Merge remote-tracking branch 'upstream/ltp-new-server'

examples/Makefile

@@ -1,4 +1,7 @@
-all: cws cws_cmdline pos par ner multi_cws_cmdline
+all: cws cws_cmdline multi_cws_cmdline \
+	pos pos_cmdline multi_pos_cmdline \
+	par \
+	ner
 
 cws: cws.cpp
 	g++ -o cws cws.cpp -I./ \
@@ -12,11 +15,33 @@ cws_cmdline: cws_cmdline.cpp
 		-I../thirdparty/boost/include \
 		-Wl,-dn -L../lib/ -lsegmentor -lboost_regex -Wl,-dy
 
+multi_cws_cmdline: multi_cws_cmdline.cpp
+	g++ -o multi_cws_cmdline multi_cws_cmdline.cpp \
+		thirdparty/tinythreadpp/tinythread.cpp \
+		-I./ \
+		-I../include/ \
+		-I../thirdparty/boost/include/ \
+		-I./thirdparty/tinythreadpp/ \
+		-Wl,-dn -L../lib/ -lsegmentor -lboost_regex -Wl,-dy -lpthread
+
 pos: pos.cpp
 	g++ -o pos pos.cpp -I./ \
 		-I../include/ \
 		-L../lib/ -lpostagger
 
+pos_cmdline: pos_cmdline.cpp
+	g++ -o pos_cmdline pos_cmdline.cpp -I./ \
+		-I../include/ \
+		-Wl,-dn -L../lib/ -lpostagger -Wl,-dy
+
+multi_pos_cmdline: multi_pos_cmdline.cpp
+	g++ -o multi_pos_cmdline multi_pos_cmdline.cpp \
+		thirdparty/tinythreadpp/tinythread.cpp \
+		-I./ \
+		-I../include/ \
+		-I./thirdparty/tinythreadpp/ \
+		-Wl,-dn -L../lib/ -lpostagger -Wl,-dy -lpthread
+
 ner: ner.cpp
 	g++ -o ner ner.cpp -I./ \
 		-I../src/ner/ \
@@ -27,21 +52,14 @@ par: par.cpp
 		-I../src/parser/ \
 		-L../lib -lparser
 
-multi_cws_cmdline: multi_cws_cmdline.cpp
-	g++ -o multi_cws_cmdline multi_cws_cmdline.cpp \
-		thirdparty/tinythreadpp/tinythread.cpp \
-		-I./ \
-		-I../include/ \
-		-I../thirdparty/boost/include/ \
-		-I./thirdparty/tinythreadpp/ \
-		-Wl,-dn -L../lib/ -lsegmentor -lboost_regex -Wl,-dy -lpthread
-
 .PHONY: clean
 
 clean:
 	rm cws
 	rm cws_cmdline
 	rm pos
+	rm pos_cmdline
 	rm ner
 	rm par
 	rm multi_cws_cmdline
+	rm multi_pos_cmdline

examples/multi_cws_cmdline.cpp

@@ -12,6 +12,7 @@
  * is not compilable under MSVC
  */
 #include <iostream>
+#include <cstring>
 #include <ctime>
 #include <vector>
 #include <list>
@@ -89,13 +90,14 @@ void multithreaded_segment( void * args) {
 }
 
 int main(int argc, char ** argv) {
-    if (argc < 2) {
-        std::cerr << "multi-cws [model path] [lexicon file]" << std::endl;
+    if (argc < 1 || (0 == strcmp(argv[1], "-h"))) {
+        std::cerr << "Example: ./multi_cws_cmdline [model path] [lexicon file]" << std::endl;
+        std::cerr << std::endl;
+        std::cerr << "This program recieve input word sequence from stdin." << std::endl;
+        std::cerr << "One sentence per line." << std::endl;
+        return -1;
     }
 
-    string sentence;
-    vector<string> result;
-
     void * engine = 0;
     if (argc == 2) {
         engine = segmentor_create_segmentor(argv[1]);
@@ -107,8 +109,6 @@ int main(int argc, char ** argv) {
         return -1;
     }
 
-    std::vector<std::string> words;
-
     int num_threads = thread::hardware_concurrency();
     std::cerr << "TRACE: Model is loaded" << std::endl;
     std::cerr << "TRACE: Running " << num_threads << " thread(s)" << std::endl;

examples/multi_pos_cmdline.cpp

@@ -0,0 +1,146 @@
+/*
+ * Multi-threaded postagger test program. The user input a line
+ * of Chinese sentence an the program will output its segment
+ * result.
+ *
+ *  @dependency package: tinythread - a portable c++ wrapper for
+ *                       multi-thread library.
+ *  @author:             LIU, Yijia
+ *  @data:               2013-09-24
+ *
+ * This program is special designed for UNIX user, for get time
+ * is not compilable under MSVC
+ */
+#include <iostream>
+#include <sstream>
+#include <cstring>
+#include <ctime>
+#include <vector>
+#include <list>
+#include <sys/time.h>
+#include <sys/types.h>
+
+#include "postag_dll.h"
+#include "tinythread.h"
+#include "fast_mutex.h"
+
+using namespace std;
+using namespace tthread;
+
+const int MAX_LEN = 1024;
+
+double get_time(void) {
+    struct timeval tv;
+    gettimeofday(&tv, NULL);
+    return tv.tv_sec + (tv.tv_usec / 1000000.0);
+}
+
+class Dispatcher {
+public:
+    Dispatcher( void * model ) {
+        _model = model;
+    }
+
+    int next(std::vector<std::string> &words) {
+        std::string line;
+        std::string word;
+        lock_guard<fast_mutex> guard(_mutex);
+        if (getline(std::cin, line, '\n')) {
+            std::stringstream S(line);
+            words.clear();
+            while (S >> word) { words.push_back(word); }
+        } else {
+            return -1;
+        }
+        return 0;
+    }
+
+    void output(const std::vector<std::string> & words,
+            const std::vector<std::string> &postags) {
+        lock_guard<fast_mutex> guard(_mutex);
+        if (words.size() != postags.size()) {
+            return;
+        }
+
+        for (int i = 0; i < words.size(); ++ i) {
+            std::cout << words[i] << "_" << postags[i];
+            std::cout << (i == words.size() - 1 ? '\n' : '|');
+        }
+        return;
+    }
+
+    void * model() {
+        return _model;
+    }
+
+private:
+    fast_mutex      _mutex;
+    void *          _model;
+    string          _sentence;
+};
+
+void multithreaded_postag( void * args) {
+    std::vector<std::string> words;
+    std::vector<std::string> postags;
+
+    Dispatcher * dispatcher = (Dispatcher *)args;
+    void * model = dispatcher->model();
+
+    while (true) {
+        int ret = dispatcher->next(words);
+
+        if (ret < 0)
+            break;
+
+        postags.clear();
+        postagger_postag(model, words, postags);
+        dispatcher->output(words, postags);
+    }
+
+    return;
+}
+
+int main(int argc, char ** argv) {
+    if (argc < 1 || (0 == strcmp(argv[1], "-h"))) {
+        std::cerr << "Usage: ./multi_pos_cmdline [model path]" << std::endl;
+        std::cerr << std::endl;
+        std::cerr << "This program recieve input word sequence from stdin." << std::endl;
+        std::cerr << "One sentence per line. Words are separated by space." << std::endl;
+        return -1;
+    }
+
+    void * engine = postagger_create_postagger(argv[1]);
+
+    if (!engine) {
+        return -1;
+    }
+
+    int num_threads = thread::hardware_concurrency();
+    std::cerr << "TRACE: Model is loaded" << std::endl;
+    std::cerr << "TRACE: Running " << num_threads << " thread(s)" << std::endl;
+
+    Dispatcher * dispatcher = new Dispatcher( engine );
+
+    double tm = get_time();
+    list<thread *> thread_list;
+    for (int i = 0; i < num_threads; ++ i) {
+        thread * t = new thread( multithreaded_postag, (void *)dispatcher );
+        thread_list.push_back( t );
+    }
+
+    for (list<thread *>::iterator i = thread_list.begin();
+            i != thread_list.end(); ++ i) {
+        thread * t = *i;
+        t->join();
+        delete t;
+    }
+
+    tm = get_time() - tm;
+    std::cerr << "TRACE: consume "
+        << tm 
+        << " seconds."
+        << std::endl;
+
+    return 0;
+}
+

examples/pos.cpp

@@ -5,6 +5,7 @@
 
 int main(int argc, char * argv[]) {
     if (argc < 1) {
+        std::cerr << "pos [model path]" << std::endl;
         return -1;
     }
 

examples/pos_cmdline.cpp

@@ -0,0 +1,82 @@
+/*
+ * Single-threaded segmentor test program. The user input a line
+ * of Chinese sentence an the program will output its segment
+ * result.
+ *
+ *  @dependency package: tinythread - a portable c++ wrapper for
+ *                       multi-thread library.
+ *  @author:             LIU, Yijia
+ *  @data:               2013-09-24
+ *
+ * This program is special designed for UNIX user, for get time
+ * is not compilable under MSVC
+ */
+#include <iostream>
+#include <sstream>
+#include <ctime>
+#include <cstring>
+#include <string>
+#include <sys/time.h>
+#include <sys/types.h>
+#include "postag_dll.h"
+
+double get_time(void) {
+    struct timeval tv;
+    gettimeofday(&tv, NULL);
+    return tv.tv_sec + (tv.tv_usec / 1000000.0);
+}
+
+int main(int argc, char * argv[]) {
+    if (argc < 1 || (0 == strcmp(argv[1], "-h"))) {
+        std::cerr << "Example: ./pos_cmdline [model path]" << std::endl;
+        std::cerr << std::endl;
+        std::cerr << "This program recieve input word sequence from stdin." << std::endl;
+        std::cerr << "One sentence per line. Words are separated by space." << std::endl;
+        std::cerr << std::endl;
+        return 1;
+    }
+
+    void * engine = postagger_create_postagger(argv[1]);
+    if (!engine) {
+        std::cerr << "WARNINIG : Failed to load model." << std::endl;
+        return -1;
+    }
+
+    std::string line;
+    std::string word;
+    std::vector<std::string> words;
+    std::vector<std::string> postags;
+
+    std::cerr << "TRACE: Model is loaded" << std::endl;
+    double tm = get_time();
+
+    while (std::getline(std::cin, line, '\n')) {
+        std::stringstream S(line);
+        words.clear();
+        while (S >> word) { words.push_back(word); }
+
+        if (words.size() == 0) { continue; }
+        int len = postagger_postag(engine, words, postags);
+        if (postags.size() != words.size()) {
+            std::cerr << "WARNINIG: Number of postags is different from number of words"
+                << std::endl;
+        }
+
+        for (int i = 0; i < len; ++ i) {
+            std::cout << words[i] << "_" << postags[i];
+            if (i+1 == len) std::cout <<std::endl;
+            else std::cout<< "|";
+        }
+    }
+
+    postagger_release_postagger(engine);
+
+    tm = get_time() - tm;
+    std::cerr << "TRACE: consume "
+        << tm 
+        << " seconds."
+        << std::endl;
+
+    return 0;
+}
+

src/ner/ner_dll.cpp

@@ -30,9 +30,6 @@ class NERWrapper : public ltp::ner::NER {
             return false;
         }
 
-        ltp::ner::rulebase::RuleBase base(model->labels);
-        decoder = new ltp::ner::Decoder(model->num_labels(), base);
-
         // beg_tag0 = model->labels.index( );
         // beg_tag1 = model->labels.index( );
 
@@ -42,6 +39,9 @@ class NERWrapper : public ltp::ner::NER {
     int recognize(const std::vector<std::string> & words,
             const std::vector<std::string> & postags,
             std::vector<std::string> & tags) {
+        ltp::ner::rulebase::RuleBase base(model->labels);
+        ltp::ner::Decoder deco(model->num_labels(), base);
+
         ltp::ner::Instance * inst = new ltp::ner::Instance;
         if (words.size() != postags.size()) {
             return 0;
@@ -54,7 +54,7 @@ class NERWrapper : public ltp::ner::NER {
 
         ltp::ner::NER::extract_features(inst);
         ltp::ner::NER::calculate_scores(inst, true);
-        decoder->decode(inst);
+        deco.decode(inst);
 
         for (int i = 0; i < words.size(); ++ i) {
             tags.push_back(model->labels.at(inst->predicted_tagsidx[i]));

src/postagger/postag_dll.cpp

@@ -27,21 +27,21 @@ class PostaggerWrapper : public ltp::postagger::Postagger {
             return false;
         }
 
-        decoder = new ltp::postagger::Decoder(model->num_labels());
-
         return true;
     }
 
     int postag(const std::vector<std::string> & words,
             std::vector<std::string> & tags) {
         ltp::postagger::Instance * inst = new ltp::postagger::Instance;
+        ltp::postagger::Decoder deco(model->num_labels());
+
         for (int i = 0; i < words.size(); ++ i) {
             inst->forms.push_back(ltp::strutils::chartypes::sbc2dbc_x(words[i]));
         }
 
         ltp::postagger::Postagger::extract_features(inst);
         ltp::postagger::Postagger::calculate_scores(inst, true);
-        decoder->decode(inst);
+        deco.decode(inst);
 
         ltp::postagger::Postagger::build_labels(inst, tags);