feat: 完善文档分块参数配置

- 支持自定义正则匹配内容保留开关及配置持久化

- 约束分块重叠长度始终小于分块长度

- 补充中英文交互文案

easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/service/impl/DocumentServiceImpl.java

@@ -720,6 +720,9 @@ public class DocumentServiceImpl extends ServiceImpl<DocumentMapper, Document> i
         if (config.getMdSplitterLevel() == null || config.getMdSplitterLevel() <= 0) {
             config.setMdSplitterLevel(RagDefaults.MD_SPLITTER_LEVEL);
         }
+        if (config.getRetainRegexMatch() == null) {
+            config.setRetainRegexMatch(Boolean.FALSE);
+        }
         return config;
     }
 
@@ -742,6 +745,10 @@ public class DocumentServiceImpl extends ServiceImpl<DocumentMapper, Document> i
         config.setChunkSize(asInteger(rawProfile.get("chunkSize"), config.getChunkSize()));
         config.setOverlapSize(asInteger(rawProfile.get("overlapSize"), config.getOverlapSize()));
         config.setRegex(asString(rawProfile.get("regex")));
+        if (rawProfile.containsKey("retainRegexMatch")) {
+            config.setRetainRegexMatch(
+                asBoolean(rawProfile.get("retainRegexMatch"), false));
+        }
         config.setRowsPerChunk(asInteger(rawProfile.get("rowsPerChunk"), config.getRowsPerChunk()));
         config.setMdSplitterLevel(asInteger(rawProfile.get("mdSplitterLevel"), config.getMdSplitterLevel()));
         return config;
@@ -763,6 +770,9 @@ public class DocumentServiceImpl extends ServiceImpl<DocumentMapper, Document> i
         if (StringUtil.hasText(source.getRegex())) {
             target.setRegex(source.getRegex());
         }
+        if (source.getRetainRegexMatch() != null) {
+            target.setRetainRegexMatch(source.getRetainRegexMatch());
+        }
         if (source.getRowsPerChunk() != null) {
             target.setRowsPerChunk(source.getRowsPerChunk());
         }
@@ -777,6 +787,8 @@ public class DocumentServiceImpl extends ServiceImpl<DocumentMapper, Document> i
         map.put("chunkSize", strategyConfig.getChunkSize());
         map.put("overlapSize", strategyConfig.getOverlapSize());
         map.put("regex", strategyConfig.getRegex());
+        map.put("retainRegexMatch", Boolean.TRUE.equals(
+            strategyConfig.getRetainRegexMatch()));
         map.put("rowsPerChunk", strategyConfig.getRowsPerChunk());
         map.put("mdSplitterLevel", strategyConfig.getMdSplitterLevel());
         return map;