perf(toolbox): 词素分析跳过空句

ba-archive · Jun 23, 2024 · b978a30 · b978a30
1 parent 7b6d590
commit b978a30
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 0 deletions.
diff --git a/apps/ba-online-toolbox/src/tools/ScenarioEditor/components/OriginalTextDisp.vue b/apps/ba-online-toolbox/src/tools/ScenarioEditor/components/OriginalTextDisp.vue
@@ -39,6 +39,7 @@ async function getSemantics() {
         message: res.message,
         type: "error",
       });
+      return;
     }
     // @ts-ignore
     useTranslationCache.setSemantic(props.text, res.tokens);

diff --git a/apps/ba-online-toolbox/src/tools/ScenarioEditor/types/Semantic.ts b/apps/ba-online-toolbox/src/tools/ScenarioEditor/types/Semantic.ts
@@ -9,6 +9,7 @@ export type SemanticUnit = {
 export enum AnthropicStatusCode {
   NORMAL,
   TOO_SHORT,
+  NO_CJK,
   NO_API_KEY,
   API_ERROR,
   JSON_PARSE_ERROR,

diff --git a/apps/ba-online-toolbox/src/tools/public/helper/AnthropicSemanticParser.ts b/apps/ba-online-toolbox/src/tools/public/helper/AnthropicSemanticParser.ts
@@ -114,6 +114,16 @@ async function parseSemantics(
     tokens: [] as SemanticUnit[],
   };
 
+  const CJKChars = new RegExp(
+    "[\u3040-\u309F\u30A0-\u30FF\u3400-\u4DBF\u4E00-\u9FFF\uF900-\uFAFF]"
+  );
+
+  if (!CJKChars.test(input)) {
+    result.status = AnthropicStatusCode.NO_CJK;
+    result.message =
+      "输入不包含有效日文字符";
+  }
+
   if (input && input.length < 10) {
     result.status = AnthropicStatusCode.TOO_SHORT;
     result.message = "输入过短，请在长度超过10个字符的文本中使用。";