Service verification method, system and equipment

Abstract

The invention belongs to the technical field of data analysis, and provides a service verification method, system and equipment in order to solve the problems of jamming, resource waste, poor real-time performance and the like existing in existing service verification, a dynamic rule dependency graph is constructed through combination of service feature data and rule definition related to a current service scene in a rule base; based on a rule dependency analysis result, combining with historical behavior data of the user, predicting through a behavior prediction model to obtain subsequent execution operations and required preloading verification rules, and preloading a priority queue based on the rules, and performing grammar analysis and compiling on the necessary verification rule script and the preloaded verification rule script of the current service scene to obtain a target code, and performing verification on the target code according to the rule execution priority sequence in combination with the service feature data. By predicting in advance and loading the verification logic as required, the service processing time is greatly shortened, and meanwhile, the waste of memory and network resources is reduced.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field related to data analysis, and particularly relates to a business verification method, system and device. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute prior art.

[0003] The existing system generally adopts a verification rule full-load mechanism, which loads all verification rules at one time during system initialization, including verification logic that is not needed in the current business scenario. At the same time, a static verification combination method is adopted, and the fixed verification rule set cannot adapt to the dynamic changing business scenario requirements, resulting in a serious lack of system flexibility. These two technical defects together cause the problem of high delay response in the business operation process, especially when the complex verification logic is temporarily loaded during the submission operation, causing obvious interaction lag phenomenon, which seriously affects the processing efficiency of the key business.

[0004] The technical root of the above problem lies in the static architecture design of the traditional system, the verification rule is strongly coupled with the core business, and the dynamic loading capability is lacking. The existing scheme is limited by the early technology stack, which cannot realize modularization splitting and does not support the dynamic import mechanism of modern ES6, resulting in the necessity of full-load of all verification rules. The deeper reason lies in the disconnection between business understanding and technical implementation: the traditional system regards verification as an independent link and fails to identify the continuity characteristics of the operation such as the link relationship of bill uploading→expense classification→budget checking, resulting in the inability to predictively load the required verification rules. In terms of performance optimization, excessive reliance on hardware upgrade and neglect of algorithm improvement, without distinguishing the priority of blocking / non-blocking verification, all verification is executed synchronously, causing delay of the key operation path. This design thinking has been unable to meet the requirements of real-time response and flexible configuration of modern business requirements. In summary, how to solve the problems of lag, delay and poor real-time performance in the existing business verification process is a problem to be solved at present. SUMMARY

[0005] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a business verification method, system and device, which greatly shortens the business processing time by predicting and loading verification logic in advance, and reduces the waste of memory and network resources.

[0006] In order to achieve the above purpose, the present application adopts the following technical scheme: In a first aspect, the present application provides a business verification method, comprising: Collecting current business key parameters and encapsulating to obtain business feature data; According to the business feature data, the verification rules are abstracted as nodes in the graph and the dependency relationships between the rules are abstracted as edges in the graph by combining the rule definitions related to the current business scenario in the rule library, a dynamic rule dependency graph is constructed, graph analysis is performed on the dynamic rule dependency graph, and a minimum rule set and a rule execution priority sequence under the current business scenario are obtained; Based on the rule dependency analysis result, the necessary verification rules and the preloaded verification rules for the current business scenario are subjected to syntax analysis and compilation to obtain target code, and the target code is executed according to the rule execution priority sequence in combination with the business feature data to perform verification. Based on the rule dependency analysis result, the necessary verification rules and the preloaded verification rules for the current business scenario are subjected to syntax analysis and compilation to obtain target code, and the target code is executed according to the rule execution priority sequence in combination with the business feature data to perform verification.

[0007] In a second aspect, the present application provides a business verification system, comprising: The acquisition module is configured to acquire and encapsulate current business key parameters to obtain business feature data; The graph construction module is configured to abstract the verification rules as nodes in the graph and the dependency relationships between the rules as edges in the graph according to the business feature data in combination with the rule definitions related to the current business scenario in the rule library, construct a dynamic rule dependency graph, perform graph analysis on the dynamic rule dependency graph, and obtain a minimum rule set and a rule execution priority sequence under the current business scenario. The prediction module is configured to analyze by a behavior prediction model based on the rule dependency analysis result in combination with user historical behavior data, predict the subsequent operations and the required preloaded verification rules, generate a rule preloaded priority queue based on the prediction result. The verification module is configured to perform syntax analysis and compilation on the necessary verification rule scripts and the preloaded verification rule scripts for the current business scenario to obtain target code, and execute the target code according to the rule execution priority sequence in combination with the business feature data to perform verification.

[0008] In a third aspect, the present application provides an electronic device comprising a memory and a processor, and computer instructions stored on the memory and running on the processor, wherein when the computer instructions are run by the processor, the method of the first aspect is completed.

[0009] The above one or more technical solutions have the following beneficial effects: In the present application, by combining the service feature data with the rules related to the current service scenario in the rule library to define the dynamic rule dependency graph, the minimum rule set can be matched according to the real-time service scenario, and the resource waste and response delay caused by loading all the verification rules are avoided; based on the rule dependency analysis result, combined with the user historical behavior data, the subsequent operation and the required preloaded verification rules are obtained by analyzing and predicting through the behavior prediction model, and through the early prediction and loading mode, the operation can be directly executed at the time of submission, solving the lag caused by temporary loading, and at the same time, only the current and predicted rules are loaded, avoiding the resource waste caused by full loading; based on the rule preloading priority queue, the necessary verification rule scripts and the preloaded verification rule scripts of the current service scenario are subjected to syntax analysis and compilation to obtain the target code, and the target code is executed according to the rule execution priority sequence in combination with the business feature data to perform verification. Through the early prediction and on-demand loading of the verification logic, the business processing time is greatly shortened, and the memory and network resource waste is reduced. The present application compiles the new verification rules into target code in real time for subsequent engine calling, without the need to restart the whole system or interrupt the current business operation, so as to ensure that the compliance adjustment takes effect immediately.

[0010] Advantages of the additional aspects of the present application will be partially given in the following description, partially will become obvious from the following description, or will be learned by the practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0011] The accompanying drawings, which form a part of the specification, are included to provide a further understanding of the application and are incorporated herein by reference. The illustrations are shown for the purpose of enabling those skilled in the art to

[0012] Figure 1 The business verification method flowchart in the embodiment one of the present application. DETAILED DESCRIPTION

[0013] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0014] It should be noted that the terms used herein are only for the purpose of describing the specific embodiments, and are not intended to limit the exemplary embodiments according to the present application.

[0015] In the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0016] Embodiment one The present embodiment discloses a business verification method, comprising: Collecting the current business key parameters and encapsulating to obtain the service feature data; According to the business feature data, the verification rules are abstracted as nodes in the graph and the dependency relationship between the rules is abstracted as edges in the graph by combining the rule definitions related to the current business scene in the rule library, a dynamic rule dependency graph is constructed, the minimum rule set and the rule execution priority sequence under the current business scene are obtained by graph analysis on the dynamic rule dependency graph; Based on the rule dependency analysis result, the subsequent operation and the required preloaded verification rule are predicted by analyzing through the behavior prediction model, and a rule preloading priority queue is generated based on the prediction result; Based on the rule preloading priority queue, the necessary verification rule script and the preloaded verification rule script for the current business scene are parsed and compiled to obtain target code, and the target code is executed according to the rule execution priority sequence to perform verification in combination with the business feature data.

[0017] In this embodiment, a dynamic rule dependency graph is constructed by combining the business feature data with the rule definitions related to the current business scene in the rule library, and the minimum rule set can be matched according to the real-time business scene, avoiding resource waste and response delay caused by full-load verification rules; based on the rule dependency analysis result, the subsequent operation and the required preloaded verification rule are predicted by analyzing through the behavior prediction model, and a rule preloading priority queue is generated based on the prediction result; by predicting and loading in advance, the operation can be directly executed at the time of submission, solving the lag caused by temporary loading, and by loading only the current and predicted rules, the resource waste caused by full-load is also avoided; based on the rule preloading priority queue, the necessary verification rule script and the preloaded verification rule script for the current business scene are parsed and compiled to obtain target code, and the target code is executed according to the rule execution priority sequence to perform verification in combination with the business feature data, by compiling the new verification rule into target code for subsequent engine calling in real time, the entire system does not need to be restarted or the current business operation does not need to be interrupted, ensuring that the compliance adjustment takes effect immediately.

[0018] The overall system architecture of this embodiment includes five main modules: business feature perception engine, rule dependency analyzer, predictive loading controller, hot update compiler and verification rule execution engine, based on the overall architecture Figure 1 A business verification method proposed in this embodiment is described in detail: Step 1: Collect the current business key parameters and encapsulate them to obtain business feature data.

[0019] As a specific implementation, the business feature perception engine serves as the input perception layer, responsible for collecting business key parameters and encapsulating the business key parameters into JSON format to provide business scene portraits for subsequent decision-making, which is the basis for dynamic adaptation. The business feature perception engine uses artificial intelligence technology to dynamically perceive business features and infer specific application scenarios, thereby providing a basis for subsequent decision-making.

[0020] Among them, the business key parameters include the stage such as the operation stage, the load, the save, the submission and the like, the amount, the type such as the business type.

[0021] In this embodiment, the extracted amount feature is logarithmically classified to determine the order of magnitude range of the amount, such as 10^3, 10^4, 10^5 and the like. The logarithmic classification of the amount helps to divide the numerical range into discrete levels, facilitating subsequent decision-making. This judgment is represented by a diamond decision block, and different processing branches are entered according to whether the amount exceeds a certain threshold.

[0022] Step 2: According to the business feature data, the verification rules are abstracted as nodes in the graph and the dependency relationship between the rules is abstracted as edges in the graph, a dynamic rule dependency graph is constructed, the dynamic rule dependency graph is analyzed by graph analysis, and the minimum rule set and the rule execution priority sequence under the current business scenario are obtained.

[0023] In this embodiment, the rules definition in the rule library that matches the current business scenario is obtained, each independent verification rule is abstracted as a node in the graph, the dependency relationship between the rules is abstracted as an edge in the graph, and a dynamic rule dependency graph is constructed. The dynamic rule dependency graph is analyzed by using a graph traversal algorithm to filter out the minimum rule set. According to the minimum rule set, the rule dependency relationship and the rule importance, a rule execution priority sequence is generated.

[0024] Specifically, according to the business feature data, a basic verification rule set suitable for the current business scenario is matched and generated from the rule library. For example, different business types and stages may correspond to different default verification rules. The basic verification set contains general rules that need to be executed regardless of the amount.

[0025] As an optional implementation, according to the result of the amount order judgment, a parallel path appears in the process. If the amount is in a lower order, for example, less than or equal to 10^3, only the ordinary rules in the basic verification set are executed. The ordinary rules in the basic verification set are usually regular verification logic and can quickly complete the verification. If the amount exceeds a certain threshold, for example, greater than 10^3, additional audit rules are executed in addition to the basic verification. Audit rules are usually more stringent or involve higher-level approval and are used to screen high-risk transactions. The additional rules and the ordinary rules are executed in parallel to ensure that all necessary verifications are covered in the case of high amount.

[0026] In this embodiment, the rule execution priority sequence is the object of the verification rule execution engine. When the hot update compiler completes the compilation, the execution engine loads the bytecode, and the execution order of the current necessary verification rules is determined according to the rule execution priority sequence, ensuring the efficiency and accuracy of the current business verification.

[0027] As a specific implementation, the rule dependency analyzer is responsible for analyzing the dependency relationship between the verification rules and generating a dynamic rule dependency graph. The rule dependency analyzer receives the business feature data from the business feature engine, constructs the rule dependency graph in combination with the rule definitions in the rule library, such as storing the rules and their dependencies using the Neo4j graph database. Through graph analysis techniques, the rule dependency analyzer can calculate the minimum rule set, i.e., the minimum rule set that actually needs to be executed in the current business scenario, avoiding redundant verification. The rule dependency analyzer outputs GraphQL queries or structured dependency information for the predictive loading controller. The rule dependency analyzer supports hot deployment of verification rules, and when the verification rules are updated, the dynamic rule dependency graph is automatically updated and the relevant modules are notified.

[0028] Specifically, after receiving the business feature data, the rule dependency analyzer filters out the candidate rule set related to the current business scenario from the rule library, and then constructs a dynamic rule dependency graph for the current scenario based on the pre / post dependency relationship between rules and data transmission requirements.

[0029] Through graph analysis techniques, the topology and node association of the dynamic rule dependency graph are calculated to obtain the minimum rule set and the rule execution priority sequence.

[0030] The minimum rule set is the minimum rule combination that meets all necessary verification requirements and has no redundant rules in the current business scenario. The graph analysis calculation logic is: taking the core verification target corresponding to the business feature data as the starting point, traversing the nodes and edges of the dynamic rule dependency graph, eliminating rules unrelated to the current scenario, eliminating redundant rules that can be replaced by other rules or do not need to be executed repeatedly, and finally obtaining a set containing only necessary verification rules.

[0031] The graph analysis technique generates a priority order for rule execution by the directional relationship of the edges in the dynamic rule dependency graph and the importance weight of the nodes, such as the amount of money and the direct impact on subsequent operations. The weight is higher than the note information verification. For example, in the high-amount reimbursement scenario, the rule dependency analyzer identifies through the graph that the invoice information integrity verification is a pre-requisite for the invoice authenticity verification, and both are core rules, so the invoice information integrity verification is set as the highest priority, the invoice authenticity verification is set as the second highest priority, and the note information verification is set as the low priority, ensuring that critical verification logic is executed first and reducing waiting time.

[0032] Step 3: Based on the rule dependency analysis results, combined with user historical behavior data, through behavior prediction model analysis, the subsequent operation and required preloaded verification rules are predicted, and a rule preloading priority queue is generated based on the prediction results.

[0033] In this embodiment, the behavior prediction model is used to analyze the matching records of the user's past business operation links, historical business scenarios and corresponding verification rules, capture the user's operation habit rules, and predict the subsequent operations and required preloaded verification rules in combination with the rule dependency analysis results of the current scenario; the priority weight is set in combination with the rule importance and historical calling frequency, the preloaded verification rules predicted are sorted according to the priority weight, and a rule preloading priority queue is generated.

[0034] In this embodiment, the rule importance is determined from the business characteristics, the rule importance is determined from the business impact degree, the compliance risk level and the verification result correlation; the historical calling frequency is determined according to the effective calling number of a single rule in a statistical period, the daily average calling number and the peak calling period frequency; the priority weight is set according to the rule importance and the historical calling frequency, the preloaded verification rules are sorted, and a rule preloading priority queue is generated.

[0035] Specifically, on the basis of minimizing the rule set, the core basis for generating the rule preloading priority queue is exactly the rule dependency relationship and the rule importance, which jointly determine the execution order of the verification rules.

[0036] For example, the behavior prediction model predicts that the next step is a save operation, and then preloads the verification rules required in the save stage. This feedback loop enables the system to dynamically optimize the subsequent verification process and reduce user waiting time. Based on the prediction results, the rule modules and resources required in the next business link are preloaded. For example, the rule script of the next stage is compiled into WASM code in advance and loaded into the memory. The preloading operation ensures that when the user actually enters the next link, the related rules are ready for immediate execution of verification, thereby improving the system response speed.

[0037] As a specific implementation, the predictive loading controller predicts the rules or modules that need to be loaded next based on the rule dependency analysis results and user historical behavior data, and prepares for loading in advance. The predictive loading controller includes an LSTM behavior prediction model that predicts the operation that may be performed next and the required rules by analyzing the user's historical operation sequence. The prediction results are used to generate a priority queue, and the predictive loading controller preloads the high-priority rule modules according to the priority queue, thereby shortening the waiting time for subsequent verification. There is a control feedback between the predictive loading controller and the rule dependency analyzer: when new rule dependency relationships are generated, the predictive loading controller can re-predict the loading requirements. The predictive loading controller outputs preloading instructions or directly loads related rules, and communicates with the hot update compiler through efficient protocols such as Protobuf to trigger the pre-compilation of rule code.

[0038] Specifically, the user historical behavior data includes a user's past business operation link, historical business scene and matching record of corresponding verification rules. The user historical behavior data serves as a basis sample for time series analysis, helping the LSTM model capture user operation habit rules, ensuring that the prediction result fits the user's actual operation logic, and avoiding invalid prediction that deviates from the actual use scene. Through standardization processing of the user historical behavior data, such as converting the user historical operation sequence, e.g., operation type and business stage, into a vector format, and mapping the structured data, such as rule type and scene label in the rule dependency analysis result, into features recognizable by the model, it is ensured that the two types of input data can participate in analysis cooperatively. The long short-term memory behavior prediction model predicts the most likely subsequent operation according to the user historical behavior data and the rule dependency analysis result, and reversely deduces the preloaded verification rules required for the operation.

[0039] Specifically, in combination with the predicted verification rules and the current rule dependency relationship, priority weights are set according to rule importance and historical calling frequency; the predicted preloaded verification rules are sorted according to the priority weights to generate a rule preloading priority queue, ensuring that high-priority rules are loaded first and avoiding waste of resources on rules with low probability of use.

[0040] For example, in combination with the characteristics of financial business, rule importance is defined from three core dimensions of business impact degree, compliance risk level and verification result relevance, and the importance score of a single rule is calculated by a weighted summation formula. The business impact degree is determined by evaluating the blocking range of rule failure on the financial business process; the verification result relevance is determined by evaluating the influence of rule verification result on subsequent rule execution; and the compliance risk level can be evaluated by referring to the financial and tax system and the enterprise internal financial system.

[0041] The historical calling frequency is quantified by three index quantities, i.e., the effective calling number of a single rule in a statistical period, the daily calling number and the peak calling period frequency, the three indexes are normalized, and then the calling frequency score is obtained by weighted summation. The statistical period can be configured according to the business scene.

[0042] The priority weight of the verification rule is determined according to the rule importance score and the historical calling frequency score.

[0043] It should be noted that the rule preloading priority queue is a priority sorting result of the verification rules required for the subsequent execution operation predicted, but also contains the priority of the necessary verification rules of the current business scenario. To ensure that the current business verification is not delayed, the necessary verification rules of the current business scenario, that is, the minimum rule set calculated by the rule dependency analyzer, are included in the head of the queue and are given the highest priority. This design makes the queue cover both the current urgently needed rules and the subsequent predicted rules, but the core sorting logic still revolves around the verification rules of the subsequent execution operation, that is, the high priority of the current necessary rules is the basis for guaranteeing the current process, and the sorting of the subsequent predicted rules is the core of realizing preloading zero waiting, and the two together constitute the complete sorting system of the queue.

[0044] After the predictive loading controller receives the minimum rule set, it will use it as a basis to only pre-load and compile the rule modules in the set, rather than loading the entire rule library, thereby greatly reducing the memory occupation and network requests.

[0045] Step 4: Based on the rule preloading priority queue, the necessary verification rule scripts and the preloaded verification rule scripts of the current business scenario are syntax parsed and compiled to obtain target code, and the target code is executed according to the rule execution priority sequence in combination with the business feature data to perform verification.

[0046] In the embodiment, the necessary verification rule scripts and the preloaded verification rule scripts of the current business scenario are syntax parsed, and the verification logic in the scripts is converted into a structured abstract syntax tree; based on the rule preloading priority queue, the abstract syntax tree is compiled into target code in the form of bytecode; when a business operation occurs, the target code is executed according to the rule dependency order to complete verification according to the real-time business context data.

[0047] As a specific implementation, the hot update compiler is responsible for real-time compiling the necessary verification rule scripts and the preloaded verification rule scripts of the current business scenario into executable code and supporting hot updates at runtime. The hot update compiler receives the preloaded verification rule module list from the predictive loading controller, uses an abstract syntax tree parser to parse the rule scripts, and then generates optimized target code such as WebAssembly bytecode. The WASM bytecode has the characteristics of cross-platform and high performance and can be quickly executed in a browser or server environment. After compilation, the hot update compiler pushes the new rule code to the verification rule execution engine to realize hot replacement of the rules without restarting the system. The entire compilation process is performed asynchronously in the background, and when the new rules are ready, the verification rule execution engine seamlessly switches to the new version of the rules, thereby realizing hot updates. The hot update technology allows updating modules without reloading the entire system, improving development and deployment efficiency.

[0048] The abstract syntax tree parser parses the verification rule script, specifically: the verification rule script is split into the smallest syntax unit, the Token sequence is combined into a structured abstract syntax tree according to the predefined verification rule syntax specification, and a hierarchical clear syntax structure is formed. Taking the example of executing invoice verification when the amount > 5000, the root node of the AST is "conditional judgment", and the child nodes are "left operand, i.e. amount", "operator, i.e. >", "right operand, i.e. 5000", and "node executed when the condition is met, i.e. invoice verification logic".

[0049] The task triggering and processing sequence of the hot update compiler is directly determined by the rule preloading priority queue generated by the predictive loading controller. The rule preloading priority queue clearly distinguishes the priority of the current necessary verification rule and the preloaded verification rule. The hot update compiler will preferentially process the scripts with high priority in the queue, for example, the current necessary rule has the highest priority and is preferentially compiled; the rule that is most likely to be used in the next step in the preloaded rule is second, and then it is compiled.

[0050] As a specific implementation, after the hot update compiler receives the preloaded rule module list, it starts a background asynchronous compilation process, and the specific operation is as follows: The abstract syntax tree parser is used to perform syntax analysis on the rule script, and the rule is converted into a structured abstract syntax tree, which facilitates subsequent code optimization and generation; Based on the syntax tree, compilation optimization is performed, such as eliminating redundant logic and optimizing condition judgment order, to generate WebAssembly bytecode. Because the WASM bytecode has the characteristics of cross-platform and high performance, it can meet the verification speed requirements of financial business; After compilation is completed, the hot update compiler performs integrity checking on the WASM bytecode, such as checking for syntax errors, missing dependencies, etc., to ensure that the generated rule code can be normally executed.

[0051] As a specific implementation, the verification rule execution engine serves as the execution core and is responsible for actually executing the verification rules and returning the results. The verification rule execution engine obtains the latest rule code such as the WASM module from the hot update compiler and loads it into the running environment. When a business operation occurs, the execution engine executes the corresponding verification logic according to the context data provided by the business feature engine in the order of rule dependencies. The verification rule execution engine traverses the dynamic rule dependency graph, triggers the execution of related rules in turn, and collects the verification results of all rules. The verification rule execution engine supports concurrent execution and priority scheduling to ensure that critical rules are executed first. Finally, the verification rule execution engine returns the verification results, such as pass or reject, to the front-end user interface or the calling party. The verification rule execution engine usually has a built-in rule execution runtime, including rule matching and execution queue management mechanisms, to efficiently handle the real-time execution of a large number of rules.

[0052] In this embodiment, after the hot update compiler completes the rule compilation, the WASM bytecode that passes the verification is pushed to the verification rule execution engine to realize rule hot replacement and loading, and the specific process is as follows: After the verification rule execution engine receives the WASM bytecode, the new rule code is loaded into the running environment, such as the rule execution container in the memory, without restarting the system, while the old rule being executed is kept to avoid affecting the current business verification; When the new rule is loaded and confirmed to be able to run normally, the verification rule execution engine automatically routes the subsequent verification request to the pre-compiled rule loaded newly through a seamless switching mechanism, and completes the hot deployment; The verification rule execution engine will initialize and prepare the preloaded rule, such as allocating computing resources in advance, establishing a context data docking channel with the business feature perception engine, to ensure that when the user triggers the pre-judgment operation, the preloaded rule can be called immediately to execute the verification, realizing zero waiting response.

[0053] After the verification rule execution engine receives the rule execution priority sequence, it will schedule the rule execution according to the sequence order, while supporting concurrent execution of high-priority rules to avoid the lag caused by traditional synchronous execution of all rules. In addition, the priority sequence can also ensure that the key verification logic is given priority to feedback the results, improving the user operation experience, such as the user not needing to wait for the execution of low-priority rules to know in advance whether the core verification is passed.

[0054] In this embodiment, the verification and preloading preparation of the current business operation are completed, and the process ends. When the user performs subsequent operations, the process will repeat the above scheme to realize dynamic verification support for the whole business process.

[0055] As an optional implementation, the verification rule is transmitted based on a differential update protocol. The differential update protocol is a key technical protocol supporting the hot update and efficient resource transmission of the verification rule. The core role is to transmit only the difference between the new and old rule codes when the verification rule is updated, rather than the full code, thereby greatly saving network traffic and shortening the update time, ensuring the real-time and efficiency of rule hot update.

[0056] As an optional implementation, the quantum compression algorithm is used to store and process the verification rules in the rule library, reducing the memory occupation by 60%. Finally, edge computing support is developed to allow part of the verification logic to be executed on the terminal device. These enhancements enable the system to maintain a verification response speed of 200ms within 10 million data volume scenarios.

[0057] Embodiment Two The purpose of this embodiment is to provide a business verification system, which comprises: The collection module is configured to collect and encapsulate current service key parameters to obtain service feature data. The graph construction module is configured to abstract the verification rules as nodes in a graph and the dependency relationship between the rules as edges in the graph according to the service feature data and the rules related to the current service scenario in the rule library, construct a dynamic rule dependency graph, perform graph analysis on the dynamic rule dependency graph, and obtain a minimum rule set and a rule execution priority sequence under the current service scenario. The prediction module is configured to analyze and predict the subsequent operation and the required preloaded verification rules by a behavior prediction model based on the rule dependency analysis result and the user historical behavior data, and generate a rule preloading priority queue based on the prediction result. The verification module is configured to perform syntax analysis and compilation on the necessary verification rule scripts and the preloaded verification rule scripts of the current service scenario to obtain target code, and execute the target code according to the rule execution priority sequence to perform verification in combination with the service feature data.

[0058] In more embodiments, there are also provided: An electronic device includes a memory and a processor, and computer instructions stored on the memory and run on the processor, when the computer instructions are run by the processor, the method described in embodiment one is completed. For brevity, this will not be repeated here.

[0059] It should be understood that in the embodiments, the processor can be a central processing unit CPU, and the processor can also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, ready-to-program gate arrays FPGA or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0060] The memory can include read-only memory and random access memory, and provide instructions and data to the processor, and a part of the memory can also include non-volatile random access memory. For example, the memory can also store device type information.

[0061] A computer readable storage medium for storing computer instructions, when the computer instructions are executed by a processor, the method described in embodiment one is completed.

[0062] The method in the embodiment one can be directly embodied as being completed by a hardware processor or being completed by a combination of hardware and software modules in the processor. The software modules can be located in a storage medium in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, a register, or the like. The storage medium is located in a memory, and a processor reads information in the memory and combines hardware to complete the steps of the above method. To avoid repetition, no further detailed description is given here.

[0063] Those skilled in the art can understand that the units and algorithm steps of the examples described in combination with the embodiments can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software manner depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0064] Although the specific embodiments of the present application are described above in combination with the drawings, the description is not a limitation on the scope of protection of the present application. Those skilled in the art should understand that various modifications or changes made by those skilled in the art on the basis of the technical solutions of the present application without creative labor are still within the scope of protection of the present application.

Claims

1. A service verification method, characterized in that: include: Collect and encapsulate key business parameters to obtain business characteristic data; Based on business feature data and rule definitions related to the current business scenario in the rule base, the verification rules are abstracted as nodes in a graph, and the dependencies between rules are abstracted as edges in the graph. A dynamic rule dependency graph is constructed, and graph analysis is performed on the dynamic rule dependency graph to obtain the minimum rule set and rule execution priority sequence for the current business scenario. Based on the rule dependency analysis results and combined with historical user behavior data, the behavior prediction model is used to analyze and predict subsequent operations and the required preload verification rules. A rule preload priority queue is generated based on the prediction results. Based on the rule preloading priority queue, the necessary verification rule scripts and preloaded verification rule scripts of the current business scenario are parsed and compiled to obtain target code, and the target code is executed according to the rule execution priority sequence in combination with the business feature data for verification.

2. A service verification method according to claim 1, characterized in that: Based on the business feature data and combined with the rule definitions related to the current business scenario in the rule base, the verification rules are abstracted as nodes in the graph, and the dependencies between the rules are abstracted as edges in the graph. A dynamic rule dependency graph is constructed and graph analysis is performed on the dynamic rule dependency graph to obtain the minimum rule set and rule execution priority sequence for the current business scenario, specifically: Obtain the rule definitions that match the current business scenario from the rule library, abstract each independent verification rule into a node in the graph, and abstract the dependencies between rules into edges in the graph to build a dynamic rule dependency graph; Use graph traversal algorithm to analyze dynamic rule dependency graph and filter out the minimum rule set; According to the minimum rule set, combined with rule dependencies and rule importance, a rule execution priority sequence is generated.

3. A service verification method according to claim 1, characterized in that: Based on the rule dependency analysis results, combined with user historical behavior data, the behavior prediction model is used to analyze and predict the subsequent operations and the required preload verification rules. Based on the prediction results, a rule preload priority queue is generated, specifically: Use behavioral prediction models to analyze the user's past business operation links, historical business scenarios, and matching records with corresponding verification rules, capture user operation habits, and combine the rule dependency analysis results of the current scenario to predict subsequent operations and the required pre-loaded verification rules; Priority weights are set based on rule importance and historical call frequency, and the predicted preload verification rules are sorted by priority weights to generate a rule preload priority queue.

4. A service verification method according to claim 1, characterized in that: Based on the rule preloading priority queue, the necessary verification rule scripts for the current business scenario and the preloaded verification rule scripts are parsed and compiled to obtain target code. The target code is executed and verified according to the rule execution priority sequence in combination with the business feature data to achieve dynamic verification of the entire business process, specifically: Perform syntax analysis on the necessary validation rule scripts and pre-loaded validation rule scripts for the current business scenario, and convert the validation logic in the scripts into a structured abstract syntax tree. Preloading the priority queue based on the rules and compiling the abstract syntax tree into target code in the form of bytecode; When a business operation occurs, the target code is executed in the order of rule dependencies based on real-time business context data to complete the verification.

5. A service verification method according to claim 3, characterized in that: Priority weights are set based on rule importance and historical call frequency. The predicted preload verification rules are sorted by priority weights to generate a rule preload priority queue. Specifically: Combined with business characteristics, determine the importance of rules based on business impact, compliance risk level, and verification result relevance; Determine the historical call frequency based on the number of valid calls, average daily call times, and peak call period frequency of a single rule within the statistical period; Priority weights are set based on rule importance and historical call frequency, and the predicted preload verification rules are sorted to generate a rule preload priority queue.

6. A service verification method according to claim 1, characterized in that: Also includes: When the user enters the next business link, the business feature data of the new scenario is obtained in real time. The dynamic rule dependency graph is rebuilt and analyzed based on the business feature data of the new scenario. Based on the results of the rule dependency analysis, the rule preloading priority queue is obtained to realize dynamic verification of the entire business process.

7. A service verification method according to claim 1, characterized in that: The rule preloading priority queue contains the priorities of the necessary verification rules for the current business scenario.

8. A service verification method according to claim 1, characterized in that: Also includes: When the validation rules in the rule base are updated, the dynamic rule dependency graph is updated, and the updated information is synchronized to the preloaded validation rule prediction and validation rule execution; The verification rules in the rule base are stored and processed using a quantized compression algorithm.

9. A business verification system, characterized in that: include: The acquisition module is configured to: collect and encapsulate current business key parameters to obtain business feature data; The graph construction module is configured to: based on the business feature data and in combination with the rule definitions related to the current business scenario in the rule library, abstract the verification rules into nodes in the graph, abstract the dependencies between the rules into edges in the graph, construct a dynamic rule dependency graph, and perform graph analysis on the dynamic rule dependency graph to obtain the minimum rule set and rule execution priority sequence for the current business scenario; The prediction module is configured to: based on the rule dependency analysis results, combined with the user's historical behavior data, analyze through the behavior prediction model, predict the subsequent execution operations and the required preload verification rules, and generate a rule preload priority queue based on the prediction results; The verification module is configured to: based on the rule preloading priority queue, perform syntax parsing and compilation on the necessary verification rule scripts and preloaded verification rule scripts of the current business scenario to obtain target code, and execute the target code according to the rule execution priority sequence in combination with the business feature data for verification.

10. An electronic device, characterized in that: The method comprises a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the method according to any one of claims 1 to 8 is completed.

Images