Intelligent monitoring and data security tracing system for donor sample collection

Abstract

The present application relates to the technical field of data security traceability, in particular to a donor sample collection intelligent monitoring and data security traceability system. The system comprises an event object generator, a semantic state diagram generator, a constraint verification engine, a temporary layer processor, an exception propagation analyzer and a traceability record generator, converts donor qualification confirmation, collection task generation, sample code binding, collection completion, sample reception, exception registration and record correction events into standardized event objects, and establishes a directed state relationship between the donor, the collection task, the sample code, the collection batch and the handover record, performs pre-requisite, time sequence, exclusive occupation and irreversible transition verification before writing into the collection file, and generates exception links and continuous traceability records through temporary layer simulation and exception propagation analysis. The present application can identify collection events with single field legality but overall state conflict, maintain collection process state causal consistency and exception data traceability.

Description

Technical Field

[0001] This invention relates to the field of data security traceability technology, specifically to an intelligent monitoring and data security traceability system for donor sample collection. Background Technology

[0002] Intelligent monitoring and data security traceability systems for donor sample collection are typically deployed in software stages such as collection registration, sample confirmation, sample handover, and anomaly recording. They are used to store data including donor eligibility status, collection tasks, sample codes, collection time, reception time, operators, and verification records. Existing conventional solutions often rely on database forms, storing relevant fields in donor number tables, collection task tables, sample code tables, and handover record tables. After donor identity verification and task selection at the collection end, the system generates or reads sample codes and then writes the collection completion information into the collection record table. Once the sample enters the reception stage, the receiving personnel retrieve the corresponding record by scanning the sample code or manually entering the code, and write it into the handover record after confirming receipt. In case of record errors or anomalies, the system typically allows data correction through supplementary entries, remarks, status modifications, or field updates during verification. This type of solution relies on field integrity checks, number duplication checks, time format checks, and permission checks to maintain basic data quality. Its data processing focuses on ensuring that fields in each table meet preset formats and that primary key relationships do not conflict. Since the records are mainly associated with each other by number, the system focuses more on whether the records can be added to the table and whether they can be queried when saving the collected data. It usually lacks unified calculation for the dependencies between collected events, state succession and exclusive occupancy.

[0003] A more conventional monitoring solution would add process node status and log auditing functions to the above form registration. The system sets collection tasks to states such as pending collection, collected, received, abnormal, and closed, and records the operator, operation time, and content before and after modification each time the status changes. Some systems also set binding states for sample codes to prevent the same sample code from being reused in the same data table; establish foreign key relationships between collection tasks and donor numbers to prevent tasks without donor records from entering subsequent processes; and create indexes between received records and sample codes to facilitate subsequent queries of the record path from collection to receipt. Anomaly handling is usually done through a separate anomaly registration table, which stores the anomaly type, anomaly description, processing status, and review opinions, and establishes a link between the sample code or task number and the collection record. This type of system can form a basic log chain, but the log chain is usually only arranged in writing time order, and the status fields are mostly used for process display or manual judgment, without abstracting donor qualifications, collection tasks, sample codes, collection batches, and handover records into a unified, computable status structure. For record correction, delayed writing, and supplementary recording events, conventional solutions typically only save modification traces without simulating the impact of the modifications on the existing process state.

[0004] The main technical problem with existing technologies lies in the lack of a consistent computational model for state-oriented causal relationships between donor sample collection events. This makes it difficult for the system to determine whether a new collection event disrupts the overall credibility of the collection process before it is written into the official archive. Specifically, donor qualification confirmation, collection task generation, sample code binding, collection completion, sample reception, anomaly registration, and record correction are carried by different records or fields. The system can only determine whether the current field exists, whether the current number is duplicated, or whether the current state allows modification, but it is difficult to calculate whether there is a causal chain break between a certain event and its predecessor, whether there is a state reversal between a certain event and its successor, or whether there is a hidden occupancy conflict between a certain event and the same task or sample code. While supplementary entry and correction operations may meet permission and format requirements for individual fields, their writing time, event occurrence time, predecessor source, and sample flow status may contradict existing records. Due to the lack of temporary simulation and anomaly propagation analysis, once an erroneous event enters the official archive, subsequent handover, anomaly closure, and source tracing will continue to reference the erroneous relationship. There is no recoverable data path between the original anomaly point and the affected record, ultimately causing a break or confusion in the source tracing chain of the collected data. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent monitoring and data security traceability system for donor sample collection, which can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The intelligent monitoring and data security traceability system for donor sample collection includes an event object generator, a semantic state graph generator, a constraint verification engine, a temporary layer processor, an anomaly propagation analyzer, and a traceability record generator. The event object generator receives collection events corresponding to donor qualification confirmation, collection task generation, sample encoding binding, collection completion, sample reception, anomaly registration, and record correction, and generates standardized event objects. The semantic state graph generator establishes directed state relationships between donors, collection tasks, sample encoding, collection batches, and handover records based on the standardized event objects. The constraint verification engine performs pre-dependency, time sequence, exclusive occupancy, and irreversible transition verification on the directed state relationships. The temporary layer processor performs state transition simulation before the standardized event objects are written into the formal collection archive. The anomaly propagation analyzer generates anomaly links based on the simulation results. The traceability record generator saves the association records between anomaly links and corresponding collection events.

[0007] Preferably, the event object generator includes event source normalization processing logic and event semantic completion logic. The event source normalization processing logic converts the collected events formed by manual input, batch import, interface synchronization, and review and correction into a unified event format. The unified event format includes event subject identifier, event object identifier, event type, event occurrence time, record writing time, preceding event pointer, operation source identifier, and record status identifier. The event semantic completion logic generates candidate preceding relationships for collected events lacking preceding event pointers based on the existing relationship between the collection task, sample encoding, and donor qualification status, and submits the candidate preceding relationships to the constraint verification engine along with the standardized event object.

[0008] Preferably, the semantic state graph generator includes node hierarchical logic and edge attribute configuration logic. The node hierarchical logic generates state nodes according to the donor layer, acquisition task layer, sample encoding layer, acquisition batch layer, and handover record layer, and configures the current state, source event, and previous state pointer for each state node. The edge attribute configuration logic generates directed edges according to the event types of qualification confirmation, task issuance, encoding binding, acquisition confirmation, handover confirmation, anomaly registration, and record correction, and writes the pre-dependency type, reversibility type, exclusive occupancy type, and time window type into the directed edges, so that a computable state association structure is formed between the same donor, the same acquisition task, and the same sample encoding.

[0009] Preferably, the temporary layer processor includes mirror graph generation logic, transition sequence construction logic, and difference output logic. The mirror graph generation logic copies and adds local nodes and directed edges related to standardized event objects based on the formal semantic state graph. The transition sequence construction logic arranges the state transitions to be simulated according to the preceding event pointer, event occurrence time, and record writing time. The difference output logic marks the node states, edge attributes, preceding dependencies, and occupancy relationships before and after the simulation, and sends the difference items corresponding to state chain breakage, reverse transition, duplicate occupancy, cross-task binding, and time causality inversion to the anomaly propagation analyzer.

[0010] Preferably, the event semantic completion logic includes a multi-candidate preceding relationship filtering logic. When there are two or more candidate preceding relationships for a standardized event object, the multi-candidate preceding relationship filtering logic generates a candidate relationship set according to four conditions: the matching relationship between the donor qualification status and the collection task status, whether the sample code has been occupied, whether the event occurrence time is within the valid range of the collection task, and whether the record correction event references the original record. For each candidate preceding relationship in the candidate relationship set, a relationship source, a relationship conflict field, and a relationship confirmation identifier are written, so that the candidate preceding relationships maintain a distinguishable data structure before entering the semantic state graph generator.

[0011] Preferably, the edge attribute configuration logic includes constraint template matching logic. The constraint template matching logic calls the corresponding constraint template according to the starting state node, target state node, and event type of the directed edge connection. The constraint template includes an irreversible transition field that prohibits rollback, an exclusive occupancy field that restricts the same sample code from being bound to multiple acquisition tasks at the same time, a time window field that restricts acquisition confirmation from being later than task issuance but earlier than handover confirmation, and a pre-dependency field that restricts record correction from pointing to the original record. After the constraint template is written into the directed edge attribute, a verification index associated with the state node is generated synchronously.

[0012] Preferably, the anomaly propagation analyzer includes forward tracing logic and backward impact localization logic. The forward tracing logic searches for the source events corresponding to donor qualification confirmation, collection task generation, and sample encoding binding along the preceding event pointer based on the difference items output by the temporary layer processor. The backward impact localization logic searches for the affected events corresponding to collection completion, sample reception, anomaly registration, and record correction along the subsequent directed edges in the semantic state graph, and generates anomaly link sets according to identity association anomalies, task flow anomalies, sample encoding anomalies, handover chain anomalies, and supplementary recording conflict anomalies.

[0013] Preferably, the difference output logic includes formal write control logic. The formal write control logic configures the write state of the newly added standardized event object according to the abnormal link set generated by the abnormal propagation analyzer. The write state includes write-allowed state, write-blocked state, write-after-review state, and read-only suspended state. When the write state is write-after-review, the formal write control logic binds the standardized event object, temporary layer difference item, and abnormal link set as a record to be reviewed. When the write state is read-only suspended state, it only saves the summary, source, and candidate predecessor relationship of the standardized event object.

[0014] Preferably, the anomaly propagation analyzer further includes anomaly link merging logic. When multiple discrepancies correspond to the same donor, the same acquisition task, or the same sample encoding, the anomaly link merging logic maps the multiple discrepancies to common source events and common affected events, respectively. The anomaly link set is merged in the order of common source events first, common sample encoding second priority, and common handover records third priority. The original type, preceding event pointer, subsequent event pointer, and temporary layer identifier of each discrepancy item are retained in the merged anomaly link set.

[0015] Preferably, the source tracing record generator includes abnormal evidence solidification logic. When the formal write control logic generates a blocked write state, a post-review write state, or a read-only suspended state, the abnormal evidence solidification logic extracts the event subject identifier, event object identifier, event type, preceding event pointer, temporary layer difference item, abnormal link set, and write state of the corresponding standardized event object, generates continuous summary records according to the event write order, and associates the continuous summary records with the donor node, the acquisition task node, the sample encoding node, and the handover record node, respectively.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention eliminates the need for isolated fields to directly enter the formal archive of collected events. Instead, the event object generator transforms them into standardized event objects containing event subject identifier, event object identifier, event type, event occurrence time, record write time, preceding event pointer, operation source identifier, and record status identifier. A semantic state graph generator then establishes directed state relationships between the donor, collection task, sample encoding, collection batch, and handover record. This processing method places donor qualification confirmation, collection task generation, sample encoding binding, collection completion, sample reception, anomaly registration, and record correction within the same state structure for calculation. The constraint verification engine can perform joint verification of new events based on preceding dependencies, temporal order, exclusive occupancy, and irreversible transitions. Before formal writing, the temporary layer processor copies local nodes and directed edges related to the new standardized event object and constructs a transition sequence based on the preceding event pointer, event occurrence time, and record write time. By simulating the differences in node states, edge attributes, preceding dependencies, and occupancy relationships before and after, it outputs the differences corresponding to state chain breaks, reverse transitions, duplicate occupancy, cross-task binding, and temporal causal inversion. The anomaly propagation analyzer generates anomaly chains along the semantic state diagram based on discrepancies, and the source tracing record generator saves the anomaly chains and corresponding acquisition events as associated records. Thus, the system can identify acquisition events with valid individual fields but conflicting overall states before data is written, and confine the impact of new events on donors, tasks, sample encoding, batches, and handover records to a temporary layer, preventing erroneous state relationships from entering the formal acquisition archive. For events resulting from supplementary entries, delayed writing, and record corrections, the system can identify the source of conflict between the original event object and the existing state diagram while maintaining the original event object, providing direct data processing improvements for source chain breaks caused by the lack of state causal consistency calculations in acquisition events.

[0017] 2. This invention establishes clearer data boundaries for the source and scope of anomalies. The event semantic completion logic generates candidate preceding relationships for acquisition events lacking preceding event pointers. The multi-candidate preceding relationship filtering logic forms a candidate relationship set based on the matching relationship between the donor's eligibility status and the acquisition task status, sample code occupancy, whether the event occurrence time falls within the allowed range of the acquisition task, and whether the record correction event references the original record. It retains the relationship source, relationship conflict fields, and pending confirmation identifiers, ensuring that subsequent graph calculations do not lose their judgment basis due to incomplete preceding relationships. The edge attribute configuration logic writes irreversible transitions, exclusive occupancy, time windows, and preceding dependencies into directed edges and generates a verification index associated with state nodes, enabling continuous identification of conflicts between the same donor, the same acquisition task, and the same sample code. The anomaly propagation analyzer searches for the source event along the preceding event pointer and for the affected event along subsequent directed edges. The anomaly link merging logic merges the differences under the same donor, the same acquisition task, or the same sample code, retaining the original type, preceding event pointer, subsequent event pointer, and temporary layer identifier. The formal write control logic is configured accordingly to allow writes, block writes, write after verification, and read-only suspension. The source tracing record generator then generates continuous summary records from the event subject identifier, event object identifier, event type, preceding event pointer, temporary layer difference items, anomaly link set, and write status, and associates them with donor nodes, acquisition task nodes, sample encoding nodes, and handover record nodes, ensuring the continuity of the source, propagation path, and processing results of the anomaly event at the data level. The association method between continuous summary records and status nodes can preserve data evidence during the blocking, verification, and suspension processes, and ensure that the records for the same anomaly in different acquisition stages remain consistent, allowing subsequent verification to reconstruct the anomaly formation process according to the record order. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the overall processing flow of the intelligent monitoring and data security traceability system for donor sample collection according to the present invention. Figure 2 This is a flowchart illustrating the standardized event object generation and semantic state diagram construction process of the present invention. Figure 3 This is a flowchart illustrating the abnormal propagation and source tracing evidence consolidation process of the temporary layer in this invention. Detailed Implementation

[0019] refer to Figure 1In one embodiment, the intelligent monitoring and data security traceability system for donor sample collection is implemented using a software data processing structure. After receiving events such as donor qualification confirmation, collection task generation, sample encoding and binding, collection completion, sample reception, anomaly registration, and record correction, the system uses an event object generator to convert collection events from different sources into standardized event objects. These standardized event objects carry an event subject identifier, an event object identifier, an event type, an event occurrence time, a record write time, a preceding event pointer, an operation source identifier, and a record status identifier. The semantic state graph generator then establishes a donor, [and other data] based on these standardized event objects. The system establishes directed state relationships between data collection tasks, sample encoding, collection batches, and handover records. The constraint verification engine performs pre-dependency, time sequence, exclusive occupancy, and irreversible transition verifications on these relationships. The temporary layer processor copies relevant local nodes and directed edges and simulates state transitions before adding standardized event objects to the formal data collection archive. The anomaly propagation analyzer generates anomaly links based on simulation differences, and the source tracing record generator associates the anomaly links with the corresponding data collection events. This ensures that the data collection data completes state causal consistency calculations before being formally written, reducing the likelihood of individual fields being valid but the overall data collection link conflicting and entering the formal archive.

[0020] In this embodiment, the system does not use the field validation results of a single data table as the basis for writing during runtime. Instead, it uses standardized event objects as a unified calculation entry point. After the acquisition task generation event is written, an acquisition task node is formed in the semantic state graph. After the donor qualification confirmation event is written, a directed edge for qualification confirmation is formed between the donor node and the acquisition task node. After the sample encoding binding event is written, a directed edge for encoding binding is formed between the acquisition task node and the sample encoding node. After the acquisition completion event is written, a directed edge for acquisition confirmation is formed between the sample encoding node and the acquisition batch node. After the sample reception event is written, a directed edge for handover confirmation is formed between the acquisition batch node and the handover record node. For each new acquisition event received by the system, a mirror state relationship is first constructed in a temporary layer. Then, the new acquisition event is inserted into the mirror state relationship for transition simulation. If the simulation result does not violate the constraints of predecessor dependency, time order, exclusive occupancy and irreversible transition, the new acquisition event obtains the formal writing condition. If the simulation result has state chain breakage, reverse transition, duplicate occupancy, cross-task binding or time causality reversal, the anomaly propagation analyzer generates an anomaly link along the predecessor event pointer and subsequent directed edges and hands it over to the formal writing control logic for processing.

[0021] refer to Figure 2In a preferred embodiment, the event object generator performs source normalization processing on collection events from different sources. Collection events formed by manual input retain the input source identifier, collection events formed by batch import retain the file batch source identifier, collection events formed by interface synchronization retain the interface call source identifier, and collection events formed by review and correction retain the original record reference source identifier. The event source normalization processing logic converts the above sources into a unified event format and writes them into the same set of fields. In the unified event format, the event subject identifier points to the donor or operation subject, the event object identifier points to the collection task, sample code, collection batch, or handover record, the event type is used to distinguish between qualification confirmation, task issuance, code binding, collection confirmation, handover confirmation, exception registration, and record correction, the event occurrence time is used to characterize the actual occurrence time of the collection action or handover action, the record writing time is used to characterize the time when the data enters the system, the preceding event pointer is used to connect the previous event that the current event depends on, the operation source identifier is used to distinguish the event source, and the record status identifier is used to distinguish the writing status such as pending simulation, written, pending review, blocked, and suspended. This enables collection events from different sources to have the same data semantic basis.

[0022] Furthermore, the event semantic completion logic is used to handle acquisition events with missing or incomplete preceding event pointers. When a standardized event object lacking a preceding event pointer is received, the event semantic completion logic calls the existing donor qualification status, acquisition task status, and sample code binding status to generate candidate preceding relationships. The candidate preceding relationships use the current event object identifier as the retrieval starting point to find event objects that occurred earlier than the current event under the same donor, the same acquisition task, or the same sample code and whose record status has not been revoked. The candidate relationship set is formed according to conditions such as whether the donor qualification status matches the acquisition task status, whether the sample code has been occupied, whether the event occurrence time is within the valid range of the acquisition task, and whether the record correction event references the original record. The candidate relationship set does not directly replace the manually confirmed preceding event pointer, but enters the constraint verification engine along with the standardized event object to participate in state simulation, thereby avoiding the inability of the event to be included in the semantic state graph calculation due to the lack of preceding relationships.

[0023] The matching value of the candidate preceding relationship is calculated according to the following formula:

[0024] in, Indicates the current event being collected. With candidate preceding events The matching value between them, in physical terms, represents the strength of the state continuity between the current collected event and the candidate preceding event. This indicates the donor eligibility matching result, with a value of 1 for a match and a value of 0 for a non-match. This indicates the matching result of the data collection task status. A value of 1 represents a match, and a value of 0 represents a non-match. This indicates the matching result of the sample encoding occupancy relationship. A value of 1 represents a match, and a value of 0 represents a non-match. This indicates whether the event occurred within the valid timeframe of the data collection task; a value of 1 indicates it occurred, and a value of 0 indicates it did not. , , and These represent the weights of the four matching conditions, and their sum is 1. For example... , , , If the four matching results of a candidate preceding event are 1, 1, 0, and 1 respectively, then The calculation results are used to sort candidate preceding relations while retaining the relation source and conflict fields.

[0025] Table 1 is a field organization table for standardized event objects, which describes the data content saved by the event object generator before different collected events enter the semantic state diagram.

[0026] Table 1. Field Organization Table for Standardized Event Objects ; In this embodiment, the preceding event pointers in Table 1 are used to connect the causal relationships between different acquisition events. The record status identifier is used to restrict the flow of standardized event objects between temporary layers, formal acquisition archives, and review records. When generating standardized event objects, the event object generator sets different candidate preceding relationship identifiers for null pointers, duplicate pointers, and undone pointers. The semantic state graph generator only accepts standardized event objects with explicit preceding event pointers or candidate preceding relationship sets. When verifying preceding dependencies, the constraint verification engine reads the fields listed in Table 1 and compares them with the node states in the current semantic state graph. This ensures that subsequent state transition simulations have complete input data and reduces the problem of misuse of synonymous fields caused by differences in event sources.

[0027] In one embodiment, the semantic state graph generator includes node hierarchical logic and edge attribute configuration logic. The node hierarchical logic generates state nodes according to the donor layer, acquisition task layer, sample encoding layer, acquisition batch layer, and handover record layer. Each state node includes a node identifier, current state, source event, previous state pointer, and associated summary identifier. The donor layer records the donor qualification confirmation status and its correspondence with the acquisition task. The acquisition task layer records the task generation status, the task validity period, and whether the task is closed. The sample encoding layer records whether the sample encoding has been bound, acquired, and received. The acquisition batch layer records the inheritance relationship between the acquisition completion event and subsequent handover events. The handover record layer records the state relationship between the sample reception event and the anomaly registration event. The edge attribute configuration logic generates directed edges according to qualification confirmation, task issuance, encoding binding, acquisition confirmation, handover confirmation, anomaly registration, and record correction, and writes the pre-dependency type, reversibility type, exclusive occupancy type, and time window type for the directed edges, thereby enabling the state nodes and directed edges to jointly express the process inheritance relationship of the acquired data.

[0028] Preferably, the edge attribute configuration logic calls the constraint template matching logic to configure verification fields for different directed edges. The irreversible transition field that prohibits rollback is used to restrict the sample code that has been collected from directly rolling back to the unbound state. The exclusive occupancy field is used to restrict the same sample code from being occupied by multiple collection tasks at the same time. The time window field is used to restrict the collection confirmation event from being after the task issuance event and before the handover confirmation event. The prerequisite dependency field is used to restrict the record correction event from pointing to the original record event. After the constraint template is written to the directed edge attributes, a verification index associated with the state node is generated synchronously. The verification index uses the combination of node identifier and edge attribute as the retrieval key, so that the constraint verification engine can call the corresponding constraint template by pressing the retrieval key when a new event enters the temporary layer, without having to traverse all collection records. This concentrates the state conflict calculation in the local area that has a causal relationship or occupancy relationship with the new event.

[0029] Table 2 is a configuration table for directed edge constraint templates, which explains how the semantic state graph generator converts collected event types into computable constraints.

[0030] Table 2 Configuration table of directed edge constraint templates ; In this embodiment, the constraint fields in Table 2 are written into the directed edge attributes when the semantic state graph is generated. When the temporary layer processor copies the local directed edges, it simultaneously copies the constraint fields and the verification index. If the newly added acquisition event is a sample receiving event, the mirror graph copies at least the sample encoding node, the acquisition batch node, the handover record node, and the acquisition confirmation edge and handover confirmation edge connecting these nodes. If the newly added acquisition event is a record correction event, the mirror graph copies at least the original record node, the candidate position of the correction record node, and the affected directed edges connected to the original record node. The copying range is determined by the preceding event pointer and the event object identifier, thereby ensuring that the simulated objects in the temporary layer cover the state relationships that may be destroyed by the newly added event.

[0031] refer to Figure 3 In one embodiment, the temporary layer processor includes mirror graph generation logic, transition sequence construction logic, and difference output logic. The mirror graph generation logic copies local nodes and directed edges related to the newly added standardized event objects based on the formal semantic state graph. The copied content includes the current state of the node, source event, previous state pointer, edge attributes, verification index, and association summary identifier. The transition sequence construction logic arranges the state transitions to be simulated according to the preceding event pointer, event occurrence time, and record writing time. If the event occurrence time is earlier than the record writing time and there are subsequent written events, the transition sequence construction logic inserts the newly added event into the corresponding historical position and simulates the replay of the subsequent state. If the event occurrence time is later than the preceding event but earlier than the subsequent event, a local replay sequence is generated according to the preceding event pointer. The difference output logic differentiates the node states, edge attributes, preceding dependencies, and occupancy relationships before and after the simulation and sends the difference items to the anomaly propagation analyzer, thereby avoiding the direct overwriting of formal collection files by supplementary recording and delayed writing.

[0032] The difference values ​​in the temporary layer are calculated according to the following formula:

[0033] in, Indicates a newly added data collection event The overall difference value created in the temporary layer has the physical meaning of representing the intensity of the conflict between the newly added acquisition event and the existing state relationships. This represents the state of a broken link; a value of 1 indicates the broken link, and a value of 0 indicates the broken link. This represents the result of the reverse transition; the value is 1 if it exists, and 0 if it does not exist. This indicates the result of duplicate usage; the value is 1 if the duplicate exists, and 0 if the duplicate does not exist. This represents the result of reversing causality in time; the value is 1 if it exists, and 0 if it does not. This indicates the result of cross-task binding; the value is 1 if the task exists, and 0 if it does not. , , , and The weights of the difference terms are summed to 1, for example... , , , , If a newly collected event exhibits state chain breaks and duplicate usage but no other differences, then The calculation result, along with the specific difference items, is written into the temporary layer difference record.

[0034] In this embodiment, the difference output logic does not determine the write state solely based on the comprehensive difference value. Instead, it outputs the comprehensive difference value in combination with the difference item type. State chain break corresponds to difference items that the preceding dependency cannot be inherited. Reverse transition corresponds to difference items where the node state rolls back from received, collected, or bound to the preceding state. Duplicate occupation corresponds to difference items where the same sample code is bound by multiple collection tasks. Cross-task binding corresponds to difference items where there is a lack of consistent task source between the current donor, collection task, and sample code. Time causality inversion corresponds to difference items where the occurrence time of the new event is inconsistent with the order of existing preceding or subsequent events. After receiving the difference item, the anomaly propagation analyzer selects the starting node for forward tracing or backward influence location based on the difference item type. This allows the anomaly analysis to be carried out on specific state conflicts, rather than blocking based solely on the overall score.

[0035] In a preferred embodiment, the multi-candidate precursor relationship filtering logic is used to enhance the feasibility of the event semantic completion logic. When there are more than two candidate precursor relationships for the same standardized event object, the multi-candidate precursor relationship filtering logic forms a qualification candidate set according to the matching relationship between the donor qualification status and the collection task status, forms a code candidate set according to whether the sample code has been occupied by other collection tasks that have not been closed, forms a time candidate set according to whether the event occurrence time is within the valid range of the collection task, and forms a correction candidate set according to whether the record correction event references the original record. For each candidate precursor relationship, a relationship source, a relationship conflict field, and a relationship confirmation identifier are written. If a candidate precursor relationship hits multiple candidate sets at the same time, all hit sources are retained and the names of mutually conflicting fields are recorded in the relationship conflict field. If a candidate precursor relationship does not hit any candidate set, it is marked as a low-match candidate relationship and restricted from entering the formal layer. This ensures that the candidate precursor relationship maintains a distinguishable, comparable, and verifiable data structure before entering the semantic state diagram.

[0036] Furthermore, the constraint template matching logic also performs template inheritance processing when configuring directed edges. When the record correction edge points to the original acquisition confirmation edge, the record correction edge inherits the time window field and irreversible transition field of the original acquisition confirmation edge, and adds an original record reference field. When the exception registration edge points to the handover confirmation edge, the exception registration edge inherits the pre-dependency field of the handover confirmation edge, and adds an exception registration source field. When the task distribution edge and the coding binding edge are both connected to the same acquisition task node and the same sample coding node, the exclusive occupancy field shares the occupancy index between the two directed edges, so that any new coding binding event must simultaneously verify the task source and coding occupancy status. This merges the constraint relationships between the donor layer, task layer and sample coding layer into the same computable edge attribute, reducing the situation where cross-node conflicts are ignored when only a single node is verified.

[0037] In one embodiment, the anomaly propagation analyzer includes forward tracing logic and backward impact localization logic. The forward tracing logic searches for the source events corresponding to donor qualification confirmation, acquisition task generation, and sample encoding binding level by level along the preceding event pointer based on the difference items output by the temporary layer processor. If the difference item is a cross-task binding, the forward tracing logic traces two or more acquisition task nodes from the sample encoding node and compares the corresponding donor nodes. If the difference item is a supplementary recording conflict, the forward tracing logic traces from the record correction node to the original record node and continues to search for the preceding events of the original record node. The backward impact localization logic searches for the affected events corresponding to acquisition completion, sample reception, anomaly registration, and record correction along the subsequent directed edges in the semantic state graph. If the subsequent directed edge has been written into the formal acquisition file, the affected event is marked as written as an impact. If the subsequent directed edge only exists in the temporary layer, the affected event is marked as an impact to be written, thereby forming an anomaly link set.

[0038] The aggregate value of the abnormal link set is calculated according to the following formula:

[0039] in, Indicates the first The aggregate value of a set of anomalous links represents, in physical terms, the comprehensive complexity of the same set of anomalous links in terms of the number of sources, the scope of impact, the version status, and the type of conflict. This indicates the number of affected events in the abnormal link set. Used to convert the number of affected events to a normalized value less than 1. This indicates whether the code involves multiple tasks using the same sample encoding; a value of 1 indicates involvement, and a value of 0 indicates non-involvement. This indicates whether subsequent events already included in the formal data collection file are involved; a value of 1 indicates involvement, and a value of 0 indicates no involvement. This indicates whether there are two or more difference item types; a value of 1 indicates the presence of such items, and a value of 0 indicates the absence of such items. , , and For aggregated weights with a sum of 1, for example , , , If a certain abnormal link set contains 3 affected events, involves multi-task occupancy, involves subsequent events already written, and the difference item type is 1, then The calculation results are used for sorting the abnormal link set and writing state configuration.

[0040] In this embodiment, the anomaly propagation analyzer classifies the set of anomaly links according to identity association anomalies, task flow anomalies, sample encoding anomalies, handover chain anomalies, and supplementary recording conflict anomalies. Identity association anomalies are triggered by the break of the preceding dependency chain between the donor node and the acquisition task node. Task flow anomalies are triggered by the state inversion between the acquisition task node, the sample encoding node, and the acquisition batch node. Sample encoding anomalies are triggered by the conflict of exclusive occupancy fields. Handover chain anomalies are triggered by the time window or preceding dependency conflict between the acquisition batch node and the handover record node. Supplementary recording conflict anomalies are triggered by the preceding reference conflict between the record correction edge and the original record edge. The classification results, forward tracing results, backward influence location results, comprehensive difference values, and aggregate values ​​are saved together. After reading the above data, the formal write control logic configures the newly added standardized event object with write-allowed, write-blocked, write-after-review, or read-only suspended states, thereby making the anomaly handling correspond to the specific state conflict type.

[0041] Furthermore, when the write state is in the write-allowed state, the formal write control logic writes the standardized event objects and the node states simulated by the temporary layer into the formal acquisition archive, and synchronously updates the current node state, source event, and previous state pointer in the semantic state diagram. When the write state is in the write-blocked state, only the summary, source, and blocking reason of the standardized event object are saved, without changing the formal semantic state diagram. When the write state is in the write-after-review state, the standardized event objects, temporary layer difference items, and abnormal link sets are bound as records to be reviewed. After the review is passed, the review result generates a new record correction event and enters the same state simulation process. When the write state is in the read-only suspended state, only the summary, source, and candidate predecessor relationship of the standardized event object are saved, and it is not allowed to be used as a predecessor event pointer for subsequent events. In this way, a controllable boundary is formed between the formal acquisition archive and the data to be reviewed.

[0042] In a preferred embodiment, the anomaly link merging logic is used to handle situations where multiple discrepancies correspond to the same donor, the same acquisition task, or the same sample code. After receiving multiple anomaly link sets output by the anomaly propagation analyzer, the anomaly link merging logic maps each discrepancy to a common source event and a common affected event. The common source event is determined by the earliest non-revoked event found by the forward tracing logic, and the common affected event is determined by the latest written event or the latest pending written event found by the backward influence location logic. If multiple discrepancies have the same common source event, they are merged according to the common source event. If the common source events are different but the sample code nodes are the same, they are merged according to the common sample code. If the common sample codes are different but the handover record nodes are the same, they are merged according to the common handover record. The merged anomaly link set retains the original type, preceding event pointer, subsequent event pointer, and temporary layer identifier of each discrepancy, thereby enabling the review record to simultaneously present the anomaly merging path and the source of a single discrepancy.

[0043] In this embodiment, the abnormal link merging logic does not delete the merged difference items, but generates a link member identifier for each difference item. The link member identifier is stored together with the identifier of the merged abnormal link set. If a certain supplementary recording conflict exception causes both time causality reversal and handover chain exception, the time causality reversal difference item still retains its event occurrence time comparison result, and the handover chain exception difference item still retains its pre-dependency comparison result between the collection batch node and the handover record node. The merged abnormal link set is only used as the upper-level object for writing state configuration, generating records to be reviewed, and generating continuous summary records. This avoids the merging process from obscuring specific state conflicts and maintains the data correspondence between the source of the exception, the spread of the exception, and the handling of the exception.

[0044] In one embodiment, the source tracing record generator includes anomaly evidence solidification logic. When the formal write control logic generates a blocked write state, a post-review write state, or a read-only suspended state, the anomaly evidence solidification logic extracts the event subject identifier, event object identifier, event type, preceding event pointer, temporary layer difference item, anomaly link set, and write state of the corresponding standardized event object, and generates continuous summary records according to the event write order. The continuous summary records are respectively associated with the donor node, the acquisition task node, the sample encoding node, and the handover record node. If the same standardized event object corresponds to multiple anomaly link sets, the continuous summary record saves the mapping relationship between the identifiers of multiple anomaly link sets and the identifier of the same event object. If the same anomaly link set corresponds to multiple difference items, the continuous summary record saves the link member identifier of each difference item. This allows the anomaly evidence to retain an independent record outside the formal archive and maintain a reference relationship with the semantic state diagram.

[0045] The consecutive summary records generate a chained summary index according to the following formula:

[0046] in, Indicates the first A chained summary index of consecutive summary records, physically representing the sequential binding identifier between the current record and the previous record. This represents the linked summary index of the previous consecutive summary record. Indicates the first The encoded field summary values ​​of the event subject identifier, event object identifier, event type, preceding event pointer, temporary layer difference item, abnormal link set, and write status in each consecutive summary record. Represents the chained recursive multiplier. Indicates the preset summary modulus. This represents the modulo operation, for example. , , hour, The calculation result is used as the first A chained summary index of consecutive summary records and associated with the corresponding state nodes.

[0047] Furthermore, the abnormal evidence solidification logic also saves the record generation time, source record status, and associated node set when generating continuous summary records. The associated node set includes at least the donor node, acquisition task node, sample encoding node, and handover record node that have a connection relationship with the abnormal link set. If the write status is blocked write status, the continuous summary record saves the blocking reason and the corresponding difference item type. If the write status is reviewed write status, the continuous summary record saves the record identifier to be reviewed and the status before review. If the write status is read-only suspended status, the continuous summary record saves the candidate preceding relationship set and the read-only suspension reason. After the continuous summary record is generated, it does not change the current status of the nodes in the formal semantic state graph, but forms a read-only reference with the relevant nodes, thereby preserving abnormal handling evidence and avoiding unconfirmed events from changing the formal acquisition link.

[0048] In a preferred embodiment, the system partitions and stores formal acquisition archives, temporary layer records, anomaly link sets, and continuous summary records. Formal acquisition archives store standardized event objects and semantic state graph nodes that have passed state simulation. Temporary layer records store newly acquired events to be simulated or those that failed simulation, as well as local mirror state relationships. Anomaly link sets store source events, affected events, and difference types generated by forward tracing and backward impact location. Continuous summary records store chained summary indexes and associated node sets during anomaly handling. Partitioning does not change the reference relationships between data objects. Each partition is interconnected through event object identifiers, preceding event pointers, anomaly link set identifiers, and chained summary indexes, thereby making the boundaries between formal data, data to be reviewed, and anomaly evidence clear and preventing supplementary or corrected data from directly overwriting confirmed states.

[0049] In this embodiment, the system can use a transactional writing approach to process formal collection archives and continuous summary records. When a new standardized event object is simulated through a temporary layer and obtains a write-allowed state, the update of formal semantic state graph nodes, the writing of formal collection archives, and the generation of chained summary indexes are bound in the same writing transaction. When a new standardized event object obtains a blocked writing state, a write-after-review state, or a read-only suspended state, the formal semantic state graph nodes are not updated, and only the continuous summary records and corresponding state references generated by the abnormal evidence solidification logic are written. If the generation of any data object in the writing transaction fails, the writing does not change the formal collection archives and retains the temporary layer records for subsequent replay, thereby keeping the state simulation results, formal archive writing results, and source tracing evidence consistent.

[0050] In one embodiment, the system adopts the same standardized processing path for record correction events as for ordinary acquisition events. After a record correction event enters the event object generator, it must carry the pointer to the original record event. If the pointer to the original record event is missing, the event semantic completion logic generates candidate original record relationships based on the same event object identifier, the same operation source identifier, and the writing time of adjacent records. The constraint template matching logic writes the pre-dependency field and the irreversible transition field of the record correction edge. The temporary layer processor inserts the record correction event after the original record node and performs local replay. The difference output logic compares the node state, edge attributes, and subsequent event inheritance relationship before and after the correction. The anomaly propagation analyzer performs backward impact location on the acquisition completion, sample reception, and anomaly registration events that the correction may affect. This ensures that record correction will not change the existing acquisition link by simply replacing fields.

[0051] Preferably, the system uses exclusive occupancy index processing for sample encoding binding events. The sample encoding node stores the current binding task, binding state, and previous state pointer in the semantic state graph. When a new sample encoding binding event enters the temporary layer, the constraint verification engine reads the exclusive occupancy field and searches whether the same sample encoding node already has an encoding binding edge for an unclosed acquisition task. If an unclosed acquisition task exists and the current acquisition task is inconsistent with the existing acquisition task, the difference output logic generates duplicate occupancy difference items and cross-task binding difference items. The anomaly propagation analyzer traces forward along the two acquisition task nodes to their respective donor qualification confirmation events and backward along the sample encoding node to find acquisition completion and sample reception events. This allows the implicit occupancy conflict of the same sample encoding between different tasks to form an anomaly link before writing.

[0052] In a preferred embodiment, the system employs a handover chain verification process for sample reception events. The preceding event pointer of the sample reception event points to the acquisition completion event. The semantic state graph generator establishes a handover confirmation edge between the acquisition batch node and the handover record node. The handover confirmation edge carries a preceding dependency field and a time window field. When simulating the sample reception event, the temporary layer processor copies the sample encoding node, the acquisition batch node, the handover record node, and the associated directed edge. The constraint verification engine compares whether the occurrence time of the sample reception event is later than the acquisition completion event and reads whether the sample encoding node is in the acquired state. If the sample reception event lacks a preceding acquisition completion relationship or occurs earlier than the acquisition completion event, the difference output logic generates a handover chain abnormal difference item. The abnormal propagation analyzer searches for abnormal registration and record correction events along the handover record node, thereby ensuring that the state continuity relationship between the reception record and the acquisition completion record remains computable.

[0053] Furthermore, the system employs object binding processing for abnormal registration events. After entering the event object generator, an abnormal registration event must contain a pointer to the registered event. The pointer to the registered event can point to a donor qualification confirmation event, a collection task generation event, a sample encoding binding event, a collection completion event, a sample reception event, or a record correction event. The semantic state graph generator establishes abnormal registration edges based on the pointers to the registered events. The constraint template matching logic writes the abnormal registration edges into the pre-dependency field and the traceability field. When simulating abnormal registration events, the temporary layer processor copies the corresponding node of the registered event and its preceding and following directed edges. If the registered event has been revoked or is in a read-only suspended state, the difference output logic generates an abnormal association broken chain difference item. The abnormal propagation analyzer merges or establishes a reference relationship between the abnormal registration event and the abnormal link set to which the registered event belongs, thereby preventing abnormal registration records from becoming isolated text records separated from specific collection events.

[0054] In one embodiment, the system processes the time sequence by simultaneously reading the event occurrence time and the record writing time. The event occurrence time is used to determine the sequential order of the acquisition process itself, and the record writing time is used to determine the order in which data enters the system. When processing supplementary recording events, the temporary layer processor does not place them directly at the end of the acquisition link simply because the record writing time is late. Instead, it determines their simulated position in the state chain based on the event occurrence time and the preceding event pointer. If the insertion of a supplementary recording event causes a change in the preceding dependency of subsequent written acquisition completion, sample reception, or anomaly registration events, the difference output logic outputs the corresponding difference item and hands the affected event over to the anomaly propagation analyzer for processing. This allows the supplementary recording data to participate in state causal calculations while preventing the supplementary recording data from altering the existing formal link without verification.

[0055] In this embodiment, the previous state pointer of a state node is used to support local replay. When the temporary layer processor copies a local node, it also copies the previous state pointer. The transition sequence construction logic restores the state of the node before the insertion of the new event according to the previous state pointer. After the simulation of the new event is completed, the node state is replayed in the order of subsequent events. If any subsequent event lacks the required preceding state in the replay, the difference output logic generates a state chain break difference item. If any subsequent event causes the node state to regress to a state before the irreversible state, the difference output logic generates a reverse transition difference item. This allows the system to locate the specific damage location of the new event to the historical link and subsequent links, rather than just recording the current event write failure.

[0056] In a preferred embodiment, the record to be reviewed corresponding to the write state after review still follows the standardized event object structure. The review result generated by the reviewer or review program does not directly modify the formally collected archive, but generates a new record correction event or anomaly closure event which enters the event object generator. The newly generated event carries the identifier of the record to be reviewed, the identifier of the original event object, and the identifier of the review source. The semantic state graph generator attaches the newly generated event to the original anomaly link set. The temporary layer processor performs state transition simulation again. If the review result can eliminate the conflict of prerequisite dependencies or the conflict of candidate prerequisite relationships, the formal write control logic updates the write state and generates a new continuous summary record. If the review result still has exclusive occupancy or time causality reversal, it continues to be blocked or read-only suspended, thereby making the review process itself enter a traceable data link.

[0057] Furthermore, when generating source tracing records, the system uses a three-way reference system for the abnormal link set, continuous summary records, and semantic state graph nodes. The abnormal link set stores the path between the source event and the affected event, the continuous summary records store the sequential binding identifier of the abnormal handling process, and the semantic state graph nodes store the current state and the pointer to the previous state. When subsequent sample encoding, collection tasks, or donor node queries are performed, the system locates the relevant collection links through the semantic state graph nodes, finds the source and scope of the abnormality through the abnormal link set, and restores the processing order of blocking, review, suspension, or formal writing through the continuous summary records. This enables the source tracing of collected data to be cross-verified from three data perspectives: node status, abnormal path, and processing evidence.

[0058] In one embodiment, the system employs irreversible transition processing for new events added after the acquisition task is closed. When the acquisition task node is in the closed state, the constraint template matching logic reads the irreversible transition fields in the task issuance edge, encoding binding edge, and acquisition confirmation edge. When the temporary layer processor simulates the new sample encoding binding event, acquisition completion event, or record correction event, it uses the acquisition task closed state as the preceding state. If the new event attempts to change the closed task back to the pending acquisition, bound, or acquired state, the difference output logic generates a reverse transition difference item and hands it over to the anomaly propagation analyzer to search for subsequent sample reception and anomaly registration events along the acquisition task node, thereby preventing the closed state from being directly rolled back by supplementary recording or correction events.

[0059] In this embodiment, the system also retains the original field set generated by the event object generator for events that are allowed to be written to. The standardized event objects saved in the formal collection archive are connected to the semantic state graph nodes through the event object identifier. If a record correction event occurs later, the correction object does not directly overwrite the original event field, but associates the original event object through the original record event pointer and forms a correction edge in the temporary layer for simulation. The traceability record generator generates continuous summary records when the correction is passed or blocked, thereby ensuring that each collection confirmation, handover reception, anomaly registration and record correction maintains an independent data trace.

[0060] In a preferred embodiment, the system can use candidate preceding relationships, temporary layer difference items, abnormal link sets, and write status as the output data content of the review interface or external query interface. However, the output process does not change any state nodes or directed edge attributes. The query request can only read read-only data associated with the event object identifier. If the query object is in a read-only suspended state, the query result includes an event object summary, operation source identifier, candidate preceding relationship set, and read-only suspension reason. If the query object is in a blocked write state, the query result includes an event object summary, blocking reason, and abnormal link set identifier. If the query object is in a post-review write state, the query result includes the record identifier to be reviewed and temporary layer difference items. This enables the system to provide the data evidence required for tracing the source, while avoiding changes to the acquisition link during the query operation.

[0061] In one embodiment, the system as a whole uses event objects, semantic state diagrams, temporary layers, anomaly link sets, and continuous summary records as the main data objects. Event objects provide a unified input structure, semantic state diagrams provide state causal expressions, temporary layers provide a simulated environment before writing, anomaly link sets provide the location of anomaly sources and impact ranges, and continuous summary records provide the binding of anomaly handling order. During operation, each newly added acquisition event forms a closed processing link through event standardization, semantic state mapping, constraint verification, temporary layer simulation, anomaly propagation analysis, write state configuration, and source tracing record generation. Thus, even when there is supplementation, correction, delayed writing, or multi-node synchronization of donor sample acquisition data, the consistent correlation between the acquisition process status, anomaly source, and processing evidence can still be maintained.

Claims

1. A smart monitoring and data security traceability system for donor sample collection, characterized in that, The system includes an event object generator, a semantic state graph generator, a constraint verification engine, a temporary layer processor, an anomaly propagation analyzer, and a source tracing record generator. The event object generator receives acquisition events corresponding to donor qualification confirmation, acquisition task generation, sample encoding binding, acquisition completion, sample reception, anomaly registration, and record correction, and generates standardized event objects. The semantic state graph generator establishes directed state relationships between donors, acquisition tasks, sample encoding, acquisition batches, and handover records based on the standardized event objects. The constraint verification engine performs pre-dependency, time sequence, exclusive occupancy, and irreversible transition checks on the directed state relationships. The temporary layer processor performs state transition simulation before the standardized event objects are written into the formal acquisition archive. The anomaly propagation analyzer generates anomaly links based on the simulation results. The source tracing record generator saves the association records between anomaly links and corresponding acquisition events.

2. The intelligent monitoring and data security traceability system for donor sample collection according to claim 1, characterized in that, The event object generator includes event source normalization processing logic and event semantic completion logic. The event source normalization processing logic converts the collected events formed by manual input, batch import, interface synchronization, and review and correction into a unified event format. The unified event format includes event subject identifier, event object identifier, event type, event occurrence time, record writing time, preceding event pointer, operation source identifier, and record status identifier. The event semantic completion logic generates candidate preceding relationships for collected events lacking preceding event pointers based on the existing relationship between collection tasks, sample codes, and donor qualification status, and submits the candidate preceding relationships to the constraint verification engine along with the standardized event object.

3. The intelligent monitoring and data security traceability system for donor sample collection according to claim 1, characterized in that, The semantic state graph generator includes node hierarchical logic and edge attribute configuration logic. The node hierarchical logic generates state nodes according to the donor layer, acquisition task layer, sample encoding layer, acquisition batch layer, and handover record layer, and configures the current state, source event, and previous state pointer for each state node. The edge attribute configuration logic generates directed edges according to the event types of qualification confirmation, task issuance, encoding binding, acquisition confirmation, handover confirmation, anomaly registration, and record correction, and writes the pre-dependency type, reversibility type, exclusive occupancy type, and time window type into the directed edges, so that a computable state association structure is formed between the same donor, the same acquisition task, and the same sample encoding.

4. The intelligent monitoring and data security traceability system for donor sample collection according to claim 1, characterized in that, The temporary layer processor includes mirror graph generation logic, transition sequence construction logic, and difference output logic. The mirror graph generation logic copies and adds local nodes and directed edges related to standardized event objects based on the formal semantic state graph. The transition sequence construction logic arranges the state transitions to be simulated according to the preceding event pointer, event occurrence time, and record writing time. The difference output logic marks the node states, edge attributes, preceding dependencies, and occupancy relationships before and after the simulation, and sends the difference items corresponding to state chain breakage, reverse transition, duplicate occupancy, cross-task binding, and time causality inversion to the anomaly propagation analyzer.

5. The intelligent monitoring and data security traceability system for donor sample collection according to claim 2, characterized in that, The event semantic completion logic includes a multi-candidate preceding relationship filtering logic. When a standardized event object has more than two candidate preceding relationships, the multi-candidate preceding relationship filtering logic generates a candidate relationship set according to four conditions: the matching relationship between the donor qualification status and the collection task status, whether the sample code has been occupied, whether the event occurrence time is within the valid range of the collection task, and whether the record correction event references the original record. For each candidate preceding relationship in the candidate relationship set, a relationship source, a relationship conflict field, and a relationship confirmation identifier are written, so that the candidate preceding relationships maintain a distinguishable data structure before entering the semantic state graph generator.

6. The intelligent monitoring and data security traceability system for donor sample collection according to claim 3, characterized in that, The edge attribute configuration logic includes constraint template matching logic. The constraint template matching logic calls the corresponding constraint template according to the starting state node, target state node and event type of the directed edge connection. The constraint template includes an irreversible transition field that prohibits rollback, an exclusive occupancy field that restricts the same sample code from being bound to multiple acquisition tasks at the same time, a time window field that restricts acquisition confirmation from being later than task issuance and earlier than handover confirmation, and a pre-dependency field that restricts record correction from pointing to the original record. After the constraint template is written into the directed edge attribute, a verification index associated with the state node is generated synchronously.

7. The intelligent monitoring and data security traceability system for donor sample collection according to claim 4, characterized in that, The anomaly propagation analyzer includes forward tracing logic and backward impact localization logic. The forward tracing logic searches for the source events corresponding to donor qualification confirmation, collection task generation, and sample encoding binding along the preceding event pointer based on the difference items output by the temporary layer processor. The backward impact localization logic searches for the affected events corresponding to collection completion, sample reception, anomaly registration, and record correction along the subsequent directed edges in the semantic state graph, and generates anomaly link sets according to identity association anomalies, task flow anomalies, sample encoding anomalies, handover chain anomalies, and supplementary recording conflict anomalies.

8. The intelligent monitoring and data security traceability system for donor sample collection according to claim 4, characterized in that, The difference output logic includes formal write control logic. The formal write control logic configures the write state of the newly added standardized event object based on the abnormal link set generated by the abnormal propagation analyzer. The write state includes write-allowed state, write-blocked state, write-after-review state, and read-only suspended state. When the write state is write-after-review, the formal write control logic binds the standardized event object, temporary layer difference item, and abnormal link set as a record to be reviewed. When the write state is read-only suspended, it only saves the summary, source, and candidate predecessor relationship of the standardized event object.

9. The intelligent monitoring and data security traceability system for donor sample collection according to claim 7, characterized in that, The anomaly propagation analyzer also includes anomaly link merging logic. When multiple discrepancies correspond to the same donor, the same acquisition task, or the same sample encoding, the anomaly link merging logic maps the multiple discrepancies to common source events and common affected events, respectively. The anomaly link set is merged in the order of common source events first, common sample encoding second priority, and common handover records third priority. The original type, preceding event pointer, subsequent event pointer, and temporary layer identifier of each discrepancy item are retained in the merged anomaly link set.

10. The intelligent monitoring and data security traceability system for donor sample collection according to claim 8, characterized in that, The source tracing record generator includes anomaly evidence solidification logic. When the formal write control logic generates a blocked write state, a post-review write state, or a read-only suspended state, the anomaly evidence solidification logic extracts the event subject identifier, event object identifier, event type, preceding event pointer, temporary layer difference item, anomaly link set, and write state of the corresponding standardized event object. It generates continuous summary records according to the event write order and associates the continuous summary records with the donor node, the acquisition task node, the sample encoding node, and the handover record node, respectively.

Images