Structure of personal recent paper database

Each personal recent paper database is a vector database.

• The original files of a user-specific recent paper database are stored in documents\tmp_papers\{user_id}
• Each recent paper database has a root node, and all paper nodes are child nodes of the root node.
• The ID of a paper node is the absolute path of the document file. Additional information includes the time (date time) when the file was stored in the recent paper repository. Timestamp information can be used for time filtering during retrieval.
• All content nodes (doc nodes) of each paper are inserted as child nodes of the corresponding paper node.
• If the member calls the paper summary tool, the corresponding literature summary text will be added as a child node to the corresponding paper node.