revert page column (#8217)

api/core/rag/datasource/vdb/vector_factory.py

@@ -27,7 +27,7 @@ class AbstractVectorFactory(ABC):
 class Vector:
     def __init__(self, dataset: Dataset, attributes: list = None):
         if attributes is None:
-            attributes = ["doc_id", "dataset_id", "document_id", "doc_hash", "page"]
+            attributes = ["doc_id", "dataset_id", "document_id", "doc_hash"]
         self._dataset = dataset
         self._embeddings = self._get_embeddings()
         self._attributes = attributes

api/core/workflow/nodes/knowledge_retrieval/knowledge_retrieval_node.py

@@ -163,9 +163,6 @@ class KnowledgeRetrievalNode(BaseNode):
             for item in all_documents:
                 if item.metadata.get("score"):
                     document_score_list[item.metadata["doc_id"]] = item.metadata["score"]
-                # both 'page' and 'score' are metadata fields
-                if item.metadata.get("page"):
-                    page_number_list[item.metadata["doc_id"]] = item.metadata["page"]
 
             index_node_ids = [document.metadata["doc_id"] for document in all_documents]
             segments = DocumentSegment.query.filter(
@@ -200,7 +197,6 @@ class KnowledgeRetrievalNode(BaseNode):
                                 "document_id": document.id,
                                 "document_name": document.name,
                                 "document_data_source_type": document.data_source_type,
-                                "page": page_number_list.get(segment.index_node_id, None),
                                 "segment_id": segment.id,
                                 "retriever_from": "workflow",
                                 "score": document_score_list.get(segment.index_node_id, None),

api/core/workflow/nodes/llm/llm_node.py

@@ -451,7 +451,6 @@ class LLMNode(BaseNode):
                 "segment_position": metadata.get("segment_position"),
                 "index_node_hash": metadata.get("segment_index_node_hash"),
                 "content": context_dict.get("content"),
-                "page": metadata.get("page"),
             }
 
             return source