Added a new way to generate unit names, using rules of a context-free grammar.

src/CMakeLists.txt

@@ -980,6 +980,7 @@ set(wesnoth-main_SRC
 	units/map.cpp
 	units/types.cpp
 	utils/sha1.cpp
+	utils/context_free_grammar_generator.cpp
 	variable.cpp
 	variable_info.cpp
 	whiteboard/action.cpp

src/SConscript

@@ -555,6 +555,7 @@ wesnoth_sources = Split("""
     units/udisplay.cpp
     units/unit.cpp
     utils/sha1.cpp
+    utils/context_free_grammar_generator.cpp
     variable_info.cpp
     variable.cpp
     whiteboard/action.cpp

src/race.cpp

@@ -191,6 +191,21 @@ unit_race::unit_race(const config& cfg) :
 		name_[FEMALE] = (cfg["name"]);
 	}
 
+	if (cfg.has_attribute("male_name_generator")) {
+		name_generator_[MALE].constructFromString(cfg["male_name_generator"]);
+	}
+	if (cfg.has_attribute("female_name_generator")) {
+		name_generator_[FEMALE].constructFromString(cfg["female_name_generator"]);
+	}
+	if (cfg.has_attribute("name_generator")) {
+		if (!name_generator_[MALE].is_initialized()) {
+			name_generator_[MALE].constructFromString(cfg["name_generator"]);
+		}
+		if (!name_generator_[FEMALE].is_initialized()) {
+			name_generator_[FEMALE].constructFromString(cfg["name_generator"]);
+		}
+	}
+
 	if(chain_size_ <= 0)
 		chain_size_ = 2;
 
@@ -202,6 +217,9 @@ unit_race::unit_race(const config& cfg) :
 std::string unit_race::generate_name(
 		unit_race::GENDER gender) const
 {
+    if (name_generator_[gender].is_initialized()) {
+        return name_generator_[gender].generate();
+    }
 	return unicode_cast<utf8::string>(
 		markov_generate_name(next_[gender], chain_size_, 12));
 }

src/race.hpp

@@ -17,8 +17,7 @@
 
 #include "config.hpp"
 #include "serialization/unicode_types.hpp"
-
-
+#include "utils/context_free_grammar_generator.hpp"
 
 typedef std::map<ucs4::string, ucs4::string > markov_prefix_map;
 
@@ -63,6 +62,7 @@ private:
 	unsigned int ntraits_;
 	markov_prefix_map next_[NUM_GENDERS];
 	int chain_size_;
+	context_free_grammar_generator name_generator_[NUM_GENDERS];
 
 	config::const_child_itors traits_;
 	config::const_child_itors topics_;

src/utils/context_free_grammar_generator.cpp

@@ -0,0 +1,123 @@
+/*
+   Copyright (C) 2016 by Ján Dugáček
+   Part of the Battle for Wesnoth Project http://www.wesnoth.org/
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2 of the License, or
+   (at your option) any later version.
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY.
+
+   See the COPYING file for more details.
+*/
+
+/**
+ *  @file
+ *  Algorithm to generate names using a context-free grammar, which allows more control
+ *  than the usual Markov chain generator
+ */
+
+#include "context_free_grammar_generator.hpp"
+#include "../log.hpp"
+#include "../random_new.hpp"
+
+context_free_grammar_generator::context_free_grammar_generator() :
+	initialized_(false)
+{
+
+}
+
+context_free_grammar_generator::~context_free_grammar_generator()
+{
+
+}
+
+bool context_free_grammar_generator::constructFromString(const std::string &source) {
+	const char* reading = source.c_str();
+	nonterminal* current = nullptr;
+	std::vector<std::string>* filled = nullptr;
+	std::string buf;
+
+	while (*reading != 0) {
+		if (*reading == '=') {
+			current = &nonterminals_[buf];
+			current->possibilities_.push_back(std::vector<std::string>());
+			filled = &current->possibilities_.back();
+			buf.clear();
+		} else if (*reading == '\n') {
+			if (filled) filled->push_back(buf);
+			filled = nullptr;
+			current = nullptr;
+			buf.clear();
+		} else if (*reading == '|') {
+			if (!filled || !current) {
+				lg::wml_error() << "[context_free_grammar_generator] Parsing error: misplaced | symbol";
+				return false;
+			}
+			filled->push_back(buf);
+			current->possibilities_.push_back(std::vector<std::string>());
+			filled = &current->possibilities_.back();
+			buf.clear();
+		} else if (*reading == '\\' && reading[1] == 'n') {
+			reading++;
+			buf.push_back('\n');
+		} else if (*reading == '\\' && reading[1] == 't') {
+			reading++;
+			buf.push_back('\t');
+		} else {
+			if (*reading == '{') {
+				if (!filled) {
+					lg::wml_error() << "[context_free_grammar_generator] Parsing error: misplaced { symbol";
+					return false;
+				}
+				filled->push_back(buf);
+				buf.clear();
+			}
+			if (*reading == '}') {
+				if (!filled) {
+					lg::wml_error() << "[context_free_grammar_generator] Parsing error: misplaced } symbol";
+					return false;
+				}
+				filled->push_back(buf);
+				buf.clear();
+			} else buf.push_back(*reading);
+		}
+		reading++;
+	}
+	if (filled) filled->push_back(buf);
+
+	initialized_ = true;
+	return true;
+}
+
+std::string context_free_grammar_generator::print_nonterminal(const std::string& name, uint32_t* seed, short seed_pos) const {
+	std::string result;
+	std::map<std::string, nonterminal>::const_iterator found = nonterminals_.find(name);
+	if (found == nonterminals_.end()) {
+		lg::wml_error() << "[context_free_grammar_generator] Warning: needed nonterminal " << name << " not defined";
+		return "!" + name;
+	}
+	const context_free_grammar_generator::nonterminal& got = found->second;
+	unsigned int picked = seed[seed_pos++] % got.possibilities_.size();
+	if (seed_pos >= seed_size) seed_pos = 0;
+	if (picked == got.last_) {
+		picked = seed[seed_pos++] % got.possibilities_.size();
+		if (seed_pos >= seed_size) seed_pos = 0;
+	}
+	const_cast<unsigned int&>(got.last_) = picked; /* The variable last_ can change, the rest must stay const */
+	const std::vector<std::string>& used = got.possibilities_[picked];
+	for (unsigned int i = 0; i < used.size(); i++) {
+		if (used[i][0] == '{') result += print_nonterminal(used[i].substr(1), seed, seed_pos);
+		else result += used[i];
+	}
+	return result;
+}
+
+std::string context_free_grammar_generator::generate() const {
+	uint32_t seed[seed_size];
+	for (unsigned short int i = 0; i < seed_size; i++) {
+		seed[i] = random_new::generator->next_random();
+	}
+	return print_nonterminal("main", seed, 0);
+}

src/utils/context_free_grammar_generator.hpp

@@ -0,0 +1,61 @@
+/*
+   Copyright (C) 2016 by Ján Dugáček
+   Part of the Battle for Wesnoth Project http://www.wesnoth.org/
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2 of the License, or
+   (at your option) any later version.
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY.
+
+   See the COPYING file for more details.
+*/
+
+#ifndef CONTEXT_FREE_GRAMMAR_GENERATOR_INCLUDED
+#define CONTEXT_FREE_GRAMMAR_GENERATOR_INCLUDED
+
+#include <string>
+#include <map>
+#include <list>
+#include <vector>
+
+class context_free_grammar_generator
+{
+private:
+
+	struct nonterminal {
+		nonterminal() : last_(1) {}
+		std::vector<std::vector<std::string> > possibilities_;
+		unsigned int last_;
+	};
+
+	std::map<std::string, nonterminal> nonterminals_;
+	bool initialized_;
+	std::string print_nonterminal(const std::string& name, uint32_t* seed, short int seed_pos) const;
+	static const short unsigned int seed_size = 20;
+
+public:
+	/** Default constructor */
+	context_free_grammar_generator();
+
+	/** Initialisation
+	 * @param source the definition of the context-free grammar to use
+	 * @returns if the operation was successful
+	 */
+	bool constructFromString(const std::string& source);
+
+	/** Generates a possible word in the grammar set before
+	 * @returns the word
+	 */
+	std::string generate() const;
+
+	~context_free_grammar_generator();
+
+	/** Checks if the object is initialized
+	 * @returns if it is initialized
+	 */
+	bool is_initialized() const {return initialized_; }
+};
+
+#endif