CN-121996206-A - Automatic task flow generating method and device, computer equipment and storage medium
Abstract
The invention relates to the technical field of multi-mode collaboration, and discloses a task flow automatic generation method, a device, computer equipment and a storage medium, wherein the method comprises the steps of obtaining multi-mode input information; extracting semantic elements of multi-mode input information, inputting the semantic elements into a preset story template, injecting domain knowledge into the story template to output a story frame, carrying out structural processing, compliance and integrity verification on the story frame, inputting a qualified structural user story into a project management system to generate an end user story, and updating a knowledge base based on the end user story to form an executable automatic task flow. By means of structuring, knowledge enhancement and compliance integrity verification and combining an automatic access program and knowledge base support based on object storage, the problems of low efficiency and insufficient automation of traditional manual writing are solved, story compliance and professionality of users are guaranteed, and convenient development requirements in the financial and medical fields are met.
Inventors
- LI QIAN
Assignees
- 平安科技(深圳)有限公司
Dates
- Publication Date
- 20260508
- Application Date
- 20260202
Claims (10)
- 1. The automatic task flow generating method is characterized by comprising the following steps: acquiring multi-mode input information; Extracting semantic elements of the multi-mode input information, inputting the semantic elements into a preset story template, and injecting domain knowledge into the story template to output a story frame, wherein the domain knowledge is stored in a knowledge base and comprises pre-acquired industry terms and compliance requirements; Carrying out structuring processing on the story frame to generate a structured user story; Compliance and integrity verification of the structured user story; Triggering a preset automatic access program based on the qualified structured user stories, and inputting the qualified structured user stories into a project management system to generate final user stories; based on the end user stories, the knowledge base is updated to form an executable automated task flow.
- 2. The task flow automatic generation method according to claim 1, wherein the extracting the semantic elements of the multimodal input information, inputting the semantic elements into a preset story template, and injecting domain knowledge into the story template to output a story frame, comprises: analyzing natural language requirements of the input information by adopting a semantic understanding model, and extracting core semantic elements; Inputting the core semantic elements into a preset story template, wherein the story template generates a data structure of a user story, and the data structure comprises fields, field types and whether filling is needed or not; Calling a domain knowledge enhancement model through a model collaboration platform, enabling the domain knowledge enhancement model to inject industry terms and compliance requirements into the story frame according to fields based on the knowledge base, and obtaining a completed story frame; based on the completed story frame, a logic association network between the core semantic elements is established, and a standardized story frame is generated and output.
- 3. The method for automatically generating a task flow according to claim 2, wherein the structuring the story frame to generate a structured user story comprises: performing field type and format limitation on the standardized story frame by adopting a structured output plug-in; Outputting a standardized story frame after limiting as the structured user story.
- 4. The method for automatically generating a task flow according to claim 2, wherein said verifying compliance and integrity of the structured user story comprises: Invoking a classification model through the model collaboration platform, and checking whether acceptance criteria of the formatted user story cover core business logic; invoking an anomaly detection model through the model collaboration platform, and identifying a compliance risk or a logic vulnerability in the formatted user story; If the two checks of the classification model and the anomaly detection model are not passed, outputting a correction suggestion to regenerate a user story; And if the two checks of the classification model and the anomaly detection model are passed, outputting a structured user story with qualified check.
- 5. The method for automatically generating a task flow according to claim 4, wherein the triggering a preset automatic access program based on the qualified structured user story, inputting the qualified structured user story into a project management system to generate an end user story, comprises: Acquiring a verification result of the structured user story, and acquiring an access certificate of the project management system through an identity authentication system of the project management system if the verification is passed; Calling an API interface, and sending the qualified structured user story to a project management system based on the access certificate; After the project management system passes the verification of the access certificate, completing the automatic login of the project management system and generating an end user story; wherein, the calling API interface, based on the access credential, sends the qualified structured user story to a project management system, further comprising: if the API interface call fails, a fault-tolerant switching mechanism is entered, and the qualified structured user story is sent to a project management system through the fault-tolerant switching mechanism.
- 6. The automatic task flow generating method according to claim 5, wherein if the API call fails, entering a fault tolerant switching mechanism, and sending the qualified structured user story to a project management system through the fault tolerant switching mechanism includes: triggering an interface automation operation script of the project management system; Based on the interface automation operation script, switching to an interface automation operation flow, and executing visual operation of user story input so as to simulate manual operation of a user; calling a visual identification plug-in, and checking whether a visual operation result is accurate; If the visual operation result is accurately checked, sending the checked qualified structured user story to a project management system; And if the visual operation result is not accurately checked, re-executing the fault-tolerant switching mechanism until the visual operation result is accurately checked.
- 7. The automated task flow generation method of claim 6, wherein the updating the knowledge base based on the end user story forms an executable automated task flow comprising: Storing historical user stories, industry terms and compliance requirement data into the knowledge base in a classified manner through an object storage compatible protocol configured by the project management system so as to realize mixed storage of structured and unstructured data; automatically acquiring historical data of the project management system by utilizing an interface automation operation flow, performing incremental update on a knowledge base, and constructing a knowledge graph; Controlling access rights of the knowledge base based on a rights management system configured by the project management system, and managing and backtracking multiple versions of the knowledge base content through a version control function of object storage; and forming an executable automatic task flow based on the knowledge base, and outputting a traceable execution log.
- 8. An automatic task flow generating device, comprising: The acquisition module is used for acquiring multi-mode input information; The system comprises an extraction module, a storage module and a storage module, wherein the extraction module is used for extracting semantic elements of the multi-mode input information, inputting the semantic elements into a preset story template, and injecting domain knowledge into the story template to output a story frame, wherein the domain knowledge is stored in a knowledge base and comprises pre-acquired industry terms and compliance requirements; the structuring module is used for carrying out structuring processing on the story frame to generate a structured user story; The verification module is used for verifying the compliance and the integrity of the structured user story; The generation module is used for triggering a preset automatic access program based on the qualified structured user stories, and inputting the qualified structured user stories into a project management system to generate final user stories; And the management module is used for updating the knowledge base based on the final user story to form an executable automatic task flow.
- 9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the task flow automatic generation method according to any one of claims 1 to 7 when the computer program is executed by the processor.
- 10. A computer-readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the task flow automatic generation method according to any one of claims 1 to 7.
Description
Automatic task flow generating method and device, computer equipment and storage medium Technical Field The present invention relates to the field of multimodal collaboration technologies, and in particular, to a method and apparatus for automatically generating a task flow, a computer device, and a storage medium. Background In the software development process, a user story is used as a core carrier for agile development, is used for accurately describing user requirements, functional scenes and service values, is a key bridge for connecting product managers, service parties and development teams, and the quality and the circulation efficiency of the key bridge directly determine the progress of software development and the suitability of a final product. The creation and management process of the traditional user stories mainly relies on manual operation, namely, the creation and management process is manually written by a product manager or business personnel according to experience, and is transmitted to a development team in an unstructured mode such as mail, documents and the like, and then is manually input into a project management system. However, the traditional scheme has a plurality of problems to be solved, namely firstly, the writing of user stories requires deep participation of business personnel, a great deal of time and cost are consumed, the demand description is fuzzy and core semantic elements are lost easily due to personal experience differences and expression habits, understanding deviation is brought to subsequent development work, secondly, the user stories output by different writers are obvious in structure, field and expression modes, the automatic analysis and circulation demands are difficult to adapt, the cooperation efficiency of development teams is seriously influenced, thirdly, the existing tool only supports the storage and display functions of the user stories, the end-to-end automation from demand input, structural generation and compliance verification to system interaction cannot be realized, the number of manual intervention links is too large, and the iteration speed of agile development is restricted. Therefore, the requirements of automatic generation and management of user stories in complex business scenes cannot be met, and the method is particularly applicable to fields with extremely high requirements on compliance, professionals and safety, such as finance, medical treatment and the like. For example, in the medical field, medical services have extremely strong specialization, involve professional knowledge such as diagnosis and treatment specifications, instrument operation standards, medical insurance policies and the like, the manually-written user stories are easy to have problems of wrong professional terms, core business logic deletion and the like, the developed medical system possibly does not meet clinical use requirements, in addition, the integrity requirement of the medical field on the user stories is extremely high, any requirement omission can influence the safety and continuity of medical procedures, the prior art lacks a specific professional knowledge enhancement and accurate verification mechanism, the generated requirements are difficult to ensure the speciality and the integrity of the user stories, and the problem of low requirement transmission efficiency is more prominent when the medical information system is developed in cooperation across hospitals and departments, so that the landing speed of the medical information system is restricted. Therefore, there is a need for a method, a device, a computer device, and a storage medium for automatically generating a full-process task flow, which is compliant and highly specialized. Disclosure of Invention The invention provides a task flow automatic generation method, a device, computer equipment and a storage medium, which are used for solving the technical problems of strong manual dependence, low automation degree and insufficient compliance and professionality in the existing user story generation and management technology. In a first aspect, a method for automatically generating a task flow is provided, including: acquiring multi-mode input information; Extracting semantic elements of the multi-mode input information, inputting the semantic elements into a preset story template, and injecting domain knowledge into the story template to output a story frame, wherein the domain knowledge is stored in a knowledge base and comprises pre-acquired industry terms and compliance requirements; Carrying out structuring processing on the story frame to generate a structured user story; Compliance and integrity verification of the structured user story; Triggering a preset automatic access program based on the qualified structured user stories, and inputting the qualified structured user stories into a project management system to generate final user stories; based on the end user stories, the k