A dynamic knowledge consistency governance and gating control method based on dependency graph

Abstract

The application discloses a dynamic knowledge consistency management and gating control method based on a dependency graph, comprising the following steps: acquiring source knowledge data, downstream dependency data and knowledge state structure information, constructing a dependency graph, detecting that the source knowledge data is added, changed, invalidated, replaced or state-switched, determining an affected subgraph, performing differential consistency verification on downstream dependency nodes in the affected subgraph, generating a control result according to the verification result and the node type, and writing the control result into a management record to perform update control, query control, output control or state-switching control. The method can improve the consistency management efficiency, control accuracy and closed-loop control capability in a dynamic knowledge environment.

Description

Technical Field

[0001] This invention relates to the fields of information processing, data governance, and access control technology, specifically to a dynamic knowledge consistency governance and gating control method based on dependency graphs. Background Technology

[0002] In scenarios such as educational informatization, enterprise system management, process management, rule management, knowledge management, and cross-system data services, there are often datasets that serve as upstream rules or master data, such as teaching plan data, course data, system and rule data, process rule data, fee rule data, and salary rule data. Downstream dependent data typically includes task data, enrollment data, grade data, statistical data, template data, settlement data, and access control data.

[0003] When upstream knowledge or rules are added, modified, become invalid, replaced, or their status changes, the following problems can easily occur in downstream dependent data: downstream dependent data still references knowledge content before the change, resulting in inconsistent reference values; derived statistical results are still generated based on rules before the change, resulting in inconsistent calculation rules; access control objects, output objects, or lifecycle objects do not switch synchronously with upstream status changes, resulting in inconsistent status attributes; new knowledge lacks access control, making it easy to write duplicate or conflicting knowledge. Existing processing methods often remain at the stage of rule detection, early warning prompts, or repair work orders, making it difficult to form a closed loop that continuously participates in the control of subsequent writing, querying, output, and status switching.

[0004] Existing technologies typically employ rule-based detection, data quality scanning, lineage analysis, or message synchronization to address these issues. While these methods can initially detect anomalies, they have significant limitations: First, they primarily rely on full scans, lacking the ability to perform differential processing on affected objects. Second, after anomaly detection, only warnings or remediation tasks are generated, and the governance results cannot continuously participate in subsequent control, resulting in insufficient closed-loop capabilities. Third, they fail to unify reference dependencies, computation dependencies, output dependencies, and state inheritance dependencies into the same governance chain. Fourth, they do not adequately utilize the version status, effective status, and applicable scope information of dynamic knowledge objects, making it difficult to adapt to the consistency governance requirements in complex scenarios.

[0005] Therefore, a new dynamic knowledge consistency governance and gating control method is needed to quickly identify affected objects when source knowledge data is added or changed, perform differential consistency verification on the objects, and automatically generate executable control results based on the verification results, thereby improving governance efficiency, control accuracy and closed-loop capability. Summary of the Invention

[0006] This invention addresses the shortcomings of existing technologies by providing a dynamic knowledge consistency governance and gating control method based on dependency graphs. This method solves the problems of mismatched downstream dependency data references, invalidation of calculation rules, asynchronous state attributes, uncontrolled access to duplicate knowledge, and inability to reuse governance results in a closed loop after changes in source knowledge data.

[0007] Another objective of this invention is to provide a closed-loop mechanism that writes governance results into governance records and continues to execute update control, query control, output control, and state switching control based on the governance records, thereby improving software governance and gating control capabilities in a dynamic knowledge environment.

[0008] To achieve the above objectives, this invention provides a dynamic knowledge consistency governance and gating control method based on dependency graphs, comprising the following steps: S1. Obtain source knowledge data, downstream dependency data, and knowledge state structure information corresponding to the source knowledge data. The knowledge state structure information is used to characterize the version status, validity status, and applicable scope information of the source knowledge data. S2. Based on the source knowledge data, the downstream dependency data, and the knowledge state structure information, construct a dependency graph describing the relationship between the source knowledge data and the downstream dependency data; S3. Based on the knowledge state structure information, detect whether the source knowledge data has been added, changed, invalidated, replaced, or switched in state. When the source knowledge data is detected to have been added, changed, invalidated, replaced, or switched in state, determine the affected subgraph based on the comprehensive influence of the dependency graph on the source knowledge data. S4. Perform differential consistency verification on the downstream dependent nodes in the affected subgraph to obtain the differential consistency verification result of the downstream dependent data corresponding to each downstream dependent node relative to the source knowledge data after the addition, change, failure, replacement or state switch. S5. Based on the differential consistency verification result and the node type of the downstream dependent node, and according to the preset control action state machine, generate the corresponding control result. The control result includes at least one of the following: early warning record, recalculation mark, recalculation task, write restriction control instruction, modification restriction control instruction, output restriction control instruction, and archive state switching instruction. S6. Write the control result, the source knowledge data corresponding identifier that triggered the control result, the downstream dependent data corresponding identifier, and the corresponding knowledge state structure information into the governance record, and execute the corresponding update control, query control, output control, or state switching control based on the governance record.

[0009] To optimize the above technical solution, the specific measures also include: Furthermore, the knowledge status structure information includes at least three of the following: version identifier, effective time, expiration time, substitution relationship, conflict relationship, scope of application information, and status identifier; The version identifier is used to distinguish different versions of the same source knowledge data; the effective time and expiration time are used to determine the effective status of the source knowledge data; the substitution relationship and conflict relationship are used to characterize the substitution and conflict constraints between different source knowledge data; the scope of application information is used to characterize the scope of the associated subjects corresponding to the source knowledge data; and the status identifier is used to characterize which of the following states the source knowledge data is currently in: editable, available, restricted, or archived.

[0010] Furthermore, the dependency graph has nodes, which include at least two types of nodes: source knowledge nodes, downstream dependency nodes, derived statistics nodes, access control nodes, and lifecycle nodes. The dependency graph has edges, and the edges include at least one of reference dependencies, computation dependencies, output dependencies, and state inheritance relationships; The reference dependency relationship is used to characterize the reference of downstream dependent data to the content of source knowledge data; the computation dependency relationship is used to characterize the downstream dependent data to perform computation based on the source knowledge data to generate derived results; the output dependency relationship is used to characterize the output of the downstream dependent data to be constrained by the source knowledge data; and the state inheritance relationship is used to characterize the state of the downstream dependent data to be affected by changes in the state of the source knowledge data.

[0011] Furthermore, the process of determining the affected subgraph based on the dependency graph in step S3 includes: Calculate the overall influence of the source knowledge node on each downstream dependent node; include downstream dependent nodes with an overall influence of not less than a preset influence threshold into the affected subgraph; The overall influence is determined by at least two of the following: reference dependency strength, computation dependency strength, output dependency strength, and state inheritance dependency strength.

[0012] Furthermore, the differential consistency check in step S4 includes: For downstream dependent nodes that have a reference dependency relationship with the source knowledge data, perform reference value consistency verification to determine whether the referenced content in the corresponding downstream dependent data is consistent with the currently valid source knowledge data; For downstream dependent nodes that have computational dependencies on the source knowledge data, perform computation rule consistency checks to determine whether the corresponding downstream dependent data is still generated based on the rules before the change. For downstream dependent nodes that have state inheritance relationships or output dependency relationships with the source knowledge data, perform state attribute consistency checks to determine whether the current state or output permissions in the corresponding downstream dependent data are consistent with the current state of the source knowledge data.

[0013] Furthermore, the process of generating the corresponding control result according to the preset control action state machine in step S5 includes: Based on the differential consistency verification results, calculate the anomaly level value of each downstream dependent node in the affected subgraph; Based on the anomaly level range where the anomaly level value is located and the node type of the corresponding downstream dependent node, select the target control result from the warning record, recalculation flag, recalculation task, write restriction control command, modify restriction control command, output restriction control command and archived state switching command.

[0014] Furthermore, step S3 also includes the processing of access control for newly added knowledge data, specifically including: When new knowledge data is received, the key attribute combination of the new knowledge data is extracted and compared with the corresponding key attribute combination of existing knowledge data in the knowledge base. When the comparison results show that there is existing knowledge data that meets the duplication or conflict conditions, the new knowledge data will be judged as having failed to be admitted, and a conflict warning record and a write restriction control instruction will be generated; when the comparison results show that there is no existing knowledge data that meets the duplication or conflict conditions, the new knowledge data will be allowed to be written to the knowledge base.

[0015] Furthermore, the downstream dependency data includes derived statistical data; When source knowledge data is added, changed, becomes invalid, replaced, or its state changes and affects statistical rules, the corresponding derived statistical nodes are included in the affected subgraph, and the consistency of calculation rules is checked on the derived statistical nodes. When it is determined that the derived statistical data corresponding to the derived statistical node is generated based on the statistical rules before the change, a recalculation mark or recalculation task is generated, and the recalculation of the derived statistical data is triggered based on the governance record.

[0016] Furthermore, the query control and output control in step S6 include: When a data query request is received, the identity identifier of the requesting entity and the ownership identifier of the target data are obtained; Based on the ownership or authorization relationship between the associated entity corresponding to the requesting entity's identity identifier and the associated entity corresponding to the target data's ownership identifier, determine whether the requesting entity has output permission; When it is determined that the requesting subject has output permission, an output control result is generated and the target data is output; When it is determined that the requesting entity does not have output permission, an output restriction control instruction is generated, and the output of target data is restricted or denied.

[0017] Furthermore, the state transition control in step S6 includes: Check whether the associated subjects applicable to the target knowledge base meet the preset completion conditions; When the detection results show that the associated subject meets the preset completion conditions, the status identifier of the target knowledge base is switched from editable to archived, and the archived status switching instruction is written into the governance record. In the archived state, query operations are allowed on the target knowledge base, while adding, modifying, and deleting operations on the target knowledge base are prohibited.

[0018] The beneficial effects of this invention are as follows: By jointly modeling knowledge state structure information and dependency graphs, this invention can more accurately identify the affected subgraphs after changes in source knowledge data; through differential consistency verification, verification is only performed on the affected subgraphs, reducing the amount of irrelevant data processing and improving governance efficiency; through the control action state machine, the verification results are directly converted into executable control actions such as recalculation, write restriction, output restriction, and archive switching; through the writing back and reuse of governance records, the governance results continue to participate in subsequent control, forming a closed loop; it is applicable to a variety of scenarios and has strong versatility. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the dynamic knowledge consistency governance and gating control method based on dependency graphs of the present invention. Figure 2 This is a schematic diagram of the dependency graph structure of the dynamic knowledge consistency governance and gating control method based on dependency graph of the present invention. Figure 3 This is a schematic diagram of the control action state machine of the dynamic knowledge consistency governance and gating control method based on dependency graphs of this invention. Detailed Implementation

[0020] The invention will now be described in further detail with reference to the accompanying drawings.

[0021] The embodiments described in this invention are merely some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0022] Implementation method one: General governance and control implementation method. For example... Figure 1-3As shown, this invention discloses a dynamic knowledge consistency governance and gating control method based on dependency graphs. In one embodiment, this invention is deployed in a dynamic knowledge governance platform, which includes a data acquisition module, a knowledge state management module, a dependency graph construction module, a change detection module, a differential verification module, a control action generation module, and a governance record management module.

[0023] Step S1, Data Acquisition and State Modeling. The system first acquires source knowledge data, downstream dependent data, and knowledge state structure information corresponding to the source knowledge data. Source knowledge data can be rule-based data, plan-based data, template-based data, course-based data, product-based data, or fee-based data, etc.; downstream dependent data can be task data, result data, statistical data, template data, permission data, access data, etc.; the knowledge state structure information is used to characterize the version status, validity status, and applicable scope information of the source knowledge data.

[0024] The source knowledge data, downstream dependency data, and knowledge state structure information can each originate from one or more data storage units, business system interfaces, message subscription channels, or rule configuration units. Specifically: source knowledge data can be obtained by accessing the main data storage table, rule configuration table, knowledge base data table, or interface services provided by the corresponding business system; downstream dependency data can be obtained by accessing the task data table, statistical result table, template data table, access control table, or the interface of the corresponding business subsystem. Knowledge state structure information can be obtained through at least one of the following methods: directly reading from the state field associated with the source knowledge data; extracting from the version record table, effective / invalid record table, substitution relationship table, conflict relationship table, or applicable scope configuration table; or generating it by parsing the version field, time field, relationship field, and state field of the source knowledge data.

[0025] In one embodiment, the system obtains source knowledge data and downstream dependency data through database queries; in another embodiment, the system obtains source knowledge data and downstream dependency data through application programming interface (API) calls; in yet another embodiment, the system triggers the reading and updating of corresponding data by receiving message notifications of new events, change events, state transition events, or failure events corresponding to the source knowledge data. It should be understood that the term "obtaining" is not limited to a one-time read, but also includes methods such as periodic fetching, event-triggered loading, cache synchronization, or targeted queries based on governance records.

[0026] In this invention, the source knowledge data correspondence identifier and the downstream dependency data correspondence identifier are used to uniquely identify the corresponding data objects and establish the correspondence between the corresponding data objects and dependency graph nodes and governance records. The source knowledge data correspondence identifier can be obtained by reading the pre-set primary key field, business code field, version number field, or system index field in the source knowledge data; the downstream dependency data correspondence identifier can be obtained by reading the pre-set primary key field, business record number field, task number field, result number field, or system index field in the downstream dependency data. When there is no readily available field in the source knowledge data or downstream dependency data that can be directly used for unique representation, the system can generate a corresponding data correspondence identifier based on the data table identifier to which the object belongs, the business object number, version information, timestamp information, or path information. In one embodiment, each source knowledge node and downstream dependency node in the dependency graph is associated with and stored with the corresponding source knowledge data correspondence identifier or downstream dependency data correspondence identifier, so that when generating governance records, the system can establish a correspondence between the control result and the source knowledge data object that triggered the control result and the affected downstream dependency data object. The purpose of writing the source knowledge data corresponding identifier and the downstream dependent data corresponding identifier in the governance record is to enable subsequent update control, query control, output control, derived statistical data recalculation or state switching control to directly locate the corresponding object without having to re-match all data objects.

[0027] In one embodiment, the knowledge state structure information includes at least three of the following: version identifier, effective time, expiration time, substitution relationship, conflict relationship, scope of application information, and status identifier. Specifically: the version identifier distinguishes different versions of the same source knowledge data; the effective time and expiration time determine the currently valid version; the substitution relationship represents the substitution chain between new and old source knowledge data objects; the conflict relationship represents the mutual exclusion or conflict constraints between different source knowledge data objects; the scope of application information characterizes the associated subject range corresponding to the source knowledge data, which may include individuals, organizations, groups, grades, majors, business units, or system entities, etc., and indicates the set of objects to which the knowledge object is applicable, constrained, served, or authorized; the status identifier characterizes whether the knowledge object is currently in an editable, available, restricted, or archived state.

[0028] This invention does not depend on a specific type of knowledge state structure. As long as the structure can support subsequent change detection, affected subgraph determination, and control result generation, it can be applied to this invention.

[0029] Step S2, Dependency Graph Construction. Based on the source knowledge data, the downstream dependency data, and the knowledge state structure information, a dependency graph describing the relationship between the source knowledge data and the downstream dependency data is constructed. In one embodiment, the nodes in the dependency graph include at least two types of nodes selected from source knowledge nodes, downstream dependency nodes, derived statistics nodes, access control nodes, and lifecycle nodes. The edges in the dependency graph include one or more of the following: reference dependencies, computation dependencies, output dependencies, and state inheritance relationships. Specifically: a reference dependency indicates that the downstream dependency data corresponding to a downstream dependency node directly references the content of the source knowledge data corresponding to a source knowledge node; a computation dependency indicates that the downstream dependency data corresponding to a downstream dependency node participates in computation or statistical generation based on the source knowledge data corresponding to a source knowledge node; an output dependency indicates that the output content of the downstream dependency data corresponding to a downstream dependency node is constrained by the source knowledge data corresponding to a source knowledge node; and a state inheritance relationship indicates that the current state of the downstream dependency data corresponding to a downstream dependency node is affected by changes in the state of the source knowledge data corresponding to a source knowledge node.

[0030] It should be noted that a dependency graph is not a simple data lineage graph, but a comprehensive relationship graph that simultaneously includes content references, rule calculations, output control, and state inheritance.

[0031] Step S3: Change detection and determination of affected subgraphs.

[0032] When the system detects that source knowledge data has been added, changed, expired, replaced, or its state has changed, it first identifies the corresponding source knowledge node, and then determines the affected subgraph based on the dependency graph. It should be noted that the determination of the "affected subgraph" in this invention does not involve a full scan of all downstream dependent nodes, but rather a quantitative judgment of each downstream dependent node based on the dependency strength represented by the edge attributes in the dependency graph before selecting the node to enter the subsequent verification process.

[0033] In one embodiment, the overall influence of the source knowledge node on each downstream dependent node is calculated; downstream dependent nodes with an overall influence not lower than a preset influence threshold are included in the affected subgraph. Specifically, the system reads the dependency relationship type, call frequency, computation participation, output constraint degree, or state transit flag corresponding to each edge from the dependency graph, and generates reference dependency strength, computation dependency strength, output dependency strength, and state inheritance dependency strength accordingly. In different business scenarios, not all dependencies exist simultaneously; therefore, the overall influence can be determined by the strength of at least two types of dependencies that actually exist. When all four types of dependencies exist, the overall influence can be jointly determined.

[0034] In one embodiment, to facilitate quantification, the reference dependency strength is... Calculate the dependence strength Output Dependency Strength Dependency strength of state inheritance Normalized to values ​​in the interval [0,1]. Among these, the reference dependency strength... The value can be determined based on the proportion of references made by downstream dependent nodes to the fields of the source knowledge node; when a downstream dependent node references all key fields of the source knowledge node, the value is 1; when it only references some key fields, the value is determined by the ratio of the number of referenced fields to the total number of key fields. Calculate dependency strength. The output dependency strength can be determined based on at least one of the following: the proportion of parameters participating in downstream dependency data calculations from source knowledge data, the proportion of call frequency, or the depth of rule calls. The strength of state inheritance dependency can be determined based on the constraints imposed on the range of output fields, output conditions, or output permissions of downstream dependent data by the source knowledge data. The weighting coefficient can be determined based on the proportion of factors influencing the current state, editing permissions, output permissions, or lifecycle state of downstream dependent nodes according to changes in the state of source knowledge data. In a preferred embodiment, when all four types of dependency strengths are normalized, the weighting coefficients are... , , , satisfy + + + =1. In one embodiment, the overall impact can be expressed as: , in: Represents source knowledge nodes For downstream dependent nodes The overall impact; This represents the i-th source knowledge node; This represents the j-th downstream dependent node; , , , Representing source knowledge nodes respectively With downstream dependent nodes The strength of reference dependencies, computation dependencies, output dependencies, and state inheritance dependencies between them; , , , These are the weighting coefficients for the corresponding dependency strengths.

[0035] It should be noted that, , , , The options can be preset based on the application scenario or dynamically adjusted based on historical governance results. It should be understood that the importance of the four types of dependencies varies across different scenarios; therefore, the weighting coefficients may not be equal. Preferably, =0.35、 =0.35、 =0.15、 =0.15, because in most knowledge governance scenarios, reference dependencies and computation dependencies have a more direct impact on determining whether a downstream node enters the affected subgraph, while output dependencies and state inheritance have a greater influence on the generation of subsequent control actions. Therefore, the weights corresponding to the strengths of the first two types of dependencies can be appropriately increased during the affected subgraph identification stage. Furthermore, the above... , , and All of these are derived from the edge attributes of the dependency graph, rather than external scoring parameters independent of the dependency graph. Therefore, the comprehensive influence is the result of the dependency graph structure information participating in the calculation, rather than an abstract mathematical description of the governance process.

[0036] In one embodiment, the system includes the corresponding downstream dependent node in the affected subgraph when the overall impact meets the following conditions: , in: This indicates the threshold for determining the affected subgraph.

[0037] In one embodiment, the affected subgraph determination threshold The pre-configured normalization threshold ranges from 0 to 1. Preferably, A value between 0.4 and 0.7 can be used; a higher threshold is used when the system prioritizes control accuracy; a lower threshold is used when the system prioritizes affecting coverage integrity. In another embodiment, the... It can also be dynamically adjusted based on the statistical results of the comprehensive influence distribution of confirmed abnormal nodes in historical governance records. In other words, downstream dependent nodes are only included in the affected subgraph when the comprehensive influence of the source knowledge node on downstream dependent nodes is not lower than a preset threshold.

[0038] In one embodiment, step S3 further includes access control processing for newly added knowledge data.

[0039] When new knowledge data is received, the system extracts the key attribute combinations of the new knowledge data and compares them with the corresponding key attribute combinations of existing knowledge data in the knowledge base to determine whether the new knowledge data meets the duplication or conflict conditions. The duplication condition indicates that the new knowledge data and existing knowledge data are completely identical in the key attribute combinations used to identify knowledge content and application boundaries, or reach a preset consistency threshold, such that they essentially point to the same knowledge item. The conflict condition indicates that although the new knowledge data and existing knowledge data do not belong to the same knowledge item, if their key attribute parts are the same, their effective time intervals, status identifiers, associated subject scopes, substitution relationships, or authorization boundaries have overlapping, mutually exclusive, or covering relationships that are not allowed to coexist.

[0040] In one embodiment, the key attribute combination includes primary attributes and constraint attributes. The primary attribute identifies the core identity of the knowledge item and may include at least one of name, code, category, or rule object identifier. The constraint attribute characterizes the applicable boundaries of the knowledge item and may include at least one of effective time interval, expiration time interval, associated subject scope, status identifier, parameter value, authorization boundary, or substitution relationship. The consistency threshold is used to determine whether newly added knowledge data and existing knowledge data substantially point to the same knowledge item. It is preferably determined by the key attribute matching ratio, with a value ranging from 0 to 1. When the key attribute matching ratio is not lower than the consistency threshold, the duplication condition is satisfied. Preferably, the consistency threshold is between 0.8 and 1.

[0041] In one embodiment, comparing the key attribute combination of newly added knowledge data with the corresponding key attribute combination of existing knowledge data can include at least one of the following methods: field value equality comparison, time interval overlap comparison, intersection and union comparison of the sets of associated subject ranges, and mutual exclusion comparison of rule constraints. Specifically, field value equality comparison is used to determine whether the newly added knowledge data and existing knowledge data are identical in name, code, category, parameter value, or other key attribute fields; time interval overlap comparison is used to determine whether the corresponding effective time intervals and expiration time intervals of the two overlap; intersection and union comparison of the sets of associated subject ranges is used to determine whether the corresponding associated subject ranges have intersecting, inclusive, or overlapping relationships; and mutual exclusion comparison of rule constraints is used to determine whether there are constraints that cannot be simultaneously satisfied under the same applicable conditions.

[0042] When the comparison result shows that the newly added knowledge data meets the duplication condition, the system determines the newly added knowledge data as duplicate knowledge and generates a conflict warning record and a write restriction control instruction; when the comparison result shows that the newly added knowledge data meets the conflict condition, the system determines the newly added knowledge data as conflict knowledge and generates a conflict warning record and a write restriction control instruction; when the comparison result shows that the newly added knowledge data neither meets the duplication condition nor the conflict condition, the system allows the newly added knowledge data to be written to the knowledge base.

[0043] In one embodiment, when new source knowledge data is added, the system first performs access control processing on the new knowledge data; when the access control result indicates that the new knowledge data is allowed to be written into the knowledge base, the system maps the new knowledge data to a new source knowledge node, and determines candidate affected nodes based on the reference dependency, computation dependency, output dependency, or state inheritance relationship formed between the new source knowledge node and existing downstream dependent nodes; downstream dependent nodes that do not have a dependency relationship with the new source knowledge node are not included in the affected subgraph.

[0044] In one embodiment, when the source knowledge data changes, the system reads the differences in content, state, or scope of application information of the source knowledge data before and after the change, and maps the differences to the edge attributes of the dependency graph to update the corresponding reference dependency strength, computation dependency strength, output dependency strength, or state inheritance dependency strength; then, the system calculates the overall impact of the change based on the updated dependency strength, and filters the affected nodes accordingly.

[0045] In one embodiment, when source knowledge data becomes invalid, the system identifies downstream dependent nodes that depend on the invalid source knowledge data based on the change in the valid state of the source knowledge data, combined with the state inheritance relationship, output dependency relationship and reference dependency relationship in the dependency graph, and calculates the comprehensive impact degree based on the strength of the corresponding dependency relationship to determine whether to include the corresponding downstream dependent node in the affected subgraph.

[0046] In one embodiment, when source knowledge data is replaced, the system identifies the source knowledge data before and after replacement based on the replacement relationship in the knowledge state structure information, and includes the source knowledge nodes corresponding to the replacement relationship in the analysis scope. For downstream dependent nodes that reference the original source knowledge data, depend on the original calculation rules, or are constrained by the original output, the system calculates the comprehensive impact degree by combining the replacement relationship and the corresponding edge attribute changes to determine the affected subgraph in the replacement process.

[0047] In one embodiment, when the source knowledge data undergoes a state transition, the system identifies downstream dependent nodes that may be affected by state propagation based on the type of change in the state identifier and the state inheritance relationship. When the state transition causes changes in the current state, editable permissions, output permissions, or lifecycle control conditions of the downstream dependent nodes, the system calculates the comprehensive impact degree based on the relevant edge attributes and includes downstream dependent nodes with a comprehensive impact degree not lower than a preset impact threshold in the affected subgraph.

[0048] Therefore, for the five types of events—addition, modification, failure, replacement, and state transition—the system can complete the identification of affected subgraphs under a unified processing framework of "source knowledge node identification—dependency extraction—comprehensive impact calculation—threshold filtering—affected subgraph determination." Step S4: Differential consistency verification. Differential consistency verification is performed on the downstream dependent nodes in the affected subgraph to obtain the differential consistency verification results of the downstream dependent data corresponding to each downstream dependent node relative to the source knowledge data after the addition, modification, failure, replacement, or state transition.

[0049] In one embodiment, differential consistency verification includes at least: Reference value consistency check: Perform reference value consistency check on downstream dependent nodes that have a reference dependency relationship with the source knowledge data to determine whether the reference content in the corresponding downstream dependent data is consistent with the currently valid source knowledge data; Consistency verification of computation rules: Perform consistency verification of computation rules for downstream dependent nodes that have computational dependencies on source knowledge data to determine whether the corresponding downstream dependent data is still generated based on the rules before the change; State attribute consistency check: Perform state attribute consistency check on downstream dependent nodes that have state inheritance relationship or output dependency relationship with source knowledge data to determine whether the current state or output permission in the corresponding downstream dependent data is consistent with the current state of source knowledge data.

[0050] In one implementation, the differential consistency score can be expressed as: , in: Indicates the first downstream dependent nodes The difference consistency score; , , These represent the consistency results of the reference values, calculation rules, and state attributes of the downstream dependent nodes, respectively. , , These are the corresponding weighting coefficients.

[0051] In one embodiment, , and All values ​​are represented using normalized numerical values ​​between 0 and 1; the higher the consistency, the closer the value is to 1. It can be determined based on the matching ratio between the target reference field set of the changed source knowledge data and the current reference field set of the downstream dependent nodes; It can be determined based on the consistency ratio between the modified rule parameter set and the rule parameter set currently used by downstream dependent nodes; The weighting coefficient can be determined based on the proportion of matching items between the changed source knowledge state and the current state, editing permissions, and output permissions of downstream dependent nodes. Preferably, when all comparison items are consistent, the corresponding result is 1; when all comparison items are inconsistent, the corresponding result is 0; when some are consistent, the value is taken as the proportion of the number of consistent items to the total number of comparison items. Preferably, the weighting coefficient... , , satisfy + + =1, preferably, =0.4、 =0.4、 =0.2. Because in dynamic knowledge governance scenarios, the consistency of referenced content and computational rules typically have a more direct impact on the correctness of downstream dependent data, while the consistency of state attributes has a greater impact on subsequent access control or lifecycle control. Therefore, it is preferable to increase the weights corresponding to the consistency results of reference values ​​and computational rules. It should be noted that the "differential consistency verification" uses the changed source knowledge data as a benchmark, performing differential comparisons on the current referenced content, current computational rules, and current state attributes of downstream dependent nodes before and after the change. Therefore, the... , and All results are quantitative assessments of the differences before and after the change, rather than static scores independent of the change process.

[0052] In one embodiment, a consistency anomaly is determined in downstream dependent nodes when the differential consistency score meets the following condition: , in: This indicates the consistency threshold.

[0053] Step S5: Control the action state machine. When the differential consistency check result indicates inconsistency, the system further generates control results based on the degree of inconsistency and node type.

[0054] The preset control action state machine includes normal state, warning state, recalculation state, restricted state, and archived state. It takes differential consistency verification results, anomaly level values, node types, and knowledge state structure information as inputs. Based on the anomaly level value's range and whether the target knowledge base meets the archive trigger conditions, the system selects and switches to the corresponding state among warning, recalculation, restricted, or archived states. It should be noted that the control results and anomaly level values ​​are not merely used to simply display the degree of anomaly, but rather serve as the input and output of the control action state machine, driving the selection and execution of control actions such as write restrictions, output restrictions, recalculation tasks, or archive switching.

[0055] In one embodiment, the system first calculates the anomaly level value of the affected nodes: , in: Indicates the first downstream dependent nodes The abnormality level value; Represents a node The result of consistent reference values; Represents a node Consistent results of calculation rules; Represents a node Consistency results of state attributes; Indicates the abnormal contribution coefficient of the reference value; This represents the coefficient of abnormal contribution in the calculation rules; This represents the contribution coefficient of anomalies in state attributes.

[0056] It should be noted that the formula uses , and The form is to convert the consistency result into the corresponding inconsistency contribution, that is, the lower the consistency, the higher the contribution to the anomaly level value. , and Different settings can be configured based on node type. For example, for derived statistics nodes, the contribution coefficient of abnormal calculation rules can be increased; for access control nodes or lifecycle nodes, the contribution coefficient of abnormal status attributes can be increased.

[0057] In one embodiment, the abnormal contribution coefficient , , satisfy + + =1. For derived statistical nodes, it is preferable to increase the anomaly contribution coefficient of the calculation rule. For access control nodes, it is preferable to increase the contribution coefficient of anomalies in state attributes. For template nodes or task nodes, it is preferable to increase the abnormal contribution coefficient of reference values. For example, for derived statistics nodes, you can set... =0.2、 =0.6、 =0.2; For access control nodes, it can be set to 0.2. =0.2、 =0.2、 =0.6.

[0058] In one embodiment, the system selects a control action based on the anomaly level value and node type, according to a preset control action state machine: , in: Indicates that for the first Control actions generated by downstream dependent nodes; Indicates warning records; Indicates a recalculation flag or recalculation task; This indicates that writing control commands, modifying control commands, or output control commands are restricted. Indicates a command to switch the storage status; , , The threshold values ​​for classifying anomalies are preset thresholds between 0 and 1, and preferably satisfy 0 < 1. < < ≤1. Preferably, A value of 0.2 to 0.4 is acceptable. A value of 0.4 to 0.7 is acceptable. A value of 0.7 to 0.9 is acceptable. That is, the system selects the corresponding control result from generating early warning records, recalculation markers or recalculation tasks, restriction control instructions or archived state switching instructions based on the range of the anomaly level value and the node type.

[0059] It should be noted that the above-mentioned control action selection relationship is not a simple hierarchical display of anomaly level values, but directly determines the type of control result generated. Furthermore, after the control action is generated, it will be written into the governance record along with the triggering event, the corresponding identifier of the source knowledge data, the corresponding identifier of the downstream dependent data, and the knowledge state structure information. This governance record will then continuously influence subsequent write control, query control, output control, recalculation of derived statistical data, and target knowledge base state switching control.

[0060] It should be noted that when the same downstream dependent node simultaneously meets multiple control conditions, the system can generate two or more control results in parallel for the same node and write the multiple control results into the governance record together.

[0061] Step S6: Governance record write-back and closed-loop control.

[0062] After a control action is generated, the system writes the control result, the source knowledge data corresponding to the triggering control result, the downstream dependent data corresponding to the triggering control result, and the corresponding knowledge state structure information into the governance record. In one embodiment, the governance record includes at least one of the following: trigger event identifier, source knowledge data corresponding identifier, downstream dependent data corresponding identifier, knowledge state structure information, control result type, processing object identifier, processing time, applicable scope information, and related subject scope information corresponding to the applicable scope information. The trigger event identifier characterizes the type of event that caused the generation of this control result; the control result type characterizes which of the following control actions it belongs to: warning, recalculation, restriction, or sealing; the processing object identifier uniquely identifies the object to be controlled subsequently; the processing time characterizes the time the governance record was generated; and the applicable scope information and related subject scope information define the applicable boundaries of the control result. The governance record is written to the governance record management module after generation and is called again during subsequent new knowledge writing, data querying, data output, derived statistical data recalculation, and target knowledge base state switching. It should be noted that the governance record not only records the control result, but also the trigger event identifier, the corresponding identifier of the source knowledge data, the corresponding identifier of the downstream dependent data, and the knowledge state structure information. Therefore, when subsequent modules perform control, they do not need to re-perform a full judgment on all source knowledge data and downstream dependent data, but can directly complete targeted control based on the governance record.

[0063] Implementation Method 2: Implementation Method for Academic Affairs Scenarios.

[0064] In one embodiment, the present invention is applied to a university academic affairs management platform. The source knowledge data includes teaching plan data and course data; the downstream dependent data includes teaching task data, student enrollment data, grade data, and derived statistical data.

[0065] The teaching plan data may include college information, major information, major specialization information, course code, course name, credits, planned semester, course nature, course category, GPA weight, and remarks. Course data may include course code, course name, credits, and the scope of related entities associated with the course. Derived statistical data may include a credit progress bar data table, a module completion data table, and a graduation achievement data table.

[0066] When the teaching plan data for a certain grade, major, or specialization changes—for example, if course credits, course categories, planned semesters, or grade point weights are modified—the system detects the change in the corresponding source knowledge nodes and determines the affected subgraph along the dependency graph. The affected subgraph includes: the teaching task node for the corresponding course; student enrollment nodes; relevant grade nodes; and derived statistical nodes corresponding to the credit progress bar data table.

[0067] Subsequently, the system performs reference value consistency checks on teaching task nodes, reference value consistency checks or status attribute consistency checks on student enrollment nodes, and calculation rule consistency checks on derived statistics nodes corresponding to the credit progress bar data table. If teaching task or student enrollment data still references old teaching plan fields, an early warning record or modification control instruction is generated; if the credit progress bar data table is still calculated according to the old rules, a recalculation mark or recalculation task is generated.

[0068] In another embodiment, when a grade query request is received, the system determines the relationship between the associated entity corresponding to the requesting entity's identity identifier and the associated entity corresponding to the target grade data's ownership identifier. If the ownership relationship is valid, the corresponding grade data is allowed to be output; if the ownership relationship is invalid, output is restricted or denied.

[0069] The attribution or authorization relationship can be jointly determined based on the requesting entity's identity identifier, the target data's attribution identifier, the scope of application information, the access control table, and the control results in the governance records. Specifically, the target data's attribution identifier is used to characterize the associated entity or scope of associated entities to which the target data belongs, and is used to match it with the scope of application information to determine whether the requesting entity has the corresponding output permissions.

[0070] In another embodiment, when new course data is received, the system performs an admission control check based on a combination of key attributes such as course name, the associated subject scope information of the course, and credits; if an existing course is detected to meet the duplicate or conflict conditions, a conflict warning record is generated and the course is refused to be written to the course knowledge base.

[0071] In an academic administration scenario, newly added knowledge data can be course data or teaching plan data. The system extracts key attribute combinations from the newly added knowledge data and compares them with corresponding key attribute combinations from existing knowledge data in the knowledge base to determine whether duplicate or conflict conditions are met.

[0072] For example, if the course name, course code, associated subject scope information, and credits of newly added course data are completely identical to existing course data, then the duplication condition is met. If the knowledge base already contains course data such as "Advanced Mathematics A, Major: Mechanical Engineering, Grade: 2024, Credits: 4," and the newly added course data contains the same content, the system determines that the newly added course data constitutes duplicate knowledge with the existing course data, generates a conflict warning record, and refuses to write it. If the newly added course data and existing course data are identical in course name and associated subject scope information, but have mutually exclusive settings in planned semester, course nature, or credit rules, and the corresponding time intervals overlap, then the conflict condition is met. For example, if the knowledge base already contains the course rule "Advanced Mathematics A, Mechanical Engineering Major, 2024, First Semester, Required, 4 Credits," and the newly added course data is recorded as "Advanced Mathematics A, Mechanical Engineering Major, 2024, First Semester, Elective, 2 Credits," the system determines that it meets the conflict condition and refuses access.

[0073] In another embodiment, when the associated entities corresponding to the teaching plan database meet preset completion conditions, such as all students in a certain grade having completed their training cycle, the system switches the state of the teaching plan database to a sealed state, where only querying is allowed and modification is not permitted. The preset completion conditions can be determined by at least one of the following: end time condition, completion markers for all associated entities, completion ratio reaching a preset threshold, and no pending downstream dependent tasks.

[0074] Implementation Method 3: Enterprise Scenario Implementation Method.

[0075] In one embodiment, the present invention is applied to an enterprise system and process knowledge platform. The source knowledge data includes system rule data, process rule data, and template rule data; the downstream dependent data includes approval task data, process template data, statistical data, and access control data.

[0076] For example, when reimbursement policies, approval process rules, or contract template rules change, the system detects the corresponding changes in source knowledge nodes and identifies the affected subgraphs along the dependency graph. The affected subgraphs may include: approval task nodes referencing old policy content; statistical nodes calculated based on old rules; template nodes referencing old clauses; and output control nodes for the employee Q&A system. The system further performs differential consistency checks and generates control results based on the anomaly level and node type. For example: when template nodes have inconsistent reference values, a restriction modification control instruction is generated; when statistical nodes have inconsistent calculation rules, a recalculation task is generated; when output control nodes have inconsistent state attributes, a restriction output control instruction is generated; and when the associated entity corresponding to the retired policy knowledge base meets the sealing conditions, a sealing state switching instruction is generated.

[0077] In the context of enterprise rules and regulations, newly added knowledge data can be expense reimbursement rules, approval process rules, or template rules. If a newly added expense reimbursement rule is completely identical to an existing rule in key attribute combinations such as rule name, applicable department scope, effective period, and reimbursement limit, it can be determined that the duplication condition is met. For example, if the knowledge base already contains a rule that states "travel reimbursement rule, applicable to the sales center, effective from January 1, 2026 to December 31, 2026, accommodation reimbursement limit 800 yuan / day," and a newly added rule still contains the same rule name, the same applicable department scope, the same effective period, and the same reimbursement threshold, the system will determine that it constitutes duplicate knowledge and refuse to write it. If a newly added rule is identical to an existing rule in rule name and applicable department scope, but sets different and incompatible control thresholds for the same reimbursement item within the same effective period, it can be determined that the conflict condition is met. For example, if an existing rule stipulates that "the maximum reimbursement for accommodation for sales center employees is 800 yuan per day," while a new rule stipulates that "the maximum reimbursement for accommodation for sales center employees is 1,000 yuan per day," and the effective time intervals of the two rules overlap and both are available, then the system will determine that there is a rule conflict between the two rules within the same applicable scope and the same time interval, generate a conflict warning record, and refuse to write it.

[0078] Implementation Method 4: Interface Output Control Implementation Method.

[0079] In one embodiment, the present invention is applied to a data interface control scenario between a school system or enterprise system and a third-party platform. The source knowledge data may be interface field authorization rules, identity status rules, or output authorization scope rules oriented towards associated subjects; the downstream dependent data may be interface exported data and access control data.

[0080] When field authorization rules change, the system detects the change in the corresponding source knowledge node and determines the affected interface output node along the dependency graph. If an interface output node is detected to still contain fields that should not be output, an output restriction control instruction is generated and the corresponding field's export is blocked. For example, in a single sign-on scenario, only student ID and name are allowed to be output, while ID card information is not allowed; for teacher information, only employee ID and name are allowed to be output.

[0081] When the identity status rules change, such as when an object's status changes to an inaccessible state, the system can generate restriction outputs or reject output control results for the relevant access control nodes based on the status inheritance relationship, thereby prohibiting them from continuing to access the relevant system.

[0082] To avoid ambiguity, the following is stated in this instruction manual: "Source knowledge data" refers to data objects that serve as the source of upstream rules, master data, or knowledge. "Downstream dependent data" refers to task, result, statistics, template, permission, or access-related data that directly or indirectly depends on source knowledge data. “Source knowledge node”, “downstream dependency node”, “derived statistics node”, “access control node” and “lifecycle node” are used to represent the structured representation of the corresponding data in the dependency graph; A "dependency graph" refers to a graph structure that describes the dependency relationships between source knowledge data and downstream dependent data. "Affected subgraph" refers to the set of downstream dependent nodes and their connections, determined by the source knowledge nodes that have changed, based on their dependencies and overall impact. "Differential consistency check" refers to the consistency check performed only on objects in the affected subgraph after the source knowledge data has changed. "Governance records" refer to data records used to record control results and their contextual information, and to be reused in subsequent controls; "Associated subject" refers to the object that corresponds to the scope of application of the source knowledge data or the target knowledge base and is subject to its constraints, services, authorization, management, or judgment. It may include individuals, organizations, groups, grades, majors, business units, or system entities. "Scope of associated subject" refers to the boundary of the set of associated subjects represented by the scope of application information. "Scope of application information" is part of the knowledge state structure information and is used to represent the scope of associated subjects corresponding to the source knowledge data or the target knowledge base.

[0083] "Preset completion conditions" refer to the conditions that trigger sealing, invalidation, or state switching, such as the end of the lifecycle, the end of the business phase, the completion of the object state, or the completion of the corresponding business cycle by the associated subject. "Duplicate condition" refers to the judgment condition that the newly added knowledge data and the existing knowledge data have the same combination of key attributes, which is sufficient to represent the same knowledge item; "conflict condition" refers to the judgment condition that the newly added knowledge data and the existing knowledge data do not represent the same knowledge item, but have overlapping, mutually exclusive or covering relationships that are not allowed to coexist in terms of effective interval, status identifier, scope of associated subject or rule constraints.

[0084] This invention achieves differential response and continuous closed-loop control for dynamic knowledge object changes by combining dependency graph modeling, affected subgraph identification, differential consistency verification, control action state machine, and governance record reuse. This is different from data governance solutions that rely solely on static rule inspection, one-time verification, or simple early warning prompts.

[0085] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A dynamic knowledge consistency governance and gating control method based on dependency graphs, characterized in that, Includes the following steps: S1. Obtain source knowledge data, downstream dependency data, and knowledge state structure information corresponding to the source knowledge data. The knowledge state structure information is used to characterize the version status, validity status, and applicable scope information of the source knowledge data. S2. Based on the source knowledge data, the downstream dependency data, and the knowledge state structure information, construct a dependency graph describing the relationship between the source knowledge data and the downstream dependency data; S3. Based on the knowledge state structure information, detect whether the source knowledge data has been added, changed, invalidated, replaced, or switched in state. When the source knowledge data is detected to have been added, changed, invalidated, replaced, or switched in state, determine the affected subgraph based on the comprehensive influence of the dependency graph on the source knowledge data. S4. Perform differential consistency verification on the downstream dependent nodes in the affected subgraph to obtain the differential consistency verification result of the downstream dependent data corresponding to each downstream dependent node relative to the source knowledge data after the addition, change, failure, replacement or state switch. S5. Based on the differential consistency verification result and the node type of the downstream dependent node, and according to the preset control action state machine, generate the corresponding control result. The control result includes at least one of the following: early warning record, recalculation mark, recalculation task, write restriction control instruction, modification restriction control instruction, output restriction control instruction, and archive state switching instruction. S6. Write the control result, the source knowledge data corresponding identifier that triggered the control result, the downstream dependent data corresponding identifier, and the corresponding knowledge state structure information into the governance record, and execute the corresponding update control, query control, output control, or state switching control based on the governance record.

2. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 1, characterized in that, The knowledge status structure information includes at least three of the following: version identifier, effective time, expiration time, substitution relationship, conflict relationship, scope of application information, and status identifier; The version identifier is used to distinguish different versions of the same source knowledge data; the effective time and expiration time are used to determine the effective status of the source knowledge data; the substitution relationship and conflict relationship are used to characterize the substitution and conflict constraints between different source knowledge data; the scope of application information is used to characterize the scope of the associated subjects corresponding to the source knowledge data; and the status identifier is used to characterize which of the following states the source knowledge data is currently in: editable, available, restricted, or archived.

3. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 1, characterized in that, The dependency graph has nodes, which include at least two of the following types: source knowledge nodes, downstream dependency nodes, derived statistics nodes, access control nodes, and lifecycle nodes. The dependency graph has edges, and the edges include at least one of reference dependencies, computation dependencies, output dependencies, and state inheritance relationships; The reference dependency relationship is used to characterize the reference of downstream dependent data to the content of source knowledge data; the computation dependency relationship is used to characterize the downstream dependent data to perform computation based on the source knowledge data to generate derived results; the output dependency relationship is used to characterize the output of the downstream dependent data to be constrained by the source knowledge data; and the state inheritance relationship is used to characterize the state of the downstream dependent data to be affected by changes in the state of the source knowledge data.

4. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 3, characterized in that, Step S3, which involves determining the affected subgraph based on the dependency graph, includes: Calculate the overall influence of the source knowledge node on each downstream dependent node; include downstream dependent nodes with an overall influence of not less than a preset influence threshold into the affected subgraph; The overall influence is determined by at least two of the following: reference dependency strength, computation dependency strength, output dependency strength, and state inheritance dependency strength.

5. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 1, characterized in that, The differential consistency check in step S4 includes: For downstream dependent nodes that have a reference dependency relationship with the source knowledge data, perform reference value consistency verification to determine whether the referenced content in the corresponding downstream dependent data is consistent with the currently valid source knowledge data; For downstream dependent nodes that have computational dependencies on the source knowledge data, perform computation rule consistency checks to determine whether the corresponding downstream dependent data is still generated based on the rules before the change. For downstream dependent nodes that have state inheritance relationships or output dependency relationships with the source knowledge data, perform state attribute consistency checks to determine whether the current state or output permissions in the corresponding downstream dependent data are consistent with the current state of the source knowledge data.

6. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 5, characterized in that, Step S5, which generates the corresponding control result according to the preset control action state machine, includes: Based on the differential consistency verification results, calculate the anomaly level value of each downstream dependent node in the affected subgraph; Based on the anomaly level range where the anomaly level value is located and the node type of the corresponding downstream dependent node, select the target control result from the warning record, recalculation flag, recalculation task, write restriction control command, modify restriction control command, output restriction control command and archived state switching command.

7. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 1, characterized in that, Step S3 also includes the processing of new knowledge data access control, specifically including: When new knowledge data is received, the key attribute combination of the new knowledge data is extracted and compared with the corresponding key attribute combination of existing knowledge data in the knowledge base. When the comparison results show that there is existing knowledge data that meets the duplication or conflict conditions, the new knowledge data will be judged as having failed to be admitted, and a conflict warning record and a write restriction control instruction will be generated; when the comparison results show that there is no existing knowledge data that meets the duplication or conflict conditions, the new knowledge data will be allowed to be written to the knowledge base.

8. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 3, characterized in that, The downstream dependency data includes derived statistical data; When source knowledge data is added, changed, becomes invalid, replaced, or its state changes and affects statistical rules, the corresponding derived statistical nodes are included in the affected subgraph, and the consistency of calculation rules is checked on the derived statistical nodes. When it is determined that the derived statistical data corresponding to the derived statistical node is generated based on the statistical rules before the change, a recalculation mark or recalculation task is generated, and the recalculation of the derived statistical data is triggered based on the governance record.

9. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 1, characterized in that, The query control and output control in step S6 include: When a data query request is received, the identity identifier of the requesting entity and the ownership identifier of the target data are obtained; Based on the ownership or authorization relationship between the associated entity corresponding to the requesting entity's identity identifier and the associated entity corresponding to the target data's ownership identifier, determine whether the requesting entity has output permission; When it is determined that the requesting subject has output permission, an output control result is generated and the target data is output; When it is determined that the requesting entity does not have output permission, an output restriction control instruction is generated, and the output of target data is restricted or denied.

10. The dynamic knowledge consistency governance and gating control method based on dependency graphs according to claim 1, characterized in that, The state transition control in step S6 includes: Check whether the associated subjects applicable to the target knowledge base meet the preset completion conditions; When the detection results show that the associated subject meets the preset completion conditions, the status identifier of the target knowledge base is switched from editable to archived, and the archived status switching instruction is written into the governance record. In the archived state, query operations are allowed on the target knowledge base, while adding, modifying, and deleting operations on the target knowledge base are prohibited.

Images