summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorPeng Wu <alexepico@gmail.com>2013-04-10 15:45:05 +0800
committerPeng Wu <alexepico@gmail.com>2013-04-10 15:45:05 +0800
commit274f51f5af2555c4aa2b95f5c45e1468be572f18 (patch)
tree95c817d79d4f5799682e3760545b424348dbde94
parent5cebe47fae1d07adc1d099c2fdbc1b06a4d526ce (diff)
downloadlibpinyin-274f51f5af2555c4aa2b95f5c45e1468be572f18.tar.gz
libpinyin-274f51f5af2555c4aa2b95f5c45e1468be572f18.tar.xz
libpinyin-274f51f5af2555c4aa2b95f5c45e1468be572f18.zip
fixes make distcheck
-rw-r--r--data/Makefile.am4
-rw-r--r--utils/storage/import_interpolation.cpp27
-rw-r--r--utils/training/gen_unigram.cpp24
3 files changed, 51 insertions, 4 deletions
diff --git a/data/Makefile.am b/data/Makefile.am
index da34394..fe1a6de 100644
--- a/data/Makefile.am
+++ b/data/Makefile.am
@@ -57,8 +57,8 @@ $(tablefiles): interpolation2.text
bigram.db: $(textual_model_data)
$(RM) $(binary_model_data)
../utils/storage/gen_binary_files --table-dir $(top_srcdir)/data
- ../utils/storage/import_interpolation < $(top_srcdir)/data/interpolation2.text
- ../utils/training/gen_unigram
+ ../utils/storage/import_interpolation --table-dir $(top_srcdir)/data < $(top_srcdir)/data/interpolation2.text
+ ../utils/training/gen_unigram --table-dir $(top_srcdir)/data
phrase_index.bin pinyin_index.bin $(binfiles): bigram.db
diff --git a/utils/storage/import_interpolation.cpp b/utils/storage/import_interpolation.cpp
index c1ac601..5a0030e 100644
--- a/utils/storage/import_interpolation.cpp
+++ b/utils/storage/import_interpolation.cpp
@@ -20,10 +20,21 @@
*/
#include <stdio.h>
+#include <locale.h>
#include <glib.h>
#include "pinyin_internal.h"
#include "utils_helper.h"
+
+static const gchar * table_dir = ".";
+
+static GOptionEntry entries[] =
+{
+ {"table-dir", 0, 0, G_OPTION_ARG_FILENAME, &table_dir, "table directory", NULL},
+ {NULL}
+};
+
+
enum LINE_TYPE{
BEGIN_LINE = 1,
END_LINE,
@@ -224,13 +235,27 @@ int main(int argc, char * argv[]){
FILE * input = stdin;
const char * bigram_filename = "bigram.db";
+ setlocale(LC_ALL, "");
+
+ GError * error = NULL;
+ GOptionContext * context;
+
+ context = g_option_context_new("- import interpolation model");
+ g_option_context_add_main_entries(context, entries, NULL);
+ if (!g_option_context_parse(context, &argc, &argv, &error)) {
+ g_print("option parsing failed:%s\n", error->message);
+ exit(EINVAL);
+ }
+
SystemTableInfo system_table_info;
- bool retval = system_table_info.load("table.conf");
+ gchar * filename = g_build_filename(table_dir, "table.conf", NULL);
+ bool retval = system_table_info.load(filename);
if (!retval) {
fprintf(stderr, "load table.conf failed.\n");
exit(ENOENT);
}
+ g_free(filename);
PhraseLargeTable2 phrase_table;
diff --git a/utils/training/gen_unigram.cpp b/utils/training/gen_unigram.cpp
index 47aee91..83a15c8 100644
--- a/utils/training/gen_unigram.cpp
+++ b/utils/training/gen_unigram.cpp
@@ -20,20 +20,42 @@
*/
#include <stdio.h>
+#include <locale.h>
#include "pinyin_internal.h"
#include "utils_helper.h"
+static const gchar * table_dir = ".";
+
+static GOptionEntry entries[] =
+{
+ {"table-dir", 0, 0, G_OPTION_ARG_FILENAME, &table_dir, "table directory", NULL},
+ {NULL}
+};
+
/* increase all unigram frequency by a constant. */
int main(int argc, char * argv[]){
+ setlocale(LC_ALL, "");
+
+ GError * error = NULL;
+ GOptionContext * context;
+
+ context = g_option_context_new("- increase uni-gram");
+ g_option_context_add_main_entries(context, entries, NULL);
+ if (!g_option_context_parse(context, &argc, &argv, &error)) {
+ g_print("option parsing failed:%s\n", error->message);
+ exit(EINVAL);
+ }
SystemTableInfo system_table_info;
- bool retval = system_table_info.load("table.conf");
+ gchar * filename = g_build_filename(table_dir, "table.conf", NULL);
+ bool retval = system_table_info.load(filename);
if (!retval) {
fprintf(stderr, "load table.conf failed.\n");
exit(ENOENT);
}
+ g_free(filename);
FacadePhraseIndex phrase_index;