JIRA demand data automatic review method and device based on dify platform and medium

Abstract

The invention relates to the technical field of computers, in particular to an automatic JIRA demand data review method and device based on a dify platform and a medium. The method comprises the steps that in the development process, a JMWE trigger plug-in is added to a node triggered by a workflow state of an issue where a demand is located; when the workflow state of the issue reaches a verification node, judging whether a null value exists in a self-defined field or not through a JMWE verifier plug-in, and writing a judgment result in the issue; then, the workflow state of the issue reaches a trigger node, if a null value exists in a self-defined field, the trigger plug-in is not triggered, and if a null value does not exist in a related field, the trigger plug-in calls an API of a dify platform to trigger demand review; on the dify platform, predefining an automatic demand review workflow, the review workflow comprising at least one structured cue word; the structured cue word is used for defining a key information field set, a review dimension, a review standard and a review result output format required by a review demand; and the review workflow generates a corresponding application program interface (API) endpoint through the dify platform.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and specifically to a method, apparatus, and medium for automatic review of JIRA requirements data based on the dify platform. Background Technology

[0002] In software development, requirements review is a crucial step in ensuring requirements quality and reducing rework later. Although JIRA, as a mainstream agile project management tool, supports Scrum and Kanban methodologies and tracks task lifecycles through workflow statuses (such as To Do and In Progress), its native functionality still has significant limitations in requirements review scenarios. Traditional manual reviews rely on JIRA's field entry and status transition mechanisms, requiring manual triggering of the review process, leading to inefficiency and a high risk of omissions. While JIRA provides basic automation features (such as triggering field changes and validators to check required fields) and supports plugins like JMWE to extend workflows (e.g., field non-empty validation or subtask status linkage), these capabilities are still limited to simple rule-driven logic and cannot achieve intelligent requirements analysis. Summary of the Invention

[0003] The purpose of this invention is to provide a method, apparatus, and medium for automatic review of JIRA requirements data based on the dify platform, aiming to solve the problems of low efficiency and insufficient intelligence in the existing requirements review process, and to realize the automation of the requirements review process.

[0004] The technical solution of the present invention:

[0005] An automated review method for JIRA requirements data based on the Dify platform includes:

[0006] During development, a JMWE trigger plugin is added to the node where the workflow status of the issue containing the requirement is triggered. When the workflow status of the issue reaches the verification node, the JMWE validator plugin checks whether a custom field has a null value and writes the judgment result to the issue. Subsequently, when the workflow status of the issue reaches the trigger node, if the custom field has a null value, the trigger plugin is not triggered; if the relevant field does not have a null value, the trigger plugin calls the API of the Dify platform to trigger the requirement review. On the Dify platform, an automated requirement review workflow is predefined, which includes at least one structured prompt. The structured prompt is used to define the set of key information fields required for reviewing the requirement, the review dimensions, the review criteria, and the review result output format. The review workflow generates the corresponding application programming interface (API) endpoint through the Dify platform.

[0007] Furthermore, when the workflow status of the issue reaches the verification node, the JMWE validator plugin determines whether a custom field has a null value, and then writes the determination result into the issue; when the workflow status of the issue reaches the trigger node, if the custom field has a null value, the trigger plugin is not triggered; if the custom field does not have a null value, the trigger plugin calls the dify API to trigger the requirement review, specifically including:

[0008] The validator plugin checks whether there are empty values ​​in the information fields of the custom fields required in the issue. If there are empty values, it writes "Yes" after the record field in the custom field; if there are no empty values, it writes "No" after the record field in the custom field.

[0009] Furthermore, the trigger plugin determines the value of the record field. If the value is yes, the requirement review is not triggered; if the value is no, the trigger plugin calls the Dify API to review the requirements under that issue. When calling, the trigger plugin takes the current value of the information field set in the task item as the input parameter, passes it to the Dify platform, and returns the result to the corresponding field.

[0010] Furthermore, the dify platform's response to receiving the API endpoint's call request specifically includes:

[0011] Execute a predefined review workflow, populate the received requirement data into the structured prompts; utilize an integrated large language model (LLM) to perform requirement review analysis based on the populated structured prompts; generate structured review results conforming to a predetermined output format; and return the structured review results via the API response.

[0012] After receiving the review results returned by the dify platform, the trigger plugin parses the review results and automatically writes the review results into the predefined result field of the JIRA requirement task item.

[0013] Furthermore, the method automatically triggers the Dify platform when the value of the predefined verification status field of the requirement task item is not empty. In response to the trigger, the Dify platform, based on the received values ​​of the key information field set, drives the integrated large-scale language model to automatically review and analyze the requirement content under the requirement task item by executing its predefined structured prompts and review workflow. The Dify platform generates a structured review result conforming to a predetermined output format and returns the result. The review result includes review comments for the predefined review items and is automatically written into the predefined result field of the JIRA requirement task item.

[0014] An apparatus for reviewing JIRA requirement data based on the dify platform, applied to the automatic review method, includes: a dify workflow module, a JIRA verification module, and a JIRA triggering module.

[0015] The `dify` workflow module is used to handle requirement reviews. As the core engine of the review process, it is pre-configured with structured prompts and an automated review workflow. It receives review requests through its generated application programming interface (API) endpoint. Upon receiving a request, the `dify` workflow module executes the pre-configured automated review workflow, filling the structured prompts with the requirement-related information contained in the request. This drives the integrated Large Language Model (LLM), which analyzes and reviews the requirement content item by item based on the filled prompts and predefined review items. The module generates a structured review result that conforms to a predetermined format and returns it through the API endpoint.

[0016] The JIRA verification module includes a JMWE verifier plugin instance configured in the JIRA target requirement type workflow verification status node. This verifier is used to check whether all key information fields required by the structured prompts of the dify workflow module in the current requirement task item are non-empty. If there are empty fields, a first identifier indicating incomplete data is written into the predefined dify review ready field of the task item. If all information fields and record fields are non-empty, a second identifier indicating complete data is written into the predefined dify review ready field.

[0017] The JIRA triggering module includes a JMWE trigger plugin instance configured in the JIRA target requirement type workflow trigger status node. It is configured to perform a trigger action in response to the value of the predefined dify review ready field being the second identifier: automatically calling the API endpoint provided by the dify workflow module, passing the information field and record field values ​​of the task item as input parameters to the dify platform; and writing the review result content into the predefined review result field of the task item after receiving the structured review result returned by the dify workflow module.

[0018] Furthermore, the device automatically reviews the requirements after the issue is constructed, based on relevant requirement information, specifically including:

[0019] When the relevant information about the requirement is complete, the review results of the requirement will be automatically written to the corresponding field in JIRA.

[0020] A computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method.

[0021] The beneficial effects of this invention are:

[0022] This invention proposes an automated JIRA requirements data review method, apparatus, and medium based on the Dify platform. It organically combines JIRA's native validator / trigger mechanism with the Dify platform's AI workflow engine. The verification module ensures the integrity of the input data, the triggering module controls the intelligent start and stop of the review process, and the Dify workflow module provides professional AI review capabilities. These three components work together to form a complete technical chain of "verification-triggering-review-feedback." This design not only solves the problems of traditional requirements review relying on manual processes and being inefficient, but also improves the professionalism and consistency of the review through the natural language processing capabilities of a large language model, providing an innovative solution for requirements management in agile development. Attached Figure Description

[0023] Figure 1 This is a flowchart of the method in an embodiment of the present invention;

[0024] Figure 2 This is a schematic diagram of the improved issue workflow in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the device in an embodiment of the present invention. Detailed Implementation

[0026] The following description of embodiments provides a more detailed explanation of the specific implementation of the present invention, including the shape and structure of each component, the relative positions and connections between the parts, the function and working principle of each part, the manufacturing process, and the operation and use methods, in order to help those skilled in the art to have a more complete, accurate, and in-depth understanding of the concept and technical solution of the present invention.

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The automatic review method for JIRA requirement data based on the Dify platform of this invention, through the JMWE plugin and the JIRA system, combined with the workflow orchestration capabilities and intelligent analysis capabilities of the Dify platform's large language model, achieves automated review of requirement data. The specific implementation methods of this invention are described in detail below.

[0028] The first embodiment of the present invention provides an automatic review method for JIRA requirements data based on the Dify platform, characterized by:

[0029] Automated review control is achieved by deploying a validator plugin at the workflow verification status node of the requirement task item (issue) and a trigger plugin at the trigger status node. When the requirement task item flows to the verification node, the validator plugin automatically checks whether there are empty values ​​in the key information fields related to the requirement review in the task item, and writes the detection results to preset fields: if there are empty values, it is marked as "yes", and if all fields are not empty, it is marked as "no".

[0030] When a task item enters the workflow trigger node, the trigger plugin performs a differentiated operation based on the flag value: if the flag is "Yes" (empty value exists), the subsequent process is not triggered; if the flag is "No" (no non-empty field), the pre-configured Dify platform API endpoint is automatically called. During the call, the plugin passes the current value of the key information fields within the task item as input parameters to the Dify platform. The Dify platform has a pre-defined automated review workflow, which includes structured prompts that clearly define the set of key fields, review dimensions, review criteria, and structured output format required for the review. After responding to the API request, the platform dynamically populates the received requirement data into the structured prompts, drives the integrated Large Language Model (LLM) to perform multi-dimensional analysis of the requirement content, generates a structured review result conforming to the predetermined format, and returns it through the API response.

[0031] The dify platform then returns the structured review results to the JIRA system via API. JIRA automatically updates the corresponding fields in the issue with the review results and adjusts the issue status and workflow progress accordingly. This technical solution, through deep integration of the JMWE plugin with the dify platform, constructs a complete automated closed loop of "verification-triggering-review-feedback," achieving an intelligent upgrade of the requirements review process. Compared to traditional manual review methods, it offers significant advantages such as high review efficiency, standardized procedures, and strong traceability.

[0032] Secondly, this invention provides an automated JIRA requirements data review device based on the Dify platform. This device integrates AI capabilities with JIRA workflow automation technology to achieve intelligent management of the entire requirements review process. The core of the device consists of three parts: a Dify workflow module, a JIRA verification module, and a JIRA triggering module, forming a closed-loop system from data verification to intelligent review.

[0033] The dify workflow module, as the core processing unit of the device, is responsible for building and managing the complete process of requirements review. This module constrains review criteria through preset review prompt templates, invokes a large language model to perform multi-dimensional analysis of the requirements, including review items such as logical completeness and technical feasibility, and finally returns the structured review results in JSON format. The module supports deep data interaction with the JIRA system, ensuring that review results can be accurately mapped to specified fields.

[0034] JIRA's validation module automates data verification during the workflow status transitions of requirement issues by integrating a validator plugin. When an issue reaches a validation node, the module automatically checks the completeness of key fields such as "Requirement Description" and "Requirement Type," and updates the "Field Contains Null Values" status field with the validation result as a Boolean value. This design adheres to the basic specifications of JIRA's workflow validation mechanism while providing decision-making support for subsequent processes through status flags.

[0035] The JIRA triggering module implements intelligent process control based on the aforementioned trigger plugin, dynamically triggering the review process according to the status flags of the verification module. When the "Field contains null values" flag is marked "No", the module calls the review service of the dify platform via API, submitting complete requirement data to the language big model for analysis; if marked "Yes", the process terminates and prompts for data completion. After the review is completed, the module automatically writes back the review conclusions generated by the big model to the corresponding fields in JIRA, including structured data such as compliance assessment and risk level, achieving seamless integration of review results with the project management system.

[0036] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0037] A second embodiment of the present invention provides an automatic review method for JIRA requirements data based on the Dify platform, such as... Figure 1 As shown, it includes the following steps:

[0038] Step 101: Configure a workflow for the requirement issue in the JIRA system, add a validator plugin to the workflow status verification node, and add a trigger plugin to the workflow status trigger node.

[0039] Step 102: When the issue workflow status reaches the verification node, the verifier plugin automatically checks the completeness of preset fields, including but not limited to key fields such as "requirement description", "requirement type" and "whether it is derived".

[0040] Step 103: Based on the verification results, update the "Field contains null values" status field in the issue. If null values ​​exist, mark it as "Yes"; otherwise, mark it as "No".

[0041] Step 104: When the workflow status reaches the trigger node, the trigger plugin determines whether to trigger the review based on the value of the "field contains null values" field. If it is "yes", the process is terminated; if it is "no", the requirement review service of the dify platform is called through the API.

[0042] Step 105: After receiving the request, the Dify platform calls the Deepseek model to intelligently review the requirements based on preset prompt constraints and workflow configuration. The review results are returned to the JIRA system in a structured data format and automatically updated to the corresponding fields of the issue.

[0043] Configure a workflow for the target project in the JIRA system. This workflow should include at least three status nodes: "Pending Review," "Under Review," and "Completed." Figure 2 As shown, configure the validator plugin on the "Pending Review" to "Under Review" status transition path, and configure the trigger plugin on the "Under Review" to "Completed" transition path. Simultaneously, create a custom "Field Contains Null Values" field in the JIRA system. This field is of boolean type and is used to record the validation results.

[0044] When developers submit a requirement issue, they must fill in mandatory fields such as "Requirement Description," "Requirement Type," and "Whether it is Derivative." When the status of a requirement issue changes from "Pending Review" to "Under Review," the validator plugin automatically checks whether the content of the mandatory fields is empty.

[0045] If any of the above fields fails validation, write "Yes" in the "Field contains null values" field, terminate the state transition, and record the specific reason for the validation failure in the system log.

[0046] If all fields pass validation, then write "No" in the "Field contains null values" field to allow the status to transition to "Under Review". When the status of the requirement issue changes from "Under Review" to "Completed", the trigger plugin executes the following processing logic:

[0047] Read the value of the "Field contains null values" field. If the value is "yes", terminate the process and generate a system alarm.

[0048] If the answer is "No", the review service of the dify platform will be called via API, and the request body will contain all relevant field data of the issue.

[0049] After receiving the request, the dify platform executes the preset review workflow: a) constructs the review context through prompt word engineering, including industry specifications, technical standards and other constraints; b) calls the API interface of the deepseek model to input the requirement description text and review criteria into the model; c) the model outputs; d) the review results are formatted into a JSON data structure.

[0050] During the results write-back phase, the dify platform returns the review results to the JIRA system via API, eventually changing the status to "Completed" and generating a review completion notification.

[0051] This invention also provides an automatic JIRA requirements data review device based on the Dify platform, such as... Figure 3 As shown, this device integrates AI capabilities with JIRA workflow automation technology to achieve intelligent management of the entire requirements review process. The core of the device consists of three parts: a dify workflow module, a JIRA verification module, and a JIRA triggering module, forming a closed-loop system from data verification to intelligent review.

[0052] The dify workflow module, as the core processing unit of the device, is responsible for building and managing the complete process of requirements review. This module constrains review criteria through preset review prompt templates, calls the deepseek large language model to perform multi-dimensional analysis of the requirements, including review items such as logical completeness and technical feasibility, and finally returns the structured review results in JSON format. The module supports deep data interaction with the JIRA system, ensuring that review results can be accurately mapped to specified fields.

[0053] JIRA's validation module automates data verification during the workflow status transitions of requirement issues by integrating a validator plugin. When an issue reaches a validation node, the module automatically checks the completeness of key fields such as "Requirement Description" and "Requirement Type," and updates the "Field Contains Null Values" status field with the validation result as a Boolean value. This design adheres to the basic specifications of JIRA's workflow validation mechanism while providing decision-making support for subsequent processes through status flags.

[0054] The JIRA trigger module implements intelligent process control based on trigger plugins, dynamically triggering the review process according to the status flags of the verification module. When the "Field contains null values" flag is marked "No", the module calls the review service of the Dify platform via API, submitting complete requirement data to the Deepseek model for analysis; if the flag is marked "Yes", the process terminates and prompts for data completion. After the review is completed, the module automatically writes back the review conclusions generated by Deepseek to the corresponding fields in JIRA.

[0055] While specific embodiments of the present invention have been described above, those skilled in the art should understand that the specific embodiments described are merely illustrative and not intended to limit the scope of the present invention. Equivalent modifications and variations made by those skilled in the art in accordance with the spirit of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An automatic review method for JIRA requirements data based on the Dify platform, characterized in that: include: During development, a JMWE trigger plugin is added to the node where the workflow status of the issue containing the requirement is triggered. When the workflow status of the issue reaches the verification node, the JMWE validator plugin checks whether the custom field has a null value and writes the judgment result to the issue. Subsequently, when the workflow status of the issue reaches the trigger node, if the custom field has a null value, the trigger plugin is not triggered. If the relevant field does not have a null value, the trigger plugin calls the API of the Dify platform to trigger the requirement review. On the Dify platform, an automated requirement review workflow is predefined, which contains at least one structured prompt word. The structured prompts are used to define the set of key information fields, review dimensions, review criteria, and review result output format required for the review requirements; the review workflow generates the corresponding application programming interface (API) endpoints through the dify platform.

2. The automatic review method according to claim 1, characterized in that: When the workflow status of an issue reaches the verification node, the JMWE validator plugin checks whether a custom field contains a null value, and then writes the result to the issue. When the workflow status of the issue reaches the trigger node, if the custom field contains a null value, the trigger plugin is not triggered; otherwise, the trigger plugin calls the dify API to trigger a requirements review, specifically including: The validator plugin checks whether there are empty values ​​in the information fields of the custom fields required in the issue. If there are empty values, it writes "Yes" after the record field in the custom field; if there are no empty values, it writes "No" after the record field in the custom field.

3. The automatic review method according to claim 2, characterized in that: The trigger plugin determines the value of the record field. If the value is yes, the requirement review is not triggered. If the value is no, the trigger plugin calls the Dify API to review the requirements under the issue. When calling, the trigger plugin takes the current value of the information field set in the task item as the input parameter, passes it to the Dify platform, and returns the result to the corresponding field.

4. The automatic review method according to claim 3, characterized in that: The dify platform responds to the API endpoint call request by receiving the request, specifically including: Execute a predefined review workflow, populate the received requirement data into the structured prompts; utilize an integrated large language model (LLM) to perform requirement review analysis based on the populated structured prompts; generate structured review results conforming to a predetermined output format; and return the structured review results via the API response. After receiving the review results returned by the dify platform, the trigger plugin parses the review results and automatically writes the review results into the predefined result field of the JIRA requirement task item.

5. The automatic review method according to claim 1 is characterized in that: The method automatically triggers the Dify platform when the value of the predefined verification status field of the requirement task item is not empty. In response to the trigger, the Dify platform, based on the received values ​​of the key information field set, drives the integrated large-scale language model to automatically review and analyze the requirement content under the requirement task item by executing its predefined structured prompts and review workflow. The Dify platform generates a structured review result conforming to a predetermined output format and returns the result. The review result includes review comments for the predefined review items and is automatically written into the predefined result field of the JIRA requirement task item.

6. A JIRA requirements data review apparatus based on the Dify platform, applied to the automatic review method according to any one of claims 1-5, characterized in that, include: The dify workflow module, the JIRA verification module, and the JIRA triggering module. The `dify` workflow module is used to handle requirement reviews. As the core engine of the review process, it is pre-configured with structured prompts and an automated review workflow. It receives review requests through its generated application programming interface (API) endpoint. Upon receiving a request, the `dify` workflow module executes the pre-configured automated review workflow, filling the structured prompts with the requirement-related information contained in the request. This drives the integrated Large Language Model (LLM), which analyzes and reviews the requirement content item by item based on the filled prompts and predefined review items. The module generates a structured review result that conforms to a predetermined format and returns it through the API endpoint. The JIRA verification module includes a JMWE verifier plugin instance configured in the JIRA target requirement type workflow verification status node. This verifier is used to check whether all key information fields required by the structured prompts of the dify workflow module in the current requirement task item are non-empty. If there are empty fields, a first identifier indicating incomplete data is written into the predefined dify review ready field of the task item. If all information fields and record fields are non-empty, a second identifier indicating complete data is written into the predefined dify review ready field. The JIRA triggering module includes a JMWE trigger plugin instance configured in the JIRA target requirement type workflow trigger status node. It is configured to perform a trigger action in response to the value of the predefined dify review ready field being the second identifier: automatically calling the API endpoint provided by the dify workflow module, passing the information field and record field values ​​of the task item as input parameters to the dify platform; and writing the review result content into the predefined review result field of the task item after receiving the structured review result returned by the dify workflow module.

7. The apparatus according to claim 6, characterized in that: The device automatically reviews the requirements based on relevant information after the issue is created, specifically including: When the relevant information about the requirement is complete, the review results of the requirement will be automatically written to the corresponding field in JIRA.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 5.

Images