Table auto-generation and editing system supporting natural language instructions

Abstract

The application discloses a table automatic generation and editing system supporting natural language instructions and belongs to the technical field of table automatic processing. In view of the technical problems that the existing table editing system has poor interactive flexibility and weak semantic adaptation capability, the system comprises a language instruction disassembling module, an operation code acquisition module, a differential result acquisition module and a table display module. The language instruction disassembling module is used for collecting natural language instructions of a user in real time and disassembling the natural language instructions into an atomic operation sequence arranged according to time dependency. The operation code acquisition module is used for carrying out vector embedding processing, column name semantic matching and disambiguation processing on the Chinese meaning of an actual physical column name in the natural language instructions and common synonyms of the actual physical column name, so as to form data operation codes corresponding to the natural language instructions. The differential result acquisition module is used for executing the data operation codes by using a restricted Python interpreter and acquiring a differential update set corresponding to a table according to an execution result. The table display module is used for generating a table view corresponding to the natural language instructions of the user by using the differential update set. The application realizes automatic conversion from natural language to table operation, and improves the semantic analysis accuracy, code execution safety and table view update efficiency.

Images