refactor: enhance document and knowledge deletion logic to remove associated files

Author: wxg0103

apps/knowledge/serializers/document.py

@@ -665,6 +665,9 @@ def cancel(self, instance, with_valid=True):
         def delete(self):
             document_id = self.data.get("document_id")
             QuerySet(model=Document).filter(id=document_id).delete()
+            source_file_ids = [doc['meta'].get('source_file_id') for doc in
+                               Document.objects.filter(id__=document_id).values("meta")]
+            QuerySet(File).filter(id__in=source_file_ids).delete()
             # 删除段落
             QuerySet(model=Paragraph).filter(document_id=document_id).delete()
             # 删除问题
@@ -1189,6 +1192,9 @@ def batch_delete(self, instance: Dict, with_valid=True):
                 BatchSerializer(data=instance).is_valid(model=Document, raise_exception=True)
                 self.is_valid(raise_exception=True)
             document_id_list = instance.get("id_list")
+            source_file_ids = [doc['meta'].get('source_file_id') for doc in
+                               Document.objects.filter(id__in=document_id_list).values("meta")]
+            QuerySet(File).filter(id__in=source_file_ids).delete()
             QuerySet(Document).filter(id__in=document_id_list).delete()
             QuerySet(Paragraph).filter(document_id__in=document_id_list).delete()
             delete_problems_and_mappings(document_id_list)

apps/knowledge/serializers/knowledge.py

@@ -29,7 +29,7 @@
 from common.utils.logger import maxkb_logger
 from common.utils.split_model import get_split_model
 from knowledge.models import Knowledge, KnowledgeScope, KnowledgeType, Document, Paragraph, Problem, \
-    ProblemParagraphMapping, TaskType, State, SearchMode, KnowledgeFolder
+    ProblemParagraphMapping, TaskType, State, SearchMode, KnowledgeFolder, File
 from knowledge.serializers.common import ProblemParagraphManage, get_embedding_model_id_by_knowledge_id, MetaSerializer, \
     GenerateRelatedSerializer, get_embedding_model_by_knowledge_id, list_paragraph, write_image, zip_dir
 from knowledge.serializers.document import DocumentSerializers
@@ -417,6 +417,9 @@ def delete(self):
             QuerySet(WorkspaceUserResourcePermission).filter(target=knowledge.id).delete()
             QuerySet(ApplicationKnowledgeMapping).filter(knowledge_id=knowledge.id).delete()
             knowledge.delete()
+            File.objects.filter(
+                source_id=knowledge.id,
+            ).delete()
             delete_embedding_by_knowledge(self.data.get('knowledge_id'))
             return True