working hash generator

concordia/hash_generator.cpp

@@ -3,17 +3,17 @@
 #include <boost/archive/binary_oarchive.hpp>
 #include <boost/archive/binary_iarchive.hpp>
 #include <boost/algorithm/string.hpp>
-#include <boost/serialization/map.hpp>
 #include <fstream>
 
-
-HashGenerator::HashGenerator(const string & wordMapFilename) throw(ConcordiaException) {
-    _wordMapFilename = wordMapFilename;
+HashGenerator::HashGenerator(const string & wordMapFilename)
+                                         throw(ConcordiaException) :
+    _wordMapFilename(wordMapFilename),
+    _wordMap(boost::shared_ptr<WordMap>(new WordMap)) {
     if (boost::filesystem::exists(_wordMapFilename)) {
         ifstream ifs(_wordMapFilename.c_str(), std::ios::binary);
         boost::archive::binary_iarchive ia(ifs);
-        ia >> _wordMap;
-        
+        boost::shared_ptr<WordMap> restoredWordMap(new WordMap);
+        ia >> *_wordMap;
     }
 }
 
@@ -25,8 +25,11 @@ vector<int> HashGenerator::generateHash(const string & sentence) {
     vector<string> tokenTexts;
     boost::split(tokenTexts, sentence, boost::is_any_of(" "));
 
-    for(vector<string>::iterator it = tokenTexts.begin(); it != tokenTexts.end(); ++it) {
+    for (vector<string>::iterator it = tokenTexts.begin();
+                                it != tokenTexts.end(); ++it) {
         string token = *it;
+        int code = _wordMap->getWordCode(token);
+        result.push_back(code);
     }
 
     return result;
@@ -35,7 +38,7 @@ vector<int> HashGenerator::generateHash(const string & sentence) {
 void HashGenerator::serializeWordMap() {
     ofstream ofs(_wordMapFilename.c_str(), std::ios::binary);
     boost::archive::binary_oarchive oa(ofs);
-    oa << _wordMap;
+    oa << *_wordMap;
 }
 
 

concordia/hash_generator.hpp

@@ -4,6 +4,8 @@
 #include <string>
 #include <map>
 #include <vector>
+#include <boost/shared_ptr.hpp>
+#include "concordia/word_map.hpp"
 #include "concordia/concordia_exception.hpp"
 
 /*!
@@ -15,7 +17,8 @@ using namespace std;
 
 class HashGenerator {
 public:
-    explicit HashGenerator(const string & wordMapFilename) throw(ConcordiaException);
+    explicit HashGenerator(const string & wordMapFilename)
+                                    throw(ConcordiaException);
 
     /*! Destructor.
     */
@@ -26,11 +29,9 @@ public:
     void serializeWordMap();
 
 private:
-    
-    map<string,int> _wordMap;
+    boost::shared_ptr<WordMap> _wordMap;
 
     string _wordMapFilename;
-
 };
 
 #endif

concordia/t/CMakeLists.txt

@@ -2,6 +2,7 @@ add_library(concordia-tests
   test_concordia.cpp
   test_concordia_config.cpp
   test_word_map.cpp
+  test_hash_generator.cpp
   )
 
 target_link_libraries(concordia-tests concordia ${LIBCONFIG_LIB} concordia-tests-common)

concordia/t/test_hash_generator.cpp

@@ -0,0 +1,57 @@
+#include <boost/filesystem.hpp>
+#include "tests/unit-tests/unit_tests_globals.hpp"
+#include <string>
+
+#include "concordia/hash_generator.hpp"
+
+#define TEST_WORD_MAP_PATH "/tmp/test_word_map.bin"
+
+using namespace std;
+
+BOOST_AUTO_TEST_SUITE(hash_generator)
+
+BOOST_AUTO_TEST_CASE( SimpleHashTest )
+{
+    if (boost::filesystem::exists(TEST_WORD_MAP_PATH)) {
+        boost::filesystem::remove(TEST_WORD_MAP_PATH);      
+    } 
+    
+    HashGenerator hashGenerator = HashGenerator(TEST_WORD_MAP_PATH);
+
+    vector<int> hash = hashGenerator.generateHash("Ala ma kota");
+    vector<int> expected;
+    expected.push_back(0);
+    expected.push_back(1);
+    expected.push_back(2);
+
+    BOOST_CHECK_EQUAL_COLLECTIONS(hash.begin(), hash.end(), expected.begin(), expected.end());
+}
+
+BOOST_AUTO_TEST_CASE( HashSerializationTest )
+{
+    if (boost::filesystem::exists(TEST_WORD_MAP_PATH)) {
+        boost::filesystem::remove(TEST_WORD_MAP_PATH);      
+    } 
+
+    HashGenerator hashGenerator1 = HashGenerator(TEST_WORD_MAP_PATH);
+    vector<int> hash1 = hashGenerator1.generateHash("Ala ma kota");
+    vector<int> expected1;
+    expected1.push_back(0);
+    expected1.push_back(1);
+    expected1.push_back(2);
+    BOOST_CHECK_EQUAL_COLLECTIONS(hash1.begin(), hash1.end(), expected1.begin(), expected1.end());
+
+    hashGenerator1.serializeWordMap();
+   
+    HashGenerator hashGenerator2 = HashGenerator(TEST_WORD_MAP_PATH);
+    vector<int> hash2 = hashGenerator2.generateHash("Ala ma psa");
+    vector<int> expected2;
+    expected2.push_back(0);
+    expected2.push_back(1);
+    expected2.push_back(3);
+    BOOST_CHECK_EQUAL_COLLECTIONS(hash2.begin(), hash2.end(), expected2.begin(), expected2.end());
+
+    boost::filesystem::remove(TEST_WORD_MAP_PATH);       
+}
+
+BOOST_AUTO_TEST_SUITE_END()

concordia/word_map.cpp

@@ -9,8 +9,7 @@ WordMap::~WordMap() {
 }
 
 int WordMap::getWordCode(const string & word) {
-    
-    if (_map.find(word) == _map.end() ) {
+    if (_map.find(word) == _map.end()) {
         int newCode = _nextFree;
         _map[word] = newCode;
         _nextFree++;
@@ -18,7 +17,6 @@ int WordMap::getWordCode(const string & word) {
     } else {
         return _map[word];
     }
-
 }
 
 

concordia/word_map.hpp

@@ -4,10 +4,13 @@
 #include <string>
 #include <map>
 #include "concordia/concordia_exception.hpp"
+#include <boost/archive/text_oarchive.hpp>
+#include <boost/archive/text_iarchive.hpp>
+#include <boost/serialization/map.hpp>
 
 
 /*!
-  Class for generating a sentence hash.
+  Class representing dictionary for word to int encoding.
 
 */
 
@@ -24,11 +27,18 @@ public:
     int getWordCode(const string & word);
 
 private:
+    friend class boost::serialization::access;
 
-    map<string,int> _map;
+    template<class Archive>
+
+    void serialize(Archive & ar, const unsigned int version) {
+        ar & _map;
+        ar & _nextFree;
+    }
+
+    map<string, int> _map;
 
     int _nextFree;
-
 };
 
 #endif