Multi-agent automatic analysis method, system, equipment and device based on complex report engine

Abstract

The invention provides a multi-agent automatic analysis method, system and device based on a complex report engine and a storage medium, and the method comprises the following steps: converting and decomposing an input problem and an industry to which the complex report engine belongs into a plurality of analysis indexes, the converted and decomposed analysis indexes are respectively input into a configuration template of the multi-agent response system; disassembling each analysis index in the configuration template into a corresponding task list; sequentially performing task execution according to the task lists disassembled from each analysis index, and outputting each task list result to a corresponding configuration template according to a preset standard output format to obtain an index conclusion of the analysis index; the index conclusions corresponding to the analysis indexes in the configuration template are integrated, and then answers corresponding to the input questions are output through a preset output template. According to the method, numerous table data in a complex report engine can be automatically analyzed, and a relatively accurate and comprehensive problem conclusion can be quickly output.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a multi-agent automatic analysis method, system, equipment and device based on a complex report engine. Background Art

[0002] Complex report engine, including forms (actual business data) and customized reports, can automatically generate reports with complex structures and formats according to pre-defined rules, templates and data sources; customized reports have the characteristics of form aggregation, report association, multi-sheet association and multi-region association. With the expansion of enterprise scale and diversification of business, a large amount of data has accumulated within the enterprise. These data are stored in different systems and databases, such as customer relationship management system (CRM), enterprise resource planning system (ERP), etc. In order to better manage and utilize these data, enterprises need a tool that can integrate these scattered data and present them in the form of intuitive and complex reports. For example, a large manufacturing enterprise stores its production data in the production management system, sales data in the sales system, and financial data in the financial system. In order to fully understand the operation status of the enterprise, such as analyzing the production cost, sales profit and inventory turnover rate of each product, a complex report engine is needed to integrate these data and generate comprehensive reports. Complex report engine is equivalent to Excel with elastic search (distributed search engine) capabilities.

[0003] Most of the reporting systems display key indicators that managers care about (for example, performance summary tables, labor efficiency statistics tables, sales funnel conversion statistics tables, etc.). Managers can quickly and directly understand these key indicators, and even see some comparisons of these indicators (for example, year-on-year / month-on-month). However, as more and more reports are generated, managers of enterprises simply cannot keep up with them; especially in complex reporting engines, reports are relatively complex, and the analysis of an indicator may involve multiple reports, multiple sheets in an Excel table, and multiple data forms, etc. It is difficult for managers to find the answers they care about at a glance; in addition, for some reports, to find the key factors that cause changes in key indicators, it is necessary to understand the relationship between report data by yourself, which takes a lot of time for users. Therefore, how to automatically analyze many tables based on the questions entered by users (such as managers and bosses of enterprises) and quickly get the answers they want to know and are more comprehensive is an urgent problem to be solved in the field of complex reporting engines. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention proposes a multi-agent automatic analysis method, system, equipment and device based on a complex report engine, which can automatically analyze numerous tables and quickly output more comprehensive problem conclusions.

[0005] The technical solution of the present invention is achieved in this way:

[0006] In one aspect, the present invention provides a multi-agent automatic analysis method based on a complex report engine, comprising the following steps:

[0007] According to the input question and the industry to which the complex report engine belongs, it is transformed and decomposed into multiple analysis indicators, and the transformed and decomposed analysis indicators are respectively input into the configuration template of the multi-agent response system;

[0008] Decompose each analysis indicator in the configuration template into a corresponding task list, which includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks;

[0009] After executing the tasks in turn according to the task list decomposed from each analysis indicator, the results of each task list are output to the corresponding configuration template according to the preset standard output format to obtain the indicator conclusion of the analysis indicator;

[0010] Integrate the indicator conclusions corresponding to each analysis indicator in the configuration template, and then output the answer corresponding to the input question using the preset output template.

[0011] Preferably, the step of disassembling each analysis indicator in the configuration template into a corresponding task list includes:

[0012] Determine an indicator type of an analysis indicator in a configuration template, wherein the indicator type includes but is not limited to a standard causal type, a customized causal type, a standard prediction type, and a customized prediction type;

[0013] Determine the indicator analysis process of the analysis indicator according to the determined indicator type, and disassemble the corresponding task list according to the preset task list of the determined indicator analysis process; the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks.

[0014] Preferably, the task list decomposed according to each analysis indicator is executed in turn, and the results of each task list are output to the corresponding configuration template according to a preset standard output format, and the step of obtaining the indicator conclusion of the analysis indicator includes:

[0015] According to the split task list, the corresponding classification prompt is constructed based on the task list questions and the output history of this round of tasks. The data agent Agent determines the programming task type of the task list according to the classification prompt;

[0016] According to the task code requirements of the programming task type, the corresponding tools are indexed and the coding prompt is constructed in combination with the historical memory information, and the Python agent writes the code according to the coding prompt;

[0017] A code verification prompt is constructed according to the task code requirements of the coding task type and the common problems of the index tool, and the verification agent verifies the code written by the Python agent according to the code verification prompt;

[0018] The python code executor agent executes the code verified by the verification agent agent;

[0019] Construct a formatting prompt according to the formatting requirements of the corresponding coding task type and the code execution results of the Python code executor agent, and the output agent Agent outputs the corresponding task list results in a structured manner according to the formatting prompt;

[0020] For the split task lists, cycle through the above steps in sequence to finally obtain the results corresponding to each task list, and then output these results to the corresponding configuration template according to the preset standard output format, and finally obtain the indicator conclusion of the analysis indicator.

[0021] Preferably, the data agent Agent determines the type of programming task in the task list according to the classification prompt:

[0022] The encoding task type includes node function category information, node query information, node change calculation information and node factor evaluation information, and a coding prompt is constructed based on the relevant information of the task list.

[0023] On the other hand, the present invention also provides a multi-agent automatic analysis system based on a complex report engine, comprising

[0024] The analysis indicator conversion module is used to convert and decompose the input question and the industry to which the complex report engine belongs into multiple analysis indicators, and input the converted and decomposed analysis indicators into the configuration template of the multi-agent response system respectively;

[0025] A task list disassembly module is used to disassemble each analysis indicator in the configuration template into a corresponding task list, which includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks;

[0026] The indicator conclusion output module is used to execute tasks in sequence according to the task list disassembled from each analysis indicator, and then output the results of each task list to the corresponding configuration template according to the preset standard output format to obtain the indicator conclusion of the analysis indicator;

[0027] The standard answer output module is used to integrate the indicator conclusions corresponding to each analysis indicator in the configuration template, and then output the answer corresponding to the input question using a preset output template.

[0028] Preferably, the task list disassembly module includes:

[0029] An indicator type determination unit, used to determine the indicator type of the analysis indicator in the configuration template, the indicator type including but not limited to a standard causal type, a customized causal type, a standard prediction type and a customized prediction type;

[0030] The task list disassembly unit is used to determine the indicator analysis process of the analysis indicator according to the determined indicator type, and disassemble the corresponding task list according to the preset task list of the determined indicator analysis process; the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks.

[0031] Preferably, the indicator conclusion output module includes:

[0032] The data agent processing unit is used to construct corresponding classification prompts based on the split task list and the task list problems and the output history of this round of tasks in turn, and the data agent determines the programming task type of the task list according to the classification prompts;

[0033] The python agent processing unit is used to index the corresponding tool and construct a coding prompt in combination with the historical memory information according to the task code requirements of the programming task type, and the python agent writes the code according to the coding prompt;

[0034] The verification agent processing unit is used to construct a code verification prompt according to the task code requirements of the coding task type and the common problems of the index tool, and the verification agent verifies the code written by the Python agent according to the code verification prompt;

[0035] The executor agent execution unit is used to execute the code verified by the verification agent agent through the python code executor agent;

[0036] The output agent processing unit is used to construct a formatting prompt according to the formatting requirements of the corresponding coding task type and the code execution result of the python code executor agent, and the output agent Agent outputs the corresponding task list result in a structured manner according to the formatting prompt;

[0037] The indicator conclusion output unit is used to cycle the above steps in sequence for the split task lists to finally obtain the results corresponding to each task list, and then output these results to the corresponding configuration template according to the preset standard output format, and finally obtain the indicator conclusion of the analysis indicator.

[0038] Preferably, in the data agent processing unit, the encoding task type includes node function category information, node query information, node change calculation information and node factor evaluation information, and a coding prompt is constructed according to the relevant information of the task list.

[0039] On the other hand, the present invention also provides a computer electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned multi-agent automatic analysis method based on a complex report engine when executing the computer program.

[0040] On the other hand, the present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the above-mentioned multi-agent automatic analysis method based on a complex report engine.

[0041] Compared with the prior art, the present invention has the following advantages: the present invention splits the complex problem input by the user into multiple analysis indicators for analysis, and each analysis indicator is further split into a series of task lists, and then the task is executed and the results of the corresponding task list are output, so as to provide a comprehensive and accurate answer to the question input by the user; and the present invention also outputs the results of each task list to the corresponding configuration template according to the preset standard output format, and after the indicator conclusions corresponding to each analysis indicator are integrated, the answer corresponding to the input question is also output in the preset output template, so that the user can quickly obtain the information he wants from a large number of reports. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 creative labor.

[0043] Figure 1 It is a flow chart of the multi-agent automatic analysis method based on a complex report engine of the present invention;

[0044] Figure 2 It is a structural block diagram of the multi-agent automatic analysis system based on the complex report engine of the present invention;

[0045] Figure 3 It is a structural diagram of multi-agent;

[0046] Figure 4 The present invention is a structural block diagram of a computer electronic device. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] See also Figure 1 The embodiment of the present invention discloses a multi-agent automatic analysis method based on a complex report engine, comprising the following steps:

[0049] S1, transform and decompose the input question and the industry to which the complex report engine belongs into multiple analysis indicators, and input the transformed and decomposed analysis indicators into the configuration template of the multi-agent response system respectively;

[0050] S2, disassemble each analysis indicator in the configuration template into a corresponding task list, wherein the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks;

[0051] S3, after executing the tasks in turn according to the task list decomposed from each analysis indicator, output the results of each task list to the corresponding configuration template according to the preset standard output format to obtain the indicator conclusion of the analysis indicator;

[0052] S4, integrating the indicator conclusions corresponding to each analysis indicator in the configuration template, and then outputting the answer corresponding to the input question using a preset output template.

[0053] Correspondingly, see Figure 2 The embodiment of the present invention also discloses a multi-agent automatic analysis system based on a complex report engine, including

[0054] The analysis indicator conversion module is used to convert and decompose the input question and the industry to which the complex report engine belongs into multiple analysis indicators, and input the converted and decomposed analysis indicators into the configuration template of the multi-agent response system respectively;

[0055] A task list disassembly module is used to disassemble each analysis indicator in the configuration template into a corresponding task list, which includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks;

[0056] The indicator conclusion output module is used to execute tasks in sequence according to the task list disassembled from each analysis indicator, and then output the results of each task list to the corresponding configuration template according to the preset standard output format to obtain the indicator conclusion of the analysis indicator;

[0057] The standard answer output module is used to integrate the indicator conclusions corresponding to each analysis indicator in the configuration template, and then output the answer corresponding to the input question using a preset output template.

[0058] In this embodiment, the multi-agent automatic analysis method based on a complex report engine takes the multi-agent automatic analysis system based on a complex report engine as the execution object of the steps. Specifically, step S1 takes the analysis indicator conversion module as the execution object of the step, step S2 takes the task list disassembly module as the execution object of the step, step S3 takes the indicator conclusion output module as the execution object of the step, and step S4 takes the standard answer output module as the execution object of the step.

[0059] A complex report engine, including forms (actual business data) and customized reports. Customized reports have the characteristics of form aggregation, report association, multi-sheet association and multi-region association. The definitions of form aggregation, report association, multi-sheet association and multi-region association are as follows:

[0060] 1) Form aggregation: various statistical aggregations based on the original form (database level);

[0061] 2) Report association: data association reference of multiple analysis reports;

[0062] 3) Multi-sheet association: data association reference of multiple sheets in a single report;

[0063] 4) Multi-area association: data association reference of multiple areas in a single sheet;

[0064] To put it simply, a complex reporting engine is actually an Excel with an embedded database application.

[0065] Among them, the structural block diagram of the multi-agent response system is as follows Figure 3As shown, when conducting specific single indicator analysis, the data in the complex reporting engine will be used as the basis.

[0066] In step S1, since the complex report engines constructed by different industries are different, the analysis indicators converted and decomposed according to the relevant industry characteristics of the complex report engine and the questions input by the user in the embodiment of the present invention will be different. For example, the decline in supermarket performance and the decline in manufacturing performance have completely different focuses of analysis, which may involve various factors such as stores, people, products, channels, etc. The content of the complex report engine is different. Therefore, it is necessary to convert and decompose into multiple analysis indicators according to the industry to which the complex report engine belongs and the user input questions during the specific implementation process, and the analysis indicators decomposed from different industries will be different. And these multiple analysis indicators will involve multiple factors. For example, when a user asks "Why did the performance decline this month?", then step 1 will be divided into multiple analysis indicators according to this question, such as performance indicators, sales type indicators, sales process indicators... etc. These indicators involve different factors in various aspects. Only by analyzing the analysis indicators of different factors can the input questions be answered comprehensively and accurately.

[0067] In an embodiment of the present invention, corresponding analysis indicators can be pre-set according to complex report engines of different industries. When a user asks a question, a specific analysis indicator is determined from the preset analysis indicators based on the question asked, and can be specifically determined through the constructed model.

[0068] Moreover, the present invention inputs these analysis indicators into the configuration template, and finally the corresponding indicator conclusions of each analysis indicator are also finally filled back into the configuration template, thereby forming a standard output result so that users can quickly obtain the information they want.

[0069] In step 2, each analysis indicator in the configuration template needs to be broken down into a series of task lists for task execution, and the output results after execution form the analysis conclusion corresponding to the analysis indicator.

[0070] Specifically, in step 2, the step of disassembling each analysis indicator in the configuration template into a corresponding task list includes:

[0071] S201, determining the indicator type of the analysis indicator in the configuration template, wherein the indicator type is not limited to a standard causal type, a customized causal type, a standard prediction type, and a customized prediction type;

[0072] S202, determining the indicator analysis process of the analysis indicator according to the determined indicator type, and disassembling the corresponding task list according to the preset task list of the determined indicator analysis process; the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks.

[0073] Correspondingly, in the multi-agent automatic analysis system based on the complex report engine, the task list disassembly module includes:

[0074] An indicator type determination unit, used to determine the indicator type of the analysis indicator in the configuration template, the indicator type including but not limited to a standard causal type, a customized causal type, a standard prediction type and a customized prediction type;

[0075] The task list disassembly unit is used to determine the indicator analysis process of the analysis indicator according to the determined indicator type, and disassemble the corresponding task list according to the preset task list of the determined indicator analysis process; the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks.

[0076] Step 2 in the multi-agent automatic analysis method based on a complex report engine is to use the task list disassembly module in the multi-agent automatic analysis system based on a complex report engine as the execution object of the step. Specifically, step S201 is to use the indicator type determination unit as the execution object of the step, and step S202 is to use the task list disassembly unit as the execution object of the step.

[0077] In step S201, after splitting the problem into multiple analysis indicators, it is also necessary to determine whether each indicator type is a standard causal type, a customized causal type, a standard prediction type, a customized prediction type, or other types. For example, the performance indicator split out of the above-mentioned "Why did the performance decline this month?" is used as an example. To analyze the performance of this month, this is a result that can be directly queried in the report engine, so it belongs to the standard causal type. Similarly, other analysis indicators can also determine the indicator type of the analysis indicator based on the constructed model or implementation method or pre-setting.

[0078] In step S202, since the corresponding indicator analysis process has been pre-set in the system for standard causal type, customized causal type, standard prediction type, customized prediction type or other types, after determining the indicator type of the analysis indicator, the corresponding task list will be disassembled according to the preset task list of the corresponding indicator analysis process. Taking the performance indicator as an example, to analyze the performance in August, you need to know what function to use for analysis, what analysis method to adopt, what tools to use, how to present the results, etc. Therefore, it is necessary to disassemble this analysis indicator into a corresponding task list for analysis, and the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks.

[0079] Specifically, in step S3, the task list decomposed according to each analysis indicator is executed in turn, and the results of each task list are output to the corresponding configuration template according to the preset standard output format. The step of obtaining the indicator conclusion of the analysis indicator includes:

[0080] S301, according to the split task list, build corresponding classification prompts based on the task list problems and the output history of this round of tasks, and the data agent Agent determines the programming task type of the task list according to the classification prompts;

[0081] S302, according to the task code requirements of the programming task type, index the corresponding tool and build a coding prompt in combination with the historical memory information, and the Python agent writes the code according to the coding prompt;

[0082] S303, constructing a code verification prompt according to the task code requirements of the coding task type and the common problems of the indexing tool, and the verification agent verifies the code written by the Python agent according to the code verification prompt;

[0083] S304, executing the code verified by the verification agent agent through the Python code executor agent;

[0084] S305, constructing a formatting prompt according to the formatting requirements of the corresponding coding task type and the code execution result of the python code executor agent, and the output agent Agent performs structured output of the corresponding task list result according to the formatting prompt;

[0085] S306, the split task lists are cycled through the above steps in sequence to finally obtain the results corresponding to each task list, and then these results are output to the corresponding configuration template according to the preset standard output format, and finally the indicator conclusion of the analysis indicator is obtained.

[0086] Correspondingly, in the multi-agent automatic analysis system based on the complex report engine, the indicator conclusion output module includes:

[0087] The data agent processing unit is used to construct corresponding classification prompts based on the split task list and the task list problems and the output history of this round of tasks in turn, and the data agent determines the programming task type of the task list according to the classification prompts;

[0088] The python agent processing unit is used to index the corresponding tool and construct a coding prompt in combination with the historical memory information according to the task code requirements of the programming task type, and the python agent writes the code according to the coding prompt;

[0089] The verification agent processing unit is used to construct a code verification prompt according to the task code requirements of the coding task type and the common problems of the index tool, and the verification agent verifies the code written by the Python agent according to the code verification prompt;

[0090] The executor agent execution unit is used to execute the code verified by the verification agent agent through the python code executor agent;

[0091] The output agent processing unit is used to construct a formatting prompt according to the formatting requirements of the corresponding coding task type and the code execution result of the python code executor agent, and the output agent Agent outputs the corresponding task list result in a structured manner according to the formatting prompt;

[0092] The indicator conclusion output unit is used to cycle the above steps in sequence for the split task lists to finally obtain the results corresponding to each task list, and then output these results to the corresponding configuration template according to the preset standard output format, and finally obtain the indicator conclusion of the analysis indicator.

[0093] Similarly, step 3 in the multi-agent automatic analysis method based on a complex report engine takes the indicator conclusion output module in the multi-agent automatic analysis system based on a complex report engine as the execution object of the step. Specifically, step S301 takes the data agent agent processing unit as the execution object of the step, step S302 takes the python agent processing unit as the execution object of the step, step S303 takes the verification agent agent processing unit as the execution object of the step, step S304 takes the actuator agent execution unit as the execution object of the step, step S305 takes the output agent agent processing unit as the execution object of the step, and step S306 takes the indicator conclusion output unit as the execution object of the step.

[0094] In this embodiment, after the corresponding task list is split out, the corresponding classification prompt, coding prompt, code verification prompt, and formatting prompt are constructed based on each item in the task list, so that the data agent Agent can judge the programming task type of the task list according to the classification prompt prompt. After determining the coding task type of the task, the python agent knows what function to quote for coding. Therefore, at this time, the python agent writes the code according to the prompt word given by the coding prompt prompt. The written code is verified by the verification agent agent according to the code verification prompt prompt, and then executed by the python code executor agent. Finally, the output agent Agent outputs the corresponding task list results to the configuration template according to the formatting prompt prompt. In this embodiment, the split task list is cycled through steps S301-S305 in sequence to finally obtain the results corresponding to each task list. These results are filled back into the configuration template as required, forming the indicator conclusion of the analysis indicator.

[0095] In step S4, after the indicator conclusions of all analysis indicators are obtained, all indicator conclusions are integrated, and finally the answers to the questions asked by the user are output using a preset output template, so that the user can quickly obtain relevant information of the questions he asked from many reports.

[0096] Specifically, in step S301, the data agent Agent determines the type of programming task in the task list according to the classification prompt:

[0097] The encoding task type includes node function category information, node query information, node change calculation information and node factor evaluation information, and a coding prompt is constructed based on the relevant information of the task list.

[0098] Correspondingly, in the data agent processing unit, the encoding task type includes node function category information, node query information, node change calculation information and node factor evaluation information, and a coding prompt is constructed according to the relevant information of the task list.

[0099] In this embodiment, the encoding task type includes information such as node function category information, node query information, node change calculation information, and node factor evaluation information, so as to facilitate the subsequent guidance of the Python agent to write code.

[0100] In summary, the present invention splits the complex problems input by the user into multiple analysis indicators for analysis, and each analysis indicator is further split into a series of task lists, and then the tasks are executed and the results of the corresponding task lists are output, so as to provide a comprehensive and accurate answer to the questions input by the user; and the present invention also outputs the results of each task list to the corresponding configuration template according to a preset standard output format, and after the indicator conclusions corresponding to each analysis indicator are integrated, the answer corresponding to the input question is also output in a preset output template, so that users can quickly obtain the information they want from numerous reports.

[0101] On the other hand, the present invention also provides a computer electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned multi-agent automatic analysis method based on a complex report engine when executing the computer program.

[0102] On the other hand, the present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the above-mentioned multi-agent automatic analysis method based on a complex report engine.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-agent automatic analysis method based on a complex report engine, characterized in that: The following steps are involved: According to the input question and the industry to which the complex report engine belongs, it is transformed and decomposed into multiple analysis indicators, and the transformed and decomposed analysis indicators are respectively input into the configuration template of the multi-agent response system; Decompose each analysis indicator in the configuration template into a corresponding task list, which includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks; After executing the tasks in turn according to the task list decomposed from each analysis indicator, the results of each task list are output to the corresponding configuration template according to the preset standard output format to obtain the indicator conclusion of the analysis indicator; Integrate the indicator conclusions corresponding to each analysis indicator in the configuration template, and then output the answer corresponding to the input question using the preset output template.

2. The multi-agent automatic analysis method based on a complex report engine according to claim 1 is characterized in that: The step of disassembling each analysis indicator in the configuration template into a corresponding task list includes: Determine an indicator type of an analysis indicator in a configuration template, wherein the indicator type includes but is not limited to a standard causal type, a customized causal type, a standard prediction type, and a customized prediction type; Determine the indicator analysis process of the analysis indicator according to the determined indicator type, and disassemble the corresponding task list according to the preset task list of the determined indicator analysis process; the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks.

3. The multi-agent automatic analysis method based on a complex report engine according to claim 1 or 2, characterized in that: The steps of executing tasks in turn according to the task list decomposed from each analysis indicator, and outputting the results of each task list to the corresponding configuration template according to the preset standard output format, and obtaining the indicator conclusion of the analysis indicator include: According to the split task list, the corresponding classification prompt is constructed based on the task list questions and the output history of this round of tasks. The data agent Agent determines the programming task type of the task list according to the classification prompt; According to the task code requirements of the programming task type, the corresponding tools are indexed and the coding prompt is constructed in combination with the historical memory information, and the Python agent writes the code according to the coding prompt; Construct a code verification prompt according to the task code requirements of the coding task type and the common problems of the index tool, and the verification agent verifies the code written by the Python agent according to the code verification prompt; The python code executor agent executes the code verified by the verification agent agent; Construct a formatting prompt according to the formatting requirements of the corresponding coding task type and the code execution results of the Python code executor agent, and the output agent Agent outputs the corresponding task list results in a structured manner according to the formatting prompt; For the split task lists, cycle through the above steps in sequence to finally obtain the results corresponding to each task list, and then output these results to the corresponding configuration template according to the preset standard output format, and finally obtain the indicator conclusion of the analysis indicator.

4. The multi-agent automatic analysis method based on a complex report engine according to claim 3 is characterized in that: The data agent Agent determines the type of programming task in the task list according to the classification prompt: The encoding task type includes node function category information, node query information, node change calculation information and node factor evaluation information, and a coding prompt is constructed based on the relevant information of the task list.

5. A multi-agent automatic analysis system based on a complex report engine, characterized in that: It includes an analysis indicator conversion module, which is used to convert and decompose the input question and the industry to which the complex report engine belongs into multiple analysis indicators, and input the converted and decomposed analysis indicators into the configuration template of the multi-agent response system respectively; A task list disassembly module is used to disassemble each analysis indicator in the configuration template into a corresponding task list, which includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks; The indicator conclusion output module is used to execute tasks in sequence according to the task list disassembled from each analysis indicator, and then output the results of each task list to the corresponding configuration template according to the preset standard output format to obtain the indicator conclusion of the analysis indicator; The standard answer output module is used to integrate the indicator conclusions corresponding to each analysis indicator in the configuration template, and then output the answer corresponding to the input question using a preset output template.

6. The multi-agent automatic analysis method based on a complex report engine according to claim 1 is characterized in that: The task list disassembly module includes: An indicator type determination unit, used to determine the indicator type of the analysis indicator in the configuration template, the indicator type including but not limited to a standard causal type, a customized causal type, a standard prediction type and a customized prediction type; The task list disassembly unit is used to determine the indicator analysis process of the analysis indicator according to the determined indicator type, and disassemble the corresponding task list according to the preset task list of the determined indicator analysis process; the task list includes but is not limited to classification tasks, encoding tasks, verification tasks, code execution tasks and formatted output tasks.

7. The multi-agent automatic analysis system based on a complex report engine according to claim 5 or 6, characterized in that: The indicator conclusion output module includes: The data agent processing unit is used to construct corresponding classification prompts based on the split task list and the task list problems and the output history of this round of tasks in turn, and the data agent determines the programming task type of the task list according to the classification prompts; The python agent processing unit is used to index the corresponding tool and construct a coding prompt in combination with the historical memory information according to the task code requirements of the programming task type, and the python agent writes the code according to the coding prompt; The verification agent processing unit is used to construct a code verification prompt according to the task code requirements of the coding task type and the common problems of the index tool, and the verification agent verifies the code written by the Python agent according to the code verification prompt; The executor agent execution unit is used to execute the code verified by the verification agent agent through the python code executor agent; The output agent processing unit is used to construct a formatting prompt according to the formatting requirements of the corresponding coding task type and the code execution result of the python code executor agent, and the output agent Agent outputs the corresponding task list result in a structured manner according to the formatting prompt; The indicator conclusion output unit is used to cycle the above steps in sequence to obtain the results corresponding to each task list, and then output these results to the corresponding configuration template according to the preset standard output format, and finally obtain the indicator conclusion of the analysis indicator.

8. The multi-agent automatic analysis system based on a complex report engine according to claim 3 is characterized in that: In the data agent processing unit, the encoding task type includes node function category information, node query information, node change calculation information and node factor evaluation information, and a coding prompt is constructed according to the relevant information of the task list.

9. A computer electronic device, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the multi-agent automatic analysis method based on a complex report engine as described in any one of claims 1 to 4 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the multi-agent automatic analysis method based on a complex report engine according to any one of claims 1 to 4 are implemented.

Images