Method and apparatus for repairing missing parameters, storage medium, and program product

Abstract

This application discloses a method, apparatus, storage medium, and program product for repairing missing parameters. The method includes: when a target system node is found to have missing parameters among multiple system nodes of a target system, acquiring the exception stack information of the target system node and acquiring the complete dependency parameters of the target system node; determining the missing parameter name of a first missing parameter of the target system node based on the exception stack information, and determining the upstream large model node of the target system node based on the complete dependency parameters, and acquiring the output data of the upstream large model node, wherein the upstream large model node is a large model node located upstream of the target system node among the multiple system nodes, and the large model node operates based on a large model; and repairing the first missing parameter based on the output data and the missing parameter name.

Description

Technical Field

[0001] This application relates to the field of business process automation, and more specifically, to a method and apparatus for repairing missing parameters, a storage medium, and a program product. Background Technology

[0002] In enterprise automation systems, workflows are widely used to orchestrate the execution of multi-stage tasks. Each node depends on the output of the preceding node as input, forming a strict data dependency chain. When a node fails to execute due to missing parameters, inconsistent field names, or incorrect data types, traditional systems typically interrupt the process and throw an exception, requiring manual intervention for troubleshooting.

[0003] In enterprise automation systems, there are still no effective solutions to problems such as the inability of system processes to be interrupted when a node experiences abnormal parameters, requiring manual intervention for troubleshooting, and poor system process stability. Summary of the Invention

[0004] This application provides a method, apparatus, storage medium, and program product for repairing missing parameters, to at least solve the problem in related technologies where, in enterprise-level automation systems, the system process can be interrupted when a node experiences parameter anomalies, requiring manual intervention for troubleshooting, resulting in poor system process stability.

[0005] According to one embodiment of this application, a method for repairing missing parameters is provided, comprising: when it is detected that a target system node has missing parameters among multiple system nodes of a target system, obtaining the exception stack information of the target system node and obtaining the complete dependency parameters of the target system node; determining the missing parameter name of a first missing parameter of the target system node based on the exception stack information, and determining the upstream large model node of the target system node based on the complete dependency parameters, and obtaining the output data of the upstream large model node, wherein the upstream large model node is a large model node located upstream of the target system node among the multiple system nodes, and the large model node operates based on a large model; and repairing the first missing parameter based on the output data and the missing parameter name.

[0006] In an exemplary embodiment, repairing the first missing parameter based on the output data and the missing parameter name includes: obtaining multiple aliases of the first missing parameter from a metadata management service based on vector retrieval according to the missing parameter name; performing synonym semantic matching on the multiple aliases and the output data based on the semantic understanding capability of the upstream large model node; and repairing the first missing parameter based on the field data of the synonym field if a synonym field is matched from the output data.

[0007] In an exemplary embodiment, after repairing the first missing parameter based on the output data and the missing parameter name, the method further includes: verifying whether all the first missing parameters have been repaired; if it is determined that a second missing parameter among the first missing parameters has not been repaired, determining a first business scenario of the target system node; matching the default value of the second missing parameter in the parameter default value library according to the first business scenario, and repairing the second missing parameter according to the default value, wherein the parameter default value library stores multiple parameter default values ​​under multiple second business scenarios, and the multiple second business scenarios include the first business scenario.

[0008] In an exemplary embodiment, before or after matching the default value of the second missing parameter in the default value library according to the first business scenario, the method further includes: obtaining input parameters of the plurality of system nodes, and classifying the plurality of input parameters into enumerated type parameters and non-enumerated type parameters; for the non-enumerated type parameters, performing cluster analysis on the non-enumerated type cluster parameters using a density clustering algorithm, determining a target parameter based on the obtained clustering results, and updating the default value library according to the target parameter, wherein the target parameter belongs to the cluster with the most class samples in the clustering results, and the target parameter is the center sample of the cluster; for the enumerated type parameters, updating the default value library according to the mode of the enumerated type parameters.

[0009] In an exemplary embodiment, after repairing the second missing parameter according to the default value, the method further includes: verifying whether all the second missing parameters have been repaired; if it is determined that a third missing parameter among the second missing parameters has not been repaired, guiding the large model to generate a target structured form according to the third missing parameter and through structured prompt information; and sending the target structured form to the target object to instruct the target object to complete the third missing parameter according to the target structured form.

[0010] In an exemplary embodiment, before sending the target structured form to a target object to instruct the target object to complete the third missing parameter based on the target structured form, the method further includes: guiding the large model to generate multiple structured forms based on the third missing parameter and the structured prompt information, wherein the multiple structured forms include the target structured form; scoring the multiple structured forms using a multi-dimensional readability scoring function to obtain multiple scoring results, wherein the multiple structured forms correspond one-to-one with the multiple scoring results; and determining the structured form with the highest scoring result among the multiple structured forms as the target structured form.

[0011] In one exemplary embodiment, scoring the plurality of structured forms using a multi-dimensional readability scoring function includes: using a multi-dimensional readability scoring function Scoring is performed on the multiple structured forms, whereby... Used to indicate structured forms, For clarity weight, For integrity weight, As a weight for the naturalness of the process, As a weight for the rationality of grouping, Used to evaluate the clarity of the structured form. Used to assess the completeness of the structured form. Used to evaluate the flow naturalness of the structured form. Used to evaluate the grouping rationality of the structured form.

[0012] According to another embodiment of this application, a device for repairing missing parameters is also provided, comprising: an acquisition module, configured to acquire anomaly stack information of the target system node and acquire complete dependency parameters of the target system node when a parameter is missing in a target system node among multiple system nodes of the target system; a determination module, configured to determine the missing parameter name of a first missing parameter of the target system node based on the anomaly stack information, and determine the upstream large model node of the target system node based on the complete dependency parameters, and acquire the output data of the upstream large model node, wherein the upstream large model node is a large model node located upstream of the target system node among the multiple system nodes, and the large model node operates based on a large model; and a repair module, configured to repair the first missing parameter based on the output data and the missing parameter name.

[0013] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, and the computer program is configured to execute the above-described method for repairing missing parameters when running.

[0014] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-described method for repairing missing parameters through the computer program.

[0015] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the methods described in various embodiments of this application.

[0016] In this embodiment, the execution status of multiple system nodes in the target system is monitored in real time. When a parameter missing anomaly is detected in a target system node, the anomaly stack information of the target system node and the complete dependency parameters are obtained. Based on the anomaly stack information, the name of the missing first parameter of the target system node is determined, and the upstream large model node of the target system node is determined based on the complete dependency parameters. The output data of the upstream large model node is obtained, wherein the large model node works based on the large model. Finally, the first missing parameter is repaired based on the output data and the missing parameter name. By adopting the above scheme, a context-aware anomaly data completion method is provided, which effectively identifies missing parameters in the workflow and generates accurate completion suggestions that conform to business semantics based on context information. This solves the problem in related technologies where, in enterprise-level automation systems, when a node has a parameter anomaly, the system process can be interrupted, requiring manual intervention for troubleshooting, resulting in poor system process stability. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 This is a hardware structure block diagram of a computer terminal for an optional method of repairing missing parameters according to an embodiment of this application.

[0019] Figure 2 This is a flowchart of an optional method for repairing missing parameters according to an embodiment of this application;

[0020] Figure 3 This is a flowchart of an optional workflow abnormal data completion method based on context awareness and interactive recommendation, according to an embodiment of this application.

[0021] Figure 4 This is a structural block diagram of a device for repairing missing parameters according to an embodiment of this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., 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 of this application described herein can be implemented in orders other than those 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 apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] The methods and embodiments provided in this application can be executed on a computer terminal or similar computing system. Taking running on a computer terminal as an example, Figure 1 This is a hardware structure block diagram of a computer terminal according to an embodiment of this application. Figure 1 As shown, a computer terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. In one exemplary embodiment, the computer terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the computer terminal described above. For example, the computer terminal may also include components that are more complex than those described above. Figure 1 The more or fewer components shown, or having the same Figure 1 Equivalent functions or ratios shown Figure 1 The functions shown have more different configurations.

[0025] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the missing parameter repair method in this embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thus implementing the above-described method. The memory 104 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to a secure text network via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0026] The transmission system 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider for the computer terminal. In one example, the transmission system 106 includes a network interface controller (NIC), which can connect to other network devices via a base station to communicate with the Internet.

[0027] This embodiment provides a method for repairing missing parameters. Figure 2 This is a flowchart of an optional method for repairing missing parameters according to an embodiment of this application. The method includes the following steps:

[0028] Step S202: If it is detected that there are missing parameters in one of the multiple system nodes of the target system, obtain the exception stack information of the target system node and obtain the complete dependency parameters of the target system node.

[0029] Step S204: Determine the missing parameter name of the first missing parameter of the target system node according to the exception stack information, and determine the upstream large model node of the target system node according to the complete dependency parameters, and obtain the output data of the upstream large model node, wherein the upstream large model node is the large model node located upstream of the target system node among the multiple system nodes, and the large model node works based on the large model;

[0030] Step S206: Repair the first missing parameter according to the output data and the missing parameter name.

[0031] The above solution monitors the execution status of multiple system nodes in the target system in real time. When a parameter missing anomaly is detected in a target system node, it obtains the anomaly stack information of that target system node and its complete dependency parameters. Based on the anomaly stack information, it determines the name of the first missing parameter of the target system node, and based on the complete dependency parameters, it determines the upstream large model node of the target system node and obtains the output data of the upstream large model node, where the large model node operates based on the large model. Finally, it repairs the first missing parameter based on the output data and the missing parameter name. This solution provides a context-aware anomaly data completion method that effectively identifies missing parameters in the workflow and generates accurate completion suggestions that conform to business semantics based on context information. This solves the problem in related technologies where, in enterprise-level automation systems, a parameter anomaly at a node can interrupt the system process, requiring manual intervention for troubleshooting, resulting in poor system process stability.

[0032] Optionally, the first missing parameter is repaired based on the output data and the missing parameter name, including: obtaining multiple aliases of the first missing parameter from the metadata management service based on the missing parameter name using vector retrieval; performing synonym semantic matching between the multiple aliases and the output data based on the semantic understanding capability of the upstream large model node; and repairing the first missing parameter based on the field data of the synonym fields if a synonym field is matched from the output data.

[0033] This embodiment focuses on repairing the first missing parameter, utilizing the advanced semantic understanding capabilities of the upstream large model node and the rich lexical alias library in the metadata management service. A series of aliases for the first missing parameter are retrieved from the metadata service, and the semantic relationship between the output data and the aliases is analyzed in depth using the large model. If a synonymous field matching the alias exists in the output data, that field is extracted to accurately repair the first missing parameter.

[0034] It should be noted that the metadata management service is a data service platform specifically designed to manage and store data, including various expressions of words and their interrelationships. It has built an alias knowledge base to support field recognition and synonym matching.

[0035] It should be noted that the upstream large model node is a key component in the workflow. It uses a large pre-trained model to process data, and its semantic understanding ability surpasses that of conventional text analysis. It can effectively identify synonyms or similar expressions, thus enhancing the intelligence of data repair.

[0036] This embodiment achieves intelligent repair of the first missing parameter through deep semantic parsing of upstream large model nodes and alias knowledge base of metadata management service, which significantly improves the accuracy and speed of workflow data completion, enhances the system's automated processing capability and business process continuity.

[0037] Optionally, after repairing the first missing parameter based on the output data and the missing parameter name, the method further includes: verifying whether all the first missing parameters have been repaired; if it is determined that there is a second missing parameter among the first missing parameters that has not been repaired, determining the first business scenario of the target system node; matching the default value of the second missing parameter in the parameter default value library according to the first business scenario, and repairing the second missing parameter according to the default value, wherein the parameter default value library stores multiple parameter default values ​​under multiple second business scenarios, and the multiple second business scenarios include the first business scenario.

[0038] For automatically completed second missing parameters that have not been repaired, a default value filling strategy is provided based on the first business scenario of the target system node and the parameter default value library. Once it is found that a second missing parameter is still missing after the first missing parameter has been repaired, the system immediately identifies the first business scenario corresponding to the current workflow node, and then searches for a matching default value in the parameter default value library. This default value is then used directly to repair the second missing parameter, ensuring that all necessary parameters are properly handled.

[0039] It should be noted that the parameter default value library maintains a set of parameter default values ​​for multiple business scenarios, providing preset and most common parameter value options for different business scenarios to deal with missing parameters and improve the success rate of workflow automation.

[0040] It should be noted that the first business scenario refers to the specific business environment in which the current workflow node (i.e., the target system node mentioned above) is located, such as contract signing, approval process, etc. This scenario information is crucial for determining the appropriate default values ​​for parameters.

[0041] This embodiment identifies business scenarios and extracts default parameter values, automatically repairing incomplete second missing parameters, significantly reducing the need for manual intervention, enhancing the stability and automation level of workflow execution, and providing the system with more robust exception handling capabilities.

[0042] Optionally, before or after matching the default value of the second missing parameter in the default value library according to the first business scenario, the method further includes: obtaining input parameters of the plurality of system nodes, and classifying the plurality of input parameters into enumerated type parameters and non-enumerated type parameters; for the non-enumerated type parameters, performing cluster analysis on the non-enumerated type cluster parameters using a density clustering algorithm, determining the target parameter based on the obtained clustering results, and updating the default value library according to the target parameter, wherein the target parameter belongs to the cluster with the most class samples in the clustering results, and the target parameter is the center sample of the cluster; for the enumerated type parameters, updating the default value library according to the mode of the enumerated type parameters.

[0043] This embodiment describes an update mechanism for a default parameter library to adapt to dynamically changing business scenarios and user preferences. Specifically, the system collects input parameters from various nodes and classifies them into two main categories: enumerated and non-enumerated types. For non-enumerated parameters, a density clustering algorithm (such as DBSCAN) is used for deep analysis to uncover parameter distribution patterns. The center sample in the cluster with the largest number of class samples is determined as the target parameter, and these target parameters are then integrated into the default parameter library to reflect the most frequently occurring parameter values. For enumerated parameters, the default parameter library is updated directly with the value that appears most frequently (the mode), ensuring that the default parameter values ​​closely match actual business needs.

[0044] It should be noted that density clustering algorithms, such as DBSCAN, are unsupervised machine learning algorithms based on point density, used to discover high-density regions in a dataset, thereby forming clusters. In this embodiment, its key role is to analyze the distribution of non-enumerated type parameters, identify the most common parameter values, and provide data basis for updating the parameter default value library.

[0045] This embodiment achieves intelligent updates to the default parameter library by subdividing parameter types and combining density clustering and mode statistics. This better reflects user preferences and data distribution trends in business scenarios, thereby providing default value suggestions that are closer to actual needs when completing data, significantly improving the system's ability to handle abnormal situations and the user experience.

[0046] In an exemplary embodiment, after repairing the second missing parameter according to the default value, the method further includes: verifying whether all the second missing parameters have been repaired; if it is determined that a third missing parameter among the second missing parameters has not been repaired, guiding the large model to generate a target structured form according to the third missing parameter and through structured prompt information; and sending the target structured form to the target object to instruct the target object to complete the third missing parameter according to the target structured form.

[0047] After automatically repairing the second missing parameter, the system further checks for a third missing parameter to be repaired. If a third missing parameter is found, it activates structured prompts to subtly guide the large model to generate a customized target structured form. This form is sent directly to the target object, explicitly instructing it to complete the third missing parameter according to the form's content, ensuring a seamless workflow.

[0048] It should be noted that the structured prompts are well-designed, clearly defined texts used to guide the large model in understanding the requirements and producing the desired results. In this embodiment, they serve to describe the details of the third missing parameter, the workflow context, and the expected output format, ensuring that the form generated by the large model accurately matches the requirements of the missing information.

[0049] It should be noted that the target structured form is a structured data input interface that is automatically generated based on missing parameter information and workflow context. It is designed specifically to complete specific parameters, making it easy for users to quickly and accurately provide the information they need.

[0050] This embodiment uses structured prompts to drive the large model to generate a dedicated completion form, specifically addressing the issue of missing third parameters that the automatic repair mechanism fails to cover. This greatly simplifies the manual completion work for users or the system, ensures data integrity and workflow continuity, and improves the efficiency and user experience of the entire process.

[0051] In an exemplary embodiment, before sending the target structured form to a target object to instruct the target object to complete the third missing parameter based on the target structured form, the method further includes: guiding the large model to generate multiple structured forms based on the third missing parameter and the structured prompt information, wherein the multiple structured forms include the target structured form; scoring the multiple structured forms using a multi-dimensional readability scoring function to obtain multiple scoring results, wherein the multiple structured forms correspond one-to-one with the multiple scoring results; and determining the structured form with the highest scoring result among the multiple structured forms as the target structured form.

[0052] The system first uses the third missing parameter and carefully designed structured prompts to generate a series of alternative structured forms from the large model, rather than simply producing a single version. Then, a multi-dimensional readability scoring function is used to evaluate these alternative forms, covering multiple indicators such as field clarity, form completeness, flow naturalness, and grouping rationality, with each form receiving a score. Finally, the system selects the form with the highest score and identifies it as the target structured form for completing the third missing parameter, ensuring both efficiency and a good user experience in the completion process.

[0053] This embodiment generates multiple structured forms and uses a multi-dimensional readability scoring function for evaluation and selection, ensuring that the target structured forms provided to users are both comprehensive and easy to operate, greatly improving the efficiency and accuracy of data completion, while optimizing the user interface and enhancing the overall user experience and satisfaction.

[0054] In an optional embodiment, the plurality of structured forms are scored using a multi-dimensional readability scoring function, including: using a multi-dimensional readability scoring function. Scoring is performed on the multiple structured forms, whereby... Used to indicate structured forms, For clarity weight, For integrity weight, As a weight for the naturalness of the process, As a weight for the rationality of grouping, Used to evaluate the clarity of the structured form. Used to assess the completeness of the structured form. Used to evaluate the flow naturalness of the structured form. Used to evaluate the grouping rationality of the structured form.

[0055] This embodiment further describes the detailed process of using a multi-dimensional readability scoring function to comprehensively evaluate the quality of multiple candidate forms when determining the optimal structured form to complete the third missing parameter. The system not only focuses on the surface completeness of the form but also examines its clarity, flow naturalness, and grouping rationality. Through weighted configuration, these dimensions are comprehensively scored, with each form receiving a clear score. This facilitates the system in selecting the target structured form with the highest overall quality, ensuring the best user experience during the completion process.

[0056] It should be noted that clarity measures whether the field labels in the form are standardized, whether they use consistent corporate terminology, reduce cognitive load, and ensure that users can easily understand the various requirements of the form.

[0057] Completeness assessment: Evaluate whether each field provides sufficient guidance for filling out, including placeholder hints, sample values, and validation rules, to ensure that users can complete the form accurately.

[0058] The process should be reasonable. The order of fields should be reviewed to ensure that it conforms to the business logic and the user's mental model, so as to ensure that the form filling process is smooth and logically consistent.

[0059] Grouping rationally determines whether semantically related fields should be grouped appropriately and visually aggregated to facilitate users' quick location and filling in of relevant items.

[0060] This embodiment uses a multi-dimensional comprehensive scoring system to accurately select the structured form that best meets user needs and business logic as the completion tool, significantly improving the efficiency and accuracy of the user's completion process. At the same time, it optimizes the form design, promotes the smooth execution of the workflow, provides users with a smooth filling experience, and ultimately improves the efficiency of data completion and the overall stability of the system.

[0061] In an optional embodiment, this embodiment proposes an optional workflow anomaly data completion scheme based on context awareness and interactive recommendation, the implementation process of which is as follows: Figure 3 As shown, it includes the following steps:

[0062] Step 1: Anomaly Detection and Missing Parameter Extraction

[0063] During workflow engine runtime, the system monitors the execution status of each node (i.e., the aforementioned multiple system nodes) in real time. When a node enters the "input validation failed" state, exception handling is triggered:

[0064] 1.1. Parse the exception stack trace and identify the missing parameter names;

[0065] 1.2 Query the input data of the current abnormal node (i.e., the target system node mentioned above) to obtain its complete list of dependency parameters (equivalent to the complete dependency parameters mentioned above). The list structure is as follows: [

[0067] {

[0068] "p_name": "Parameter name: such as contract_sign_time",

[0069] "p_desc": Natural language description: such as "time stamp of the customer signing the contract",

[0070] "p_type": "Data type: such as datetime or yyyy-MM-dd HH:mm:ss",

[0071] "p_required": Whether it is required: Boolean value.

[0072] } ]

[0074] 1.3 Extract the structured information of the missing parameters from the complete dependency parameter list based on the missing parameter name; at the same time, the system accesses the output cache of all upstream large model nodes (equivalent to the aforementioned upstream large model nodes), obtains all outputs, and returns the data structure, including all field names, values, types, and descriptions, with the same structure as the parameter structure in the complete dependency parameter list.

[0075] Step 2: Perform synonym extraction and repair on the output of large model nodes: This step aims to re-extract parameters from the output data of the upstream large model nodes based on the set of missing field aliases. If the naming or semantic equivalence is satisfied, synonym mapping is performed to repair the missing fields.

[0076] 2.1 Missing Field Alias ​​Set Query: Retrieve the set of field aliases for missing parameter names from the metadata management service using vector retrieval (equivalent to the multiple aliases mentioned above). For example: ["sign_time","contract_time","execution_ts","signing time","contract signing time"].

[0077] 2.2 Synonym Field Matching and Repair: Perform semantic matching of synonyms for each upstream field (i.e., the output data mentioned above). Matching method: Based on the semantic understanding capability of the large model and based on the prompt word engineering, the input is given to the large model. If the large model extracts the corresponding synonym field management matching information, then field repair is performed; otherwise, skip to step 3.

[0078] The matching process is illustrated below:

[0079] You are an intelligent data extraction and rewriting engine.

[0080] #Target parameters

[0081] Name: ${p.name}

[0082] Description: ${p.desc}

[0083] Type: ${p.type}

[0084] Acceptable field names (alias set): ${A(p)}

[0085] # Output fields provided

[0086] ${upstream_fields_table}

[0087] #Original User Issues

[0088] ${question}

[0089] #Task

[0090] Based on the target parameter alias set, the original user question is matched to extract parameter values. If no match is found, null is output.

[0091] Output only the extraction results; no explanation is required.

[0092] Step 3: Fill and repair based on the scenario-based parameter default value library (equivalent to the parameter default value library mentioned above):

[0093] For parameters that are still incomplete (equivalent to the second missing parameter mentioned above), the system queries the "Scenario-Parameter Default Value Library," which maintains default value rules for common business scenarios, as shown in Table 1:

[0094] Table 1

[0095]

[0096] On the other hand, the default value library for scenarios and parameters contains automatic update logic, which is based on user preference analysis. The analysis method is as follows:

[0097] (1) Collect all input results of each node in the user workflow and distinguish each input parameter according to enumeration type and non-enumeration type;

[0098] (2) For non-enumerated type parameters, the DBSCAN algorithm (distance calculation can be measured by embedding vector similarity, edit distance, etc.) is used to perform cluster analysis on each parameter, extract the parameter distribution, and the cluster with the most samples is taken as the current user preference. Based on the above distance calculation method, the sample with the lowest average distance from the current sample to other samples is selected as the class center of the class, and the data table is updated based on the sample data as the default value.

[0099] (3) For enumeration type parameters, the mode is used for statistics and the mode is used as the default value.

[0100] Step 4: If automatic repair fails, a structured form will be generated for manual completion.

[0101] The form generation method based on the large model prompt word project is as follows:

[0102] 4.1 Constructing Multi-Parameter Joint Prompts (Prompt Engineering): Using structured prompts to guide the output of complete form configurations for large models, rather than field by field. The prompt template is as follows:

[0103] You are a smart form design expert. Based on the following information with multiple missing parameters, design a highly available and easy-to-fill autocomplete interface.

[0104]

Parameter List

[0105] {% for p in missing_params %}

[0106] - Parameter {{loop.index}}:

[0107] - Name: {{p.name}}

[0108] - Description: {{p.desc}}

[0109] - Type: {{p.type}}

[0110] - Required: {{p.required}}

[0111] - Alias: {{p.aliases|join(", ")}}

[0112] - Business domain: {{p.domain}}

[0113] {% endfor %}

[0114] [Contextual Information]

[0115] - Workflow name: {{workflow_name}}

[0116] - Current node: {{current_node}}

[0117] - Upstream output fields: {{upstream_fields|join(", ")}}

[0118] - User role: {{user_role}}

[0119]

Task Requirements

[0120] Please output a JSON object with the following structure:

[0121] {

[0122] "fields": [

[0123] {

[0124] "param_name": "string",

[0125] "control_type": "text|date|datetime|select|radio|switch|...",

[0126] "label": "Display Text",

[0127] "placeholder": "prompt message",

[0128] "required": true|false,

[0129] "options": [...], / / Enumerate options

[0130] "validation_rules": ["regular expression or validation type"],

[0131] "depends_on": "another field name", / / Optional: Explicit / implicit dependency

[0132] "group": "group name" / / e.g., "contract information"

[0133] }

[0134] ],

[0135] "layout_order": ["field1", "field2", ...],

[0136] "field_groups": {

[0137] "group1": { "title": "Basic Information", "fields": ["name", "time"]}

[0138] },

[0139] "submit_button_label": "Start execution"

[0140] }

[0141] Output only JSON, do not interpret it.

[0142] It should be noted that the following key points should be considered when constructing the prompt word project:

[0143] (1) The system collects all unrepaired parameters, each of which includes:

[0144] A. Name, description, type, and whether it is required;

[0145] B. Business domain and formatting prompts;

[0146] C. Collection of field aliases.

[0147] (2) The generated form results have the following characteristics:

[0148] A. Logical relationship: 'Signing date' should follow 'Signing party';

[0149] B. Reasonable grouping: 'Contact Information' includes 'Name', 'Phone Number', and 'Email'.

[0150] 4.2 Form Generation Effectiveness Evaluation: To optimize the manual completion experience, this embodiment proposes a multi-dimensional readability scoring function to evaluate the quality of different form layout schemes. Given a layout scheme... Its readability score is defined as:

[0151] .

[0152] The weights are configured as follows (data is for illustrative purposes only):

[0153] =0.2: Clarity;

[0154] =0.3: Completeness;

[0155] =0.2: Flow;

[0156] =0.3: Grouping rationality.

[0157] In one optional embodiment, the system generates three permutation schemes by default. Select max( (Equivalent to the target structured form mentioned above) Render the form.

[0158] 4.2.1 Clarity – The degree of standardization of label terminology.

[0159] Evaluate whether the field labels in the form use standardized terminology consistent within the enterprise / system to avoid cognitive burden caused by synonyms.

[0160] Entry structure:

[0161] {

[0162] "Contract Signing Time": ["Contract Signing Time", "Contract Effective Time", "Signing Completion Time"]

[0163] }

[0164] Label definition: Labels are displayed in the form.

[0165]

[0166] n represents the total number of missing parameters. The main terms are from the metadata management system. Semantic similarity is calculated using Sentence-BERT encoding to determine cosine similarity.

[0167] Example: The parameter contract_sign_time (contract signing time) gives a score of 1.0 when using "contract signing time" and 0.7 when using "contract signing time" if it is in the alias library.

[0168] 4.2.2 Integrity - Hints, Examples, and Verification Rule Integrity:

[0169] Assess whether each field provides adequate guidance on how to fill it out: .

[0170] See Table 2 for parameter definitions.

[0171] Table 2

[0172]

[0173] Example: A field has placeholder and example but no validation → Score 0.4 + 0.3 = 0.7.

[0174] 4.2.3. Process Self-Awakening - Does the field order conform to business logic?

[0175] Evaluate whether the order of the fields aligns with the user's mental model and business processes.

[0176]

[0177] Where R is a predefined set of ordered relation pairs, for example:

[0178] R=[

[0179] ("Signing Date", "Signing Party"), # The date should precede the person's name.

[0180] ("Department", "Department Head"), # Department first, then Head

[0181] ("Approval Status", "Approval Comments") # Status first, then comments ]

[0183] =1 if and only if the field a Appeared b after;

[0184] Total number of relations.

[0185] If all the order is correct → If half is out of order → .

[0186] 4.2.4 Grouping rationality - Is the grouping reasonable?

[0187] Determine whether to visually aggregate semantically related fields.

[0188] .

[0189] Where k is the number of groups. If and only if the fields within the group belong to the same business subdomain (such as "Contract Information" or "Approval Configuration").

[0190] Clustering algorithms (DBSCAN + Sentence Embedding) can be used to predict reasonable groupings.

[0191] 4.3 Interaction Logic Injection and Front-end Rendering: The front-end parses the structured information and renders the UI for the user to fill in.

[0192] 4.4 User Interaction and Resumption of Execution:

[0193] After the user submits the request, the system executes the following:

[0194] 4.4.1 Format validation (e.g., whether the date is valid);

[0195] 4.4.2 Type conversion (e.g., string to datetime);

[0196] 4.4.3 Inject data into the node input context;

[0197] 4.4.4 Restart the node to execute.

[0198] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0199] Figure 4 This application discloses a missing parameter repair device, comprising:

[0200] The acquisition module is used to acquire the exception stack information of the target system node and the complete dependency parameters of the target system node when it is detected that there are missing parameters in the target system node among multiple system nodes of the target system.

[0201] The determination module is used to determine the missing parameter name of the first missing parameter of the target system node based on the exception stack information, and to determine the upstream large model node of the target system node based on the complete dependency parameters, and to obtain the output data of the upstream large model node, wherein the upstream large model node is the large model node located upstream of the target system node among the plurality of system nodes, and the large model node works based on the large model;

[0202] The repair module is used to repair the first missing parameter based on the output data and the missing parameter name.

[0203] The aforementioned device monitors the execution status of multiple system nodes in the target system in real time. When a parameter missing anomaly is detected in a target system node, the anomaly stack information of that target system node and its complete dependency parameters are obtained. Based on the anomaly stack information, the name of the first missing parameter of the target system node is determined, and based on the complete dependency parameters, the upstream large model node of the target system node is determined, and the output data of the upstream large model node is obtained. The large model node operates based on the large model. Finally, the first missing parameter is repaired based on the output data and the missing parameter name. This approach provides a context-aware anomaly data completion method that effectively identifies missing parameters in the workflow and generates accurate and business-semantic-compliant completion suggestions based on contextual information. This solves the problem in related technologies where, in enterprise-level automation systems, a parameter anomaly at a node can interrupt the system process, requiring manual intervention for troubleshooting, resulting in poor system process stability.

[0204] In an exemplary embodiment, the above-described repair module is further configured to obtain multiple aliases of the first missing parameter from the metadata management service based on the missing parameter name according to vector retrieval; perform synonym semantic matching on the multiple aliases and the output data based on the semantic understanding capability of the upstream large model node; and repair the first missing parameter according to the field data of the synonym field if a synonym field is matched from the output data.

[0205] In an exemplary embodiment, the repair module is further configured to verify whether all the first missing parameters have been repaired; if it is determined that there is a second missing parameter among the first missing parameters that has not been repaired, determine the first business scenario of the target system node; match the default value of the second missing parameter in the parameter default value library according to the first business scenario, and repair the second missing parameter according to the default value, wherein the parameter default value library stores multiple parameter default values ​​under multiple second business scenarios, and the multiple second business scenarios include the first business scenario.

[0206] In an exemplary embodiment, the above-mentioned repair module is further configured to obtain input parameters of the plurality of system nodes and classify the plurality of input parameters into enumerated type parameters and non-enumerated type parameters; for the non-enumerated type parameters, perform cluster analysis on the non-enumerated type cluster parameters using a density clustering algorithm, determine the target parameter based on the obtained clustering results, and update the parameter default value library based on the target parameter, wherein the target parameter belongs to the cluster with the most class samples in the clustering results, and the target parameter is the center sample of the cluster; for the enumerated type parameters, update the parameter default value library based on the mode of the enumerated type parameters.

[0207] In an exemplary embodiment, the above-mentioned repair module is further configured to verify whether all the second missing parameters have been repaired; if it is determined that a third missing parameter among the second missing parameters has not been repaired, the module guides the large model to generate a target structured form based on the third missing parameter and through structured prompt information; and sends the target structured form to the target object to instruct the target object to complete the third missing parameter based on the target structured form.

[0208] In an exemplary embodiment, the repair module is further configured to guide the large model to generate multiple structured forms based on the third missing parameter and the structured prompt information, wherein the multiple structured forms include the target structured form; score the multiple structured forms using a multi-dimensional readability scoring function to obtain multiple scoring results, wherein the multiple structured forms correspond one-to-one with the multiple scoring results; and determine the structured form with the highest scoring result among the multiple structured forms as the target structured form.

[0209] In one exemplary embodiment, the aforementioned repair module is further configured to use a multi-dimensional readability scoring function. Scoring is performed on the multiple structured forms, whereby... Used to indicate structured forms, For clarity weight, For integrity weight, As a weight for the naturalness of the process, As a weight for the rationality of grouping, Used to evaluate the clarity of the structured form. Used to assess the completeness of the structured form. Used to evaluate the flow naturalness of the structured form. Used to evaluate the grouping rationality of the structured form.

[0210] Embodiments of this application also provide a storage medium including a stored program, wherein the program executes any of the methods described above when it is run.

[0211] Optionally, in this embodiment, the storage medium may be configured to store program code for performing the following steps:

[0212] S1, when it is detected that there are missing parameters in one of the multiple system nodes of the target system, obtain the exception stack information of the target system node and obtain the complete dependency parameters of the target system node;

[0213] S2, determine the missing parameter name of the first missing parameter of the target system node according to the exception stack information, and determine the upstream large model node of the target system node according to the complete dependency parameters, and obtain the output data of the upstream large model node, wherein the upstream large model node is the large model node located upstream of the target system node among the multiple system nodes, and the large model node works based on the large model;

[0214] S3, Repair the first missing parameter based on the output data and the missing parameter name.

[0215] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0216] Specific examples in this embodiment can be found in the examples described in the above embodiments and exemplary implementations, and will not be repeated here.

[0217] Embodiments of this application also provide an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0218] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0219] S1, when it is detected that there are missing parameters in one of the multiple system nodes of the target system, obtain the exception stack information of the target system node and obtain the complete dependency parameters of the target system node;

[0220] S2, determine the missing parameter name of the first missing parameter of the target system node according to the exception stack information, and determine the upstream large model node of the target system node according to the complete dependency parameters, and obtain the output data of the upstream large model node, wherein the upstream large model node is the large model node located upstream of the target system node among the multiple system nodes, and the large model node works based on the large model;

[0221] S3, Repair the first missing parameter based on the output data and the missing parameter name.

[0222] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0223] Embodiments of this application also provide a computer program product, including a non-volatile computer-readable storage medium storing the computer program product, wherein the computer program, when executed by a processor, implements the steps of the methods described in various embodiments of this application.

[0224] Optionally, in this embodiment, the computer program described above can be configured to perform the following steps when executed by the processor:

[0225] S1, when it is detected that there are missing parameters in one of the multiple system nodes of the target system, obtain the exception stack information of the target system node and obtain the complete dependency parameters of the target system node;

[0226] S2, determine the missing parameter name of the first missing parameter of the target system node according to the exception stack information, and determine the upstream large model node of the target system node according to the complete dependency parameters, and obtain the output data of the upstream large model node, wherein the upstream large model node is the large model node located upstream of the target system node among the multiple system nodes, and the large model node works based on the large model;

[0227] S3, Repair the first missing parameter based on the output data and the missing parameter name.

[0228] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0229] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing systems. They can be centralized on a single computing system or distributed across a network of multiple computing systems. Optionally, they can be implemented using program code executable by a computing system, thereby storing them in a storage system for execution by the computing system. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0230] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for repairing missing parameters, characterized in that, include: If, when it is detected that a target system node has missing parameters among multiple system nodes of the target system, the exception stack information of the target system node is obtained, as well as the complete dependency parameters of the target system node are obtained. The missing parameter name of the first missing parameter of the target system node is determined according to the exception stack information, and the upstream large model node of the target system node is determined according to the complete dependency parameters, and the output data of the upstream large model node is obtained. The upstream large model node is the large model node located upstream of the target system node among the multiple system nodes, and the large model node works based on the large model. The first missing parameter is repaired based on the output data and the name of the missing parameter.

2. The method for repairing missing parameters according to claim 1, characterized in that, Repairing the first missing parameter based on the output data and the missing parameter name includes: Based on the missing parameter name, retrieve multiple aliases of the first missing parameter from the metadata management service using vector retrieval; Based on the semantic understanding capabilities of the upstream large model nodes, synonym semantic matching is performed on the multiple aliases and the output data; If a synonym field for one of the multiple aliases is found in the output data, the first missing parameter is repaired based on the field data of the synonym field.

3. The method for repairing missing parameters according to claim 1, characterized in that, After repairing the first missing parameter based on the output data and the missing parameter name, the method further includes: Verify whether all the first missing parameters have been repaired; If it is determined that a second missing parameter exists in the first missing parameter and has not been repaired, the first business scenario of the target system node is determined; According to the first business scenario, the default value of the second missing parameter is matched in the default value library, and the second missing parameter is repaired according to the default value. The default value library stores multiple default values ​​of parameters under multiple second business scenarios, including the first business scenario.

4. The method for repairing missing parameters according to claim 3, characterized in that, Before or after matching the default value of the second missing parameter in the default value library according to the first business scenario, the method further includes: Obtain the input parameters of the multiple system nodes, and classify the multiple input parameters into enumerated type parameters and non-enumerated type parameters; For the non-enumerated type parameters, a density clustering algorithm is used to perform cluster analysis on the non-enumerated type clustering parameters. The target parameter is determined based on the obtained clustering results. The default parameter library is updated based on the target parameter. The target parameter belongs to the cluster with the most class samples in the clustering results. The target parameter is the center sample of the cluster. For the enumeration type parameter, update the parameter default value library according to the mode of the enumeration type parameter.

5. The method for repairing missing parameters according to claim 3, characterized in that, After repairing the second missing parameter according to the default value, the method further includes: Verify whether all the second missing parameters have been repaired. If it is determined that a third missing parameter exists in the second missing parameter and has not been repaired, the large model is guided to generate a target structured form based on the third missing parameter and structured prompt information. The target structured form is sent to the target object to instruct the target object to complete the third missing parameter based on the target structured form.

6. The method for repairing missing parameters according to claim 5, characterized in that, Before sending the target structured form to the target object to instruct the target object to complete the third missing parameter based on the target structured form, the method further includes: Based on the third missing parameter, the large model is guided to generate multiple structured forms through the structured prompt information, wherein the multiple structured forms include the target structured form; The multiple structured forms are scored using a multi-dimensional readability scoring function to obtain multiple scoring results, wherein each of the multiple structured forms corresponds one-to-one with the multiple scoring results; The structured form with the highest score among the multiple structured forms is determined as the target structured form.

7. The method for repairing missing parameters according to claim 6, wherein the plurality of structured forms are scored using a multi-dimensional readability scoring function, including: Through a multi-dimensional readability scoring function Scoring is performed on the multiple structured forms, whereby... Used to indicate structured forms, For clarity weight, For integrity weight, As a weight for the naturalness of the process, As a weight for the rationality of grouping, Used to evaluate the clarity of the structured form. Used to assess the completeness of the structured form. Used to evaluate the flow naturalness of the structured form. Used to evaluate the grouping rationality of the structured form.

8. A device for repairing missing parameters, characterized in that, include: The acquisition module is used to acquire the exception stack information of the target system node and the complete dependency parameters of the target system node when it is detected that there are missing parameters in the target system node among multiple system nodes of the target system. The determination module is used to determine the missing parameter name of the first missing parameter of the target system node based on the exception stack information, and to determine the upstream large model node of the target system node based on the complete dependency parameters, and to obtain the output data of the upstream large model node, wherein the upstream large model node is the large model node located upstream of the target system node among the plurality of system nodes, and the large model node works based on the large model; The repair module is used to repair the first missing parameter based on the output data and the missing parameter name.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method of any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images