Knowledge retrieval and quality control calculation collaborative report generation system for environmental protection detection

Abstract

The invention provides an environmental protection detection-oriented knowledge retrieval and quality control calculation collaborative report generation system, and establishes a detection report automatic generation technical route taking knowledge retrieval and quality control calculation as a core, taking an evidence chain and a structured constraint as constraints, and taking automatic compliance verification and audit tracing as a guarantee. A verifiable engineering path is provided for the problems of clause conflict, model instability, difficulty in covering a complex scene by a template and the like, and compared with a traditional template filling type scheme and a pure model generation type scheme, the scheme has remarkable advantages in the aspects of consistency, traceability and compliance; and rapid adaptation and long-term operation and maintenance of multiple detection scenes are supported by versioning knowledge and pluggable rules, so that good tool support can be provided for environmental protection detection work, and environmental management work can be promoted to be better carried out.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of environmental protection detection, in particular to a report generation system for knowledge retrieval and quality control calculation cooperation of environmental protection detection. BACKGROUND

[0002] Environmental protection detection is an important process in environmental management, but the application inventors have found that in existing environmental detection and testing services, the generation of environmental protection detection reports generally relies on manual information collection among multiple source systems, including information such as commissioning contracts and sample information, sampling records and on-site detection data, laboratory instrument raw data, national and local standard clauses and limits, historical reports and experience templates, and due to the large differences in data structure, semantic expression, time granularity and quality requirements of various types of information, manual alignment and verification is time-consuming, inefficient and prone to errors. When the report involves multiple quality control indicators such as parallel sample relative deviation, blank deduction, standard addition recovery rate, detection limit and quantification limit, calibration curve and linear correlation coefficient, existing template filling type automation cannot guarantee the consistency of the determination logic and the standard clauses.

[0003] In addition, as environmental protection detection scenarios expand to oil and gas recovery, non-mobile source emissions, wastewater and exhaust gas, soil and groundwater, and other fields, the demand for dynamic adaptation of report templates and determination rules is increasing, and existing technologies lack an end-to-end technical route that can uniformly carry out "knowledge retrieval-rule calculation-evidence chain-constrained generation-compliance verification-signing audit", making it difficult to meet the requirements of traceability and consistency of supervision. SUMMARY

[0004] The application provides a report generation system for knowledge retrieval and quality control calculation cooperation of environmental protection detection, which establishes a detection report automatic generation technical route with knowledge retrieval and quality control calculation as the core, evidence chain and structured constraints as the constraints, and automatic compliance verification and audit traceability as the guarantee. It provides a verifiable engineering path for problems such as clause conflicts, model instability, and template difficulty in covering complex scenarios. Compared with traditional template filling and pure model generation schemes, it has significant advantages in consistency, traceability and compliance, and supports rapid adaptation and long-term operation in multiple detection scenarios with versioned knowledge and pluggable rules. This can provide good tool support for environmental protection detection work and promote the better development of environmental management work.

[0005] In a first aspect, the application provides a report generation system for knowledge retrieval and quality control calculation cooperation of environmental protection detection, the system comprising: The data access and governance module is configured to configure the identification information and stage permissions of each participant in the whole process of environmental protection detection, and to perform data access and governance to form a unified data view and establish associations between different objects. The knowledge retrieval and clause positioning module is configured to perform knowledge retrieval and clause positioning based on the unified data view, and generate an executable rule list; The quality control calculation module is configured to perform quality control calculation based on the executable rule list to generate a structured quality control result; The evidence chain construction module is configured to dynamically construct an evidence chain package according to a report conclusion related process; The constrained report generation module is configured to generate a report draft based on the evidence chain package under the constraints of a report Schema and a template; The automatic compliance verification module is configured to perform compliance verification on the report draft and locate field abnormal information; The man-machine collaborative review module is configured to display report generation results, compliance verification conclusions and evidence chain references, and make manual review and revision; The version freezing and signing module is configured to perform version freezing operation when the compliance verification passes and the review state is signable, and bind the corresponding electronic signature; The export and archiving module is configured to index the report and the evidence chain, export corresponding PDF files, Word files and structured data exchange packages, and perform archiving; The whole-process audit and backtracking module is configured to record key events and difference snapshots in the whole process, and perform traceability retrieval when needed.

[0006] In a second aspect, the application provides a report generation method for knowledge retrieval and quality control calculation cooperation of environmental protection detection, the method is applied to a report generation system, and the method comprises the following steps: Data access and governance are configured with identification information and stage permissions of each participant in the whole process of environmental protection detection, and data access and governance are performed to form a unified data view, and an association between different objects is established; Knowledge retrieval and clause positioning are performed based on the unified data view, and an executable rule list is generated; The quality control calculation module performs quality control calculation based on the executable rule list to generate a structured quality control result; The evidence chain construction module dynamically constructs an evidence chain package according to a report conclusion related process; The constrained report generation module generates a report draft based on the evidence chain package under the constraints of a report Schema and a template; The automatic compliance verification module performs compliance verification on the report draft and locates field abnormal information; The man-machine collaborative review module displays report generation results, compliance verification conclusions and evidence chain references, and makes manual review and revision; The version freezing and signing module performs version freezing operation when the compliance verification passes and the review state is signable, and binds the corresponding electronic signature; The export and archiving module indexes the report and evidence chain, exports corresponding PDF files, Word files and structured data exchange packages, and archives them; The whole-process auditing and backtracking module records key events and difference snapshots in the whole process, and performs source retrieval when needed.

[0007] In a third aspect, the present application provides a computer-readable storage medium, which stores a plurality of instructions adapted to be loaded by a processor to execute the method provided in the second aspect of the present application.

[0008] From the above, the present application has the following beneficial effects: For the generation of high-quality environmental protection detection reports, the present application establishes a detection report automatic generation technology route with knowledge retrieval and quality control calculation as the core, evidence chain and structured constraints as the constraints, and automatic compliance verification and audit tracking as the guarantee. It provides a verifiable engineering path for problems such as clause conflict, model instability, and template difficulty in covering complex scenarios. Compared with traditional template filling and pure model generation schemes, it has significant advantages in consistency, traceability and compliance. It supports rapid adaptation and long-term operation in multiple detection scenarios with versioned knowledge and pluggable rules. In this way, it can provide good tool support for environmental protection detection work and promote environmental management work to be better carried out. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0010] Figure 1 A system architecture diagram of a report generation system of the present application for knowledge retrieval and quality control calculation cooperation of environmental protection detection; Figure 2 A flowchart of a report generation method of the present application for knowledge retrieval and quality control calculation cooperation of environmental protection detection. DETAILED DESCRIPTION

[0011] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0012] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules is not necessarily limited to those explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices. The naming or numbering of steps appearing in this application does not imply that the steps in the method flow must be performed in the chronological / logical order indicated by the naming or numbering. The execution order of named or numbered process steps can be changed according to the desired technical purpose, as long as the same or similar technical effect is achieved.

[0013] The module division described in this application is a logical division. In practical applications, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual coupling, direct coupling, or communication connections may be through interfaces, and the indirect coupling or communication connections between modules may be electrical or other similar forms, none of which are limited in this application. Moreover, the modules or sub-modules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed across multiple circuit modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution in this application.

[0014] This application provides a report generation system for environmental monitoring that integrates knowledge retrieval and quality control calculations, targeting high-quality non-mobile source detection. Figure 1 The diagram illustrates a system architecture for a report generation system based on knowledge retrieval and quality control calculation for environmental monitoring, as described in this application. From a functional module perspective, this system can include modules such as data access and governance, knowledge retrieval and clause location, quality control calculation, evidence chain construction, constrained report generation, automatic compliance verification, human-machine collaborative review, version freezing and issuance, export and archiving, and full-process auditing and backtracking. These modules correspond to the major stages of system operation: data access and governance, knowledge retrieval and clause location, quality control calculation, evidence chain construction, constrained report generation, automatic compliance verification, human-machine collaborative review, version freezing and issuance, export and archiving, and full-process auditing and backtracking.

[0015] Specifically, the modules / stages can be connected using a unified domain data model and message interface. The output of each stage is an information object with version identifier and hash verification to ensure the data verifiability and reproducibility across stages.

[0016] In terms of specific hardware implementation, the system in this application may involve specific deployment forms such as server-side (involving backend / cloud servers), server-side and mobile terminals (user-side devices), and local processing devices.

[0017] Specifically, user-side devices can include terminal devices such as personal digital assistants (PDAs), tablets, laptops, all-in-one computers, smartphones, and smart bracelets. Local processing devices can include desktop computers, all-in-one computers, and other devices deployed locally / on-site that are not easily relocated.

[0018] As an example, the system in this application can be specifically applied to the server-side and mobile-side collaborative scenarios of a Laboratory Information Management System (LIMS). The server-side is responsible for version management of enterprise / laboratory master data, knowledge and rules, constraint generation and compliance verification, human-machine collaborative review and audit traceability, while the mobile-side is responsible for the collection and transmission of on-site sampling, barcode and location, images and videos, and in-situ / portable testing data.

[0019] Furthermore, the operational roles on the server side can be further divided into platform administrators, lab administrators, and general users, and roles such as reviewers, signatories, compliance officers, and auditors can be introduced as enhancements when necessary.

[0020] Specifically, the platform administrator can be configured to have platform management and security policy configuration permissions, and can complete the registration and review of institutions and laboratories, the distribution of roles and access control policies, log retention and disaster recovery policy settings on the server side; Laboratory administrators can be configured to have permissions such as information configuration, plan and batch management, quality control rules and template management, and interface and instrument binding management. They can create projects and commissions, maintain sample and batch master data, and configure the visibility scope and effectiveness strategy of knowledge and rule versions on the server side. General users can be specifically configured to include testing analysts and sampling personnel, with corresponding permissions such as sample and project execution, data entry and verification, on-site task processing and feedback. The reviewer can be specifically configured to have report review and revision permissions; The issuer can be specifically configured to have report version freezing and electronic signature permissions; The compliance officer can be specifically configured to have the authority to adapt to compliance rules and conduct random checks and reviews; Auditors can be specifically configured to have read-only traceability permissions across reports and versions.

[0021] To improve system reliability, in terms of hardware, the system of this application may preferably adopt a microservice architecture and containerized deployment, message bus or service orchestration to achieve high system performance such as cross-module communication. Key modules have idempotent processing and breakpoint resume capabilities. Rules and knowledge are stored in versioned form and support canary release. The generation module has load balancing and circuit breaker degradation strategies. When the model is unavailable or the output does not meet the constraints for a long time, it will automatically fall back to the template filling mode to ensure business continuity.

[0022] To ensure consistency across devices, the corresponding computer programs can be configured to load rules and knowledge versions via drivers. During operation, key intermediate results and configuration snapshots are signed and hashed to ensure the comparability and legal validity of experiments reproduced at different points in time.

[0023] In terms of application scenarios, this application can specifically involve environmental monitoring scenarios such as oil and gas recovery, non-mobile source emissions, wastewater and exhaust gas, soil and groundwater, and can also be combined with industrial IoT, Internet data services or artificial intelligence (AI) to carry out specific solution applications.

[0024] The following section will describe the high-quality processing performance of the report generation system for environmental monitoring reports, based on the detailed description of the above modules.

[0025] (I) Data Access and Governance Module The data access and governance module is used to configure the identification information and stage permissions of each participating entity throughout the environmental monitoring process, as well as to perform data access and governance to form a unified data view and establish the relationship between different objects.

[0026] As can be seen, the data access and governance module is located on the access side of the system and specifically involves three major aspects of data processing.

[0027] On the one hand, for the data access and governance module here, the identification information of each participating entity can specifically include the identification information of the client, the identification information of the sampling unit, the identification information of the testing laboratory, and the identification information of the regulatory agency, etc. The stage permissions of each participating entity include viewing permissions and modification permissions, etc.

[0028] Among these features, when necessary, a dual-person review strategy and a waiting window strategy can be set for relevant review and issuance operations to better manage and apply stage permissions.

[0029] As an example of the stage permissions of each participating entity, in specific applications, the system of this application can have the following corresponding permission configuration contents: Configure the client's identification information viewing permissions during the delegation process (or delegation stage, the same applies below), as well as restricted viewing permissions during the report generation process; Configure viewing and modification permissions for sampling unit identification information during the sampling process, as well as read-only permissions during the report generation process; Configure the viewing and modification permissions of the testing laboratory identification information in the sample receiving and login stage, the preparation and testing stage, and the quality control and verification stage, as well as the restricted permissions in the issuance stage; Configure read-only permissions and sampling review permissions for regulatory agency identification information during the report generation and archiving audit stages.

[0030] As is easily understood, the aforementioned permissions can be dynamically adjusted through policy issuance without altering the business semantics, in order to adapt to the positions and regulatory requirements of different institutions.

[0031] On the other hand, the data access and governance module can access multi-source data, including commission information, sample information, sampling records, instrument raw data, historical reports and regulatory standards knowledge, and can form a unified data view through governance operations such as field standardization, unit conversion, timestamp sample identification alignment, missing value imputation and outlier identification. The unified data view includes three core objects: sample entities, test item entities and measurement entities.

[0032] Specifically, the sample entity may involve elements such as unique sample identifier, sample matrix, source and storage conditions; the detection item entity may involve elements such as item code, standard method number, target substance and limit system; and the measurement entity may involve elements such as raw reading, processed reading, detection limit and quantitation limit reference, calibration batch, instrument identification and environmental parameters.

[0033] On the other hand, for the data access and governance module, the relationship between different objects (mainly the numerous data accessed above) can be established by foreign keys and timestamps. Specifically, it can be a many-to-one relationship or a one-to-many relationship, and specifically, it can be a structured relationship in the form of sample-project-batch. This helps to ensure that any conclusion can be traced back to a specific batch, specific instrument and specific operation time in practical applications, forming a good scenario backtracking effect.

[0034] (II) Knowledge Retrieval and Clause Location Module The knowledge retrieval and clause location module is used to perform knowledge retrieval and clause location based on a unified data view, and generate a list of executable rules.

[0035] As can be seen, the input to the knowledge retrieval and clause location module can involve the unified data view formed by the preceding data access and governance module, which is easy to view and process.

[0036] As an example, in some cases, the sample matrix, test items, applicable standard versions, operating conditions, and batch contexts involved in the unified data view can be entered.

[0037] Furthermore, for the knowledge retrieval and clause location module, a knowledge graph can be used to organize standards and methodologies for knowledge retrieval and clause location. The nodes of a knowledge graph can specifically include standards, clauses, limits, methods, interference items, quality control rules and versions. The edges of the graph include relationships such as applicable, reference, substitution, effective, applicable scenario, and conflict. Furthermore, to resolve conflicts or coverage issues, version selection and conflict resolution strategies can be introduced. Specifically, these strategies can prioritize candidate clauses based on time priority, regulatory priority, and localization adaptation priority, selecting the most applicable clause while retaining unselected clauses as alternative evidence for audit explanation.

[0038] Thus, after processing by the knowledge retrieval and clause location module, a list of rules and clause metadata with executable semantics can be output, namely the list of executable rules. It is expressed using an interpretable rule description language and has the characteristics of parameterization, composability, and traceability.

[0039] Among them, the knowledge graph can also be replaced by a combination of high-performance retrieval index and rule base.

[0040] (III) Quality Control Calculation Module The quality control calculation module is used to perform quality control calculations based on a list of executable rules to generate structured quality control results.

[0041] As can be seen, the input of the quality control calculation module can involve the list of executable rules generated by the knowledge retrieval and clause location modules to construct batch-level and project-level calculation topologies, map relevant content, such as measurement entities, to formula parameters, complete the calculation of different aspects of pre-set requirements, and generate structured quality control results according to the allowable range and judgment logic defined by the clauses.

[0042] Specifically, for the quality control calculation module, the quality control calculations performed may include the following processing: Perform specific control calculations including relative deviation of parallel samples, blank subtraction, spike recovery, limit of detection, limit of quantitation, and linearity of calibration curve, and generate corresponding structured quality control results according to the allowable range and judgment logic defined in the terms.

[0043] The formulas, thresholds, and boundary conditions for the series of indicators involved are as follows: The relative deviation of parallel samples can be calculated by dividing the absolute value of the difference between the two values ​​by the mean and multiplying by a percentage in the case of two measurements. In the case of multiple measurements, the difference between the maximum and minimum values ​​can be divided by the average and multiplied by a percentage. If necessary, the relative standard deviation form can be used as specified in the standard. When the measured value is in the low concentration range, the relaxed threshold can be used within the allowable range of the clause. Blank subtraction can specifically use the average or representative value of the blank sample as the background quantity, subtract it from the sample measurement value, and apply a negative value truncation or lower limit zeroing strategy to the subtracted result. The specific strategy selection can be provided by the metadata of the clauses involved in the list of executable rules. Specifically, the recovery rate can be calculated by comparing the difference between the concentration of the spiked sample and the concentration of the unspecified sample with the theoretical spiked concentration. This recovery rate is then compared with the upper and lower limits set in the clause. If the upper or lower limits are exceeded, a retest or explanation of the reason must be inserted in the subsequent generation stage. The detection limit and quantitation limit can be estimated using a method based on the standard deviation of the blank sample and the regression slope, or converted according to the signal-to-noise ratio threshold specified in the methodology. The specific algorithm and parameter values ​​can be indicated by the list of executable rules. The linearity of the calibration curve can be obtained by fitting a least-squares straight line to obtain the correlation coefficient and comparing it with a threshold. It supports extended forms such as piecewise linear and weighted regression. When the threshold is not met, the batch curve is determined to be invalid and the issuance process is terminated.

[0044] In addition, to suppress occasional anomalies and improve robustness, outlier identification strategies can be introduced, such as robustness thresholds based on the absolute deviation of the median or extreme value identification based on the Grubbs test. When this strategy is enabled, it can also be configured so that the removal of outliers must be recorded in the chain of evidence and can be overridden by the reviewer.

[0045] (iv) Evidence Chain Construction Module The evidence chain construction module is used to dynamically construct evidence chain packages based on the processing related to the conclusions of a report.

[0046] Understandably, the input to the evidence chain construction module involves the work content of multiple processing modules / stages, and may also involve subsequent processing modules / stages, in order to dynamically construct an evidence chain package related to the processing of a certain report conclusion. This evidence chain package is also a major part of the environmental monitoring report processing work, and most of the content in the report is extracted from this evidence chain package.

[0047] In terms of details, the evidence chain construction module can be specifically encapsulated into an evidence chain package based on a certain report conclusion, including original data fragment identifiers, clause identifiers, rule identifiers, formula text, parameter values, calculation intermediates, judgment boundaries, exception descriptions, coverage reasons, generation pre-prompts, and parameter version processing dynamics. This evidence chain package can generate hash digests using content addressing, and establish source-purpose (or "source-purpose") reference relationships at the field level to ensure traceability. Any report field can be traced back to the original data and clause source. Furthermore, the evidence chain package can simultaneously record version numbers and timestamps to better support consistency comparison and reproduction across generation.

[0048] (v) Constrained Report Generation Module Used to generate report drafts based on the evidence chain package, under report schema constraints and template constraints.

[0049] As mentioned earlier, the evidence chain package provides the main content of the environmental monitoring report. Based on this, the constrained report generation module can be used to generate the specific report draft. This involves both report schema constraints and template constraints. Thus, with the evidence chain package as the core context and read-only context, and with the double constraints superimposed, the corresponding text generation model (collectively referred to as the deep learning model configuration) is called to generate the specific report text and table fragments.

[0050] Among them, the report schema constraint is the constraint mechanism of the report XML document, which is used to define and describe the structure and content of the report XML document. This is a constraint configured at the technical level, while the template constraint is a constraint configured from the perspective of the report template itself.

[0051] In terms of details, for the constrained report generation module, the report schema constraints can be expressed in JSONSchema or equivalent form to express data structure constraints, including the existence, type, value range and business rules of nodes such as header meta information, sample list, project list, quality control table, statistical table, conclusion paragraph and remarks paragraph; Template constraints define the visual presentation and formatting details using parametric layouts; The report draft generation process employs a pipelined mechanism of structured context package - constrained generation - parsing and correction.

[0052] Specifically, the structured context package may include a domain terminology list, a sentence pattern library, and example corpora to reduce the risk of terminology drift and style inconsistency.

[0053] The parsing and correction process can perform syntax and semantic checks on the model output according to the report schema constraints, and trigger automatic retry or template placeholder rollback when the constraints are not met.

[0054] To control the impact of free generation on compliance, generated models may be authorized to rewrite language only within the conclusion and explanatory paragraphs, and may not rewrite the values, units, and clause numbers from the evidence chain package. This restriction can be enforced through placeholders and read-only fragment mechanisms.

[0055] (vi) Automatic Compliance Verification Module The automatic compliance verification module is used to perform compliance verification on draft reports and locate abnormal information in fields.

[0056] As is easy to understand, after the initial draft of the report is generated, relevant compliance verification can be carried out to further optimize the draft report and ensure that a higher quality environmental monitoring report can be output. The processing involved can be completed by the automatic compliance verification module here.

[0057] Furthermore, the automatic compliance verification module can specifically verify the generated results in dimensions such as clause consistency, numerical boundaries, unit conversion, significant figure rules, rounding rules, table format, paragraph format, and terminology consistency.

[0058] Specifically, the corresponding ones are: Clause consistency is achieved by comparing the clause numbers referenced in the generated text with the set of clauses returned during the knowledge retrieval phase; Numerical boundary checks are constrained by the upper and lower limits output by the rule engine, and unit conversions are based on the unit conversion table. Rounding rules are subject to the terms and conditions or localization policies; Terminology consistency is achieved by matching and correcting against domain terminology lists; The other rules concerning valid numbers, table formats, and paragraph formats are more basic content compliance verification strategies that are easily understood from the text.

[0059] Regarding the specific compliance checks mentioned above, when any check fails, the system can generate abnormal information locating the field and block the issuance process, while recording the reason for the inconsistency and revision suggestions in the evidence chain.

[0060] (vii) Human-machine collaborative review module The human-machine collaborative review module is used to display the report generation results, compliance verification conclusions and evidence chain citations, and to perform manual review and revision.

[0061] It is understandable that this application system may also involve a result presentation stage to display the report generation results (i.e., environmental monitoring reports), compliance verification conclusions, and evidence chain citations involved in the previous processing.

[0062] Of course, in specific applications, other processing results or processes can also be displayed as needed.

[0063] The display screens (including touch screens) involved can be the system's own, external display devices, or other devices outside the system. This is quite flexible, similar to the coherent data sources that the system can access.

[0064] In this way, by presenting the results, manual review and revision can be easily carried out while meeting the viewing needs.

[0065] In terms of details, the human-machine collaborative audit module can provide differentiated view display. Auditors can revise, annotate, and roll back fields. Auditors can leave reasons for revisions. Revisions of numerical fields must be accompanied by alternative evidence or rule version descriptions. The system will version the revision records and link them to the evidence chain, and write them to the audit log.

[0066] To ensure consistency, changes to numerical fields can be configured to require the replacement or supplementation of supporting evidence; otherwise, the system will remain in an uninterpreted state and will not be allowed to enter a frozen state.

[0067] This phase also allows for parallel collaboration across roles, with role permissions controlling visibility and operable fields, and writing critical node operations to audit logs.

[0068] (viii) Version Freeze and Issuance Module The version freeze and issuance module is used to perform a version freeze operation and bind the corresponding electronic seal when all compliance checks have passed and the audit status is ready for issuance.

[0069] Understandably, once it is confirmed that the environmental monitoring report generated by the system is usable, that is, all compliance checks have passed and the audit status is ready for issuance, then it can be frozen and issued, thus implementing the subsequent operations of the automatically generated report.

[0070] In terms of details, the version freezing and issuance module can legally validate reports through electronic signatures and trusted timestamps, while the frozen objects can specifically include the report's structured data, visual layout, and evidence chain snapshots.

[0071] Electronic signatures can also employ a signature mechanism with equivalent legal effect.

[0072] (ix) Export and Archive Module The export and archive module is used to index reports and chains of evidence, export corresponding PDF files, Word files, and structured data exchange packages, and archive them.

[0073] Similarly, the export and archiving module processing here is also one of the follow-up operations of the automatically generated report.

[0074] In terms of details, the export and archive modules can export results in the form of PDF files, Word files, and structured data exchange packages. The structured data exchange packages can specifically include hash digests and signature information that can be verified by third-party systems. Archives are a basic setting in data management work, which can be used for spot checks.

[0075] (x) Full-process audit and retrospective module The end-to-end audit and backtracking module is used to record key events and difference snapshots throughout the entire process, and to perform source retrieval when needed.

[0076] Similarly, the full audit and retrospective module here is also one of the follow-up operations of automated report generation, in order to promote better environmental monitoring report processing services in practical applications.

[0077] In terms of details, the full-process audit and backtracking module can write key events, including data access, rule calculation, generation, verification, revision and issuance, as well as difference snapshots, into an immutable full-link audit log to support granular traceability retrieval by report, field, rule or clause dimension, so as to better ensure the operability of regulatory spot checks and internal reviews.

[0078] It should be noted that source tracing does not specifically refer to the processing involved in generating an environmental monitoring report. In practical applications, it is usually a process initiated some time after the completion of the environmental monitoring report, under the need for source tracing such as random sampling inspections.

[0079] In this way, through the main system operations described above, the certainty of knowledge retrieval terms and the numerical certainty of the rule engine are first solidified into an evidence chain, and then the constrained generation expressive capabilities are used for linguistic output. By leveraging automatic compliance verification, human-machine collaborative review, and freeze issuance to form a self-consistent closed loop, an end-to-end technical route that can uniformly support "knowledge retrieval - rule calculation - evidence chain - constrained generation - compliance verification - issuance audit" is created. This enables the provision of automated, intelligent, and high-quality environmental monitoring report processing services, which can well meet the regulatory requirements for traceability and consistency.

[0080] In addition to the main system configurations mentioned above, this application system may also involve other configuration aspects.

[0081] For example, in application scenarios involving both server-side and mobile devices, corresponding to the aforementioned data access and governance module, the system can also perform mobile task push and data feedback, with the following processing content: The server can push on-site sampling and in-situ testing tasks to the mobile terminal based on batch plans, personnel schedules, and geofencing. The mobile terminal can perform testing operations such as barcode scanning, positioning, watermark photo acquisition, video acquisition, instrument Bluetooth reading acquisition, and serial port reading acquisition, and cache them offline. When the network is available, the data will be automatically transmitted back to the server.

[0082] Among them, the offline caching mechanism on mobile devices can be replaced by a breakpoint resume and priority queue mechanism.

[0083] In addition to the constrained report generation module, the system can also perform template and terminology management operations. Laboratory administrators and model engineers can maintain versioned versions of the corresponding report templates, terminology lists, and sentence structure libraries as needed. Template changes trigger regression verification to verify consistency with the schema constraints of the schema-side report.

[0084] Meanwhile, the system can also perform exception handling and rollback operations for the corresponding constrained report generation module. When the text generation model is unavailable or the output does not meet the report schema constraints for a long time, the system can automatically roll back to the template placeholder filling mode to maintain the determinism of the downstream issuance process and insert a rollback flag in the report conclusion section for auditing purposes.

[0085] For example, the system can also handle the following tasks in the data access and governance module: Batch and project lists can be imported from external systems or created within the system. Specifically, external systems can be regulatory filing systems or databases of historical laboratory information management systems. Import sample and batch Gantt views from external project scheduling tools, or manage schedules within the system using a business process view; On the mobile device, fill in the sampling and in-situ testing items according to the preset inspection item template. The inspection item template includes items such as witness content, control points, example diagrams and videos, testing area type, number of testing points and number of test points in the group. On the server side, the task list is filtered by conditions including type, category, status and executor, and a statistical dashboard is displayed showing the number and quality of tasks undertaken by each participating entity and its department.

[0086] For example, regarding the content display operations involved in the human-machine collaborative review module, the system can also handle the following: Displays measured data points from historical batches or tasks, and supports comparison with the current batch; Display a two-dimensional sampling point layout or laboratory floor plan, and bind measurement data and evidence chain references to sample or measurement point elements; Display sample and batch Gantt views to show progress, or display quality control dashboards to show the real-time status of indicators such as parallel samples, blanks, and spiked recoveries; As the status of the main plan nodes and batches changes, the node status is updated in real time to any of the following: not started, in progress, or completed, and reminders and review processes are triggered as needed.

[0087] For example, the system may also have the following processing capabilities: Perform identity authentication and multi-factor verification on user accounts logging into the system; Review the registration application of the laboratory or institution; Set performance scores for each type of task and link them to quality indicators, including quality control pass rate and rework rate; Manage personnel, organizational, project information, task and report data; Bind the instruments, methods, and barcode rules to the electronic signature medium; Based on the evidence chain and templates, commonly used reports and ledgers, including daily monitoring reports, quality control summaries, retesting suggestion lists, report review records, and issuance registration, are automatically generated.

[0088] In conclusion, regarding the above solutions, this application establishes a technical route for automatically generating environmental monitoring reports, centered on knowledge retrieval and quality control calculations, constrained by evidence chains and structured constraints, and guaranteed by automatic compliance verification and audit traceability. It provides a verifiable engineering path to address issues such as clause conflicts, model instability, and the difficulty of templates covering complex scenarios. Compared to traditional template-filling and pure model-generating solutions, it has significant advantages in consistency, traceability, and compliance. Furthermore, versioned knowledge and pluggable rules support rapid adaptation and long-term operation and maintenance across multiple monitoring scenarios. This provides excellent tool support for environmental monitoring work, enabling better environmental management.

[0089] The above is an introduction to the report generation system for knowledge retrieval and quality control calculation collaboration in environmental protection testing provided in this application. Based on this, this application also provides a report generation method for knowledge retrieval and quality control calculation collaboration in environmental protection testing, focusing on the system's control process. Obviously, this method is applied to the report generation system for knowledge retrieval and quality control calculation collaboration in environmental protection testing. This report generation system can specifically include modules such as data access and governance, knowledge retrieval and clause location, quality control calculation, evidence chain construction, constrained report generation, automatic compliance verification, human-machine collaborative review, version freezing and issuance, export and archiving, and full-process audit and backtracking.

[0090] In this context, the report generation method for knowledge retrieval and quality control calculation collaboration in environmental monitoring provided in this application refers to... Figure 2 The diagram shown is a flowchart of a report generation method for environmental monitoring that combines knowledge retrieval and quality control calculations, specifically including the following steps S201 to S210: Step S201: Configure the identification information and stage permissions of each participating entity in the entire process of environmental monitoring, and perform data access and governance to form a unified data view, and establish the relationship between different objects; Step S202: Knowledge retrieval and clause location are performed based on the unified data view, and an executable rule list is generated. Step S203: The quality control calculation module performs quality control calculations based on the list of executable rules to generate structured quality control results; Step S204: The evidence chain construction module dynamically constructs an evidence chain package based on the processing related to the conclusion of a certain report; Step S205: The constrained report generation module generates a draft report based on the evidence chain package, under the constraints of report schema and template. Step S206: The automatic compliance verification module performs compliance verification on the report draft and locates abnormal field information; Step S207: The human-machine collaborative audit module displays the report generation results, compliance verification conclusions, and evidence chain citations, and performs manual review and revision. Step S208: When all compliance checks are passed and the audit status is ready for issuance, the version freeze and issuance module performs the version freeze operation and binds the corresponding electronic signature. Step S209: The export and archiving module exports the report and evidence chain index, the corresponding PDF files, Word files and structured data exchange packages, and archives them. Step S210: The full audit and backtracking module records key events and difference snapshots throughout the entire process and performs source retrieval when necessary.

[0091] As an exemplary embodiment, for the data access and governance module, the identification information of each participating entity specifically includes the client identification information, sampling unit identification information, testing laboratory identification information, and regulatory agency identification information, and the stage permissions of each participating entity include viewing permissions and modification permissions.

[0092] As another exemplary embodiment, the data access and governance module specifically accesses multi-source data including commission information, sample information, sampling records, instrument raw data, historical reports and regulatory standard knowledge, and forms a unified data view through governance operations including field standardization, unit conversion, timestamp sample identification alignment, missing value imputation and outlier identification. The unified data view includes three core objects: sample entity, test item entity and measurement entity. The associations between different objects are established using foreign keys and timestamps, specifically many-to-one or one-to-many associations.

[0093] As another exemplary embodiment, for the knowledge retrieval and clause location module, standards and methodologies are organized in the form of a knowledge graph to perform knowledge retrieval and clause location. The nodes of the knowledge graph include standards, clauses, limits, methods, interference items, quality control rules and versions, and the edges include relationships such as applicability, reference, substitution, effectiveness, applicable scenarios and conflicts. To resolve conflict or coverage issues, a version selection and conflict resolution strategy is introduced. The version selection and conflict resolution strategy prioritizes candidate clauses based on time priority, regulatory priority, and localization adaptation priority, selects the most applicable clause, and retains the unselected clauses as alternative evidence for audit explanation.

[0094] As another exemplary embodiment, for the quality control calculation module, the quality control calculation includes the following processing: Perform specific control calculations including relative deviation of parallel samples, blank subtraction, spike recovery, limit of detection, limit of quantitation, and linearity of calibration curve, and generate corresponding structured quality control results according to the allowable range and judgment logic defined in the terms.

[0095] As another exemplary embodiment, the evidence chain construction module is specifically encapsulated into an evidence chain package based on a certain report conclusion, including original data fragment identifiers, clause identifiers, rule identifiers, formula text, parameter values, calculation intermediate quantities, judgment boundaries, exception descriptions, overriding reasons, generation of pre-prompt prompts, and parameter version processing dynamics. The evidence chain package generates a hash digest using content addressing, establishes a source-purpose reference relationship at the field level, and records both the version number and timestamp.

[0096] As another exemplary embodiment, for the constrained report generation module, the report schema constraint expresses data structure constraints in JSONSchema or equivalent form, including the existence, type, value range and business rules of nodes in header meta information, sample list, project list, quality control table, statistical table, conclusion section and remarks section; Template constraints define the visual presentation and formatting details using parametric layouts; The report draft generation process employs a pipelined mechanism of structured context package - constrained generation - parsing and correction.

[0097] As another exemplary embodiment, the automatic compliance verification module specifically performs system verification on the generated results in dimensions including clause consistency, numerical boundaries, unit conversion, significant figure rules, rounding rules, table format, paragraph format, and terminology consistency.

[0098] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the report generation method for knowledge retrieval and quality control calculation collaboration in environmental monitoring, as described above, can be found in [reference needed]. Figure 1 The description of the report generation system for knowledge retrieval and quality control calculation collaboration for environmental protection testing in the corresponding embodiment will not be repeated here.

[0099] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0100] Therefore, this application provides a computer-readable storage medium storing a plurality of instructions that can be loaded by a processor to execute the present application. Figure 2The steps of the report generation method for knowledge retrieval and quality control calculation collaboration for environmental monitoring in the corresponding embodiment can be found in the following example. Figure 2 The description of the report generation method for knowledge retrieval and quality control calculation collaboration for environmental protection testing in the corresponding embodiment will not be repeated here.

[0101] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0102] Because of the instructions stored in the computer-readable storage medium, the present application can be executed as described above. Figure 2 The corresponding embodiment's report generation method for knowledge retrieval and quality control calculation collaboration in environmental monitoring can therefore achieve the results of this application. Figure 2 The beneficial effects that the report generation method for knowledge retrieval and quality control calculation collaboration for environmental protection testing can achieve in the corresponding embodiment are detailed in the preceding description and will not be repeated here.

[0103] The foregoing has provided a detailed description of the report generation system, method, and computer-readable storage medium for knowledge retrieval and quality control calculation collaboration in environmental monitoring provided by this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the core ideas of this application; furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A report generation system for environmental monitoring that integrates knowledge retrieval and quality control calculations, characterized in that, The system includes: The data access and governance module is used to configure the identification information and stage permissions of each participating entity throughout the environmental monitoring process, as well as to perform data access and governance to form a unified data view and establish the relationship between different objects; The knowledge retrieval and clause location module is used to perform knowledge retrieval and clause location based on the unified data view, and generate a list of executable rules; The quality control calculation module is used to perform quality control calculations based on the executable rule list to generate structured quality control results; The evidence chain construction module is used to dynamically construct an evidence chain package based on the processing related to the conclusions of a report; The constrained report generation module is used to generate a draft report based on the evidence chain package, under report schema constraints and template constraints. The automatic compliance verification module is used to perform compliance verification on the draft report and locate abnormal information in the fields; The human-machine collaborative review module is used to display the report generation results, compliance verification conclusions and evidence chain citations, and to perform manual review and revision; The version freeze and issuance module is used to perform a version freeze operation and bind the corresponding electronic seal when all compliance checks have passed and the audit status is ready for issuance. The export and archiving module is used to index reports and chains of evidence, export corresponding PDF files, Word files, and structured data exchange packages, and archive them; The end-to-end audit and backtracking module is used to record key events and difference snapshots throughout the entire process, and to perform source retrieval when needed.

2. The method according to claim 1, characterized in that, For the data access and governance module, the identification information of each participating entity specifically includes the client identification information, sampling unit identification information, testing laboratory identification information, and regulatory agency identification information. The stage permissions of each participating entity include viewing permissions and modification permissions.

3. The method according to claim 1, characterized in that, The data access and governance module specifically accesses multi-source data including commission information, sample information, sampling records, instrument raw data, historical reports and regulatory standards knowledge, and forms a unified data view through governance operations including field standardization, unit conversion, timestamp sample identification alignment, missing value imputation and outlier identification. The unified data view includes three core objects: sample entities, detection item entities and measurement entities. The associations between different objects are established using foreign keys and timestamps, specifically many-to-one or one-to-many associations.

4. The method according to claim 1, characterized in that, For the knowledge retrieval and clause location module, a knowledge graph is used to organize standards and methodologies to perform the knowledge retrieval and clause location. The knowledge graph nodes include standards, clauses, limits, methods, interference items, quality control rules and versions, and the graph edges include relationships such as applicability, reference, substitution, effectiveness, applicable scenarios and conflicts. To resolve conflict or coverage issues, a version selection and conflict resolution strategy is introduced. This strategy prioritizes candidate clauses based on time priority, regulatory priority, and localization adaptation priority, selects the most applicable clause, and retains unselected clauses as alternative evidence for audit explanation.

5. The method according to claim 1, characterized in that, For the quality control calculation module, the quality control calculation includes the following processing: Perform specific control calculations including relative deviation of parallel samples, blank subtraction, spike recovery, limit of detection, limit of quantitation, and linearity of calibration curve, and generate the corresponding structured quality control results according to the allowable range and judgment logic defined in the clause.

6. The method according to claim 1, characterized in that, For the evidence chain construction module, the evidence chain is specifically encapsulated into an evidence chain package based on the processing dynamics of a certain report conclusion, including the original data fragment identifier, clause identifier, rule identifier, formula text, parameter value, calculation intermediate quantity, judgment boundary, exception description, coverage reason, generation of pre-prompt and parameter version. The evidence chain package generates a hash digest using content addressing, establishes a source-purpose reference relationship at the field level, and records the version number and timestamp.

7. The method according to claim 1, characterized in that, For the constrained report generation module, the report schema constraints are expressed in JSON Schema or equivalent form to represent data structure constraints, including the existence, type, value range and business rules of nodes in header metadata, sample list, project list, quality control table, statistical table, conclusion section and remarks section; The template constraints are defined by parametric layout and formatting details; The report draft generation process employs a pipelined mechanism of structured context package - constrained generation - parsing and correction.

8. The system according to claim 1, characterized in that, The automatic compliance verification module specifically performs system verification on the generated results in dimensions including clause consistency, numerical boundaries, unit conversion, significant figure rules, rounding rules, table format, paragraph format, and terminology consistency.

9. A report generation method for environmental monitoring that integrates knowledge retrieval and quality control calculations, characterized in that... The method is applied to a report generation system, and the method includes: The data access and governance configuration includes the identification information and stage permissions of each participating entity throughout the environmental monitoring process, as well as the execution of data access and governance to form a unified data view and establish the relationship between different objects; Knowledge retrieval and clause location are performed based on the unified data view, and an executable rule list is generated. The quality control calculation module performs quality control calculations based on the executable rule list to generate structured quality control results; The evidence chain construction module dynamically constructs an evidence chain package based on the processing related to the conclusions of a given report. The constrained report generation module generates a draft report based on the evidence chain package, under report schema constraints and template constraints. The automatic compliance verification module performs compliance verification on the draft report and locates abnormal information in the fields; The human-machine collaborative audit module displays the report generation results, compliance verification conclusions, and evidence chain citations, and allows for manual review and revision. The version freeze and issuance module executes the version freeze operation and binds the corresponding electronic seal when all compliance checks pass and the audit status is ready for issuance. The export and archiving module indexes reports and chains of evidence, exports corresponding PDF files, Word files, and structured data exchange packages, and archives them; The full-process audit and backtracking module records key events and discrepancies snapshots throughout the entire process and performs source retrieval when necessary.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to execute the method of claim 9.

Images