Merge branch 'main' of github.com:a-ejiri/open_jtalk into 1.11

tsukumijima · Aug 5, 2024 · 213b605 · 213b605
2 parents 33d2eb6 + eedc127
commit 213b605
Show file tree

Hide file tree

Showing 3 changed files with 23 additions and 12 deletions.
diff --git a/src/njd_set_accent_phrase/njd_set_accent_phrase.c b/src/njd_set_accent_phrase/njd_set_accent_phrase.c
@@ -243,7 +243,7 @@ void njd_set_accent_phrase(NJD * njd)
 
          /* Rule 19 */
          if (strcmp(NJDNode_get_pos(node), NJD_SET_ACCENT_PHRASE_MEISHI) == 0 && strcmp(NJDNode_get_pos_group1(node), NJD_SET_ACCENT_PHRASE_HIJIRITSU) == 0) {
-               if (strcmp(NJDNode_get_pos_group1(node->prev), NJD_SET_ACCENT_PHRASE_JIRITSU) == 0 || strcmp(NJDNode_get_pos(node->prev), NJD_SET_ACCENT_PHRASE_JODOUSHI) == 0
+               if (strcmp(NJDNode_get_pos(node->prev), NJD_SET_ACCENT_PHRASE_JODOUSHI) == 0
                   || strcmp(NJDNode_get_pos(node->prev), NJD_SET_ACCENT_PHRASE_RENTAISHI) == 0
                   || strcmp(NJDNode_get_pos(node->prev), NJD_SET_ACCENT_PHRASE_JOSHI) == 0 
                   || strcmp(NJDNode_get_pos(node->prev), NJD_SET_ACCENT_PHRASE_DOUSHI) == 0 ) 
@@ -255,10 +255,15 @@ void njd_set_accent_phrase(NJD * njd)
 
          /* Rule 21 */
          if (strcmp(NJDNode_get_pos(node->prev), NJD_SET_ACCENT_PHRASE_MEISHI) == 0
-             && strcmp(NJDNode_get_pos_group1(node->prev), NJD_SET_ACCENT_PHRASE_HIJIRITSU) == 0
-             && strcmp(NJDNode_get_pos(node), NJD_SET_ACCENT_PHRASE_MEISHI) == 0)
+             && (strcmp(NJDNode_get_pos_group1(node->prev), NJD_SET_ACCENT_PHRASE_HIJIRITSU) == 0 ||strcmp(NJDNode_get_pos_group1(node->prev), NJD_SET_ACCENT_PHRASE_DAIMEISHI) == 0) 
+             && (strcmp(NJDNode_get_pos(node), NJD_SET_ACCENT_PHRASE_MEISHI) == 0 && strcmp(NJDNode_get_pos(node), NJD_SET_ACCENT_PHRASE_IPPAN) == 0))
+            NJDNode_set_chain_flag(node, 0);
+
+         /* Rule 22 */
+         if (strcmp(NJDNode_get_pos(node->prev), NJD_SET_ACCENT_PHRASE_KANDOUSHI) == 0)
+            NJDNode_set_chain_flag(node, 0);
+         if (strcmp(NJDNode_get_pos(node), NJD_SET_ACCENT_PHRASE_KANDOUSHI) == 0)
             NJDNode_set_chain_flag(node, 0);
-
       }
    }
 }

diff --git a/src/njd_set_accent_phrase/njd_set_accent_phrase_rule_utf_8.h b/src/njd_set_accent_phrase/njd_set_accent_phrase_rule_utf_8.h
@@ -70,18 +70,21 @@ NJD_SET_ACCENT_PHRASE_RULE_H_START;
   Rule 16 「*,*,*,姓」の後の「*,*,*,名」は別のアクセント句に
   Rule 17 「名詞」の後の「*,*,*,名」は別のアクセント句に
   Rule 18 「*,接尾」は前にくっつける
-  Rule 19 「名詞,非自立」は「動詞」or「*,自立」or「助動詞,体言接続」or 「助詞」or「連体詞」に接続する場合に前にくっつける
+  Rule 19 「名詞,非自立」は「動詞」or「助動詞,体言接続」or 「助詞」or「連体詞」に接続する場合に前にくっつける
   Rule 20 「固有名詞」は前とくっつけない
-  Rule 21 「名詞,非自立」の後の「名詞」は別のアクセント句に
+  Rule 21 「名詞,非自立」or「名詞,代名詞」の後の「名詞,一般」は別のアクセント句に
+  Rule 22 「感動詞」は単独のアクセントに
 */
 
 #define NJD_SET_ACCENT_PHRASE_MEISHI "名詞"
+#define NJD_SET_ACCENT_PHRASE_DAIMEISHI "代名詞"
 #define NJD_SET_ACCENT_PHRASE_KEIYOUSHI "形容詞"
 #define NJD_SET_ACCENT_PHRASE_DOUSHI "動詞"
 #define NJD_SET_ACCENT_PHRASE_FUKUSHI "副詞"
 #define NJD_SET_ACCENT_PHRASE_SETSUZOKUSHI "接続詞"
 #define NJD_SET_ACCENT_PHRASE_RENTAISHI "連体詞"
 #define NJD_SET_ACCENT_PHRASE_JODOUSHI "助動詞"
+#define NJD_SET_ACCENT_PHRASE_KANDOUSHI "感動詞"
 #define NJD_SET_ACCENT_PHRASE_JOSHI "助詞"
 #define NJD_SET_ACCENT_PHRASE_KIGOU "記号"
 #define NJD_SET_ACCENT_PHRASE_KOYUU_MEISHI "固有名詞"
@@ -90,6 +93,7 @@ NJD_SET_ACCENT_PHRASE_RULE_H_START;
 #define NJD_SET_ACCENT_PHRASE_SETSUBI "接尾"
 #define NJD_SET_ACCENT_PHRASE_JIRITSU "自立"
 #define NJD_SET_ACCENT_PHRASE_HIJIRITSU "非自立"
+#define NJD_SET_ACCENT_PHRASE_IPPAN "一般"
 #define NJD_SET_ACCENT_PHRASE_RENYOU "連用"
 #define NJD_SET_ACCENT_PHRASE_SETSUZOKUJOSHI "接続助詞"
 #define NJD_SET_ACCENT_PHRASE_SAHEN_SETSUZOKU "サ変接続"

diff --git a/src/njd_set_accent_type/njd_set_accent_type.c b/src/njd_set_accent_type/njd_set_accent_type.c
@@ -142,6 +142,8 @@ void njd_set_accent_type(NJD * njd)
    char rule[MAXBUFLEN];
    int add_type = 0;
    int mora_size = 0;
+   int top_acc = 0;
+   int top_mora_size = 0;
 
    if (njd == NULL || njd->head == NULL)
       return;
@@ -153,12 +155,17 @@ void njd_set_accent_type(NJD * njd)
          /* store the top node */
          top_node = node;
          mora_size = 0;
+         top_acc = NJDNode_get_acc(node);
+         top_mora_size = NJDNode_get_mora_size(node);
+
       } else if (node->prev != NULL && NJDNode_get_chain_flag(node) == 1) {
          /* get accent change type */
          get_rule(NJDNode_get_chain_rule(node), NJDNode_get_pos(node->prev), rule, &add_type);
 
          /* change accent type */
-         if (strcmp(rule, "*") == 0) {  /* no chnage */
+         if (strcmp(NJDNode_get_chain_rule(top_node), "P2") == 0 && top_acc != 0 && top_mora_size > top_acc) {
+            NJDNode_set_acc(top_node, top_acc);
+         } else if (strcmp(rule, "*") == 0) {  /* no chnage */
          } else if (strcmp(rule, "F1") == 0) {  /* for ancillary word */
          } else if (strcmp(rule, "F2") == 0) {
             if (NJDNode_get_acc(top_node) == 0)
@@ -184,11 +191,6 @@ void njd_set_accent_type(NJD * njd)
                NJDNode_set_acc(top_node, 0);
             else
                NJDNode_set_acc(top_node, mora_size + NJDNode_get_acc(node));
-         } else if (strcmp(rule, "P2") == 0) {
-            if (NJDNode_get_acc(node) == 0)
-               NJDNode_set_acc(top_node, mora_size + 1);
-            else
-               NJDNode_set_acc(top_node, mora_size + NJDNode_get_acc(node));
          } else if (strcmp(rule, "P6") == 0) {
             NJDNode_set_acc(top_node, 0);
          } else if (strcmp(rule, "P14") == 0) {