A risk assessment method and device, a storage medium, an electronic device and a product
By combining a large language model with a knowledge base and attention mechanism, a solution evaluation model is developed to automatically evaluate business change solutions. This solves the problem of low efficiency in traditional manual evaluation and enables efficient and accurate identification of change risks and improvement suggestions.
Patent Information
- Application Number
- CN202511310274.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-12
- Publication Date
- 2026-01-09
- Estimated Expiration
- 2045-09-12
AI Technical Summary
Traditional business change assessments rely on human experience, which is inefficient and has low accuracy and reliability, and cannot effectively identify potential risks during the change process.
By leveraging a large language model and combining change reference information from a knowledge base, and through a solution evaluation model that generates prompt words and incorporates an attention mechanism, the system automatically assesses whether the initial change plan meets the target change requirements, generates evaluation results, and provides improvement suggestions.
It improved the efficiency and accuracy of change plan evaluation, significantly reduced operational risks, and enhanced the reliability and security of evaluation results.
Smart Images

Figure CN120803887B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] One or more embodiments of the present specification relate to the technical field of computer technology, and in particular to a risk assessment method and device, a storage medium, an electronic device and a product. BACKGROUND
[0002] In a fast iteration software development environment, business changes for businesses such as e-commerce transactions, financial payments, cloud computing services, etc. have become the norm, and the change content covers multiple aspects such as new function online, defect repair, configuration update, etc. It is an important technical means for continuous optimization and iterative upgrade of business application implementation.
[0003] Before implementing changes to the target business, it is usually necessary to develop a change plan based on business goals, technical architecture and risk prediction, and to comprehensively evaluate the feasibility, security and potential impact of the change plan, in order to reduce various risks that may occur during the change process.
[0004] However, the traditional change plan evaluation method mainly relies on manual experience, which is not only inefficient, but also limited by the knowledge reserve and focus of the evaluators, resulting in low accuracy and reliability of the evaluation results. It cannot accurately detect the risks that may exist in the implementation process of the business change plan, and it is difficult to provide reliable protection for business changes. SUMMARY
[0005] Therefore, one or more embodiments of the present specification provide technical solutions as follows:
[0006] According to a first aspect of one or more embodiments of the present specification, a risk assessment method is provided, comprising:
[0007] obtaining an initial change plan for a target business and software design information corresponding to the target business; wherein the initial change plan includes change content and implementation process of the business change to be implemented to the target business;
[0008] According to the initial change plan, determine the change reference information matched therewith in the pre-constructed knowledge base;
[0009] fill the change reference information, the software design information and the initial change scheme into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change scheme and prompt text in the prompt word template, and input the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets target change requirements reflected by the change reference information under the constraint of the software design information under the indication of the prompt text, to evaluate the initial change scheme and obtain an evaluation result; wherein the target change requirements include at least one of change release requirements, change monitoring requirements and change recovery requirements.
[0010] According to the evaluation result, determine the change risk of implementing business change based on the initial change scheme for the target business, wherein the change risk includes the abnormality of the target business caused by implementing the initial change scheme without meeting the target change requirements.
[0011] According to a second aspect of one or more embodiments of the present specification, a risk evaluation device is provided, comprising:
[0012] The acquisition module is configured to acquire an initial change scheme for a target business and software design information corresponding to the target business; wherein the initial change scheme includes change content and implementation process of business change to be implemented on the target business.
[0013] The matching module is configured to determine change reference information matched with the initial change scheme in a pre-constructed knowledge base according to the initial change scheme.
[0014] The evaluation module is configured to fill the change reference information, the software design information and the initial change scheme into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change scheme and prompt text in the prompt word template, and input the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets target change requirements reflected by the change reference information under the constraint of the software design information under the indication of the prompt text, to evaluate the initial change scheme and obtain an evaluation result; wherein the target change requirements include at least one of change release requirements, change monitoring requirements and change recovery requirements.
[0015] A determining module is configured to determine a change risk of implementing a business change of the target business based on the initial change scheme according to the evaluation result, wherein the change risk includes an abnormality of the target business caused by implementing the initial change scheme without meeting the target change demand.
[0016] According to a third aspect of one or more embodiments of the present specification, an electronic device is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor implements the steps of the above method by running the executable instructions.
[0017] According to a fourth aspect of one or more embodiments of the present specification, a computer readable storage medium is provided, which stores computer instructions, and the instructions are executed by a processor to implement the steps of the above method.
[0018] According to a fifth aspect of one or more embodiments of the present specification, a computer program product is provided, comprising computer program / instructions, which are executed by a processor to implement the steps of the above method.
[0019] As can be seen from the above embodiments, the present specification generates prompt words by changing reference information, software design information and initial change scheme, and performs full-process automatic processing of inputting evaluation model, which effectively solves the pain points of traditional manual evaluation: the present scheme takes the change demand reflected by the relevant change reference information in the knowledge base as a reference, takes the software design information as the context constraint of the target business, relies on the evaluation model to evaluate whether the initial change scheme meets the target change demand, and the whole process does not need to rely on the subjective judgment of the evaluation personnel, which greatly improves the evaluation efficiency and significantly improves the accuracy and reliability of the evaluation result. BRIEF DESCRIPTION OF DRAWINGS
[0020] Figure 1 FIG. 1 is a schematic diagram of an architecture of a risk evaluation service system provided by an exemplary embodiment;
[0021] Figure 2 FIG. 2 is a schematic diagram of a flow of a risk evaluation method provided by an exemplary embodiment;
[0022] Figure 3 FIG. 3 is a whole flowchart of a risk evaluation provided by an exemplary embodiment;
[0023] Figure 4 FIG. 4 is a schematic diagram of an interaction process of a business change system provided by an exemplary embodiment;
[0024] Figure 5 FIG. 5 is a schematic diagram of a structure of a device provided by an exemplary embodiment;
[0025] Figure 6is a block diagram of a risk assessment device provided by an example embodiment. DETAILED DESCRIPTION
[0026] In order for those skilled in the art to better understand the technical solutions in the specification, the technical solutions in the specification will be clearly and completely described below in combination with the drawings in the specification. Obviously, the described embodiments are only part of the embodiments of the specification, not all. Based on the embodiments in the specification, all other embodiments obtained by those of ordinary skill in the art without creative labor should be within the scope of protection of the specification.
[0027] In business scenarios such as financial technology, e-commerce platforms, enterprise-level SaaS systems, etc., publishing business changes (such as new function online, defect repair, configuration update, interface iteration, data structure adjustment, permission system upgrade, etc.) is the norm to support business iteration, but also contains great risks, such as service interruption, data error, user experience degradation, transaction process jam, compliance verification failure, cross-system interaction anomaly, etc.
[0028] Traditional change scheme review relies on manual experience, which has significant shortcomings: low efficiency, narrow coverage, easy to miss, and difficult to effectively utilize historical experience and complex system knowledge. Moreover, the current industry lacks a solution that can systematically and automatically assess and optimize the core change requirements (such as change resilience) of the change scheme using a large language model (LLM), which cannot quickly identify potential risks in changes, nor can it provide actionable optimization suggestions based on historical change data.
[0029] Based on this, the specification provides a business change scheme that uses the natural language understanding and reasoning capabilities of a large language model (LLM) to automatically assess whether the change scheme meets the target change requirements by combining change reference information retrieved from the knowledge base based on semantics, identify deficiencies and provide specific improvement suggestions, thereby improving the safety, reliability and efficiency of publishing changes and reducing operational risks.
[0030] Figure 1 is an architecture diagram of a service system for risk assessment provided by an example embodiment. As shown in Figure 1 The system can include a server 11, a network 12, and several electronic devices such as a personal computer (PC) 13, a mobile phone 14, etc.
[0031] The server 11 can be a physical server containing a single host, or the server 11 can be a virtual server carried by a host cluster. In operation, the server 11 can run a server-side program of an application to implement a function of the application, such as when the server 11 runs a program of a business change service, a corresponding business change service platform can be implemented.
[0032] The PC 13 and the mobile phone 14 are only some types of electronic devices that can be used by a user. In fact, the user can obviously also use electronic devices of other types, such as a tablet device, a notebook computer, a personal digital assistant (PDA), a wearable device (such as smart glasses, a smart watch, etc.), and the like, without limitation. In operation, the electronic device can run a client-side program of an application to implement a function of the application, such as when the electronic device runs a program of a business change service, a client of the business change service can be implemented. The application program of the client of the business change service described above can be started and run on the electronic device. The client-side program can be a native application installed on the electronic device, or the client-side program can be an applet, a fast application, or other similar forms. Of course, when a web page technology such as HTML5 or the like is used, the function can be implemented through a page displayed by a browser, which can be a standalone browser application or a browser module embedded in some application.
[0033] The network 12 through which the electronic devices such as the PC 13 and the mobile phone 14 interact with the server 11 can be a wired or wireless network based on a communication mode supported by the corresponding electronic device, without limitation. For example, the PC 13 can support both wired and wireless communication, and wired or wireless communication can be used as needed. The mobile phone 14 usually only supports wireless communication, and wireless communication can be used.
[0034] Based on the service system architecture described above, a risk assessment method is provided in the present specification, as shown in the following. Figure 2
[0035] Figure 2 FIG. 1 is a flow diagram of a risk assessment method provided in an example embodiment, including the following steps:
[0036] S201: Obtain an initial change scheme for a target business and software design information corresponding to the target business; wherein the initial change scheme includes change content and an implementation process of a business change to be implemented on the target business.
[0037] In this specification, the execution subject for performing the risk assessment method can be a designated device such as a server, of course, a cloud server cluster, an edge computing node, an embedded terminal device, a special-purpose hardware with data processing capability (such as an encryption server in the financial industry, a transaction processing terminal in the e-commerce field), and other devices. For ease of description, the following will take a server as an execution subject to illustrate the business change method provided in this specification.
[0038] The business change can be a change issued for an online business (such as a commodity transaction business of an e-commerce platform, an online credit business of a financial institution, a collaborative office business, an online consultation business), an offline business (such as a cash register business of a retail store, a warehouse management business of a logistics network, a counter service business of an offline bank network, a store operation business of a chain restaurant), or an online or offline business. These changes can include: online of new modules at the function level, optimization of existing functions and defect repair; adjustment of business rules, optimization of parameter thresholds, and update of permission configuration at the configuration level; data structure adjustment, interface iteration upgrade, and data migration synchronization at the data and interface level; server resource expansion, offline device firmware update, and system running environment optimization at the hardware and environment level.
[0039] Of course, other business changes corresponding to other businesses can also be included in actual applications, which are not limited in this specification.
[0040] Before the business change is issued, a user (such as an operation and maintenance personnel) usually first formulates an initial change scheme, which is used to describe the change content and implementation process of the business change to be implemented to the target business.
[0041] For example, for the online business of the e-commerce platform-function level-new module online, the change content of the initial change scheme is: online the function modules of "commodity display", "real-time interactive ordering", and "commission settlement"; the implementation process is: deploy the modules in the test environment 3 days in advance and complete the function and stress test, deploy the production environment at 00:00-02:00 on the online day, perform 1-hour gray verification after deployment, and if there is no exception, open full volume, and prepare the rollback plan at the same time.
[0042] For example, for the offline business of the retail store-hardware and environment level-firmware update of the cash register device, the initial change scheme will explicitly specify the change content as updating the firmware version that fixes the payment security vulnerability, and the implementation process is: select 2 stores as pilot stores, complete the device update of all stores before the next day's business if there is no problem in the pilot, test 3 simulated transactions for each device after updating, ensure the payment function is normal, and at the same time, prepare the old firmware installation package for rollback.
[0043] In the present specification, the timing at which the server acquires the initial change scheme can be various, for example, the server can acquire the initial change scheme after the user uploads the initial change scheme to the business change system. For another example, when the user completes the initial change scheme online in the business change system and clicks the "submit for review" button, the system automatically synchronizes the scheme data to the server, and the server acquires the scheme in real time. Of course, the user can also transmit the initial change scheme file to the designated server through email, file transfer protocol (FTP) and the like, and the server acquires the initial change scheme after monitoring the completion of file transmission and verifying the format compliance.
[0044] In addition, in order to ensure that accurate software context information is provided to the evaluation model during the process of business change, the server can also acquire software design information of the target business, which can be directly extracted from the system analysis documentation of the target business.
[0045] The system analysis documentation refers to the detailed technical description of the design documentation, interface documentation, data dictionary, dependency graph, etc. of the specific system involved in the change to be released. The software design information extracted therefrom can include: system architecture description of the target business (such as microservice module division logic, module interaction timing diagram of core business process, database hierarchical architecture, etc.), dependency list (such as software running dependent third-party component version, interface list of cross-system interaction, hardware resource dependency, data dependency, etc.) and the like. Of course, other software design information related to the software design of the target business can also be included in actual application, which will not be enumerated one by one in the present specification.
[0046] S202: According to the initial change scheme, determine the change reference information matched therewith in the pre-constructed knowledge base.
[0047] After acquiring the initial change scheme, the server can determine the change reference information matched therewith in the pre-constructed knowledge base. In the present specification, a variety of change reference information is pre-stored in the knowledge base, which can include historical change schemes for reflecting different change requirements, failed change schemes that cause change failure in the process of trying business change in the past, and preferred change schemes that are verified to have excellent effect, controllable risk and up-to-standard efficiency. In addition, the change reference information can also include some typical architecture documents, such as software architecture diagram, microservice dependency document, payment module interaction flowchart, system API access specification, distributed database sharding design document, server cluster expansion architecture record, etc., thereby providing technical basis for change evaluation.
[0048] In practical applications, the change reference information is usually stored in the form of a vector in the material information library. In the process of retrieving the change reference information, the server can first perform semantic analysis on the initial change scheme, thereby extracting structured change data and converting it into a corresponding semantic vector (such as generating a fixed-dimension vector by using a semantic analysis engine such as a natural language processing model to extract feature representations). Then, the similarity (such as cosine similarity, Euclidean distance, etc.) between the feature representation corresponding to the change data and the feature representation corresponding to each change reference information in the knowledge base is calculated, and one or more change reference information with a similarity higher than a preset threshold is filtered out, or the change reference information is sorted according to the similarity, and one or more change reference information before a preset position is filtered out.
[0049] S203: filling the change reference information, the software design information, and the initial change scheme into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change scheme, and prompt text in the prompt word template, and inputting the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets the target change requirement reflected by the change reference information under the constraint of the software design information under the indication of the prompt text, to evaluate the initial change scheme and obtain an evaluation result; wherein the target change requirement includes at least one of a change release requirement, a change monitoring requirement, and a change recovery requirement.
[0050] S204: determining a change risk of implementing business change on the target business based on the initial change scheme according to the evaluation result, wherein the change risk includes an abnormality of the target business caused by implementing the initial change scheme without meeting the target change requirement.
[0051] The server can generate a prompt word based on the change reference information, the software design information, and the initial change scheme, to instruct the scheme evaluation model to evaluate the initial change scheme through the prompt word.
[0052] In practical applications, the server can fill the change reference information, the software design information, and the initial change scheme into a preset prompt word template, which is provided with corresponding prompt text for prompting the scheme evaluation model to refer to the target change requirement reflected by which information and to take which information as a constraint, thereby generating a prompt word based on the change reference information, the software design information, the initial change scheme, and the prompt text in the prompt word template. An example of the prompt word will be described in detail below, and the present specification will not be described in detail here.
[0053] The scheme evaluation model can be a large language model based on an attention mechanism. Through token-level attention distribution, precise semantic analysis is achieved, the prompt words are disassembled into discrete tokens, the correlation weights between different tokens are calculated, and the feature vectors are weighted and summed to make the determined feature representation accurately express the semantic information of the prompt words, assisting the model in identifying core change requirements in the change reference information, key constraint clauses in the software design information, implementation points in the initial change scheme, and generation instructions expressed by the prompt text.
[0054] In this process, the model first disassembles the input prompt words into discrete tokens (which can be understood as the smallest semantic unit of text, such as words, characters, or subwords). Then, the correlation weights between different tokens are calculated through the self-attention (Self-Attention) mechanism. This step can mine the semantic correlation between words in the text. Then, the feature vectors corresponding to each token are weighted and summed according to these weights, so that the final feature representation accurately represents the semantic information of the prompt words.
[0055] In addition, the scheme evaluation model can also be a domain-specific fine-tuned model, a multi-modal evaluation model, and a rule and learning hybrid model, and other artificial intelligence models, wherein:
[0056] The domain-specific fine-tuned model is obtained by fine-tuning a general large language model based on professional data in a specific field (such as software development). For example, in the software development field, after fine-tuning a large amount of software requirement documents, design specifications, code change records, etc., the model has a more accurate understanding of professional terms in this field (such as "microservice architecture", "API interface"), business logic (such as "user registration process", "order payment link"), and change scenarios (such as "function addition in version iteration", "online bug fixing"). When used for scheme evaluation, it can more efficiently identify the core elements in software change tasks, such as whether a software module refactoring scheme meets the design principles of the field, whether there are compatibility problems with the existing system, etc., making up for the shortcomings of general models in field details.
[0057] Multimodal evaluation models can simultaneously process various input types, including text, images, and audio. In complex solution evaluation scenarios, it is often necessary to combine multiple types of information. For example, when evaluating a software adaptation solution for a smart hardware product, there may be not only textual functional requirements, but also image information such as hardware appearance diagrams and interface schematics, and even voice instructions for product operation. Multimodal evaluation models can fuse and analyze these different types of information, identify the structural features of the hardware from images, extract key operational logic points from voice, and then combine them with textual requirements to comprehensively evaluate whether the software adaptation solution can meet the various requirements of the hardware, ensuring the rationality and feasibility of the solution across multiple dimensions.
[0058] The rule-based hybrid model combines manually defined rules with the adaptive capabilities of machine learning. On one hand, humans can formulate explicit rules based on industry standards and regulatory requirements, such as rules regarding data encryption and transaction security verification in financial software solution evaluation. When a solution involves these aspects, the model first makes a preliminary judgment based on these rules. On the other hand, through machine learning, the model can learn hidden patterns and regularities from a large amount of historical solution evaluation data. These patterns may be difficult to accurately cover with human rules, such as users' potential preferences for feature priorities in different business scenarios. In actual evaluation, the model first uses the rules to check rigid constraints, and then uses the learned patterns to evaluate the flexibility, innovation, and other flexible indicators of the solution. This ensures that the solution meets basic specifications while adapting to complex and ever-changing actual needs, making the evaluation results more comprehensive and accurate.
[0059] Before using the above scheme evaluation model, the server can first train or fine-tune it to ensure that the scheme evaluation model can accurately identify the semantic information of keywords and obtain evaluation results. The training or fine-tuning process of the above scheme evaluation model is as follows:
[0060] Construct a structured training dataset, which includes:
[0061] Sample input data includes: sample change plans (such as version iteration plans of historical systems, functional module adjustment plans, etc.), sample software design information (such as system architecture documents, interface specifications, performance constraints, etc.), and sample change reference information (such as historical change plans that specify change release windows, monitoring indicators that must be covered, and recovery time requirements).
[0062] Tag data: The actual evaluation results of the sample change plan by developers or professional evaluation teams are used as tags. The tag format can be set according to the scenario, such as binary judgment results, quantitative scores, comprehensive evaluation reports including sub-items such as "release feasibility", "monitoring integrity" and "recovery effectiveness", etc.
[0063] The model training process is as follows:
[0064] The server constructs a training prompt according to the same logic as actual reasoning - fills the sample change scheme, sample software design information, sample change reference information into the preset prompt template to generate an input text containing evaluation instructions.
[0065] The training prompt is input into the scheme evaluation model to be trained to obtain the predicted evaluation result output by the model. This process is the same as the actual reasoning process, and the specific content will be described in detail in the actual reasoning process below. This specification will not be described in detail here.
[0066] The loss value is calculated according to the label data (actual evaluation result) and the predicted evaluation result. For example, cross-entropy loss can be used for binary judgment, mean square error loss can be used for evaluation, and weighted mixed loss function can be used for multi-dimensional evaluation.
[0067] The optimization goal is to minimize the loss value, and the model parameters (such as attention weight matrix, feature fusion layer parameters, etc.) are iteratively adjusted through the back propagation algorithm until the evaluation indicators (such as accuracy, F1 score, score error rate) on the validation set reach the preset standard, and the training process is completed.
[0068] Through this training mechanism, the scheme evaluation model can learn the matching rules of change requirements and design constraints from historical cases, and gradually improve the evaluation accuracy of new change schemes.
[0069] Of course, in actual application, the above scheme evaluation model can also not be trained, but a mature large language model is directly used as a risk evaluation model to evaluate the initial change scheme.
[0070] Specifically, the above target change requirement can include a change requirement that meets the Change Resilience Triad (CRT), which includes a change release requirement corresponding to Progressive Deployment, a change monitoring requirement corresponding to Comprehensive Observability, and a change recovery requirement corresponding to Rapid Rollback / Circuit-Breaker Capability.
[0071] Among them, the change release requirement needs to ensure that the change is gradually released online in batches, flows, and regions to control the impact range and facilitate problem discovery and rapid loss-making;
[0072] Change monitoring requirements need to ensure that there are sufficient monitoring indicators, logs, and tracking data during or after the implementation of changes to real-time understand the change status, business impact, and system health;
[0073] Change recovery requirements need to ensure that when changes cause unexpected problems, there is the ability to quickly, reliably, and one-click or with minimal manual intervention to roll back to a stable version, or to fuse abnormal traffic / functionality.
[0074] In addition, in order to ensure the accuracy of the evaluation results, software design information can be considered in the process of evaluating the initial change scheme. The software design information can provide the software context information of the target business for the scheme evaluation model, and use it as a constraint to assist the model to realize multi-dimensional verification, such as:
[0075] At the architecture adaptability level, it is determined whether the gradual deployment scheme is compatible with the existing service registration and discovery mechanism, such as whether the gray strategy can rely on the existing gateway to implement traffic splitting; at the resilience element verification level, combined with the coverage blind area of the software architecture evaluation observability scheme, the effectiveness of the fault isolation design is determined according to the service fuse component; at the risk inference level, the typical problems in the same architecture in the history of changes are associated, and the resource occupation estimation of the current scheme is integrated to identify potential bottlenecks.
[0076] Through the deep integration of software design information, the evaluation results are closely bound to the technical characteristics of the business system, significantly improving the accuracy of risk prediction and the operability of improvement suggestions.
[0077] In addition, the scheme evaluation model can also use software design information to extract system component dependency relationships, deployment topology, and other core information as the underlying basis for evaluating the initial change scheme. For example, by analyzing the call links between microservices in the software design information, it is determined whether the newly added interface in the change scheme will break the stability of the existing service dependency; according to the database sharding strategy, it is verified whether the data operation involved in the change conforms to the sharding rules to avoid cross-sharding transaction risks; combined with performance benchmarking indicators (such as maximum concurrency per node), it is evaluated whether the resource occupation caused by the change scheme is within the system bearing range, thereby strengthening the technical adaptability judgment of the evaluation.
[0078] Thus, an evaluation result for indicating whether the initial change information meets the target change requirements reflected by the change reference information under the constraints of the software design information can be obtained.
[0079] In the present specification, the change reference information can include several historical change schemes, and each historical change scheme can correspond to one or more change requirements. If a single historical scheme cannot cover all target change requirements (i.e., cannot completely map the three elements of change resilience or other demand dimensions), the server can determine multiple historical change schemes whose demand coverage ranges are complementary, and together completely reflect the full demand required by the change reference information, providing comprehensive historical experience reference for the evaluation of the change scheme.
[0080] Specifically, the scheme evaluation model can perform semantic analysis on the prompt word, and based on the semantic analysis result, under the indication of the prompt text, reference is made to historical change schemes reflecting corresponding change release requirements to evaluate whether the implementation process meets the change release requirements of gradual release under the constraint of software design information. The semantic analysis result can be a feature representation determined by the scheme evaluation model based on an attention mechanism.
[0081] It should be noted that different historical change schemes can be referred to when evaluating different change requirements, and of course, the same historical change scheme that can reflect all change requirements can also be referred to. In addition, in actual application, the evaluation target can be determined according to the actual evaluation target, such as evaluating only one or two of the change release requirements, change monitoring requirements, change recovery requirements, or evaluating all change requirements.
[0082] For a failed change scheme included in the change reference information, the scheme evaluation model can extract relevant information that causes the change to fail from the failed change scheme (such as the fact that the gray release strategy does not adapt to the registration center mechanism, causing abnormal traffic splitting, the fact that the monitoring indicator does not cover the core business link, causing fault discovery delay, the fact that the rollback script is not compatible with the historical data format, causing data consistency destruction, etc.), and then detect whether the initial change scheme involves the relevant information to evaluate whether the initial change scheme has a change risk that causes the change to fail.
[0083] The change risk can include an abnormality that occurs in the target business when the initial change scheme is implemented without meeting the at least one target change requirement. For example, if a failed change scheme fails because "no gray data isolation rule is set in the database read-write separation scenario, causing cross-pollution of new and old version data", the scheme evaluation model can extract this relevant information, and if it is found that the initial change scheme also does not mention the gray data isolation design in the read-write separation scenario, it can be determined that there is a similar failure risk.
[0084] When the initial change scheme is evaluated by the scheme evaluation model, one or more of the change requirements can be evaluated, and an evaluation result (i.e., whether a certain evaluation requirement is met) for each change requirement or an overall evaluation result (i.e., whether all change requirements are met) can be given.
[0085] Of course, in actual application, the above evaluation result can also be embodied in the form of a specific score, so as to more accurately quantify the degree to which the initial change scheme meets the change requirements.
[0086] Specifically, the scheme evaluation model can determine a final score according to the five dimensions of the degree to which the initial change scheme meets the change publishing requirement, the degree to which the initial change scheme meets the change monitoring requirement, the degree to which the initial change scheme meets the change recovery requirement, whether there is a situation that leads to change failure, and the matching situation with the software design information. The higher the score, the higher the degree to which the change requirement is met, and the lower the change risk brought by the implementation of the change scheme.
[0087] Among them, the server can set a weight for each dimension, and each dimension is divided into two cases of "satisfy" and "not satisfy". If it is satisfied, the value is 1, and if it is not satisfied, the value is 0. When the weight corresponding to each dimension is the same, that is, 0.2, if the results of all dimensions are satisfied, that is, the change publishing requirement, the change monitoring requirement, the change recovery requirement are met, there is no situation that leads to change failure, and it matches the software design information, the final score is 1*5*0.2*100%=100. If the result of one of the dimensions is not satisfied, the final score is 1*4*0.2*100%=80.
[0088] For example, the server can first determine whether the initial change scheme will lead to change failure. If it will lead to change failure, the score is directly output as 0, and the other dimensions are not scored. If it will not lead to change failure, the results of each dimension and the weight are weighted and summed to obtain the final score.
[0089] In addition, the server can also set a set score threshold (such as 60), and if it is lower than the threshold, it is determined that the initial evaluation scheme does not meet the requirements as a whole.
[0090] Further, before using the above scheme evaluation model, the server can first train it. In this process, the server can obtain a sample change scheme, sample software design information, and sample change reference information, and obtain the actual evaluation result of the sample change scheme as a label. The actual evaluation result can be the expert evaluation result made by the developer for the sample change scheme (such as whether a certain change reference information or score is met).
[0091] The server can then construct the prompt word in the same way as above, output a predicted evaluation result through the to-be-trained scheme evaluation model, and then determine a loss value according to the expert evaluation result and the predicted evaluation result, and minimize the loss value as an optimization objective for the scheme evaluation model.
[0092] Of course, in actual applications, the scheme evaluation model can also not be trained, but an already open-source evaluation model can be directly applied to evaluate the initial change scheme using the prompt word. After obtaining the evaluation result, the scheme evaluation model can also refer to the preferred change scheme in the change reference information, and generate improvement suggestions for the initial change scheme according to the evaluation result.
[0093] For example, if the evaluation result is “not meeting the change release requirements corresponding to the progressive deployment”, the scheme evaluation model can determine the batch release process verified by practice from the preferred change scheme, and extract the key control nodes (such as the automated verification link after each batch release, the gray scale threshold of traffic switching) as improvement basis to generate specific suggestions.
[0094] For ease of understanding, the present specification provides a whole flowchart of risk assessment, as shown in Figure 3 .
[0095] Figure 3 is a whole flowchart of risk assessment provided by an exemplary embodiment.
[0096] The server can obtain an initial change scheme and software design information, and perform semantic analysis on the initial change scheme based on a preset analysis engine to obtain structured text, and extract software design information from a system analysis document;
[0097] The server can then search in the knowledge base based on the analysis result (structured text) of the initial change scheme, to obtain change reference information;
[0098] The server can further construct a prompt word based on the analysis result of the initial change scheme, the change reference information, and the software design information, and input the prompt word into a scheme evaluation model, so that the scheme evaluation model generates an evaluation result. The content of the evaluation result can include whether the initial change scheme meets the target change requirement, whether the initial change scheme has change risks, and improvement suggestions for the initial change scheme.
[0099] After generating the evaluation result, the server can determine the risk of implementing business change based on the initial change scheme for the target business.
[0100] Exemplarily, the above prompt word can be as follows:
[0101] As a professional software change plan evaluation expert, please conduct a comprehensive evaluation of the initial change plan based on the following three types of key information:
[0102] Change reference information (target change requirements):
[0103] Requirement type: Change release requirements (incremental release), change monitoring requirements (data observability), change recovery requirements (data recoverability).
[0104] Change content: [Upgrade the existing "user points system" function and add a "points exchange coupon" function];
[0105] Implementation process: [Adopt an incremental release strategy: first release to 10% of users for 24 hours, then expand to 50% for 12 hours, and finally release to all users; Real-time monitoring of core indicators: interface call success rate (≥99.9%), exchange operation response time (≤500ms), and points deduction and coupon issuance consistency (100% match); Emergency recovery capability: if there are indicator abnormalities during the gray period, the system can be rolled back to the original version within 10 minutes, and the user points data after rollback should be consistent with that before the change.]
[0106] Software design information (design side constraints):
[0107] Existing system architecture: [Front-end and back-end separation architecture, front-end uses xxx framework, back-end uses xxx microservices, order data is stored in xxx database, file export service depends on existing "file processing microservices"];
[0108] Design specification constraints: [Back-end interfaces must comply with xxx, request parameters must be validated, Excel file export must use xxx toolkit (version ≥3.1.0), and new code must comply with the company's development specifications];
[0109] Potential conflict points: [For example: the existing file-service service has a maximum of 800 lines of files processed at a time, and the "order time" field in the t_order table in the xxx database is of type datetime, which needs to be consistent with the Excel export format].
[0110] Initial change plan:
[0111] Change content: [Add a "exchange coupon" button to the "my points" page, display the exchange rules popup after clicking, and call the back-end exchange interface after confirmation; Add points exchange interface; Add daily exchange limit logic];
[0112] Implementation process: [Perform progressive release according to demand requirements, realize user shunting through configuration; Add log printing in the exchange interface, collect logs through ELK and configure alarms; Execute version rollback script through Jenkins].
[0113] Based on the above information, output the evaluation results, which should include the following content:
[0114] Determine whether the initial change scheme meets the target change requirements reflected in the change reference information under the constraints of software design information.
[0115] Determine whether the initial change scheme has the risk of causing change failure.
[0116] Finally, give a summary conclusion: whether the initial change scheme "meets the target change requirements reflected in the change reference information under the constraints of software design information, and based on the evaluation results and the preferred change scheme in the change reference information, give improvement suggestions for the initial change scheme.
[0117] In actual application, the evaluation results output by the scheme evaluation model may not meet the preset evaluation conditions, such as incomplete information of evaluation results, obvious judgment errors in evaluation results, and lack of operability in improvement suggestions, etc. Therefore, if the evaluation results do not meet the preset evaluation conditions, a review task is automatically created, and the original evaluation results output by the model, the change reference information and the key fragments of the software design information are attached to the task, and sent to the review personnel with the target business technical background. After obtaining the review results returned by the review personnel, the initial change scheme is adjusted according to the review opinions.
[0118] The server can adjust the initial change scheme based on the evaluation results and improvement suggestions to obtain the target change scheme.
[0119] In the process of implementing business changes for the target business, the server can generate change code for the target business based on the adjusted change scheme, and then implement business changes for the target business using the change code.
[0120] Specifically, the server can input the change scheme and the related original code of the target business into the preset code generation model to parse the change scheme through the code generation model, and generate change code for the target business based on the parsing results on the basis of the related original code, and then implement business changes for the target business using the change code.
[0121] The relevant original code can be local code associated with the module or interface that is implemented with changes (such as source code of a user point management module and a coupon generation module related to changes in an integral exchange function, or code segments of an order creation module and a payment processing module related to order flow adjustment), and of course, can be all the codes of the target business. The code generation model can first identify the function points and interaction logic that need to be modified in combination with the change scheme, and then refer to the existing architecture and coding specifications of the relevant original code to identify the code units (such as interface methods, data processing logic, and page interaction components) that need to be added, modified, or deleted, thereby obtaining incremental change code (including newly added function implementation code, adjustment segments of existing code, and corresponding configuration updates) compatible with the relevant original code.
[0122] In addition, the server can also obtain initial change code of the target business, which can be written by an engineer, and of course, can also be generated by an automatic tool (such as a code generation plug-in or a low-code platform). After inputting the adjusted change scheme and the change code into the preset code generation model, the code generation model is used to optimize the change code (such as adjusting the code structure to adapt to the existing architecture of the target business, supplementing exception handling logic to improve robustness, and simplifying redundant code to improve execution efficiency), and then the optimized change code is used to implement business changes on the target business.
[0123] It should be noted that the above-mentioned code generation model can also be constructed using LLM. The model can be used as a built-in function module of the scheme evaluation model and be linked with the evaluation logic (for example, after the evaluation finds that the scheme needs to supplement rollback code, the built-in model directly generates the corresponding code segment); or the model can be independently run and receive the change scheme and the relevant original code through an interface to complete the code generation or optimization task independently.
[0124] Further, the above-mentioned business change method can be implemented through a corresponding business change system. In order to facilitate understanding, the present specification provides an interaction process diagram of a business change system, as shown in Figure 4 .
[0125] Figure 4 is an interaction process diagram of a business change system provided by an exemplary embodiment, which includes the following steps:
[0126] S401: A user (developer) submits a document of an initial change scheme on a user end;
[0127] S402: After the server receives the initial change scheme, the server performs semantic analysis on the initial change scheme;
[0128] S403: Based on the analysis result, the server queries change reference information related thereto in a knowledge base;
[0129] S404: The server generates a prompt word based on the change reference information returned by the knowledge base, the initial change scheme, and the software design information, and inputs the scheme evaluation model;
[0130] S405: The scheme evaluation model evaluates the initial change scheme based on the prompt word to obtain an evaluation result;
[0131] S406: The scheme evaluation model returns the generated evaluation result to the server;
[0132] S407: The server integrates the evaluation result and feeds back to the user.
[0133] As can be seen from the above method, the present specification realizes deep semantic understanding and complex logical reasoning of the change scheme with the help of the core capability of LLM, effectively overcomes the limitations that static list, keyword matching and simple rule engine cannot parse the natural language context, and are difficult to identify potential dependencies and risks, and can accurately capture the details, intentions and hidden problems of the change scheme.
[0134] At the same time, by fusing the knowledge base and the system analysis document, the short board of limited knowledge of artificial review, inability to fully associate historical lessons and current system complexity is made up, and the problem of lack of field knowledge support of other schemes is solved, so that the evaluation is more suitable for the actual business environment of the enterprise.
[0135] On this basis, the present specification realizes the automation and scaling of change evaluation, breaks through the bottleneck of low efficiency of artificial review and difficulty in dealing with a large number of high-frequency changes, and LLM can further evaluate the rationality, feasibility and completeness of the CRT elements on the basis of judging whether the elements are mentioned, generate specific and context-related improvement suggestions, and avoid the deficiency that other schemes can only give pass / fail or simple prompts.
[0136] Figure 5 FIG. 1 is a schematic structural diagram of a device provided by an exemplary embodiment. As shown in FIG. 1, the device includes a processor 10, a memory 20, and a communication interface 30. Figure 5As shown, the device 500 is comprised primarily of a communication interface 501, a user interface 502, a processor 503, and a data store 504, which are interconnected and communicate with each other through a system bus, network, or other connection mechanism 505. The communication interface 501 enables the device 500 to communicate with other devices, access networks, and transport networks through analog or digital modulation. For example, the communication interface 501 can include a chipset and antenna for wireless communication with a radio access network or access point. In addition, the communication interface 501 can be a wired interface such as an Ethernet, Token Ring, or USB port, or a wireless interface such as Wifi, Bluetooth, Global Positioning System (GPS), or a wide-area wireless interface (e.g., WiMAX or LTE). Of course, the communication interface 501 can support other forms of physical layer interface and standard or proprietary communication protocol. The communication interface 501 can also include multiple physical communication interfaces, such as a Wifi interface, a Bluetooth interface, and a wide-area wireless interface.
[0137] The user interface 502 includes receiving user input and providing output to a user. Thus, the user interface 502 can include input components such as a keypad, keyboard, touch- sensitive or presence-sensitive panel, computer mouse, trackball, joystick, microphone, still camera, and video camera, and output components such as a display screen (which can be combined with a touch-sensitive panel), CRT, LCD, LED, display using DLP technology, printer, other known or future developed equivalent devices. The user interface 502 can also generate audible output through a speaker, speaker jack, audio output port, audio output device, earphones, and other known or future developed equivalent devices. In certain embodiments, the user interface 502 can include software, circuitry, or other form of logic that is capable of transmitting data to and receiving data from external user input / output devices. Additionally or alternatively, the device 500 can support remote access from other devices through the communication interface 501 or another physical interface (not shown). The user interface 502 can be configured to receive user input, the location and movement of which can be indicated by an indicator or cursor described herein. The user interface 502 can also be configured as a display device for rendering or displaying a text segment.
[0138] The processor 503 can include one or more general processors and / or specialized processors.
[0139] The data store 504 can include one or more volatile and / or non-volatile storage components, and can be integrated in whole or in part with the processor 503. The data store 504 can include removable and / or non-removable components.
[0140] The processor 503 can execute program instructions 511 (e.g., compiled or interpreted program logic and / or machine code) stored in the data storage 504 to implement the various functionality described herein. The data storage 504 can include a non-transitory computer-readable medium having stored thereon program instructions that, when executed by the device 500, enable the device 500 to carry out any of the methods, processes, or functions disclosed in the specification and / or drawings. Execution of the program instructions 511 by the processor 503 can cause the processor 503 to utilize data 506.
[0141] For example, the program instructions 511 can include an operating system 510 (e.g., an operating system kernel, device drivers, and / or other modules) installed on the device 500 as well as one or more application programs 509 (e.g., a browser, a social application, or a game application). Similarly, the data 506 can include operating system data 508 and application data 507. The operating system data 508 is primarily accessible to the operating system 510, while the application data 507 is primarily accessible to the one or more application programs 509. The application data 507 can be in a file system that is visible or hidden to a user of the device 500.
[0142] The application programs 509 can communicate with the operating system 510 through one or more application programming interfaces (APIs). These APIs facilitate the application programs 509 in reading and / or writing application data 507, communicating or receiving information via the communication interface 501, receiving or displaying information on the user interface 502, and the like.
[0143] In some terminology, the application programs 509 can be referred to simply as “apps.” Furthermore, the application programs 509 can be downloaded to the device 500 through one or more online application stores or application markets. However, the application programs can also be installed on the device 500 through other means, such as through a web browser or a physical interface (e.g., a USB port) on the device 500.
[0144] Reference is made to Figure 6 , the risk assessment device can be applied to the device as shown in Figure 5 , to realize the technical solutions of the specification. Among them, the risk assessment device can include:
[0145] The acquisition module 601 is configured to acquire an initial change scheme for a target service and software design information corresponding to the target service; wherein the initial change scheme includes change content and implementation process of service changes to be implemented on the target service;
[0146] The matching module 602 is configured to determine change reference information matched with the initial change scheme in a pre-constructed knowledge base according to the initial change scheme;
[0147] The evaluation module 603 is configured to fill the change reference information, the software design information, and the initial change scheme into a preset prompt word template, to generate a prompt word based on the change reference information, the software design information, the initial change scheme, and prompt text in the prompt word template, and to input the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines, under the indication of the prompt text, whether the initial change scheme meets target change requirements reflected by the change reference information under the constraint of the software design information, to evaluate the initial change scheme and obtain an evaluation result; wherein the target change requirements include at least one of change release requirements, change monitoring requirements, and change recovery requirements.
[0148] The determination module 604 is configured to determine, according to the evaluation result, a change risk of implementing business change on the target business based on the initial change scheme, wherein the change risk includes an exception of the target business caused by implementing the initial change scheme without meeting the target change requirements.
[0149] Optionally, the change reference information includes a historical change scheme.
[0150] The evaluation of the initial change scheme specifically includes at least one of the following:
[0151] With a historical change scheme for reflecting corresponding change release requirements as a reference, whether the implementation process meets gradual change release requirements under the constraint of the software design information is evaluated;
[0152] With a historical change scheme for reflecting corresponding change monitoring requirements as a reference, whether the change content can generate externally observable data meeting the change monitoring requirements after being implemented under the constraint of the software design information is evaluated;
[0153] With a historical change scheme for reflecting corresponding change recovery requirements as a reference, whether the implementation process has a data rollback operation path meeting the change recovery requirements under the constraint of the software design information is evaluated.
[0154] Optionally, the change reference information includes a failed change scheme.
[0155] The evaluation module 603 is specifically configured to extract, from the failed change scheme, related information causing change failure;
[0156] Detect whether the initial change scheme involves the related information, to evaluate whether the initial change scheme has a risk of causing change failure.
[0157] Optionally, the change reference information comprises a preferred change scheme.
[0158] The evaluation module 603 is further configured to generate an improvement suggestion for the initial change scheme according to the evaluation result with reference to the preferred change scheme.
[0159] The apparatus further comprises a change module 605 configured to adjust the initial change scheme according to the improvement suggestion to obtain an adjusted change scheme, and determine a target change code for the target service according to the adjusted change scheme, and implement service change for the target service by using the target change code.
[0160] Optionally, the change module 605 is configured to input the adjusted change scheme and the related original code of the target service into a preset code generation model, analyze the change scheme by using the code generation model, and generate the target change code for the target service based on the analysis result and the related original code.
[0161] Optionally, the change module 605 is configured to input the adjusted change scheme and the related original code of the target service into a preset code generation model, analyze the change scheme by using the code generation model, and generate the target change code for the target service based on the analysis result and the related original code.
[0162] Optionally, the change module 605 is specifically configured to obtain an initial change code for the target service, input the adjusted change scheme and the initial change code into a preset code generation model, and optimize the initial change code by using the code generation model to obtain the target change code.
[0163] For ease of description, the above apparatus is described in various modules or units in terms of functions. Of course, the functions of each module or unit can be implemented in one or more software and / or hardware in implementing one or more of the present description, and the modules with the same function can be combined or integrated into another system, or some features can be ignored or not implemented.
[0164] Based on the same idea as the above method, the present description also provides an electronic device, comprising: a processor; a memory for storing processor executable instructions; wherein the processor implements the steps of the method according to any one of the above embodiments by running the executable instructions.
[0165] Based on the same idea as the above method, the specification also provides a computer readable storage medium, which stores computer instructions, and the instructions are executed by a processor to implement the steps of the method according to any one of the above embodiments.
[0166] Based on the same idea as the above method, the specification also provides a computer program product, which includes computer program / instructions, and the instructions are executed by a processor to implement the steps of the method according to any one of the above embodiments.
[0167] Those skilled in the art can understand that:
[0168] In the specification, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, product or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or further includes elements inherent to such a process, method, product or device. Without more limitations, it does not exclude the presence of other same or equivalent elements in the process, method, product or device including the elements.
[0169] In the specification, "one", "a" and "the" do not necessarily mean singular, but also include plural.
[0170] In the specification, the first, second, etc. ordinal numbers do not necessarily represent the order, and are often used for the purpose of distinguishing objects. For example, the first server and the second server usually refer to two servers. In order to distinguish the two servers, they are expressed as the first server and the second server. Of course, sometimes the two servers can be the same server.
[0171] In the specification, unless specifically stated, "receiving and sending of data" is not necessarily direct receiving and sending, but can be indirect receiving and sending. For example, A receives data sent by B, which can be understood as A directly receiving data sent by B, or A indirectly receiving data sent by B through C and other subjects. Similarly, B sends data to A, which can be understood as B directly sending data to A, or B indirectly sending data to A through C and other subjects. Here, C can be one subject, or two or more subjects.
[0172] In this specification, unless expressly stated otherwise, the connection relationship between structures can be a direct connection relationship or an indirect connection relationship. For example, when describing "A is connected with B", unless it is expressly stated that A is directly connected with B, it should be understood that A can be directly connected with B or indirectly connected with B; for another example, when describing "A is on B", unless it is expressly stated that A is directly on B (AB is adjacent and A is on B), it should be understood that A can be directly on B or A can be indirectly on B (there are other elements between AB and A is on B). By analogy.
[0173] The present specification uses specific terms to describe the embodiments of the present specification. As "one embodiment", "an embodiment", and / or "some embodiments" means a certain feature, structure, or characteristic in relation to at least one embodiment of the present specification. Therefore, it should be emphasized and noted that the "an embodiment" or "one embodiment" or "one alternative embodiment" mentioned in different positions in the present specification does not necessarily refer to the same embodiment. In addition, the skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples without contradiction.
[0174] Although one or more embodiments of the present specification provide method steps as described in the embodiments or flowcharts, it can be understood that the order of steps listed in the embodiments or flowcharts is only one of the many execution orders, and does not represent the only execution order. Therefore, when the method steps are involved in the claims, the adjustment of the order of the steps or the parallelism between the steps is also within the protection scope of the claims.
Claims
1. A risk assessment method, characterized by, The method comprises the following steps: acquiring an initial change scheme for a target service and software design information corresponding to the target service; wherein the initial change scheme comprises change content and an implementation process of a service change to be implemented on the target service; determining change reference information matched with the initial change scheme in a pre-constructed knowledge base according to the initial change scheme; filling the change reference information, the software design information and the initial change scheme into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change scheme and prompt text in the prompt word template, and inputting the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets target change requirements reflected by the change reference information under the constraint of the software design information under the indication of the prompt text, to evaluate the initial change scheme and obtain an evaluation result; wherein the target change requirements comprise at least one of change release requirements, change monitoring requirements and change recovery requirements; determining a change risk of implementing a service change on the target service based on the initial change scheme according to the evaluation result, wherein the change risk comprises an abnormality of the target service caused by implementing the initial change scheme without meeting the target change requirements.
2. The method of claim 1, wherein, The change reference information comprises historical change schemes; The evaluation of the initial change scheme specifically comprises at least one of the following: with reference to a historical change scheme reflecting corresponding change release requirements, evaluating whether the implementation process meets gradual change release requirements under the constraint of the software design information; with reference to a historical change scheme reflecting corresponding change monitoring requirements, evaluating whether the change content can generate externally observable data meeting the change monitoring requirements after being implemented under the constraint of the software design information; with reference to a historical change scheme reflecting corresponding change recovery requirements, evaluating whether the implementation process has a data rollback operation path meeting the change recovery requirements under the constraint of the software design information.
3. The method of claim 1, wherein, The change reference information comprises a failed change scheme; The evaluation of the initial change scheme specifically comprises: extracting related information causing change failure from the failed change scheme; detecting whether the initial change scheme involves the related information to evaluate whether the initial change scheme has a risk of causing change failure.
4. The method of claim 1, wherein, The change reference information comprises a preferred change scheme; The evaluation of the initial change scheme further comprises: with reference to the preferred change scheme, generating improvement suggestions for the initial change scheme according to the evaluation result.
5. The method of claim 4, wherein, The method further comprises: adjusting the initial change scheme according to the improvement suggestions to obtain an adjusted change scheme; determining target change code for the target service according to the adjusted change scheme, and implementing a service change on the target service by using the target change code.
6. The method of claim 5, wherein, According to the adjusted change scheme, a target change code for the target service is determined, specifically including: The adjusted change scheme and the related original code of the target service are input into a preset code generation model, so as to parse the change scheme by the code generation model, and generate the target change code for the target service based on the parsing result and the related original code.
7. The method of claim 5, wherein, According to the adjusted change scheme, a target change code for the target service is determined, specifically including: An initial change code for the target service is obtained; The adjusted change scheme and the initial change code are input into a preset code generation model, so as to optimize the initial change code by the code generation model, and obtain the target change code.
8. A service change apparatus characterized by comprising: Comprise: An acquisition module is configured to acquire an initial change scheme for a target service and software design information corresponding to the target service, wherein the initial change scheme includes change content and an implementation process of a business change to be implemented on the target service; A matching module is configured to determine change reference information matched with the initial change scheme in a pre-constructed knowledge base according to the initial change scheme; An evaluation module is configured to fill the change reference information, the software design information, and the initial change scheme into a preset prompt word template, to generate a prompt word based on the change reference information, the software design information, the initial change scheme, and prompt text in the prompt word template, and to input the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets target change requirements reflected by the change reference information under the constraint of the software design information under the indication of the prompt text, to evaluate the initial change scheme and obtain an evaluation result; wherein the target change requirements include at least one of change release requirements, change monitoring requirements, and change recovery requirements; A determination module is configured to determine a change risk of implementing a business change on the target service based on the initial change scheme according to the evaluation result, wherein the change risk includes an exception of the target service caused by implementing the initial change scheme without meeting the target change requirements.
9. An electronic device comprising: A processor; A memory for storing processor-executable instructions; characterized in that the processor executes the executable instructions to implement the steps of the method of any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, A computer program / instruction stored thereon, which is executed by a processor to implement the steps of the method of any one of claims 1-7.
11. A computer program product comprising computer programs / instructions, characterized in that, The computer program / instruction is executed by a processor to implement the steps of the method of any one of claims 1-7.
Citation Information
Patent Citations
Risk assessment method and device for code change and electronic equipment
CN119556965A
Intelligent code change recognition system, method and device based on large language model and medium
CN120596354A