IT intelligent research and development method and system based on GPT

Abstract

The invention discloses an IT intelligent research and development method and system based on GPT. The method comprises the steps of constructing a task template, configuring analysis tasks of different scene tasks in a research and development process, blocking a task stage, generating a task stage, subtasks and task targets. Receiving demand data, matching the demand data with the current task template, and outputting demand content; respectively configuring sub-tasks of a task blocking stage and a task generating stage based on the demand content; and executing the sub-task in the task blocking stage, carrying out quality inspection on the output of the current sub-task, and if the quality inspection is qualified, entering a task generation stage, and generating output content. The task content adaptive to the target is generated based on the multi-stage task processing framework of the task template, so that the accuracy and quality of task execution are ensured, efficient collaboration and automation of the research and development process are realized, the task execution efficiency and the result reliability are improved, and the task execution efficiency is improved. The problems of difficult demand analysis, loose task circulation, poor configuration adaptability and the like in the traditional research and development process are solved.

Description

Technical Field

[0001] The present invention relates to the field of GPT-assisted R&D technology, and in particular to a GPT-based IT intelligent R&D method and system. Background Art

[0002] With the rapid development of artificial intelligence technology, natural language processing capabilities based on large language models (such as GPT) have shown great application potential in various industries. However, in the field of IT R&D, the traditional R&D process still has the following significant problems: low efficiency of demand analysis and difficult to ensure quality, demand documents are often complex and diverse, manual analysis is time-consuming and prone to misunderstanding, resulting in a decline in the quality of subsequent tasks (such as R&D and testing). The current tools on the market mainly focus on demand management and storage, and lack intelligent demand analysis and content optimization capabilities. The R&D task chain is loose, and there is a lack of connection between quality inspection and generation. In the R&D process, the logical review, code generation, and test case writing of demand documents are often handled independently, lacking a unified task flow mechanism. The existing methods lack a task phase-based quality inspection and generation task linkage mechanism, which can easily lead to low demand quality or task results that do not meet the target requirements. Task configuration lacks flexibility and scalability. The R&D task configuration in different scenarios is complex. Existing tools usually use a fixed task configuration method, which is difficult to quickly adapt to the actual needs of different projects. There is limited support for dynamic input data, which makes it difficult to effectively adapt to demand changes or multi-scenario task expansion requirements. The automated R&D support capabilities are insufficient. Although GPT has certain applications in code generation and test case generation, most of them are isolated functional implementations and lack a systematic task management and execution framework. In summary, the IT R&D field urgently needs an intelligent R&D method based on GPT that can efficiently analyze requirements, reasonably allocate and execute tasks, and achieve a high degree of automation and intelligence in the R&D process, so as to significantly improve R&D efficiency and quality. Summary of the invention

[0003] In view of this, the present invention proposes an IT intelligent R&D method and system based on GPT, which can adapt to different application scenarios. The present invention provides the following technical solutions:

[0004] A GPT-based IT intelligent R&D method, the method comprising: constructing a task template, configuring the analysis task, blocking task stage, generating task stage, subtasks and task objectives of different scenario tasks in the R&D process through the task template; receiving demand data, matching the current task template, parsing the demand data through the analysis task to output the demand content; configuring the subtasks of the blocking task stage and the generating task stage respectively based on the demand content; executing the subtasks of the blocking task stage, and performing quality inspection on the output of the current subtask, if the quality inspection is qualified, entering the generating task stage, executing the subtasks of the generating task stage on the output of the subtask that has passed the quality inspection, and generating output content based on the task objective.

[0005] Optionally, the method of constructing a task template and configuring analysis tasks, task stages, subtasks and task objectives of tasks in different scenarios in the R&D process through the task template includes: identifying the current application scenario and constructing a task template based on the application scenario; based on the task template, setting the blocking task stage and the generating task stage in sequence, and configuring multiple subtasks for each task stage; associating an executable GPT unit for each subtask; and completing the configuration of the task objectives for each GPT unit through one or more of parameter control, input constraints and organization, session control, task planning and output definition control.

[0006] Optionally, the method for configuring the subtasks of the blocking task stage and the generating task stage respectively based on the requirement content includes: configuring the task goal of the blocking task stage to quality check the integrity, correctness and quality of the requirement content; configuring multiple subtasks of the blocking task stage, and configuring each subtask to be associated with a corresponding GPT unit, and the execution instructions of each GPT unit to achieve a judgment on the integrity, correctness and quality of the requirement content; configuring the task goal of the generating task stage to output a result that meets the requirement content; configuring multiple subtasks of the generating task stage, and configuring each subtask to be associated with a corresponding GPT unit, and the execution instructions of each GPT unit to achieve outputting a result that meets the requirement content.

[0007] Optionally, the method of executing the subtasks of the blocking task stage and performing quality inspection on the output of the current subtask, if the quality inspection is qualified, entering the generation task stage, executing the subtasks of the generation task stage on the output of the subtask that has passed the quality inspection, and generating output content based on the task goal includes: calling the GPT units corresponding to multiple subtasks in the blocking task stage to perform GPT quality inspection on the demand data according to the demand content and output the quality inspection results; if there is a subtask output that fails the quality inspection, triggering the modification task to the sending end of the demand data; receiving the modified demand data; calling the GPT unit corresponding to the subtask that fails the quality inspection to perform GPT quality inspection on the modified demand data according to the demand content and output the quality inspection results until the quality inspection is qualified; if the quality inspection is qualified, calling the GPT units corresponding to multiple subtasks in the generation task stage to process the demand data and output the target content according to the task goal.

[0008] Optionally, the method of constructing a task template and configuring analysis tasks, blocking task stages, generation task stages, subtasks and task objectives of tasks in different scenarios in the R&D process through the task template includes: constructing a task template based on the current application scenario, the task template having a dynamic input interface; inputting placeholders representing variable parameters in the dynamic input interface; selectively binding the placeholders with the analysis tasks, blocking task stages, generation task stages, subtasks and task objectives pre-configured in the task template to dynamically adjust the task template through the variable parameters at the dynamic input placeholders; dynamically configuring the analysis tasks, blocking task stages, generation task stages, subtasks and task objectives of tasks in different scenarios in the R&D process according to the dynamically changing task templates.

[0009] Optionally, the GPT unit is a custom processing unit to implement a task execution role and task execution instructions of a user-defined GPT unit.

[0010] The present invention further discloses an IT intelligent R&D system based on GPT, including: a template construction module, used to construct a task template, and configure the analysis task, blocking task stage, generation task stage, subtasks and task goals of different scenario tasks in the R&D process through the task template; a data processing module, used to receive demand data, match the current task template, and parse the demand data through the analysis task to output the demand content; a task configuration module, used to configure the subtasks of the blocking task stage and the generation task stage respectively based on the demand content; a task execution module, used to execute the subtasks of the blocking task stage, and perform quality inspection on the output of the current subtask. If the quality inspection is qualified, the generation task stage is entered, and the subtasks of the generation task stage are executed on the output of the subtask that has passed the quality inspection, and the output content is generated based on the task goal.

[0011] The present invention further discloses a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned GPT-based IT intelligent R&D method is implemented.

[0012] The present invention further discloses an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned GPT-based IT intelligent R&D method when executing the program.

[0013] The present invention further discloses a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, it implements the above-mentioned GPT-based IT intelligent R&D method.

[0014] This application constructs a multi-stage task processing framework based on task templates, reasonably configures analysis tasks, blocking task stages and generation task stages, and intelligently parses demand data to generate task content that adapts to the target; by introducing a flow mechanism for sub-task decomposition, quality inspection and generation tasks, the accuracy and quality of task execution are ensured; combined with flexible template configuration and optimized management of GPT units, efficient collaboration and automation of the R&D process are achieved, which greatly improves the efficiency of task execution and the reliability of results, and solves the problems of difficult demand analysis, loose task flow and poor configuration adaptability in traditional R&D processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] For the purpose of illustration and not limitation, the present invention is now described in conjunction with the embodiments of the present invention and the accompanying drawings, in which:

[0016] Figure 1 It is a flowchart of the IT intelligent R&D method based on GPT in an embodiment of the present invention;

[0017] Figure 2 It is a structural diagram of an IT intelligent R&D system based on GPT in an embodiment of the present invention;

[0018] Figure 3 is a schematic structural diagram of an electronic device in an embodiment of the present invention;

[0019] Figure 4 It is a schematic diagram of the task processing flow of the blocking task phase and the generating task phase in an implementation manner of the present invention. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the implementation mode of the present application will be clearly and completely described below in conjunction with the drawings in the implementation mode of the present application. Obviously, the described implementation mode is only a part of the implementation mode of the present application, not all the implementation modes. Based on the implementation mode in the present application, all other implementation modes obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present application.

[0021] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] In the present application, the terms "upper", "lower", "left", "right", "front", "back", "top", "bottom", "inner", "outer", "middle", "vertical", "horizontal", "lateral", "longitudinal" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings. These terms are mainly used to better describe the present application and its implementation methods, and are not used to limit the indicated devices, elements or components to have a specific orientation, or to be constructed and operated in a specific orientation.

[0023] In addition, some of the above terms may be used to express other meanings in addition to indicating orientation or positional relationship. For example, the term "on" may also be used to express a certain dependency or connection relationship in some cases. For those of ordinary skill in the art, the specific meanings of these terms in this application can be understood according to the specific circumstances. In addition, the term "plurality" should mean two or more than two.

[0024] It should be noted that, in the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other. The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0025] refer to Figure 1 This embodiment discloses a GPT-based IT intelligent R&D method, which includes:

[0026] S100: Build a task template, and configure the analysis tasks, blocking task stages, generation task stages, subtasks and task objectives of different scenario tasks in the R&D process through the task template.

[0027] S110: Identify the current application scenario and build a task template based on the application scenario;

[0028] S120: Based on the task template, a blocking task phase and a generating task phase are sequentially set, and each task phase is configured with a plurality of subtasks;

[0029] S130: Associating an executable GPT unit with each subtask;

[0030] S140: Complete the configuration of the task objectives for each GPT unit through one or more of parameter control, input constraints and organization, session control, task planning, and output definition control.

[0031] When constructing a template, a dynamic input interface is reserved for the template. A placeholder representing a variable parameter is input in the dynamic input interface. The placeholder is optionally bound to the analysis task, blocking task stage, generation task stage, subtask, and task target configured in the task template in advance, so as to dynamically adjust the task template through the variable parameters at the dynamic input placeholder, and dynamically configure the analysis task, blocking task stage, generation task stage, subtask, and task target of different scenario tasks in the R&D process according to the dynamically changing task template. Based on the dynamically adjusted task template, the system can automatically configure the analysis task, quality inspection subtask, and content generation task in the R&D process, avoiding the cumbersome and lagging task configuration in traditional R&D. Through parameterized template configuration and dynamic input, the system reuses the same template in multiple scenarios, greatly reducing the task configuration and development costs. The dynamic input placeholder ensures the consistency of the analysis task and subtask configuration in the task template, avoiding logical deviations or execution errors caused by parameter changes. The task target outputs data or content that meets the requirements according to the dynamically adjusted template, ensuring that the generated results are highly matched with the actual requirements.

[0032] For example, users can use placeholders in the input content to expand the content that needs to be replaced. For example, for code modification and adding attributes, the instructions configured in the system are as follows: Please add a new database field $<forEachJoin(list=fieldList,expre=item,separator=",")> Make code changes to add fields. Make sure that all changes that need to be made are correct. Add field annotations to the corresponding Pojo attributes, where the placeholder expression is: $<forEachJoin(list=fieldList,expre=item,separator=",")> , expand according to the expre expression, and connect each element in fieldList through . Use the form of $<> to organize and replace the placeholder. The placeholder supports calculation through the calculation expression, which is expressed as a regular expression: (1): / / ${varName}private static final Pattern placeholderPattern=Pattern.compile("\\$\\{(? <placeholder>[^}]+)\\}"); / / $<foreachJoin(list=list,expre=item_expre,seprator=",")>;(2):private static final Pattern functionPattern=Pattern.compile("\\$<(? <function>[^>]+)>"). (1) indicates variable replacement, and (2) indicates function calculation.

[0033] S200: Receive demand data, match it with the current task template, and parse the demand data through the analysis task to output the demand content.

[0034] In the IT development scenario, developers have many practical needs, such as parsing of requirement documents, code generation, and generation of test cases. When building a task template, identify the current development scenario to build the corresponding task template. For different needs, developers have different requirements for the final output of the task. Therefore, different task templates have different analysis tasks. For example, in the scenario of requirement document parsing, its analysis task is to perform content analysis on the current requirement document, perform logical defect detection on the requirement document in conjunction with the blocking task stage, and build task goals based on the current task template. Then, in the scenario of requirement document parsing, its task goal is to generate a requirement summary, challenging issues and defect issues faced, etc. The task goals here are exemplary and are not limited in this embodiment.

[0035] For another example, in the code generation scenario, the actual requirements are to add new database fields, adjust business logic, write API interface code, etc. When building a task template, it is necessary to identify the current code generation scenario and build the corresponding task template accordingly. Analysis tasks include analysis of existing code structure or requirement change content, such as parsing change records in the code repository or new field information in the requirement document. Check the integrity of code logic and dependencies between modules. Cooperate with the blocking task phase to perform quality inspection subtasks, such as checking whether code changes comply with established coding standards (such as variable naming rules and comment integrity). Verify the logical correctness, modularity and security of the code. Based on the current task template, define the task objectives in the code generation scenario, such as automatically generating code snippets that meet the requirements (such as database operation code, API interface code). Provide a quality inspection report for code changes to ensure that the generated code meets business requirements and technical specifications. Similarly, in the test case generation scenario, testers need to generate corresponding test cases based on the requirement document or code changes to ensure the correctness and stability of system functions. When building a task template, it is necessary to identify the current test scenario to design a suitable task template. The analysis task includes content analysis of the requirement document or code changes, for example: extracting the core functional points and boundary conditions in the requirement document. Identify the test logic and input parameters that need to be covered from the code changes. Cooperate with the blocking task phase to perform quality inspection subtasks, for example: check whether the extracted test content completely covers the requirements or code logic. Verify whether the generated test cases meet the test specifications (such as clear use case steps and clear expected results). Based on the current task template, define the task objectives in the test case generation scenario, for example: automatically generate a complete set of test cases covering functional requirements and boundary conditions. Provide logical check results for test cases and generate high-quality test documents. The task objectives here are exemplary and are not limited to this embodiment.

[0036] S300: Based on the requirements, configure the subtasks of the blocking task phase and the generating task phase respectively. Figure 4 , specifically:

[0037] The task goal of the blocking task stage is to check the completeness, correctness and quality of the required content; multiple subtasks are configured in the blocking task stage, and each subtask is configured to be associated with the corresponding GPT unit, and the execution instructions of each GPT unit, so as to realize the judgment of the completeness, correctness and quality of the required content; the task goal of the generation task stage is to output the result that meets the required content; multiple subtasks are configured in the generation task stage, and each subtask is configured to be associated with the corresponding GPT unit, and the execution instructions of each GPT unit, so as to realize the output result that meets the required content.

[0038] For the configuration of subtasks, the quality inspection target needs to be decomposed into multiple subtasks according to the complexity and characteristics of the demand content. For the example in step S200, the subtasks can be configured as follows: Completeness check subtask: Verify whether the demand content covers all business requirements and boundary conditions. Logical correctness check subtask: Check whether there are contradictions or errors in the logical description of the demand content. Format and consistency check subtask: Ensure that the format of the demand document is standardized and the content is clear and consistent.

[0039] After the subtask configuration is completed, configure the corresponding GPT unit for each subtask, set its task role and execution instructions: Task role: for example, "Requirement Reviewer". Execution instructions: for example, "Please check whether the requirement document covers all functional points and list the missing requirements". For another example, in the code generation scenario example of step S200, the task role can be set to the code annotation role as software development architect, the code review role as software development architect, and the code scoring role as software development architect, and the corresponding execution instructions are set: summarize the code changes, not more than 30 Chinese characters, conduct code review on the code changes, and refer to Google's code review specifications and standards for code review. Score the code changes. The scoring criteria refer to the best practices in the software field. They are code readability, code reusability, modularity, performance optimization, documentation and comments, naming conventions, and code security. The scoring adopts a 10-point system and retains two decimal places.

[0040] S400: Execute the subtasks of the blocking task phase and perform quality inspection on the output of the current subtask. If the quality inspection is qualified, enter the generation task phase, execute the subtasks of the generation task phase on the output of the subtask that has passed the quality inspection, and generate output content based on the task goal. Figure 4 , specifically:

[0041] Call the GPT units corresponding to the multiple subtasks in the blocking task stage to perform GPT quality inspection on the demand data according to the demand content and output the quality inspection results; if there is a subtask output that fails the quality inspection, trigger the modification task to the sending end of the demand data; receive the modified demand data; call the GPT units corresponding to the subtasks that failed the quality inspection to perform GPT quality inspection on the modified demand data according to the demand content and output the quality inspection results until the quality inspection is qualified; if the quality inspection is qualified, call the GPT units corresponding to the multiple subtasks in the generation task stage to process the demand data and output the target content according to the task objectives.

[0042] Through the configured blocking task stage and generating task stage, intelligent quality inspection of demand data and efficient generation of target content are realized. The blocking task stage calls the GPT units corresponding to multiple subtasks to perform quality inspection on demand data item by item to ensure the integrity, correctness and quality of the demand. By automatically triggering the modification task, a closed-loop process of quality inspection-modification-re-quality inspection is formed to effectively improve the accuracy and consistency of demand data. Only after the quality inspection is qualified, the generation task stage is entered to avoid erroneous requirements from flowing into the subsequent task stage and reduce the risk of rework. The subtasks in the generation task stage can be directly executed based on the demand data that has passed the quality inspection, reducing human intervention and improving the efficiency of task execution. The GPT unit in the generation task stage is called to generate content that meets the task objectives (such as demand summary, test case or code) according to the demand data, and ensure that the output content meets business requirements and technical standards. At the same time, it supports dynamic revision and re-quality inspection of unqualified requirements to ensure that the system can quickly adapt to complex or frequently changing demand scenarios. Through the automated quality inspection and generation process, errors caused by manual operations are reduced, and the workload of manual review and modification is reduced.

[0043] This implementation method gives an example to explain the above scheme. For example, after the product manager writes the demand document as demand data, the configured task template is called to run the analysis task and output the demand content. Based on the demand content, the subtasks of the blocking task stage and the generating task stage are configured respectively, the subtasks of the blocking task stage are executed, and the output of the current subtask is quality inspected, and each item of the product is determined to pass the quality inspection according to the quality inspection results. If the quality inspection is not passed, the modification task will be triggered to the corresponding product manager, and the product manager will repair the product demand defects according to the instructions. After the repair, the system automatically triggers the quality inspection task, and the cycle is repeated until the quality inspection passes. The tester will complete the test case corresponding to the demand according to the task assistance, which greatly reduces the work of manually writing test cases. The test only needs to review the generated results and synchronize the accurate content to the work platform. Developers use this task template to complete time-consuming, labor-intensive and error-prone code modification tasks, such as code corrections brought by adding fields. It can also help developers review each code change and give review opinions. For another example, in the iterative maintenance work of the project, operations that need to add data fields are often encountered. This operation usually requires mechanical modification of a large amount of code, which is prone to errors. At this time, when using the GPT unit to complete the code change, the original input will first be submitted to the GPT unit in the form of input and instructions, and the purpose and output format will be described in the instructions. Because the output of the GPT unit is limited, there is usually a lot of original code, and it is unrealistic and error-prone to output the modified complete content. On this basis, the system greatly reduces the amount of code generation by executing the output of custom diff, and applies the diff content to the original content to make changes, which can both control the scale of code changes and improve accuracy.

[0044] In summary, by constructing a multi-stage task processing framework based on task templates, rationally configuring analysis tasks, blocking task stages and generation task stages, and intelligently parsing the demand data, task content that adapts to the target is generated; by introducing the flow mechanism of sub-task decomposition, quality inspection and generation tasks, the accuracy and quality of task execution are ensured; combined with flexible template configuration and optimized management of GPT units, efficient collaboration and automation of the R&D process are achieved, which greatly improves the efficiency of task execution and the reliability of results, and solves the problems of difficult demand analysis, loose task flow and poor configuration adaptability in the traditional R&D process.

[0045] refer to Figure 2 This embodiment further discloses an IT intelligent R&D system based on GPT, including:

[0046] The template construction module 21 is used to construct a task template, and configure the analysis tasks, blocking task stages, generation task stages, subtasks and task objectives of tasks in different scenarios in the R&D process through the task template; it is also used to identify the current application scenario and build a task template based on the application scenario; based on the task template, the blocking task stage and the generation task stage are set in sequence, and each task stage corresponds to multiple subtasks; an executable GPT unit is associated with each subtask; for each GPT unit, the configuration of the task objective is completed through one or more methods of parameter control, input constraints and organization, session control, task planning and output definition control; it is also used to build a task template based on the current application scenario, and the task template has a dynamic input interface; a placeholder representing a variable parameter is input in the dynamic input interface; the placeholder is selectively bound to the analysis task, blocking task stage, generation task stage, subtask and task objective configured in advance in the task template, so as to realize dynamic adjustment of the task template through the variable parameters at the dynamic input placeholder; the analysis tasks, blocking task stage, generation task stage, subtasks and task objectives of tasks in different scenarios in the R&D process are dynamically configured according to the dynamically changing task template.

[0047] The data processing module 22 is used to receive demand data, match the current task template, parse the demand data by analyzing the task, and output the demand content.

[0048] The task configuration module 23 is used to configure the subtasks of the blocking task stage and the generating task stage respectively based on the requirement content; it is also used to configure the task goal of the blocking task stage to quality check the integrity, correctness and quality of the requirement content; configure multiple subtasks of the blocking task stage, and configure each subtask to be associated with the corresponding GPT unit, and the execution instructions of each GPT unit to achieve the judgment of the integrity, correctness and quality of the requirement content; configure the task goal of the generating task stage to output the result that meets the requirement content; configure multiple subtasks of the generating task stage, and configure each subtask to be associated with the corresponding GPT unit, and the execution instructions of each GPT unit to achieve the output of the result that meets the requirement content.

[0049] The task execution module 24 is used to execute the subtasks of the blocking task stage and perform quality inspection on the output of the current subtask. If the quality inspection is qualified, it enters the task generation stage, executes the subtasks of the task generation stage on the output of the subtask that has passed the quality inspection, and generates output content based on the task objectives; it is also used to call the GPT units corresponding to multiple subtasks in the blocking task stage to perform GPT quality inspection on the demand data according to the demand content and output the quality inspection results; if there is a subtask output that fails the quality inspection, it triggers the modification task to the sending end of the demand data; receives the modified demand data; calls the GPT unit corresponding to the subtask that fails the quality inspection to perform GPT quality inspection on the modified demand data according to the demand content and outputs the quality inspection results until the quality inspection is qualified; if the quality inspection is qualified, calls the GPT units corresponding to multiple subtasks in the task generation stage to process the demand data and output the target content according to the task objectives.

[0050] Figure 3 A schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, such as Figure 3 As shown, the electronic device 50 includes: a processor 501 (processor), a memory 502 (memory) and a bus 503;

[0051] The processor 501 and the memory 502 communicate with each other via the bus 503 ; the processor 501 is used to call program instructions in the memory 502 to execute the methods provided by the above-mentioned method implementation methods.

[0052] This embodiment provides a non-transitory computer-readable storage medium, which stores computer instructions. The computer instructions enable a computer to execute the methods provided by the above-mentioned method embodiments.

[0053] A person skilled in the art can understand that all or part of the steps for implementing the above-mentioned method implementation method can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium, which, when executed, executes the steps of the above-mentioned method implementation method; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various storage media that can store program codes.

[0054] The device implementation described above is merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present implementation scheme. Those of ordinary skill in the art may understand and implement it without creative effort.

[0055] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each implementation method or some parts of the implementation method.

[0056] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.< / function> < / placeholder>

Claims

1. A GPT-based IT intelligent R&D method, characterized in that: The method comprises: Build task templates, and use them to configure analysis tasks, blocking task stages, generation task stages, subtasks, and task objectives for different scenario tasks in the R&D process; Receive demand data, match it with the current task template, parse the demand data through analysis tasks, and output the demand content; Configure the subtasks of the blocking task phase and the generating task phase respectively based on the requirements; Execute the subtasks of the blocking task stage and perform quality inspection on the output of the current subtask. If the quality inspection is qualified, enter the generation task stage, execute the subtasks of the generation task stage on the output of the subtask that has passed the quality inspection, and generate output content based on the task objectives.

2. The GPT-based IT intelligent R&D method according to claim 1, characterized in that: The method of constructing a task template and configuring the analysis tasks, task stages, subtasks and task objectives of different scenario tasks in the R&D process through the task template includes: Identify the current application scenario and build a task template based on the application scenario; Based on the task template, set the blocking task stage and the generating task stage in sequence, and configure multiple subtasks for each task stage; Associate an executable GPT unit with each subtask; For each GPT unit, the task objectives are configured through one or more of parameter control, input constraints and organization, session control, task planning, and output definition control.

3. The GPT-based IT intelligent R&D method according to claim 1, characterized in that: The method for respectively configuring the subtasks of the blocking task phase and the generating task phase based on the demand content includes: The task goal of the configuration blocking task phase is to check the completeness, correctness and quality of the content of the quality inspection requirements; Configure multiple subtasks in the blocking task phase, and configure the corresponding GPT unit associated with each subtask and the execution instructions of each GPT unit to judge the completeness, correctness and quality of the required content; The task goal of the configuration generation task stage is to output results that meet the requirements; Configure multiple subtasks in the generation task phase, and configure each subtask to associate with the corresponding GPT unit and the execution instructions of each GPT unit to output results that meet the required content.

4. The GPT-based IT intelligent R&D method according to claim 1, characterized in that: The method of executing the subtasks of the blocking task stage and performing quality inspection on the output of the current subtask, and if the quality inspection is qualified, entering the generation task stage, executing the subtasks of the generation task stage on the output of the subtask that has passed the quality inspection, and generating output content based on the task target includes: Call the GPT units corresponding to the multiple subtasks in the blocking task phase to perform GPT quality inspection on the required data according to the required content and output the quality inspection results; If there is a subtask output that fails the quality inspection, the modification task is triggered to the sender of the demand data; receiving modified demand data; Call the GPT unit corresponding to the subtask that failed the quality inspection to perform GPT quality inspection on the modified demand data according to the demand content and output the quality inspection results until the quality inspection is qualified; If the quality inspection is qualified, the GPT units corresponding to multiple subtasks in the generation task phase are called to process the required data and output the target content according to the task objectives.

5. The GPT-based IT intelligent R&D method according to claim 1, characterized in that: The method of constructing a task template and configuring the analysis tasks, blocking task stages, generating task stages, subtasks and task objectives of different scenario tasks in the R&D process through the task template includes: Build a task template based on the current application scenario, and the task template has a dynamic input interface; Enter the placeholders representing the variable parameters in the dynamic input interface; The placeholder is selectively bound to the analysis task, blocking task phase, generating task phase, subtask and task target configured in advance in the task template, so as to dynamically adjust the task template by dynamically inputting variable parameters at the placeholder; Dynamically configure the analysis tasks, blocking task stages, generation task stages, subtasks and task objectives of different scenario tasks in the R&D process according to dynamically changing task templates.

6. The GPT-based IT intelligent R&D method according to claim 2, characterized in that: The GPT unit is a custom processing unit to implement the task execution role and task execution instructions of the user-defined GPT unit.

7. An IT intelligent R&D system based on GPT, characterized in that: include: The template construction module is used to build task templates. Through the task templates, the analysis tasks, blocking task stages, generation task stages, subtasks and task objectives of different scenario tasks in the R&D process are configured; The data processing module is used to receive the demand data, match the current task template, parse the demand data through the analysis task, and output the demand content; The task configuration module is used to configure the subtasks of the blocking task phase and the generating task phase respectively based on the required content; The task execution module is used to execute the subtasks of the blocking task stage and perform quality inspection on the output of the current subtask. If the quality inspection is qualified, it will enter the generation task stage, execute the subtasks of the generation task stage on the output of the subtask that has passed the quality inspection, and generate output content based on the task objectives.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images