From 1326788b7e88051375c54b9eef5f3d19457a7db4 Mon Sep 17 00:00:00 2001
From: Peng Wu <alexepico@gmail.com>
Date: Thu, 5 May 2011 11:32:55 +0800
Subject: add cmd line options to prune k mixture model

---
 utils/training/gen_ngram.cpp             |  8 ++++----
 utils/training/prune_k_mixture_model.cpp | 27 +++++++++++++++++++++------
 2 files changed, 25 insertions(+), 10 deletions(-)

(limited to 'utils')

diff --git a/utils/training/gen_ngram.cpp b/utils/training/gen_ngram.cpp
index 4ac75e3..9db5d58 100644
--- a/utils/training/gen_ngram.cpp
+++ b/utils/training/gen_ngram.cpp
@@ -41,14 +41,14 @@ int main(int argc, char * argv[]){
 
     setlocale(LC_ALL, "");
     while ( i < argc ){
-	if ( strcmp("--help", argv[i] ) == 0){
+	if ( strcmp("--help", argv[i]) == 0){
 	    print_help();
             exit(0);
-	}else if ( strcmp("--skip-pi-gram-training", argv[i] ) == 0) {
+	}else if ( strcmp("--skip-pi-gram-training", argv[i]) == 0 ){
 	    train_pi_gram = false;
-	}else if ( strcmp("--skip-unigram-training", argv[i] ) == 0) {
+	}else if ( strcmp("--skip-unigram-training", argv[i]) == 0 ){
 	    train_unigram = false;
-	}else if ( strcmp("--bigram-file", argv[i] ) == 0){
+	}else if ( strcmp("--bigram-file", argv[i]) == 0){
             if ( ++i >= argc ) {
                 print_help();
                 exit(EINVAL);
diff --git a/utils/training/prune_k_mixture_model.cpp b/utils/training/prune_k_mixture_model.cpp
index 09243f3..8845648 100644
--- a/utils/training/prune_k_mixture_model.cpp
+++ b/utils/training/prune_k_mixture_model.cpp
@@ -30,7 +30,7 @@ static guint32 g_prune_k = 3;
 static parameter_t g_prune_poss = 0.99;
 
 void print_help(){
-    printf("prune_k_mixture_model <FILENAME>\n");
+    printf("prune_k_mixture_model -k <INT> --CDF <FLOAT>  <FILENAME>\n");
 }
 
 bool prune_k_mixture_model(KMixtureModelMagicHeader * magic_header,
@@ -78,14 +78,29 @@ bool prune_k_mixture_model(KMixtureModelMagicHeader * magic_header,
 }
 
 int main(int argc, char * argv[]){
+    int i = 1;
     const char * bigram_filename = NULL;
 
     setlocale(LC_ALL, "");
-    if ( 2 != argc ){
-        print_help();
-        exit(EINVAL);
-    } else {
-        bigram_filename = argv[1];
+    while ( i < argc ){
+        if ( strcmp("--help", argv[i]) == 0 ){
+            print_help();
+            exit(0);
+        } else if ( strcmp("-k", argv[i]) == 0 ){
+            if ( ++i >= argc ){
+                print_help();
+                exit(EINVAL);
+            }
+            g_prune_k = atoi(argv[i]);
+        } else if ( strcmp("--CDF", argv[i]) == 0 ){
+            if ( ++i >= argc ){
+                print_help();
+                exit(EINVAL);
+            }
+            g_prune_poss = atof(argv[i]);
+        } else {
+            bigram_filename = argv[i];
+        }
     }
 
     /* TODO: magic header signature check here. */
-- 
cgit