Merge pull request HIT-SCIR#272 from HIT-SCIR/develop

Develop
Kitter · Dec 17, 2017 · 5ed1662 · 5ed1662
2 parents f44254e + 93b33d7
commit 5ed1662
Show file tree

Hide file tree

Showing 5 changed files with 47 additions and 30 deletions.
diff --git a/doc/ltpserver.rst b/doc/ltpserver.rst
@@ -56,6 +56,7 @@ LTP Server在轻量级服务器程序mongoose基础上开发。在编译LTP源
 
 其中较为重要的参数包括：
 
+- last-stage: 特别的，可以使用 "|" 来分割多个最终目标。例如需要ner和parser就设置为 "ner|dp"。
 - port：指定LTP server监听的端口
 - threads：指定LTP server运行的线程数，线程数影响并发的处理能力
 - log-level：指定日志级别，TRACE级别最低，显示日志信息最详细。INFO级别最高，显示日志最粗略。WARN与ERROR级日志默认显示。

diff --git a/doc/ltptest.rst b/doc/ltptest.rst
@@ -109,6 +109,8 @@ Pipeline与last-stage参数
 
 默认情况下，LTP将进行至语义角色标注的分析。但是，对于一部分用户，某些分析并不必要。举例来讲，如果用户只需进行词性标注，则ltp_test的pipeline分析只需进行到pos，`last-stage`用来指明分析的最后状态。同时，如果`last-stage`指定为pos，句法分析、命名实体识别和语义角色标注的模型将不被加载。
 
+注： last-stage: 特别的，可以使用 "|" 来分割多个最终目标。例如需要ner和parser就设置为 "ner|dp"。
+
 .. _xxxcmdline-reference-label:
 
 xxx_cmdline

diff --git a/src/console/ltp_test.cpp b/src/console/ltp_test.cpp
@@ -127,17 +127,22 @@ int main(int argc, char *argv[]) {
     }
   }
 
-  std::string last_stage = LTP_SERVICE_NAME_DEFAULT;
+  std::string last_stage = "all";
   if (vm.count("last-stage")) {
     last_stage = vm["last-stage"].as<std::string>();
-    if (last_stage != LTP_SERVICE_NAME_SEGMENT
-        && last_stage != LTP_SERVICE_NAME_POSTAG
-        && last_stage != LTP_SERVICE_NAME_NER
-        && last_stage != LTP_SERVICE_NAME_DEPPARSE
-        && last_stage != LTP_SERVICE_NAME_SRL
-        && last_stage != LTP_SERVICE_NAME_ALL) {
-      std::cerr << "Unknown stage name:" << last_stage << ", reset to '" LTP_SERVICE_NAME_DEFAULT "'" << std::endl;
-      last_stage = LTP_SERVICE_NAME_DEFAULT;
+    vector<string> stages = ltp::strutils::split_by_sep(last_stage, "|");
+
+    for (int j = 0; j < stages.size(); ++j) {
+      if (stages[j] != LTP_SERVICE_NAME_SEGMENT
+          && stages[j] != LTP_SERVICE_NAME_POSTAG
+          && stages[j] != LTP_SERVICE_NAME_NER
+          && stages[j] != LTP_SERVICE_NAME_DEPPARSE
+          && stages[j] != LTP_SERVICE_NAME_SRL
+          && stages[j] != "all") {
+        std::cerr << "Unknown stage name:" << last_stage << ", reset to 'all'" << std::endl;
+        last_stage = "all";
+        break;
+      }
     }
   }
 

diff --git a/src/ltp/Ltp.cpp b/src/ltp/Ltp.cpp
@@ -2,6 +2,7 @@
 #include <ctime>
 #include <map>
 #include <string>
+#include <utils/strutils.hpp>
 
 #include "xml4nlp/Xml4nlp.h"
 #include "splitsnt/SplitSentence.h"
@@ -50,19 +51,22 @@ bool LTP::load(const std::string& last_stage,
     const std::string& srl_model_file) {
 
   size_t target_mask = 0;
-  if (last_stage == LTP_SERVICE_NAME_SEGMENT) {
-    target_mask = kActiveSegmentor;
-  } else if (last_stage == LTP_SERVICE_NAME_POSTAG) {
-    target_mask = (kActiveSegmentor|kActivePostagger);
-  } else if (last_stage == LTP_SERVICE_NAME_NER) {
-    target_mask = (kActiveSegmentor|kActivePostagger|kActiveNER);
-  } else if (last_stage == LTP_SERVICE_NAME_DEPPARSE) {
-    target_mask = (kActiveSegmentor|kActivePostagger|kActiveParser);
-  } else if (last_stage == LTP_SERVICE_NAME_SRL) {
-    target_mask = (kActiveSegmentor|kActivePostagger|kActiveParser|kActiveSRL);
-  } else if (last_stage == "all") {
-    target_mask =
-      (kActiveSegmentor|kActivePostagger|kActiveNER|kActiveParser|kActiveSRL);
+  vector<string> stages = ltp::strutils::split_by_sep(last_stage, "|");
+  for (int j = 0; j < stages.size(); ++j) {
+    if (stages[j] == LTP_SERVICE_NAME_SEGMENT) {
+      target_mask |= kActiveSegmentor;
+    } else if (stages[j] == LTP_SERVICE_NAME_POSTAG) {
+      target_mask |= (kActiveSegmentor | kActivePostagger);
+    } else if (stages[j] == LTP_SERVICE_NAME_NER) {
+      target_mask |= (kActiveSegmentor | kActivePostagger | kActiveNER);
+    } else if (stages[j] == LTP_SERVICE_NAME_DEPPARSE) {
+      target_mask |= (kActiveSegmentor | kActivePostagger | kActiveParser);
+    } else if (stages[j] == LTP_SERVICE_NAME_SRL) {
+      target_mask |= (kActiveSegmentor | kActivePostagger | kActiveParser | kActiveSRL);
+    } else if (stages[j] == "all") {
+      target_mask |=
+              (kActiveSegmentor | kActivePostagger | kActiveNER | kActiveParser | kActiveSRL);
+    }
   }
 
   size_t loaded_mask = 0;

diff --git a/src/server/ltp_server.cpp b/src/server/ltp_server.cpp
@@ -109,14 +109,19 @@ int main(int argc, char *argv[]) {
   std::string last_stage = "all";
   if (vm.count("last-stage")) {
     last_stage = vm["last-stage"].as<std::string>();
-    if (last_stage != LTP_SERVICE_NAME_SEGMENT
-        && last_stage != LTP_SERVICE_NAME_POSTAG
-        && last_stage != LTP_SERVICE_NAME_NER
-        && last_stage != LTP_SERVICE_NAME_DEPPARSE
-        && last_stage != LTP_SERVICE_NAME_SRL
-        && last_stage != "all") {
-      std::cerr << "Unknown stage name:" << last_stage << ", reset to 'all'" << std::endl;
-      last_stage = "all";
+    vector<string> stages = ltp::strutils::split_by_sep(last_stage, "|");
+
+    for (int j = 0; j < stages.size(); ++j) {
+      if (stages[j] != LTP_SERVICE_NAME_SEGMENT
+          && stages[j] != LTP_SERVICE_NAME_POSTAG
+          && stages[j] != LTP_SERVICE_NAME_NER
+          && stages[j] != LTP_SERVICE_NAME_DEPPARSE
+          && stages[j] != LTP_SERVICE_NAME_SRL
+          && stages[j] != "all") {
+        std::cerr << "Unknown stage name:" << last_stage << ", reset to 'all'" << std::endl;
+        last_stage = "all";
+        break;
+      }
     }
   }