fix: 修复高级分段空格分段错误 (#688)

2024-07-02 14:59:31 +08:00 · 2024-07-02 14:59:31 +08:00 · 78294fc641
commit 78294fc641
parent 3a7a5ac50a
1 changed files with 1 additions and 1 deletions
--- a/apps/dataset/serializers/document_serializers.py
+++ b/apps/dataset/serializers/document_serializers.py
@ -790,7 +790,7 @@ class DocumentSerializers(ApiMixin, serializers.Serializer):
                    {'key': '#####', 'value': "(?<=\\n)(?<!#)##### (?!#).*|(?<=^)(?<!#)##### (?!#).*"},
                    {'key': '######', 'value': "(?<=\\n)(?<!#)###### (?!#).*|(?<=^)(?<!#)###### (?!#).*"},
                    {'key': '-', 'value': '(?<! )- .*'},
-                    {'key': '空格', 'value': '(?<!\\s)\\s(?!\\s)'},
+                    {'key': '空格', 'value': '(?<! ) (?! )'},
                    {'key': '分号', 'value': '(?<!；)；(?!；)'}, {'key': '逗号', 'value': '(?<!，)，(?!，)'},
                    {'key': '句号', 'value': '(?<!。)。(?!。)'}, {'key': '回车', 'value': '(?<!\\n)\\n(?!\\n)'},
                    {'key': '空行', 'value': '(?<!\\n)\\n\\n(?!\\n)'}]