndentLevels = new Stack<Position>();
/*----------------------------------------------------------------------------*/
public GOLDParser() {
super();
setCurrentScope(new Scope("GLOBAL"));
}
/*----------------------------------------------------------------------------*/
public GOLDParser(File cgtFile, String rulesPackage, boolean trimReductions) {
try {
loadTables(cgtFile);
loadRuleHandlers(rulesPackage);
setTrimReductions(trimReductions);
setCurrentScope(new Scope("GLOBAL"));
if (ruleHandlers.size() == 0) {
throw new IllegalStateException(FormatHelper.formatMessage("messages", "error.handlers_none", rulesPackage));
}
} catch (Exception e) {
addErrorMessage(FormatHelper.formatMessage("messages", "error.table_unloadable", e.getMessage()));
}
}
/*----------------------------------------------------------------------------*/
public GOLDParser(InputStream cgtFile, String rulesPackage, boolean trimReductions) {
try {
loadTables(cgtFile);
loadRuleHandlers(rulesPackage);
setTrimReductions(trimReductions);
setCurrentScope(new Scope("GLOBAL"));
if (ruleHandlers.size() == 0) {
throw new IllegalStateException(FormatHelper.formatMessage("messages", "error.handlers_none", rulesPackage));
}
} catch (Exception e) {
addErrorMessage(FormatHelper.formatMessage("messages", "error.table_unloadable", e.getMessage()));
}
}
/*----------------------------------------------------------------------------*/
public void clear() {
restart();
getScopes().clear();
clearProgramVariables();
getErrorMessages().clear();
setCurrentScope(new Scope("GLOBAL"));
}
/*----------------------------------------------------------------------------*/
/**
* Clears the variable map of all values for all scopes
*/
public void clearProgramVariables() {
for (Scope scope : scopes.values()) {
scope.getVariables().clear();
}
}
/**
* Clears (removes) the specified variable from the variable map starting at
* the current scope. If the variable is found, it is removed and no further
* scope processing is performed.
* @param name the variable to remove.
*/
public void clearProgramVariable(String name) {
String key = ignoreCase ? name.toUpperCase() : name;
Scope scope = currentScope;
while (scope != null) {
if (scope.getVariables().containsKey(key)) {
scope.getVariables().remove(key);
break;
}
scope = scope.getParent();
}
}
/**
* Gets the specified variable from the map starting at the current scope. If
* the variable is found, the Variable is immediately returned and no further
* scope processing is performed.
* @param name of the variable to retrieve
* @return a Variable instance or null if the variable isn't found
*/
public Variable getProgramVariable(String name) {
String key = ignoreCase ? name.toUpperCase() : name;
Scope scope = currentScope;
while (scope != null) {
if (scope.getVariables().containsKey(key)) {
return scope.getVariables().get(key);
}
scope = scope.getParent();
}
return null;
}
/**
* Sets the specified variable in the current scope's variable map, replacing
* any existing variable in the map.
* @param name of the variable to set
* @param value to set, null is not recommended (use clearProgramVariable() instead)
*/
public void setProgramVariable(String name, Variable value) {
if (currentScope == null) {
currentScope = new Scope(); // create new global scope
}
currentScope.getVariables().put(ignoreCase ? name.toUpperCase() : name, value);
}
/*----------------------------------------------------------------------------*/
public Map<String, Scope> getScopes() {
if (scopes == null) {
scopes = new TreeMap<String, Scope>();
}
return scopes;
}
public Scope getCurrentScope() {
return currentScope;
}
public Scope setCurrentScope(Scope newScope) {
Scope oldScope = currentScope;
currentScope = newScope;
if (currentScope != null && !getScopes().containsKey(newScope.getName())) {
getScopes().put(newScope.getName(), newScope);
}
return oldScope;
}
/*----------------------------------------------------------------------------*/
public boolean getGenerateTree() {
return generateTree;
}
public void setGenerateTree(boolean generateTree) {
this.generateTree = generateTree;
}
/*----------------------------------------------------------------------------*/
/**
* Parser and rule handlers should throw ParseExceptions using this method because
* it also adds the message to the list of parser errors. Rule handlers that
* throw ParserExceptions instead of calling this method cannot communicate the reason
* the exception was thrown.
* @param msg to add to the errors list
* @throws ParserException
*/
public void raiseParserException(String msg) throws ParserException {
addErrorMessage(msg);
throw new ParserException(msg);
}
/**
* Parser and rule handlers should throw ParseExceptions using this method because
* it also adds the message to the list of parser errors. Rule handlers that
* throw ParserExceptions instead of calling this method cannot communicate the reason
* the exception was thrown.
* @param msg to add to the errors list
* @param throwable
* @throws ParserException
*/
public void raiseParserException(String msg, Throwable cause) throws ParserException {
addErrorMessage(msg);
throw new ParserException(msg, cause);
}
/*----------------------------------------------------------------------------*/
public List<String> getErrorMessages() {
return messages;
}
public String getErrorMessage() {
int errorCount = getErrorMessages().size();
StringBuilder sb = new StringBuilder();
for (String line : getErrorMessages()) {
sb.append(line);
if (errorCount > 1) {
sb.append("\r\n");
}
}
return sb.toString();
}
public void addErrorMessage(String message) {
messages.add(message);
}
public boolean isReady() {
return messages.size() == 0;
}
/*----------------------------------------------------------------------------*/
public boolean isIgnoreCaseOfVariables() {
return ignoreCase;
}
public void setIgnoreCaseOfVariables(boolean ignoreCaseOfVariableNames) {
this.ignoreCase = ignoreCaseOfVariableNames;
}
/*----------------------------------------------------------------------------*/
/**
* Initialize the parser engine using the specified grammar (CGT) file
* @param grammar file to load
* @return true if grammar file was loaded
* @throw IOException if an invalid file was specified or the file could not be opened.
*/
public boolean setup(File grammarFile) throws IOException {
if (grammarFile == null) {
throw new IOException(FormatHelper.formatMessage("messages", "error.cgt_missing"));
}
return loadTables(grammarFile);
}
/*----------------------------------------------------------------------------*/
public boolean parseSourceStatements(String sourceStatements) {
getErrorMessages().clear();
if (sourceStatements == null || sourceStatements.trim().length() == 0) {
addErrorMessage(FormatHelper.formatMessage("messages", "error.no_source"));
return false;
}
return parseSourceStatements(new StringReader(sourceStatements));
}
public boolean parseSourceStatements(Reader sourceReader) {
boolean accept = false;
getErrorMessages().clear();
if (sourceReader == null) {
addErrorMessage(FormatHelper.formatMessage("messages", "error.no_source_reader"));
return false;
}
getErrorMessages().clear();
// Enable support for indentation sensitive grammars
useIndentVirtualTerminals = getSymbolByName(VT_INDENT_INCREASE) != null;
open(sourceReader);
boolean done = false;
while (!done) {
ParseMessage response = parse();
switch (response) {
case ACCEPT:
done = accept = processAccept();
break;
case GROUP_ERROR:
done = processGroupError();
break;
case INTERNAL_ERROR:
done = processInternalError();
break;
case LEXICAL_ERROR:
done = processLexicalError();
break;
case NOT_LOADED_ERROR:
done = processNotLoadedError();
break;
case REDUCTION:
done = processReduction();
break;
case SYNTAX_ERROR:
done = processSyntaxError();
break;
case TOKEN_READ:
done = processTokenRead();
break;
case UNDEFINED:
done = true;
addErrorMessage(FormatHelper.formatMessage("messages", "error.invalid_result"));
break;
}
}
return accept;
}
protected boolean processAccept() {
root = getCurrentReduction();
return true;
}
protected boolean processGroupError() {
addErrorMessage(FormatHelper.formatMessage("messages", "error.group_runaway",
getCurrentPosition().getLineAsString(),
getCurrentPosition().getColumnAsString()));
return true;
}
public boolean processInternalError() {
addErrorMessage(FormatHelper.formatMessage("messages", "error.internal",
getCurrentPosition().getLineAsString(),
getCurrentPosition().getColumnAsString()));
return true;
}
protected boolean processLexicalError() {
String tokenStr = getCurrentToken().toString();
addErrorMessage(FormatHelper.formatMessage("messages", "error.lexical",
getCurrentPosition().getLineAsString(),
getCurrentPosition().getColumnAsString(),
tokenStr));
return true;
}
protected boolean processNotLoadedError() {
addErrorMessage(FormatHelper.formatMessage("messages", "error.table_not_loaded"));
return true;
}
/**
* Base parser builds a tree of Reduction objects
* Override to process reductions
* @return Boolean to indicate if processing should stop (true) or continue (false).
*/
protected boolean processReduction() {
if (!generateTree && ruleHandlers.size() > 0) {
try {
Reduction reduction = createInstance();
setCurrentReduction(reduction);
} catch (Throwable t) {
addErrorMessage(t.getMessage());
return true;
}
}
return false;
}
protected boolean processSyntaxError() {
String tokenStr = getCurrentToken().toString();
addErrorMessage(FormatHelper.formatMessage("messages", "error.syntax",
getCurrentPosition().getLineAsString(),
getCurrentPosition().getColumnAsString(),
tokenStr,
getExpectedSymbols().toString()));
return true;
}
protected boolean processTokenRead() {
return false;
}
/*----------------------------------------------------------------------------*/
public String getParseTree() {
StringBuilder tree = new StringBuilder();
if (generateTree && root != null) {
tree.append("+-").append(root.getParent()).append("\r\n");
drawReduction(tree, root, 1);
} else {
tree.append(FormatHelper.formatMessage("messages", "error.tree_unavailable"));
}
return tree.toString();
}
private void drawReduction(StringBuilder tree, Reduction reduction, int indent) {
StringBuilder indentStr = new StringBuilder();
for (int i = 0; i < indent; i++) {
indentStr.append("| ");
}
for (Token token : reduction) {
switch (token.getType()) {
case NON_TERMINAL:
tree.append(indentStr).append("+-").append(token.asReduction().getParent()).append("\r\n");
drawReduction(tree, token.asReduction(), indent + 1);
break;
default:
tree.append(indentStr).append("+-").append(token.asString()).append("\r\n");
break;
}
}
}
/*----------------------------------------------------------------------------*/
public void loadRuleHandlers(String packageName) {
try {
ruleHandlers.clear();
List<Class> list = ResourceHelper.findClassesInPackage(packageName);
for (Class clazz : list) {
Annotation[] annotations = clazz.getAnnotations();
for (Annotation annotation : annotations){
if (annotation instanceof ProcessRule){
ProcessRule a = (ProcessRule)annotation;
for (String rule : a.rule()) {
ruleHandlers.put(rule, clazz);
}
}
}
}
} catch (Exception e) {
addErrorMessage("loadRuleMappings: " + e.getMessage());
}
}
public List<String> validateHandlersExist() throws ParserException {
List<String> errors = new ArrayList<String>();
for (Production production : productionTable) {
String rule = production.toString();
if (ruleHandlers.get(rule) == null && ruleHandlers.get(rule.replace("'", "")) == null) {
errors.add(FormatHelper.formatMessage("messages", "error.handler_none", rule));
}
}
return errors;
}
public void validateHandlerExists(Production rule) throws ParserException {
if (ruleHandlers.get(rule.toString()) == null) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_none", rule));
}
}
protected Reduction createInstance() throws ParserException {
Reduction currentReduction = getCurrentReduction();
String ruleName = currentReduction.getParent().toString();
return createInstance(ruleName);
}
protected Reduction createInstance(String ruleName) throws ParserException {
Reduction reduction = null;
// Look up the handler for the rule and construct an instance of the class
Class clazz = ruleHandlers.get(ruleName);
if (clazz == null) {
clazz = ruleHandlers.get(ruleName.replace("'", "")); // Try removing single quotes
}
if (clazz != null) {
Constructor con = null;
try {
con = clazz.getConstructor(GOLDParser.class);
} catch (SecurityException e) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_security", e));
} catch (NoSuchMethodException e) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_constructor", clazz.getSimpleName()));
}
try {
reduction = (Reduction)con.newInstance(this);
} catch (IllegalArgumentException e) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_create", ruleName), e);
} catch (InstantiationException e) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_create", ruleName), e);
} catch (IllegalAccessException e) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_create", ruleName), e);
} catch (InvocationTargetException e) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_create", ruleName), e);
} catch (Throwable t) {
raiseParserException(FormatHelper.formatMessage("messages", "error.handler_create", ruleName), t);
}
} else {
throw new ParserException(FormatHelper.formatMessage("messages", "error.handler_none", ruleName));
}
return reduction;
}
/*----------------------------------------------------------------------------*/
/**
* Helper function to return the current ident virtual token index levels. This
* method supports the ability indent location values for languages like Python
* where indent levels must match.
*/
private String getIndentLevels() {
StringBuilder sb = new StringBuilder();
for (int i = ivtIndentLevels.size() - 1; i >= 0; i--) {
if (sb.length() > 0) {
sb.append(",");
}
sb.append(ivtIndentLevels.get(i).getColumnAsString());
}
return sb.toString();
}
/**
* Return next token - insert IndentIncrease/IndentDecrease tokens as needed to support
* grammars where indentation is important (e.g. Python, etc.)
* @return Token
*/
@Override
protected Token nextToken() {
Token token;
// Indent virtual terminals
if (useIndentVirtualTerminals) {
if (ivtTokens.isEmpty()) {
// Get next token from stream
token = produceToken();
// Token's position in source - initialize indentLevel stack
Position position = token.getPosition();
if (ivtIndentLevels.isEmpty()) {
ivtLine = position.getLine();
ivtIndentLevels.push(position);
}
// Trigger on change of line number
if (position.getLine() != ivtLine) {
ivtLine = position.getLine();
int ivtColumn = ivtIndentLevels.peek().getColumn();
if (position.getColumn() > ivtColumn) {
ivtTokens.push(token);
ivtIndentLevels.push(position);
token = new Token(getSymbolByName(VT_INDENT_INCREASE), getIndentLevels(), position);
} else if (token.getPosition().getColumn() < ivtColumn) {
ivtTokens.push(token);
while (!ivtIndentLevels.isEmpty() && ivtIndentLevels.peek().getColumn() > position.getColumn()) {
ivtIndentLevels.pop();
ivtTokens.push(new Token(getSymbolByName(VT_INDENT_DECREASE), getIndentLevels(), position));
}
token = ivtTokens.pop();
}
}
} else {
// Return token pushed on the stack because of insertion of IndentIncrease/IndentDecrease insertion
token = ivtTokens.pop();
}
} else {
// Not an indentation sensitive grammar, just get the next token from the stream
token = produceToken();
}
return token;
}
}