path: root/RGens/src/main/java/bjc/rgens/server/ServerGrammarReader.java

package bjc.rgens.server;

import com.mifmif.common.regex.Generex;

import java.io.InputStream;
import java.util.Random;
import java.util.function.BiFunction;
import java.util.function.Predicate;

import bjc.utils.data.IPair;
import bjc.utils.data.Pair;
import bjc.utils.funcdata.FunctionalStringTokenizer;
import bjc.utils.funcdata.FunctionalList;
import bjc.utils.funcdata.FunctionalMap;
import bjc.utils.funcdata.IList;
import bjc.utils.funcdata.IMap;
import bjc.utils.funcutils.ListUtils;
import bjc.utils.gen.WeightedGrammar;
import bjc.utils.parserutils.RuleBasedConfigReader;

/**
 * Read a grammar from a stream
 * 
 * @author ben
 *
 */
public class ServerGrammarReader {
	static {
		setupReader();
	
		initPragmas();
	}

	private static RuleBasedConfigReader<ReaderState> reader;

	private static Random numgen = new Random();

	private static IMap<String, WeightedGrammar<String>> exportedRules;

	public static void setExportedRules(IMap<String, WeightedGrammar<String>> rules) {
		exportedRules = rules;
	}

	/**
	 * Read a grammar from a stream
	 * 
	 * @param inputStream
	 *            The stream to load the grammar from
	 * 
	 * @return A grammar read from the stream
	 * 
	 */
	public static IPair<WeightedGrammar<String>, IList<String>>
		fromStream(InputStream inputStream) {
			ReaderState initialState = new ReaderState();
	
			WeightedGrammar<String> gram = reader.fromStream(inputStream, initialState).getGrammar();
	
			IMap<String, IList<String>> vars = new FunctionalMap<>();
	
			Predicate<String> specialPredicate = (strang) -> {
				if(strang.matches("\\[\\S+\\]")) {
					return true;
				}
	
				return false;
			};
	
			BiFunction<String, WeightedGrammar<String>, IList<String>> 
				specialAction = (strang, gramm) -> {
					return handleSpecialRule(vars, strang, gramm);
				};
	
			gram.configureSpecial(specialPredicate, specialAction);
	
			IList<String> exports = initialState.getExports();
	
			if(gram.getInitialRule() != null && !exports.contains(gram.getInitialRule())) {
				exports.add(gram.getInitialRule());
			}
	
			return new Pair<>(gram, exports);
	}

	private static void initPragmas() {
		reader.addPragma("debug", (tokenizer, state) -> {
			debugGrammar(state);
		});
	
		reader.addPragma("uniform", (tokenizer, state) -> {
			state.toggleUniformity();
		});
	
		reader.addPragma("initial-rule", ServerGrammarReader::initialRule);
	
		reader.addPragma("remove-rule", ServerGrammarReader::removeRule);
	
		reader.addPragma("prefix-with", ServerGrammarReader::prefixRule);
		reader.addPragma("suffix-with", ServerGrammarReader::suffixRule);
	
		reader.addPragma("regex-rule", ServerGrammarReader::handleRegexRule);
	
		reader.addPragma("range-rule", ServerGrammarReader::handleRangeRule);
	
		reader.addPragma("export-rule", (tokenizer, state) -> {
			String ruleName = tokenizer.nextToken();
	
			state.addExport(ruleName);
		});
	}

	private static void setupReader() {
		reader = new RuleBasedConfigReader<>(null, null, null);
	
		reader.setStartRule((tokenizer, stateTokenPair) -> {
			stateTokenPair.doWith((initToken, state) -> {
				state.startNewRule(initToken);
	
				doCase(tokenizer, state);
			});
		});
	
		reader.setContinueRule((tokenizer, state) -> {
			doCase(tokenizer, state);
		});
	
		reader.setEndRule((tokenizer) -> {
			tokenizer.setCurrentRule(null);
		});
	}

	private static void debugGrammar(ReaderState state) {
		System.out.println("Printing rule names: ");

		for (String currentRule : state.getRuleNames().toIterable()) {
			System.out.println("\t" + currentRule);
		}

		System.out.println();
	}

	private static IList<String> handleSpecialRule(IMap<String, IList<String>> vars, String strang,
			WeightedGrammar<String> gramm) {
		IList<String> retList = new FunctionalList<>();
	
		if(strang.matches("\\[\\[\\S+\\]\\]")) {
			if(strang.matches("\\[\\[\\S+:=\\S+\\]\\]")) {
				doDefineExpandedVariable(vars, strang, gramm);
			} else if(strang.matches("\\[\\[\\S+=\\S+\\]\\]")) {
				doDefineVariable(vars, strang);
			} else if(strang.matches("\\[\\[\\S+\\]\\]")) {
				if(GrammarServerEngine.debugMode) {
					if(strang.contains("+")) 
						System.out.println("Double-triggering no-space rule " + strang);
					else 
						System.out.println("Triggered alternate no-space rule " + strang);
				}
				
				doNoSpaceRule(strang, gramm, retList);
			} else if (strang.contains("+")) {
				if(GrammarServerEngine.debugMode) 
						System.out.println("Triggered alternate no-space rule " + strang);

				doNoSpaceRule(strang, gramm, retList);
			} else {
				// @FIXME notify the user they did something wrong
				retList.add(strang);
			}
		} else {
			if(strang.matches("\\[\\$\\S+\\-\\S+\\]")) {
				retList = doExpandVariableReference(vars, strang, gramm);
			} else if(strang.matches("\\[\\$\\S+\\]")) {
				String varName = strang.substring(2, strang.length());
	
				retList = vars.get(varName);
			} else if (exportedRules.containsKey(strang) && 
					exportedRules.get(strang) != gramm &&
					!gramm.hasRule(strang)) {
				// Only pick external rules if they are both 
				// a) in a different grammar
				// b) not shadowed in the current grammar
				WeightedGrammar<String> exportGram = exportedRules.get(strang);
	
				retList = exportGram.generateGenericValues(strang, (s) -> s, " ");
			} else if (strang.contains("+")) {
				doNoSpaceRule(strang, gramm, retList);
			} else {
				// @FIXME notify the user they did something wrong
				retList.add(strang);
			}
		}
	
		return retList;
	}

	private static void handleRangeRule(FunctionalStringTokenizer tokenizer, ReaderState state) {
		String ruleName = tokenizer.nextToken();

		int start = Integer.parseInt(tokenizer.nextToken());
		int end = Integer.parseInt(tokenizer.nextToken());

		state.addSpecialRule(ruleName, () -> {
			int genNum = numgen.nextInt((end - start) + 1) + start;
			
			return new FunctionalList<>(Integer.toString(genNum));
		});
	}

	private static void handleRegexRule(FunctionalStringTokenizer tokenizer, ReaderState state) {
		String ruleName = tokenizer.nextToken();

		IList<String> regx = tokenizer.toList();
		Generex regex = new Generex(ListUtils.collapseTokens(regx));

		state.addSpecialRule(ruleName, () -> {
			return new FunctionalList<>(regex.random().split(" "));
		});
	}

	private static int readOptionalProbability(FunctionalStringTokenizer tokenizer, ReaderState state) {
		if (state.isUniform()) {
			return 1;
		}

		return Integer.parseInt(tokenizer.nextToken());
	}

	private static void initialRule(FunctionalStringTokenizer tokenizer, ReaderState state) {
		String initialRuleName = tokenizer.nextToken();
	
		state.setInitialRule(initialRuleName);
	}

	private static void prefixRule(FunctionalStringTokenizer tokenizer, ReaderState state) {
		String ruleName = tokenizer.nextToken();
		String prefixToken = tokenizer.nextToken();
	
		int additionalProbability = readOptionalProbability(tokenizer, state);
	
		state.prefixRule(ruleName, prefixToken, additionalProbability);
	}

	private static void removeRule(FunctionalStringTokenizer tokenizer, ReaderState state) {
		String ruleName = tokenizer.nextToken();
	
		state.deleteRule(ruleName);
	}

	private static void suffixRule(FunctionalStringTokenizer tokenizer, ReaderState state) {
		String ruleName = tokenizer.nextToken();
		String suffixToken = tokenizer.nextToken();

		int additionalProbability = readOptionalProbability(tokenizer, state);

		state.suffixRule(ruleName, suffixToken, additionalProbability);
	}

	private static void doCase(FunctionalStringTokenizer tokenizer, ReaderState state) {
		int ruleProbability = readOptionalProbability(tokenizer, state);
	
		state.addCase(ruleProbability, tokenizer.toList());
	}

	private static void doDefineExpandedVariable(IMap<String, IList<String>> vars, String strang, WeightedGrammar<String> gramm) {
		String[] varParts = strang.split(":=");
	
		String varName = varParts[0].substring(2);
		String ruleName = varParts[1].substring(0, varParts[1].length() - 2);
	
		IList<String> varValue = gramm.generateGenericValues(
				ruleName, (s) -> s, " ");
	
		vars.put(varName, varValue);
	}

	private static void doDefineVariable(IMap<String, IList<String>> vars, String strang) {
		String[] varParts = strang.split("=");
	
		String varName = varParts[0].substring(2);
		String varValue = varParts[1].substring(0, varParts[1].length() - 2);
	
		vars.put(varName, new FunctionalList<>(varValue));
	}

	private static IList<String> doExpandVariableReference(IMap<String, IList<String>> vars, String strang,
			WeightedGrammar<String> gramm) {
		IList<String> retList;
		String[] varParts = strang.substring(1, strang.length()).split("-");
	
		StringBuilder actualName = new StringBuilder("[");
	
		for(String varPart : varParts) {
			if(varPart.startsWith("$")) {
				IList<String> varName = vars.get(varPart.substring(1));
	
				if(varName.getSize() != 1) {
					// @FIXME notify the user they did something wrong
				}
	
				actualName.append(varName.first() + "-");
			} else {
				actualName.append(varPart + "-");
			}
		}
	
		// Trim trailing -
		actualName.deleteCharAt(actualName.length() - 1);
	
		retList = gramm.generateGenericValues(actualName.toString(), (s) -> s, " ");
		return retList;
	}

	private static void doNoSpaceRule(String strang, WeightedGrammar<String> gramm, IList<String> retList) {
		if(!GrammarServerEngine.debugMode) {
			IList<String> ruleValue = gramm.generateGenericValues(
					strang, (s) -> s.trim(), "");
	
			retList.add(ListUtils.collapseTokens(ruleValue));
		} else {
			// if(!gramm.hasRule(strang))
			// 	System.out.println("Warning: Possible unexpanded rule " + strang);

			retList.add(strang);
		}
	}
}