dump-name

xd2333 · Jun 9, 2024 · f401f39 · f401f39
1 parent d490e73
commit f401f39
Show file tree

Hide file tree

Showing 4 changed files with 31 additions and 3 deletions.
diff --git a/GalTransl/Frontend/GPT.py b/GalTransl/Frontend/GPT.py
@@ -84,7 +84,7 @@ async def doLLMTranslateSingleFile(
                     gptapi = CSakuraTranslate(
                         projectConfig, eng_type, endpoint, proxyPool
                     )
-                case "rebuildr" | "rebuilda":
+                case "rebuildr" | "rebuilda" | "dump-name":
                     gptapi = CRebuildTranslate(projectConfig, eng_type)
                 case _:
                     raise ValueError(f"不支持的翻译引擎类型 {eng_type}")
@@ -114,6 +114,16 @@ async def doLLMTranslateSingleFile(
             except Exception as e:
                 LOGGER.error(f"文件 {file_name} 加载翻译列表失败: {e}")
                 return False
+
+            # 导出人名表功能
+            if "dump-name" in eng_type:
+                global name_dict
+                for tran in trans_list:
+                    if tran.speaker and type(tran.speaker) == str:
+                        if tran.speaker not in name_dict:
+                            name_dict[tran.speaker] = 0
+                        name_dict[tran.speaker] += 1
+                return True
 
             # 2、翻译前处理
             for i, tran in enumerate(trans_list):
@@ -250,6 +260,11 @@ async def doLLMTranslate(
         LOGGER.info(f"当前使用 {workersPerProject} 个Sakura worker引擎")
     else:
         endpoint_queue = None
+
+    # 人名表初始化
+    if "dump-name" in eng_type:
+        global name_dict
+        name_dict = {}
 
     file_list = get_file_list(projectConfig.getInputPath())
     if not file_list:
@@ -279,3 +294,14 @@ async def doLLMTranslate(
     # await atqdm.gather(*tasks)
     await gather(*tasks)  # run
     progress_bar.close()
+
+    if "dump-name" in eng_type:
+        import csv
+        proj_dir = projectConfig.getProjectDir()
+        name_dict = dict(sorted(name_dict.items(), key=lambda item: item[1], reverse=True))
+        with open(joinpath(proj_dir, "人名替换表.csv"), "w", encoding="utf-8", newline="") as f:
+            writer = csv.writer(f)
+            writer.writerow(["JP_Name", "CN_Name", "Count"])  # 写入表头
+            for name, count in name_dict.items():
+                writer.writerow([name, "", count])
+            LOGGER.info(f"name已保存到'人名替换表.csv'（UTF-8编码，用Emeditor编辑），填入CN_Name后可用于后续翻译name字段。")
diff --git a/GalTransl/Name.py b/GalTransl/Name.py
@@ -18,5 +18,6 @@ def load_name_table(name_table_path: str) -> Dict[str, str]:
         # Skip the header
         next(reader)
         for row in reader:
-            name_table[row[0]] = row[1]
+            if row[1] !=  "":
+                name_table[row[0]] = row[1]
     return name_table
diff --git a/GalTransl/Runner.py b/GalTransl/Runner.py
@@ -135,7 +135,7 @@ def print_plugin_list(plugin_manager: PluginManager):
 
     # proxyPool初始化
     proxyPool = CProxyPool(cfg) if cfg.getKey("internals.enableProxy") else None
-    if proxyPool and translator != "Rebuild":
+    if proxyPool and translator not in ["rebuildr","rebuilda","dump-name","showplugs"]:
         await proxyPool.checkAvailablity()
         if not proxyPool.proxies:
             raise Exception("没有可用的代理，请检查代理设置")

diff --git a/GalTransl/__init__.py b/GalTransl/__init__.py
@@ -69,6 +69,7 @@
     "sakura-009": "为翻译轻小说/视觉小说开展大规模训练的本地模型。适用v0.9版prompt，不支持GPT字典",
     "rebuildr": "重建结果 用译前译后字典通过缓存刷写结果json -- 跳过翻译和写缓存",
     "rebuilda": "重建缓存和结果 用译前译后字典刷写缓存+结果json -- 跳过翻译",
+    "dump-name": "导出name字段，生成人名替换表，用于翻译name字段",
     "showplugs": "显示全部插件列表",
 }
 LANG_SUPPORTED = {