Online data professional analysis report automatic generation system and method driven by large language model

Abstract

The invention discloses an online data professional analysis report automatic generation system and method driven by a large language model, and the system comprises a data base storage system which is composed of a structured database, an unstructured document library and a professional data format storage library, and a basic OLAP (Online Analysis and Query) service platform center is constructed based on the data base storage system. The OLAP service platform center provides a series of agents and supports special report generation service and intelligent question and answer service, and the data base provides necessary data needed by the electric power meteorological large model agents to generate special reports. The problems that at present, templates and data are filled into a complete professional meteorological service analysis report through a special report data filling program according to set logic, although an informatization platform improves the automation level of special report compiling to a certain degree, a large number of scenes needing professional predictors to conduct research, judgment and analysis are solved; and complete requirements still cannot be met.

Description

Technical Field

[0001] The present invention relates to the technical field of special report production, and in particular to a system and method for automatically generating online data professional analysis reports driven by a large language model. Background Art

[0002] Power meteorology is a sub-field of meteorological professional services. Power meteorology professional service personnel need to provide users with regular or on-demand, daily or customized professional meteorological service analysis reports. These reports are oriented to different regions and time periods, covering a large number of concepts and statistics in the power meteorology professional field. They are highly professional and customized professional data analysis reports.

[0003] At present, the common solution for compiling these professional meteorological service analysis reports is to compile various report templates based on an information platform that brings together various types of power meteorological data, compile statistical logic according to template requirements, combine template text description rules, and use special report data filling programs to fill in templates and data according to established logic to form complete professional meteorological service analysis reports. Although the information platform has improved the level of automation in special report compilation to a certain extent, it still cannot meet the complete needs when faced with a large number of scenarios that require professional forecasters to conduct research and analysis.

[0004] To solve the above problems, this application proposes a system and method for automatically generating online data professional analysis reports driven by a large language model. Summary of the Invention

[0005] In response to the problems in the related art, the present invention provides a system and method for automatically generating online data professional analysis reports driven by a large language model, which can meet the user's usage needs.

[0006] To this end, the specific technical solutions adopted in the present invention are as follows: A large language model-driven automatic generation system and method for online data professional analysis reports includes a data base storage system consisting of a structured database, an unstructured document library, and a professional data format storage library, and thereby constructs a basic OLAP (online analysis query) service platform center. The OLAP service platform center provides a series of intelligent agents to support special report generation services and intelligent question-and-answer services. The data base provides the necessary data for the power meteorology large model intelligent agent to generate special reports. The data base storage system stores various business data, image products, etc. required for the content of professional analysis reports, and is provided with data labels and prompt word templates according to the content category of the professional analysis report.

[0007] As a further solution of the present invention, the data tags are presented as a Web API integrated service within the system and stored and managed in a database.

[0008] As a further solution of the present invention, the prompt word template includes role definition, special report generation task description, example input data, example output special report fragment content and mandatory specifications, and each part of the prompt word template is replaceable, and its various field configurations are stored in the database and can be dynamically changed to provide more rich features.

[0009] As a further solution of the present invention, the architecture of the power meteorology large model intelligent agent is a multi-agent architecture consisting of a drawing single agent, a special report dialogue single agent, a meteorological data query single agent, a common task dialogue single agent and a problem type judgment single agent.

[0010] As a further solution of the present invention, the special report generation service includes a set of special report template configuration solutions and a matching special report rendering engine.

[0011] As a further solution of the present invention, the steps of generating a special report are as follows: S1: The user initiates a special report generation request through the application interface and enters the special report generation task process; S2: After the system accepts the special report generation request, it calls the response special report template configuration file based on the input parameters and requires the user to fill in the necessary parameters to determine the specific values of the configuration file; S3: After configuration is completed, according to the template configuration items, traverse the special report data tag items contained in the template, call the tag generation interface, and generate the special report content corresponding to the tag according to the configuration request; S4: A data tag, including a name, input parameter configuration, call data interface configuration, large language model configuration, and prompt word template. The prompt word template includes the large language model role definition, generation rule definition, special report content examples, and mandatory specifications. It also provides data placeholders to fill in the data request return results generated based on the configuration. The prompt word is then input into the large language model service to obtain a response from the large language model. S5: After all data labels are generated according to the above steps, the special report rendering engine is called to generate a special report file that meets the template requirements and is returned to the front desk for the user to retrieve.

[0012] The beneficial effects of the present invention are: The present invention sets up an intelligent entity of the large-scale electric power and meteorological model with certain understanding and reasoning capabilities, which can generate logical results according to prompt examples and requirements, so that special report generation no longer requires special editing of business logic, and the prompt word template provides rich flexibility. The expression of the content required for the special report can be arbitrarily modified according to the understanding of human natural language, and different special report contents can be generated based on similar data. In addition, it is also convenient to create new special reports based on the combination of existing data labels and prompt word templates, rather than the previous need to define a specific template form. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0014] Figure 1 This is a schematic diagram of the specific architecture of the power meteorology big model intelligent agent of the system and method for automatically generating online data professional analysis reports driven by a big language model according to an embodiment of the present invention; Figure 2 2. It is a schematic diagram of the application architecture of the system and method for automatically generating online professional data analysis reports driven by a large language model according to an embodiment of the present invention; Figure 3 2. It is a schematic diagram of the business architecture of the system and method for automatically generating online professional data analysis reports driven by a large language model according to an embodiment of the present invention; Figure 4 2 is a schematic diagram of the technical architecture of a system and method for automatically generating online data professional analysis reports driven by a large language model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0015] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention. They are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. By referring to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0016] According to an embodiment of the present invention, a system and method for automatically generating online data professional analysis reports driven by a large language model are provided.

[0017] Please refer to the instruction manual Figure 1-4According to an embodiment of the present invention, a large language model-driven online data professional analysis report automatic generation system and method includes a data base storage system composed of a structured database, an unstructured document library, and a professional data format repository. This system forms a basic OLAP (online analytical query) service platform center. The OLAP service platform center provides a series of intelligent agents to support special report generation and intelligent question-and-answer services. The special report generation service includes a set of special report template configuration solutions and a supporting special report rendering engine. The data base provides the necessary data for the power and meteorology large model intelligent agent to generate special reports. The data base storage system stores various business data, image products, and other information required for professional analysis report content. Data labels and prompt word templates are set according to the content category of the professional analysis report. The data labels are presented within the system as a Web API integrated service and stored and managed in a database. The prompt word template includes a role definition, a special report generation task description, sample input data, sample output special report fragment content, and mandatory specifications. Each part of the prompt word template is replaceable, and its field configuration is stored in the database and can be dynamically modified to provide more rich features. By setting up an intelligent entity of the power meteorology large model with certain understanding and reasoning capabilities, logical results can be generated according to prompt examples and requirements, so that special report generation no longer requires special editing of business logic. Moreover, the prompt word template provides rich flexibility, and the expression of the content required for the special report can be arbitrarily modified according to the understanding of human natural language, and different special report contents can be generated based on similar data. In addition, it is also convenient to create new special reports based on the combination of existing data labels and prompt word templates, rather than the previous need to define specific template forms.

[0018] In one embodiment, please refer to the appendix of the specification. Figure 1 As a further solution of the present invention, the architecture of the large-scale power meteorology model agent is a multi-agent architecture consisting of a drawing agent, a special report dialogue agent, a meteorological data query agent, a common task dialogue agent, and a problem type identification agent. The drawing agent uses the pandas and matplotlib libraries to plot the grid data based on meteorological grid data from various regions, and then encapsulates the grid data into a function tool for use by the large-scale model. Once the large-scale model recognizes the user's drawing requirements, it automatically passes in appropriate function parameters based on the user's requirements, performs the drawing, and returns the image and its interpretation. The special report dialogue agent cleans the data of PDF special report files in the power and meteorology field, extracts the text content, and vectorizes the content based on the jina-embedding-v3 model, storing it in the milvus vector database. To facilitate the use of text vectorization, the jina-embedding-v3 model is deployed locally and the service interface is opened. If the large model recognizes that the user needs to have a special report dialogue, it will call the embedding service interface to convert the user's question into a vector, perform similarity matching in the vector database, find the special report content that best meets the user's needs, and provide an answer. The meteorological data query agent is based on the meteorological grid files of each region. It creates a meteorological data query API interface for the region as the data base of the large model and encapsulates it as a tool call function. If the large model recognizes that the user needs to query meteorological data, it will automatically call the function and pass in appropriate parameters to obtain the corresponding meteorological data and answer the user's questions. Among them, the common task dialogue single agent retains the original large-model dialogue function and passes in context history memory, so that it can better serve users according to the current dialogue environment.

[0019] In one embodiment, please refer to the appendix of the specification. Figure 1 、 Figure 2 、 Figure 3 and Figure 4 As a further solution of the present invention, the steps for generating a special report are as follows: S1: The user initiates a special report generation request through the application interface and enters the special report generation task process; S2: After the system accepts the special report generation request, it calls the response special report template configuration file based on the input parameters and requires the user to fill in the necessary parameters to determine the specific values of the configuration file; S3: After configuration is completed, according to the template configuration items, traverse the special report data tag items contained in the template, call the tag generation interface, and generate the special report content corresponding to the tag according to the configuration request; S4: A data tag, including a name, input parameter configuration, call data interface configuration, large language model configuration, and prompt word template. The prompt word template includes the large language model role definition, generation rule definition, special report content examples, and mandatory specifications. It also provides data placeholders to fill in the data request return results generated based on the configuration. The prompt word is then input into the large language model service to obtain a response from the large language model. S5: After all data labels are generated according to the above steps, the special report rendering engine is called to generate a special report file that meets the template requirements and is returned to the front desk for the user to retrieve.

[0020] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A system and method for automatically generating online data professional analysis reports driven by a large language model, characterized by: It includes a data base storage system composed of a structured database, an unstructured document library and a professional data format storage library, and builds a basic OLAP (online analytical query) service platform center based on this. The OLAP service platform center provides a series of intelligent entities to support special report generation services and intelligent question and answer services. The data base provides the necessary data for the power meteorological large model intelligent entity to generate special reports. The data base storage system stores various business data, image products, etc. required for professional analysis report content, and is equipped with data labels and prompt word templates according to the content category of the professional analysis report.

2. The system and method for automatically generating online professional data analysis reports driven by a large language model according to claim 1 are characterized by: The data tags are presented as a Web API integrated service within the system and are stored and managed in the database.

3. The system and method for automatically generating online professional data analysis reports driven by a large language model according to claim 1 is characterized by: The prompt word template includes role definition, special report generation task description, sample input data, sample output special report fragment content and mandatory specifications, and each part of the prompt word template is replaceable. Its various field configurations are stored in the database and can be dynamically changed to provide more rich features.

4. The system and method for automatically generating online professional data analysis reports driven by a large language model according to claim 1 is characterized by: The architecture of the power meteorology large model intelligent agent is a multi-agent architecture consisting of a drawing single agent, a special report dialogue single agent, a meteorological data query single agent, a common task dialogue single agent and a problem type judgment single agent.

5. The system and method for automatically generating online professional data analysis reports driven by a large language model according to claim 1 is characterized by: The special report generation service includes a set of special report template configuration solutions and a supporting special report rendering engine.

6. The system and method for automatically generating online data professional analysis reports driven by a large language model according to claim 1 is characterized in that: The steps for generating a special report are as follows: S1: The user initiates a special report generation request through the application interface and enters the special report generation task process; S2: After the system accepts the special report generation request, it calls the response special report template configuration file based on the input parameters and requires the user to fill in the necessary parameters to determine the specific values of the configuration file; S3: After configuration is completed, according to the template configuration items, traverse the special report data tag items contained in the template, call the tag generation interface, and generate the special report content corresponding to the tag according to the configuration request; S4: A data tag, including a name, input parameter configuration, call data interface configuration, large language model configuration, and prompt word template. The prompt word template includes the large language model role definition, generation rule definition, special report content examples, and mandatory specifications. It also provides data placeholders to fill in the data request return results generated based on the configuration. The prompt word is then input into the large language model service to obtain a response from the large language model. S5: After all data labels are generated according to the above steps, the special report rendering engine is called to generate a special report file that meets the template requirements and is returned to the front desk for the user to retrieve.

Images