Material information management method and device, computer device, and storage medium

Abstract

The application discloses a material information management method and device, computer equipment and a storage medium. The method comprises the following steps: receiving a material access request from an intelligent agent and obtaining a material code therein; based on the dispatch rule information configured for the intelligent agent, performing rule matching and entity library positioning on the material code to determine the corresponding material category entity library; based on the material code, obtaining candidate material data sources from the material master data, the parameter library and the corresponding material category entity library, and when there are multiple candidate material data sources, determining a target data source according to a preset conflict priority; generating target material information based on the target data source and returning the intelligent agent; and in response to a property validation request submitted by the intelligent agent, calling corresponding property checking rules to perform cross-consistency checking on the material property data to be verified, and processing according to the checking result. The application can reduce the risk of material query result deviation, error information being accepted and abnormal data update maintenance.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to methods, apparatus, computer equipment and storage media for material information management. Background Technology

[0002] In the information management process of manufacturing enterprises, material information is a crucial foundation for R&D design, procurement management, production planning, inventory management, and supply chain collaboration. As enterprises expand their product types and business scale, the types, brands, models, specifications, and matching relationships of materials continue to increase. In their daily operations, enterprises typically need to query, use, maintain, and verify material data.

[0003] In the current material information management process, material data is repeatedly used in different business processes. In practice, inconsistencies may occur in the name, model, category, specifications, or related attribute information of the same material, and query results may be unstable. When relevant personnel need to make business judgments based on this information, the accuracy and consistency of the material information will affect the results of subsequent business processing.

[0004] As AI agents are increasingly applied to material retrieval and data maintenance within enterprises, they can generate query results or maintenance suggestions based on user input. However, in practical applications, the material information output by the agent may still deviate from the actual material situation. For example, unstable query results, inconsistencies in material attribute information, and incorrect acceptance of relevant information can all affect the reliability of material retrieval or data maintenance.

[0005] Therefore, in scenarios where intelligent agents participate in material information query and maintenance, how to reduce the risks of material query result deviation, acceptance of erroneous information, and abnormal updates of maintenance data has become a technical problem that urgently needs to be solved in the intelligent management of material information in manufacturing enterprises. Summary of the Invention

[0006] This application provides a material information management method, apparatus, computer equipment, and storage medium, aiming to reduce the risks of material query result deviation, erroneous information being accepted, and abnormal updates of maintenance data in scenarios where intelligent agents participate in material information query and maintenance.

[0007] In a first aspect, embodiments of this application provide a material information management method, including:

[0008] Receive a material access request from the intelligent agent and obtain the material code carried in the material access request;

[0009] Based on the scheduling rule information configured for the intelligent agent, the material code is matched with the scheduling rule and located in the entity library to determine the material category entity library corresponding to the material code. The material master data is associated with the material category entity library through the material code, and the material master data is also associated with a parameter library for providing material parameter information.

[0010] Based on the material code, candidate material data sources related to the material access request are obtained from the material master data, the parameter library, and the corresponding material category entity library;

[0011] The target data source is determined based on the candidate material data source, wherein when there are multiple candidate material data sources, the target data source is determined from the multiple candidate material data sources according to a preset conflict priority;

[0012] Based on the target data source, extract the field data corresponding to the material access request, generate target material information, and return the target material information to the agent;

[0013] In response to receiving the attribute activation request submitted by the intelligent agent, the material attribute data to be verified carried in the attribute activation request is obtained. The material attribute data to be verified is the attribute data generated by the intelligent agent based on the target material information and requested to be written to or updated in the material data.

[0014] Invoke the attribute verification rules configured corresponding to the material category entity library, and perform cross-consistency verification on at least two related basic attributes in the attribute data of the material to be verified;

[0015] Based on the verification results of the cross-consistency check, the corresponding processing is performed on the material attribute data to be verified.

[0016] Secondly, embodiments of this application also provide a material information management device, comprising:

[0017] The request receiving module is used to receive material access requests from the intelligent agent and obtain the material code carried in the material access request;

[0018] The scheduling and positioning module is used to perform scheduling rule matching and entity library positioning on the material code based on the scheduling rule information configured for the intelligent agent, so as to determine the material category entity library corresponding to the material code. The material master data is associated with the material category entity library through the material code, and the material master data is also associated with a parameter library for providing material parameter information.

[0019] The data source acquisition module is used to acquire, based on the material code, candidate material data sources related to the material access request from the material master data, the parameter library, and the corresponding material category entity library;

[0020] A data source determination module is used to determine a target data source based on the candidate material data sources, wherein when there are multiple candidate material data sources, the target data source is determined from the multiple candidate material data sources according to a preset conflict priority;

[0021] The information generation module is used to extract field data corresponding to the material access request based on the target data source, generate target material information, and return the target material information to the intelligent agent;

[0022] The attribute acquisition module is used to, in response to receiving an attribute activation request submitted by the intelligent agent, acquire the material attribute data to be verified carried in the attribute activation request, wherein the material attribute data to be verified is attribute data generated by the intelligent agent based on the target material information and requests to write or update the material data;

[0023] The attribute verification module is used to invoke the attribute verification rules configured corresponding to the material category entity library, and perform cross-consistency verification on at least two related basic attributes in the attribute data of the material to be verified; and

[0024] The data processing module is used to perform corresponding processing on the material attribute data to be verified based on the verification results of the cross-consistency check.

[0025] Thirdly, embodiments of this application also provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0026] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor, can implement the above-described method.

[0027] The beneficial technical effects of this application's embodiments are as follows: After receiving a material access request from an intelligent agent, this application's embodiments obtain the material code carried in the material access request, and perform scheduling rule matching and entity library location on the material code based on scheduling rule information configured for the intelligent agent. This allows the intelligent agent's material access request to be guided to the material category entity library corresponding to the material code, thereby reducing material query result deviations caused by unstable query object pointing or access range deviations. Furthermore, based on the material code, this application obtains candidate material data sources from the material master data, parameter library, and corresponding material category entity library. When there are multiple candidate material data sources, a target data source is determined according to a preset conflict priority. Then, field data is extracted based on the target data source to generate target material information. This reduces the risk of different material information being mixed or incorrectly accepted by the intelligent agent, improving the consistency and stability of the returned material information. Furthermore, upon receiving an attribute activation request submitted by the agent, this application invokes the attribute verification rules configured corresponding to the material category entity library, performs cross-consistency verification on at least two related basic attributes in the material attribute data to be verified, and performs corresponding processing on the material attribute data to be verified based on the verification results. This ensures that the material attribute data to be written or updated is subject to rule constraints before taking effect, thereby reducing the risk of abnormal data updates and improving the data reliability when the agent participates in material information query and maintenance. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 A flowchart illustrating the material information management method provided in this application embodiment;

[0030] Figure 2 A schematic block diagram of a material information management device provided in the embodiments of this application;

[0031] Figure 3 A schematic block diagram of a computer device provided in an embodiment of this application. Detailed Implementation

[0032] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0033] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0034] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0035] It should also be further understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0036] In the material information management process of manufacturing enterprises, material master data, parameter data, entity database data, and historical data are often distributed across different data files, business systems, or file directories. When an intelligent agent receives a user's material query, selection, or data update task, directly retrieving and generating results from multiple data sources based on natural language prompts can easily lead to problems such as inaccurate entity database location, conflicts between fields from different data sources, incorrect combinations of material attributes, and unverified data being written to the database. Especially in scenarios where there are strong correlations between basic attributes such as material codes, brands, models, interface types, and specifications, simply relying on the intelligent agent to generate judgment results may result in erroneous attributes being contaminated back into the material database.

[0037] Based on this, embodiments of this application provide a material information management method. This method uses material coding to drive scheduling rule matching and entity library location, ensuring that a smart agent's material access requests are first directed to the corresponding material category entity library. When multiple candidate material data sources exist, the target data source is determined through a preset conflict priority. When the smart agent further submits an attribute activation request based on the returned material information, cross-consistency checks are performed on the material attribute data to be verified, and the results determine whether it should be activated, pending review, or otherwise processed. Thus, the smart agent no longer directly and unrestrictedly reads, generates, and writes material information; instead, material data retrieval and attribute activation are completed under the scheduling, arbitration, and verification mechanisms of the material information management platform.

[0038] See Figure 1 This embodiment provides a material information management method, which can be executed by a material information management platform. The material information management platform can be deployed on an enterprise local server, a private data platform, a cloud server, or an internal enterprise material management system. The intelligent agent can be an artificial intelligence agent used for material querying, material selection, material maintenance, or data entry assistance, or it can be an automated script, intelligent question-and-answer tool, or business processing agent program capable of calling material data according to user instructions. The method includes the following steps.

[0039] Step S101: Receive a material access request from the intelligent agent and obtain the material code carried in the material access request.

[0040] In practice, material access requests can be generated by an agent based on user input, or triggered by a user interface, third-party business systems, or automated tasks. For example, an engineer might input "query the matching cable for a certain barcode reader" or "obtain a replacement part for a certain model of motor" into the agent. The agent can then convert this natural language request into a material access request carrying a material code and submit it to the material information management platform. The material code is an identifier used to identify target materials within the material information system. It can be represented as an internal part number, material master data code, SKU code, or other data identifiers that can uniquely or relatively uniquely locate the material.

[0041] After receiving a material access request, the material information management platform can first parse the request message, interface parameters, or task text to extract the material code. If the request also carries auxiliary information such as material name, brand, model, or material category, the platform can also temporarily store this auxiliary information as a basis for subsequent entity database location, candidate data source filtering, or attribute verification.

[0042] Step S102: Based on the scheduling rule information configured for intelligent agents, perform scheduling rule matching and entity library location on the material code to determine the material category entity library corresponding to the material code.

[0043] Scheduling rule information is rule-based information used by intelligent agents or material information management platforms to identify data access paths. It can be stored in the form of configuration files, routing maps, rule index files, directory navigation files, or interface rule tables. This scheduling rule information is not limited to a specific file format; as long as it can express the correspondence between material codes and material category entity libraries, it can be used as the scheduling rule information in this embodiment.

[0044] Material master data is linked to the material category entity library through material codes. Material master data primarily stores basic identification information and core fields of materials, such as material code, material category, material model, and material name. The parameter library provides material parameter information, such as specifications, interface parameters, dimensional parameters, performance parameters, or category parameter templates. The material category entity library stores entity data, business rules, reference materials, or extended data related to a specific type of material. For example, different categories such as barcode readers, connectors, wire harnesses, motors, and industrial control computers can each correspond to different material category entity libraries.

[0045] When performing scheduling rule matching, the material information management platform can query scheduling rule information based on the category field, prefix field, code segment, or mapping relationship in the material code, thereby determining the material category entity library corresponding to the material code. In this way, the agent does not need to blindly search through all material data, but the platform first limits the access scope to the data area that matches the material code according to the rules, reducing the possibility of cross-category false detection and erroneous calls.

[0046] Step S103: Based on the material code, obtain the candidate material data source related to the material access request from the material master data, parameter library and corresponding material category entity library.

[0047] After completing the entity database location, the material information management platform does not simply read results from a single database. Instead, based on the content of the material access request, it retrieves potentially relevant data from the material master data, parameter database, and the corresponding material category entity database. Candidate material data sources can be basic field records in the material master data, parameter definitions or templates in the parameter database, or extended attributes, rule records, historical records, or business data in the material category entity database.

[0048] For example, when an access request involves "querying basic material information," the material master data may become the primary candidate material data source; when an access request involves "querying interface parameters" or "comparing specification parameters," the parameter information in the parameter library may become a candidate material data source; when an access request involves matching relationships, rule validation, or extended descriptions, the data in the corresponding material category entity library can also be listed as a candidate material data source. By simultaneously acquiring these candidate material data sources, the platform can comprehensively determine in subsequent steps which data source is more suitable for generating the target material information.

[0049] Step S104: Determine the target data source based on the candidate material data source.

[0050] When there is only one candidate material data source, the material information management platform can directly identify that candidate material data source as the target data source. When there are multiple candidate material data sources, the platform determines the target data source from among the multiple candidate material data sources according to a preset conflict priority. The preset conflict priority can be configured based on the authority of the data source, field completeness, update time, business applicability, or data source level.

[0051] For example, for the same material model, different data sources may all record the interface type or specification parameters. However, one data source may originate from the material master data, another from the entity database extended record, and yet another from the historical archive. The platform determines the preferred data source by pre-setting conflict priorities, preventing the agent from automatically choosing among multiple inconsistent fields, thereby reducing the risk of unstable returned results or referencing incorrect data.

[0052] Step S105: Extract the field data corresponding to the material access request based on the target data source, generate the target material information, and return the target material information to the agent.

[0053] After identifying the target data source, the material information management platform extracts relevant field data from the target data source based on the query purpose indicated by the material access request. Field data may include material code, material name, brand, model, material category, interface type, specifications, matching relationships, substitution relationships, verification status, or data source, etc. The platform can organize the extracted field data into structured target material information, such as tabular data, JSON data, field sets, report fragments, or contextual information that can be further processed by the intelligent agent.

[0054] Once the target material information is returned to the agent, the agent can generate natural language responses, selection suggestions, attribute maintenance suggestions, or other business outputs for the user based on this information. Because the target material information comes from a scheduled and arbitrated data source, the information foundation obtained by the agent is relatively more stable and traceable, reducing the risk of it drawing erroneous conclusions by piecing together data from different sources.

[0055] Step S106: In response to receiving the attribute activation request submitted by the agent, obtain the material attribute data to be verified carried in the attribute activation request.

[0056] In some scenarios, the agent not only queries material information but may also generate new material attribute data based on the target material information and request that this attribute data be written to or updated in the material data. For example, the agent might generate the interface type, model description, specifications, or compatibility relationships of a material based on product manuals, historical records, and existing parameter libraries, and submit an attribute activation request. The attribute activation request indicates that the agent wants the attribute data to be included in the material information management system, rather than simply displayed to the user as temporary information.

[0057] After receiving an attribute activation request, the material information management platform retrieves the material attribute data to be verified from the request. This data is generated by the agent based on the target material information and is used to request writing or updating of the material data. This data may be a single field, such as model number, brand, or interface type, or it may be a combination of multiple fields, such as a combination of brand and model number, a combination of code prefix and material category, or a combination of interface type and specification parameters. Because these fields often have business or engineering constraints, they cannot be directly written to by the agent and must be processed through a subsequent verification process.

[0058] It should be noted that when an attribute activation request carries only a single field to be updated, the material information management platform can read at least one basic attribute related to that field from existing records in the target material information, material master data, or the corresponding material category entity library. The field to be updated and the read basic attribute are then used together as cross-consistency verification objects. Therefore, the material attribute data to be verified can include multiple fields submitted by the agent, or fields submitted by the agent and related fields obtained by the platform based on the material code or target material information.

[0059] Step S107: Call the attribute verification rules configured corresponding to the material category entity library, and perform cross-consistency verification on at least two related basic attributes in the attribute data of the material to be verified.

[0060] Attribute verification rules can be stored in a rule set corresponding to the material category entity library, or they can be dynamically loaded by the material information management platform based on material category, parameter library, and business rules. Basic attributes are fundamental fields used to determine the identity, category, or technical characteristics of a material, such as brand, model, coding prefix, material category, interface type, and specifications. At least two basic attributes must be related, meaning these attributes are not isolated but can be mutually verified through business rules, coding rules, parameter rules, or entity library rules.

[0061] For example, for a given material, there should be a true correspondence between the brand and the model number; the code prefix and the material category should conform to the enterprise coding rules; the interface type and the specification parameters should conform to the parameter template for that type of material; and there may also be preset constraints between the model number field and the matching relationship. During verification, the platform can extract relevant basic attributes from the attribute data of the material to be verified and compare the actual combination relationships of these basic attributes with the constraint relationships in the attribute verification rules. If the actual combination relationship matches the constraint relationship, it indicates that the attribute data has high credibility; if the comparison results are inconsistent, it indicates that the attribute data generated by the intelligent agent may contain fictitious model numbers, incorrect categories, mismatched parameters, or incorrect associations.

[0062] This cross-compliance check differs from simply checking whether a single field is empty or formatted correctly. It focuses more on whether multiple basic attributes can corroborate each other, thereby identifying data that appears complete in form but does not actually conform to material business rules. For example, a model number field may conform to text format, but if the model number does not belong to the corresponding brand, or if the model number does not match the material category, an anomaly can be detected through cross-compliance of the basic attributes.

[0063] Step S108: Based on the verification result of the cross-consistency check, perform corresponding processing on the material attribute data to be verified.

[0064] The material information management platform determines whether the material attribute data to be verified can be effective based on the verification results. If the cross-consistency verification passes, the platform can write the material attribute data to be verified into the corresponding material category entity library, or mark it as data that can be used to update the corresponding material category entity library. At this time, the attribute data can become a usable data basis for subsequent material queries, parameter matching, rule judgment, or agent responses.

[0065] If cross-compliance verification fails, the platform can prevent the attribute data from being directly written into the official material data and mark it as deviation data, abnormal data, or data pending review. The platform can also generate verification records for abnormal fields, abnormal attribute combinations, and corresponding attribute verification rules for subsequent manual review or rule optimization. In this way, even if the agent generates seemingly reasonable but actually erroneous attribute data, the data will not bypass the platform's verification and directly enter the material database, thereby reducing the risk of erroneous data contaminating the enterprise's material information system.

[0066] Through the above steps, material access, target material information generation, attribute activation requests, and attribute verification processing form a complete closed loop for data retrieval and activation. Material coding ensures that access requests are directed to the correct data area, scheduling rules constrain the agent's data access path, multi-candidate material data source arbitration provides a clearer data source basis for target material information, and cross-consistency verification ensures that the attribute data generated by the agent is validated before entering the material data system. Therefore, the material information management platform can leverage the agent's retrieval and generation capabilities while technically constraining the agent's data access and attribute writing, thereby improving the accuracy, traceability, and reliability of the material information returned results.

[0067] In some implementations, before responding to a material access request, the material information management platform can pre-write the partitioned material master data and establish a routing mapping between material codes, material category entity libraries, and scheduling rule information. When the agent subsequently submits a material access request, the platform can extract the category identification field based on the material code in the request and use the routing mapping to locate the corresponding material category entity library. The generation of this pre-set mapping and the entity library location process after the access request is triggered are explained below in conjunction with steps S201 to S206.

[0068] Step S201: Obtain material master data and write the material master data into the corresponding data area according to the field usage.

[0069] The material information management platform can obtain material master data from existing business systems, historical material ledgers, material coding tables, procurement records, or R&D design documents. Material master data can include material codes, material categories, material models, material attributes, and historical records, change records, or usage records related to the material.

[0070] After acquiring the material master data, the platform can partition and write the data according to the role of different fields in material information management. Specifically, material codes, material categories, and material models can be written to the index data area for quick location of target materials based on material codes; material attributes can be written to the feature data area for parameter matching, attribute verification, or agent-generated target material information; and historical records, change records, or usage records can be written to the archive data area for tracing the changes in material information. Through this partitioning method, material master data is no longer simply piled up in a single data table, but forms a searchable, comparable, and traceable data organization structure.

[0071] Step S202: Based on the correspondence between material codes and material categories, generate a routing mapping relationship from material codes to the material category entity library.

[0072] After completing the writing of the material master data partitions, the material information management platform can further read the correspondence between material codes and material categories, and generate routing mapping relationships accordingly. The routing mapping relationship indicates which material category entity library a given material code or a code category field should point to. For example, one code prefix can correspond to barcode reader materials, another code prefix can correspond to connector materials, and yet another code prefix can correspond to wire harness materials. Based on these correspondences, the platform associates the material code, category identification field, material category entity library path, or material category entity library identifier.

[0073] Step S203: Write the routing mapping relationship into the scheduling rule information so that the scheduling rule information can be used to match scheduling rules and locate the entity library for material codes.

[0074] After generating the routing mapping relationship, the material information management platform can write this mapping relationship into the scheduling rule information configured for the intelligent agent. The scheduling rule information can be read by the intelligent agent or invoked by the platform upon receiving a request from the intelligent agent. Its function is to transform the association between material codes and material category entity libraries into executable scheduling rules. In some implementations, the scheduling rule information may include material code parsing rules, the mapping relationship between category identification fields and entity library identifiers, entity library access paths, data source priority prompts, etc.

[0075] Step S204: When locating the entity library based on scheduling rule information, extract the category identification field from the material code.

[0076] After obtaining the material code carried in the material access request, the material information management platform can first parse the material code to obtain a category identification field that represents the material category. The category identification field can be a prefix field, category code field, material group field, or other coded fragments that reflect the material's category affiliation. In some implementations, the material code does not necessarily directly represent the category with a fixed-length field; the platform can also split the material code using pre-established code parsing rules, or combine it with auxiliary fields such as material name, model, and brand carried in the material access request to determine the category identification field corresponding to the material code.

[0077] Step S205: Based on the category identification field, query the corresponding entity library path or entity library identifier in the scheduling rule information.

[0078] After obtaining the category identification field, the material information management platform uses this field as the query key to find the corresponding entity library path or entity library identifier in the scheduling rule information. The scheduling rule information can store the correspondence between category identification fields and material category entity libraries; for example, one category identification field might correspond to a barcode reader entity library, while another might correspond to a connector entity library. The entity library path can be the data directory where the entity library is located, the database access path, the interface address, or the file path; the entity library identifier can be the entity library number, entity library name, entity library index number, or other data identifiers that can uniquely locate the entity library.

[0079] Step S206: Locate the corresponding material category entity library based on the entity library path or entity library identifier, and call at least one data module in the material category entity library that is related to the material access request to obtain the candidate material data source related to the material access request.

[0080] After the query is completed, the material information management platform locates the corresponding material category entity library based on the entity library path or entity library identifier. This location action can manifest as opening the corresponding directory, connecting to the corresponding database, calling the corresponding interface, or loading the corresponding entity library index file. Once located, the platform can call relevant data modules based on the specific content of the material access request, such as attribute data modules, business rule modules, relationship data modules, or reference resource modules. The data records, rule records, parameter records, or reference records obtained through these calls can all serve as candidate material data sources for subsequent screening and arbitration when determining the target data source.

[0081] Through the initialization and routing process described above, a stable mapping relationship is established between the material master data, the material category entity library, and the scheduling rule information. The material code is not only used to identify the material itself, but also to drive the access path selection, enabling the agent's data calls to enter the correct material category range from the beginning, thereby reducing the risk of cross-category false detection and path confusion.

[0082] In some implementations, there are multiple material category entity libraries, each corresponding to a different material category. Each material category entity library is organized according to a unified modular structure, which includes at least an asset module, a rules module, and a reference resource module. The asset module stores structured parameters, material attribute data, and confirmed material records or entity data for the corresponding material category; the rules module stores attribute verification rules, parameter compliance rules, substitute material rules, conflict resolution rules, or field mapping rules for the corresponding material category; and the reference resource module stores original material documents, product manuals, supplier information, drawings, or field source information for the corresponding material category.

[0083] By adopting a unified modular structure for entity libraries of different material categories, the material information management platform can call attribute data, rule data, and reference materials according to the same module semantics after locating the target material category entity library based on the material code. This reduces the risk of false detection, misuse, and incorrect writing of intelligent agents caused by inconsistent data paths, chaotic rule calls, or missing original evidence between different material categories.

[0084] After locating the material category entity library, the same material code may simultaneously correspond to multiple candidate sources, including material master data, parameter library, and entity library extended data. To avoid the agent directly accepting conflicting or incomplete data, the material information management platform needs to perform hierarchical arbitration of candidate material data sources and retain output source records while generating target material information. The following sections, in conjunction with steps S301 to S306, explain the process of candidate data source arbitration, field supplementation and merging, and output record generation.

[0085] Step S301: Obtain the data source identifier, field integrity information, and update time information of the candidate material data source respectively.

[0086] After obtaining one or more candidate material data sources, the material information management platform can first extract source and quality information for each candidate material data source. The data source identifier indicates that the candidate material data source originates from material master data, parameter database, extended data of the material category entity database, historical archives, or other data levels. Field integrity information indicates whether the candidate material data source contains the target fields required for the material access request, such as brand, model, material category, interface type, specifications, matching relationships, or substitution relationships. Update time information indicates the time when the candidate material data source was most recently written to, updated, or confirmed.

[0087] For example, for the same material code, the material master data may contain the material name and model, but lack the interface type; the parameter library may contain the interface type and specification parameters; and the entity library extended data may contain matching relationships or historical selection records. By obtaining the data source identifier, field integrity information, and update time information of these candidate material data sources, the platform can determine the authority, completeness, and timeliness of each candidate data source during subsequent arbitration.

[0088] Step S302: When there are multiple candidate material data sources, determine the data level to which each candidate material data source belongs based on the data source identifier.

[0089] When multiple candidate material data sources exist, the material information management platform can determine the data level of each candidate material data source based on the data source identifier. The data level reflects the priority position of data within the material information management system. For example, material master data can serve as the priority data level for basic identity information and core fields; the parameter library can serve as a supplementary data level for material parameters, parameter templates, and parameter definitions; and the extended data of the material category entity library can serve as an extended data level for business rules, matching relationships, substitution relationships, or entity scenario data.

[0090] By differentiating data levels, the platform can avoid comparing all candidate data sources at the same level without distinction. For example, when a complete and valid material model field already exists in the material master data, the platform usually does not need to select another model field from the extended data to cover that field; and when the material master data lacks interface parameters or specification parameters, the parameter library or entity library extended data can serve as a supplementary source to participate in generating target material information.

[0091] Step S303: Based on the preset conflict priority of master data priority, parameter library supplementation, and entity library expansion, and combined with field integrity information and update time information, determine the target data source.

[0092] After determining the data level of candidate material data sources, the material information management platform can arbitrate according to a preset conflict priority. In this embodiment, the preset conflict priority can be reflected as master data priority, parameter library supplementation, and entity library expansion. That is, for basic identity fields, the platform prioritizes the material master data; for parameter fields missing in the master data, the platform can supplement them from the parameter library; for fields related to specific business rules, supporting relationships, or entity scenarios, the platform can further use expanded data from the material category entity library.

[0093] When multiple candidate material data sources belong to the same data level, the platform can further combine field completeness information and update time information for selection. For example, if two candidate material data sources both come from the parameter library, the platform can prioritize the data source with more complete target fields; if the field completeness is the same, it can prioritize the data source with a more recent update time or the one that has most recently passed review and confirmation. Through this combination of hierarchical priority and intra-level comparison, the platform can determine the target data source that is more suitable for the current material access request from multiple candidate material data sources, rather than simply relying on the agent to make its own judgment or randomly select a data source.

[0094] In this embodiment, the target data source can be the primary source data source used to generate target material information, or it can be a set of target field data sources determined after field-level arbitration. When some target fields in the material master data are missing or do not meet the field integrity requirements, the platform can determine supplementary field data sources from the parameter library or the corresponding material category entity library, and record the data sources of different fields in the output record.

[0095] Step S304: When the target field is missing in the material master data that is determined to be the target data source, obtain the supplementary data corresponding to the target field from the parameter library or the corresponding material category entity library, and merge the supplementary data with the material master data to generate the target material information.

[0096] In some cases, although material master data is identified as the target data source, its fields may not fully meet the material access request. For example, a user might request to query the interface type and compatibility relationships of a certain material, while the material master data only stores the material code, material name, and model. In this situation, the platform can retrieve supplementary data from the parameter library or the corresponding material category entity library based on the missing target fields.

[0097] During field merging, the material information management platform can use the basic identity fields in the material master data as the backbone, and supplement the corresponding field positions with specification parameters, interface types, or parameter template results from the parameter library. For business rules or related relationship fields, extended data can be further obtained from the material category entity library and merged. After field merging, the target material information obtained by the platform retains the basic identity information of the material master data and can also supplement the parameters or business fields required for material access requests.

[0098] Step S305: Generate an output record associated with the target material information.

[0099] To enable intelligent agents and subsequent users to identify the source of each field in the target material information, the material information management platform can synchronously generate output records when generating the target material information. The output records can include at least one of the following: the data source corresponding to the target material information, the conflict priority source identifier of the target data source, and the verification status of the target material information. For example, the brand and model in the target material information can be marked as originating from the material master data, the interface type can be marked as originating from the parameter library, and the matching relationship can be marked as originating from the extended data of the material category entity library. If a field is determined by a conflict priority rule of master data priority, parameter library supplementation, and entity library extension, the platform can also write the priority source identifier corresponding to that field into the output record.

[0100] Step S306: After performing cross-consistency verification, write the verification result of cross-consistency verification into the output record or the verification record associated with the output record.

[0101] In some scenarios, the target material information itself may have already undergone data source arbitration and field supplementation, but the agent may still generate material attribute data to be verified based on the target material information. To facilitate the agent and user in distinguishing the credibility status of different data, the platform can write the verification status of the target material information or the cross-consistency check result of the material attribute data to be verified into the output record, or into a verification record associated with the output record. For example, for target material information used only for querying and display, the output record can mark that the information comes from a confirmed data source, a supplementary data source, or a data source pending review; for material attribute data to be verified that the agent requests to write or update, the output record can mark that its cross-consistency check has passed, failed, or requires manual confirmation.

[0102] The material information management platform can return target material information along with output records to the intelligent agent. Upon receiving this result, the intelligent agent can generate user-oriented query results, selection suggestions, or maintenance recommendations based on the target material information. Simultaneously, it can identify the data source, priority source, or verification status of the target material information based on the output records. Through the aforementioned candidate material data source arbitration, field merging, and output record mechanisms, the platform can establish a rule-based and traceable selection mechanism among multiple data sources, preventing the intelligent agent from mixing data from different sources and generating conclusions without traceability.

[0103] In some implementations, after the target material information is returned to the agent, it can be displayed to the user only as a query result. Only when the agent generates, based on the target material information, the material attribute data to be written or updated in the material data is the agent submitted an attribute activation request to the material information management platform. Upon receiving this attribute activation request, the material information management platform does not directly write or update the material data. Instead, it calls the attribute verification rules corresponding to the current material category entity library to perform cross-consistency verification of the associated attributes in the material attribute data to be verified. The following describes the basic attribute extraction, constraint relationship comparison, and verification result processing process in conjunction with steps S401 to S404.

[0104] Step S401: Extract at least two basic attributes from the material attribute data to be verified.

[0105] The material information management platform can extract basic attributes from the material attribute data to be verified, based on the material category corresponding to the attribute activation request, to determine the material's identity, category, or technical compatibility. Basic attributes can include at least two of the following: brand, model, code prefix, material category, interface type, and specification parameters. During extraction, the platform first identifies the field names in the material attribute data to be verified, and then, combined with the field mapping relationships in the corresponding material category entity library, unifies fields from different sources or with different expressions under a unified basic attribute name. For example, "communication interface" and "communication method" can be unified as interface type, and "part number prefix" and "code segment" can be unified as code prefix.

[0106] Step S402: Determine the constraint relationship between at least two basic attributes based on the attribute verification rules, and compare the actual combination relationship of at least two basic attributes with the constraint relationship.

[0107] Attribute verification rules record the constraints that must be met between basic attributes such as brand and model, code prefix and material category, interface type and specification parameters, and model and matching relationship. After obtaining at least two basic attributes, the platform can call the attribute verification rules in the corresponding material category entity library to determine the consistency conditions that the current combination of basic attributes should meet, and compare the actual combination relationship in the attribute data of the material to be verified with the consistency conditions. If the actual combination relationship falls within the range allowed by the attribute verification rules, the platform can determine that the cross-consistency verification has passed; if the actual combination relationship violates the corresponding constraint relationship, the platform can determine that the cross-consistency verification has failed.

[0108] Step S403: If the cross-consistency check passes, write the material attribute data to be verified into the corresponding material category entity library, or mark the material attribute data to be verified as data that can be used to update the corresponding material category entity library.

[0109] When the verification results indicate that the basic attributes in the material attribute data to be verified can corroborate each other, the platform can allow the attribute data to enter the subsequent effective process. In one implementation, the platform can directly write the material attribute data to be verified into the asset module or attribute data area of ​​the corresponding material category entity library; in another implementation, the platform can first mark the attribute data as data that can be used to update the corresponding material category entity library, and then write it into the formal data area after meeting the conditions for batch review, version release, or data synchronization. This approach can utilize the material attribute data generated or organized by the intelligent agent while avoiding the direct contamination of the entity library by data that has not been verified by rules.

[0110] Step S404: If the cross-consistency check fails, generate a verification record to characterize the abnormal fields, abnormal attribute combinations and corresponding attribute verification rules, and write the material attribute data to be verified into the audit queue.

[0111] When the actual combination relationship is inconsistent with the attribute verification rules, the platform does not directly write the attribute data of the material to be verified into the formal entity database. Instead, it generates a verification record. The verification record may include the abnormal field name, abnormal field value, abnormal attribute combination, triggered attribute verification rule, request source, submitting agent identifier, and verification time. Subsequently, the platform can write the attribute data of the material to be verified along with the verification record into the pending review queue so that human reviewers can determine whether the anomaly stems from an agent generation error, expired original data, inaccurate field mapping, or an adjustment needed to the attribute verification rules.

[0112] Through the aforementioned cross-consistency verification process, the platform can identify anomalies from the combined relationships between multiple basic attributes, rather than simply checking whether a single field is empty or formatted correctly. Therefore, data that is formally complete but substantively does not conform to material business rules can be intercepted before it takes effect, thereby improving the reliability of the material attribute writing and updating process.

[0113] If the aforementioned cross-consistency check fails, the abnormal data still needs to enter a traceable and feedback-enabled control process to prevent erroneous attributes from bypassing the check and being written into the entity database. Simultaneously, it ensures that manual corrections can be used as a basis for rule updates. The following sections, using steps S501 to S504, explain the process of manual confirmation, correction record accumulation, and rule update candidate generation.

[0114] Step S501: In response to receiving the manual confirmation result for the verification record, perform correction, rejection, or keep pending review on the material attribute data to be verified based on the manual confirmation result.

[0115] Auditors can manually verify the material attribute data to be verified based on verification records, original material documents, historical records, or supplier information. If the manual verification result indicates that the anomaly stems from field recognition errors or agent generation deviations, but can still be considered a valid material attribute after correction, the platform can implement the correction for the material attribute data to be verified based on the manual verification result. If the manual verification result indicates that the attribute data does not match the actual material information, the platform can refuse its implementation. If the existing data is insufficient to determine its correctness, the platform can keep the material attribute data to be verified in a pending review state, awaiting further supplementary data.

[0116] Step S502: Record the correction entries for the material attribute data to be verified by the user.

[0117] During manual verification, if the reviewer modifies the brand, model, code prefix, material category, interface type, specifications, or other fields, the platform can record the corresponding correction entry. The correction entry can include information such as the field value before correction, the field value after correction, the reason for correction, the material category, the corresponding attribute verification rules, the reviewer's identifier, and the correction time. By recording correction entries, the platform can consolidate one-time manual corrections into a basis for subsequent rule optimization.

[0118] Step S503: When the cumulative number of similar correction items under the same material category reaches a preset threshold, generate candidate items for attribute verification rule update or attribute definition update.

[0119] The platform can categorize and statistically analyze correction entries based on material category, field type, cause of anomaly, or triggered attribute verification rules. When the cumulative number of similar correction entries under the same material category reaches a preset threshold, it indicates that existing attribute verification rules or attribute definitions may not be able to cover new material data. For example, if a new interface type is manually confirmed as valid multiple times in the same type of barcode reader, the platform can generate candidate options for expanding the parameter compliance range; if a brand's newly added model is repeatedly misjudged as abnormal, the platform can generate candidate options for updating brand and model constraint relationships.

[0120] Step S504: After the attribute verification rule update candidate or attribute definition update candidate is approved, update the attribute verification rule or attribute definition in the corresponding material category entity library.

[0121] After generating candidate items, the platform can submit them to rule maintainers or data administrators for review. Upon approval, the platform updates the attribute verification rules or attribute definitions in the corresponding material category entity library, ensuring that subsequent identical or similar material attribute data can be verified according to the updated rules. If a candidate item fails review, the platform can retain the original rules and continue to treat the relevant corrections as data to be observed. Through this anomaly control mechanism, the platform can not only handle single anomalies but also continuously optimize the rules and attribute definitions in the material category entity library using manual correction results.

[0122] As explained in step S105 of the aforementioned material information management method, the material information management platform can generate target material information based on the target data source and return the target material information to the intelligent agent. For manufacturing R&D selection, procurement replacement, or supply chain adjustment scenarios, material access requests can be further expressed as alternative material query requests. The following section, in conjunction with steps S601 to S605, further explains the generation process of target material information in the alternative material query scenario.

[0123] Step S601: Obtain the brand, model, and specification parameters of the material to be queried based on the target data source.

[0124] When the material information management platform identifies a material access request as a substitute material query request, it can first read the brand, model, and specifications of the material to be queried based on the target data source determined in the preceding steps. The brand and model are used to identify the material being queried, while the specifications are used to determine whether different materials are substitutable. Specifications may include parameters related to material selection, such as interface type, dimensions, electrical parameters, communication method, performance indicators, and usage conditions.

[0125] For example, after receiving a request from an engineer to "check if there are alternative materials for a certain model of barcode reader," the intelligent agent can convert the request into an alternative material query request and submit it to the material information management platform. The platform locates the corresponding material category entity library based on the material code and extracts fields such as the barcode reader's brand, model, interface type, communication method, installation dimensions, and compatible cable from the target data source. Through this data extraction, the material to be queried is transformed into a set of parameters that can be used for comparison.

[0126] Step S602: Determine the candidate alternative materials corresponding to the material to be queried based on the entity library of the corresponding material category.

[0127] After obtaining the basic information of the material to be queried, the material information management platform can search the corresponding material category entity database for candidate materials that are in the same material category, similar material category, or have a matching substitution relationship with the material to be queried. Candidate substitute materials can come from substitution relationship records in the entity database, historical selection records, brand matrix rules, manually confirmed substitution records, or parameter similarity search results.

[0128] In some implementations, the platform can first narrow down the search scope by material category, and then filter candidate alternative materials by brand, model, core specifications, or compatibility. For example, for barcode reader materials, the platform can prioritize candidate materials with similar interface types, reading performance, communication methods, and installation compatibility conditions; for connectors or wire harnesses, the platform can prioritize candidate materials with matching parameters such as interface type, number of pins, cable specifications, and compatible devices. The candidate alternative materials obtained in this way are not freely recommended by the intelligent agent based on experience, but rather verifiable data from the corresponding material category entity library.

[0129] Step S603: Call the substitute material rules and parameter mapping template in the material category entity library to perform parameter alignment between the material to be queried and the candidate substitute materials.

[0130] Once candidate alternative materials are identified, the material information management platform calls upon the alternative material rules and parameter mapping templates in the material category entity library to align the parameters of the queried material with those of the candidate alternative materials. The parameter mapping template is used to unify field names and parameter expressions across different brands, models, or data sources. For example, if one data source uses a "communication interface" field and another uses a "communication method" field, the platform can use the parameter mapping template to map both to a unified comparison field.

[0131] The substitution rules specify which parameters must be consistent, which parameters are allowed to differ, and which parameters require manual confirmation. For core parameters that directly affect assembly, communication, performance, or safety, the platform can require candidate substitution materials to be consistent with the material being queried or to meet preset compatibility conditions; for non-critical parameters, the platform can allow differences within a certain range. After parameter alignment, the platform obtains parameter comparison results between the material being queried and each candidate substitution material, allowing for item-by-item comparison.

[0132] Step S604: Calculate the compatibility score of the candidate alternative material relative to the material to be queried based on the parameter alignment results.

[0133] After parameter alignment is completed, the material information management platform calculates a compatibility score based on the parameter comparison results. The compatibility score reflects the degree of matching between candidate alternative materials and the material being queried in terms of core parameters, interface adaptation, performance indicators, matching relationships, and brand strategy. For candidate alternative materials that are completely identical or meet the conditions for direct replacement, their compatibility score can be high; for candidate alternative materials with some identical core parameters but non-critical differences, their compatibility score can be in the middle range; for candidate alternative materials with differences in key parameters that require further evaluation before use, their compatibility score can be low.

[0134] In some implementations, the platform can also adjust the compatibility score by incorporating manual verification of replacement records. If a replacement relationship has been manually confirmed and used in actual selection for a long time, the candidate replacement material can obtain higher credibility; if the candidate replacement material is only obtained by automatic matching of parameter similarity and has not been manually confirmed, the platform can mark it as needing review in the score results. This can prevent the agent from giving unreliable replacement suggestions based solely on superficial parameter similarity.

[0135] Step S605: Generate a list of alternative materials based on the compatibility score, and return the list of alternative materials as at least part of the target material information to the agent.

[0136] The material information management platform sorts, classifies, or filters candidate alternative materials based on compatibility scores, generating an alternative material list. This list may include information such as the material code, brand, model, key parameters, compatibility score, substitution level, constraints, and whether manual confirmation is required for each candidate alternative material. The platform returns this alternative material list as at least part of the target material information to the agent, which then generates selection recommendations for engineers, purchasing personnel, or other users based on the alternative material list.

[0137] When generating the alternative material list, the platform can also write information such as the data source of candidate alternative materials, compatibility score basis, parameter alignment results, and whether manual confirmation is required into the output record associated with the target material information. Through the above alternative material query process, the material information management platform transforms the process of "agent recommending alternative materials" into a verifiable calculation process based on entity library data, parameter comparison templates, and alternative material rules. Since the candidate alternative materials come from the entity library of the corresponding material category, and the parameter comparison process undergoes unified mapping and compatibility scoring, the alternative suggestions returned by the agent are no longer just natural language inferences, but structured results with data sources, parameter basis, and rule basis, thereby helping to improve the accuracy and traceability of alternative material recommendations.

[0138] The above embodiments mainly describe the process by which the material information management platform responds to intelligent agent access requests based on existing material master data, parameter database, and material category entity database. To ensure that subsequent access requests have access to continuously supplemented data sources, the platform can also perform pre-entry processing on newly imported product manuals, supplier information, BOM lists, drawings, or technical agreements. The following sections, in conjunction with steps S701 to S706, explain the feature detection, structured parameter extraction, and pre-writing conflict check process for original material documents.

[0139] Step S701: Receive the original material documents.

[0140] The material information management platform can receive raw material documents from user uploads, agent submissions, business system synchronization, or file directory scanning. Raw material documents can be product manuals, PDF documents, Word documents, Excel lists, images, drawings, or other documents containing material information. Upon receiving a raw material document, the platform can first save its original source information, such as upload source, file name, receipt time, submission recipient, and associated material code, for subsequent traceability.

[0141] In some implementations, the original material documents may not yet clearly correspond to a specific material category, or may only contain scattered information such as brand, model, parameter table, interface description, or matching relationships. The platform does not directly write this raw content into the formal material data, but first converts the document content into verifiable data through feature sniffing and structured extraction processes.

[0142] Step S702: Perform feature sniffing on the original material document to obtain document features.

[0143] The material information management platform can parse the original material documents and analyze the high-frequency technical terms in the parsing results. These high-frequency technical terms can be recurring material category names, interface names, specification terms, model terms, function terms, or parameter terms in the document. For example, terms such as "resolution," "reading speed," "RS-232C," "M12 interface," and "cable length" can reflect the material categories and technical content involved in the document.

[0144] In addition to text parsing, the platform can also detect structural features in original material documents. Structural features can include at least one of the following: table column headers, paragraph headings, data formats, and text information in drawings. For example, when an Excel BOM contains column headers such as "Material Code," "Quantity," or "Tag Number," the platform can identify that the document may belong to the Bill of Materials category; when a product manual contains a continuous parameter table, the platform can identify that it may contain extractable specification parameters; when interface annotations or connection relationships are obtained from text recognition in drawings or images, the platform can use this information as part of the document features.

[0145] After the above processing, high-frequency technical terms and structural features together constitute document features. Compared with relying solely on file names or manual classification, document features can more objectively reflect the content type and material category tendency of the original document.

[0146] Step S703: Based on the matching results between document features and category-parameter-template mapping relationship, determine the material category and document type corresponding to the original material document.

[0147] The material information management platform matches document characteristics with a pre-maintained category-parameter-template mapping relationship. This mapping relationship records the correspondence between different material categories and their common parameter fields, parameter templates, and document types. For example, barcode reader materials can correspond to parameter fields such as reading distance, interface type, reading speed, and matching cables; connector materials can correspond to parameter fields such as pin count, interface type, and cable specifications.

[0148] The platform determines the material category and document type corresponding to the original material document based on the matching results. If the document features highly match the parameter template of a certain material category, the platform can classify the document into that material category. If the document contains features of multiple categories, the platform can generate candidate material categories according to the degree of matching, and use the category with the highest matching degree as the preliminary result, or submit multiple candidate categories for subsequent review. The document type can be basic material product information, BOM list, drawing connection relationships, supplier information, or other types.

[0149] Step S704: Call the corresponding parameter template according to the material category and document type to perform structured parameter extraction on the original material document and obtain structured parameters.

[0150] After determining the material category and document type, the material information management platform calls the corresponding parameter template to perform structured parameter extraction on the original material document. The parameter template specifies which fields need to be extracted, how field names should be standardized, what data format field values ​​should use, and what relationship should exist between field values ​​and material category.

[0151] When a parameter template matching the material category and document type exists, the platform extracts structured parameters from the original material document according to the field definitions in the parameter template, and generates a confidence score for each field in the structured parameters. The confidence score may be related to the field source, text recognition results, table matching results, field format conformity, or consistency of multiple repetitions. For fields with high confidence, the platform can prioritize them for subsequent conflict pre-checking; for fields with low confidence, the platform can mark them as requiring manual confirmation.

[0152] When no parameter template matches the material category and document type, the platform can generate a temporary parameter template based on common parameter dimensions and mark the structured parameters extracted from the temporary parameter template as data to be improved. The temporary parameter template can be used to handle the initial warehousing requirements of unknown document types or new material categories, but its extraction results usually need to be approved before they can be used as formal parameter data.

[0153] Step S705: Based on the attribute verification rules or parameter compliance rules configured in the corresponding material category entity library, perform conflict pre-check on the extracted structured parameters.

[0154] After the structured parameters are extracted, the material information management platform can call the attribute verification rules or parameter compliance rules configured in the corresponding material category entity library to perform conflict pre-checks on the extracted structured parameters. Conflict pre-checks are used to determine whether there are any problems such as parameter exceeding limits, brand non-compliance, abnormal encoding format, category mismatch, or conflicting field combination relationships before the structured parameters are officially written into the entity library.

[0155] For example, if the extracted interface type does not belong to the allowed interface range for the material category, the platform can mark the field as a conflict field; if the extracted model and brand do not match, the platform can identify it as an abnormal attribute combination according to the attribute verification rules; if the material code format does not conform to the corresponding category rules, the platform can prevent the structured parameter from being directly entered into the database. Through conflict pre-detection, the platform can discover data problems in advance during the original material document entry stage, rather than exposing errors when the agent queries or writes them later.

[0156] Step S706: Write, return, or manage anomalies in the structured parameters and original material documents based on the conflict pre-inspection results.

[0157] When the structured parameters pass the conflict pre-detection, the material information management platform can write the structured parameters into the corresponding material category entity library and save the original material document as a reference in the same entity library. Specifically, the structured parameters can be entered as material attribute data into the data area of ​​the entity library, while the original material document can be saved as a reference in the reference resource area of ​​the corresponding entity library. This way, when the intelligent agent queries material information subsequently, it can not only read the structured parameters but also trace the source of the original document as needed.

[0158] When a structured parameter fails the conflict pre-check, the platform can write the failed field and its corresponding conflict rule to the pending review directory and trigger the anomaly control mechanism for structured parameters. This anomaly control mechanism can be executed in accordance with the aforementioned anomaly control, manual confirmation, and rule update process. After review and confirmation, the structured parameter can be corrected and written, rejected, or kept in a pending review state.

[0159] Through the aforementioned raw material document processing steps, the material information management platform can transform unstructured or semi-structured documents into verifiable, locatable, and traceable structured parameters. Since document feature detection first identifies the material category and document type, structured extraction is then performed based on corresponding parameter templates for field processing, and conflict pre-detection further prevents abnormal parameters from being directly entered into the database, this process reduces the workload of manual classification, extraction, and archiving, while also minimizing the risk of erroneous document data entering the material information system.

[0160] After processing in steps S701 to S706, the valid information in the original material document can be transformed into structured parameters that pass conflict pre-detection and used as candidate material data sources for subsequent material access requests. When the agent subsequently accesses the corresponding material category entity library, the platform can use these structured parameters for field supplementation, parameter comparison, or source tracing. The following uses PDF product manuals and Excel bills of materials as examples to illustrate a specific implementation method of the original material document entry process; this example is only used to illustrate the cooperation between feature sniffing, structured extraction, and conflict pre-detection, and does not limit the specific format of the original material document.

[0161] In this scenario, R&D engineers or materials maintenance personnel upload a PDF product manual for a specific product to the materials information management platform. Upon receiving the manual, the platform first parses the document, extracting textual content such as the product name, model number, interface name, specifications, and communication method. It also identifies frequently occurring technical terms. For example, if terms like "reading distance," "resolution," "communication interface," "RS-232C," and "M12 interface" appear repeatedly in the document, the platform can use these terms as part of the document's characteristics.

[0162] At the same time, the platform can also detect structural features in PDF product manuals. For example, if the product manual contains parameter tables, the platform can identify the table headers, parameter names, parameter values, and units; if the product manual contains interface diagrams or wiring instructions, the platform can obtain the interface text, cable identification, or connection relationship descriptions in the diagrams through text recognition. These high-frequency technical terms and structural features together constitute the document characteristics of the product manual.

[0163] The platform then matches document features with the category-parameter-template mapping relationship. If the matching result shows that the product manual corresponds to the parameter template of the barcode reader material, the platform can determine that the material category corresponding to the document is barcode reader, and identify its document type as product basic information or product technical manual. Next, the platform calls the parameter template of the barcode reader material, extracts structured parameters such as model, interface type, reading performance, power supply parameters, and adapter cable from the product manual, and generates a confidence score for each field.

[0164] For example, if the model number field appears consistently multiple times in the document title, parameter table, and header, its confidence score can be higher; if the interface type appears only once in the image text recognition result, the platform can assign a lower confidence score and mark it as requiring further confirmation. Through the above processing, the natural language content and table content in the PDF product manual are converted into structured parameters that can be incorporated into the material information system.

[0165] For Excel bills of materials, the platform can process them in a similar manner. Upon receiving the Excel bill of materials, the platform examines the table headers, such as fields like "Material Code," "Model," "Quantity," "Track Number," and "Brand," and determines whether the document belongs to the bill of materials category based on the table data format. The platform can then extract structured parameters such as material code, model, brand, and quantity from the Excel bill of materials and associate these parameters with the corresponding material category entity library based on the material code.

[0166] After the structured parameters are extracted, the platform does not immediately write them into the official entity database. Instead, it first calls the attribute verification rules or parameter compliance rules configured in the corresponding material category entity database to perform a conflict pre-check. For parameters extracted from PDF product manuals, the platform can verify whether the brand and model match, whether the interface type falls within the allowed interface range for that material category, and whether the specifications meet the parameter template requirements. For data extracted from Excel bills of materials, the platform can verify whether the material code format conforms to the coding rules, whether the material code matches the material category, and whether the model field and brand field form a valid combination.

[0167] When the structured parameters pass the conflict pre-detection, the platform can write the structured parameters into the corresponding material category entity library and save the original PDF product manual or Excel bill of materials as a reference to the corresponding entity library. In this way, when the intelligent agent queries the material information later, it can not only read the structured parameters, but also trace back to the original document source based on the output records.

[0168] When structured parameters are written into the corresponding material category entity library, the material information management platform can establish a binding relationship between the structured parameters and the material code, model, brand, or other material identifiers extracted from the original material document, and update the corresponding index mapping relationship. Subsequently, when the intelligent agent accesses the corresponding material category entity library based on the material code, the platform can include the structured parameters in the candidate material data source according to the index mapping relationship.

[0169] The structured parameters, once written, can serve as candidate material data sources to respond to subsequent material access requests. They can also participate in field supplementation, parameter comparison, or source tracing during the generation of target material information. When a structured parameter fails the conflict pre-check, the platform can write the failed field, the corresponding conflict rule, and the original document source into the pending review directory and trigger the anomaly control mechanism. For example, if the model number and brand relationship identified in the PDF product manual are inconsistent, the platform can mark the field as pending review; if the material code prefix in the Excel bill of materials does not match the material category, the platform can prevent the data from being directly entered into the database. After confirmation by the reviewers, the abnormal fields can be corrected, rejected for entry into the database, or kept in a pending review status.

[0170] This specific scenario demonstrates that the material information management platform can transform raw material documents such as PDF product manuals and Excel bills of materials into structured parameters and perform rule validation before writing them into the entity database. Because the inbound process retains the original document source, field confidence levels, and conflict pre-detection results, the agent can obtain a more reliable and traceable data foundation when subsequently accessing this material information, thereby reducing data quality issues caused by manual classification, manual extraction, and erroneous inbound entries.

[0171] This application also provides a material information management device. This device can be deployed within the aforementioned material information management platform, or it can function as a software component within the material information management platform. The device includes a request receiving module, a scheduling and positioning module, a data source acquisition module, a data source determination module, an information generation module, an attribute acquisition module, an attribute verification module, and a data processing module.

[0172] The request receiving module is used to receive material access requests and obtain the material code carried in the material access request. The specific processing procedure of this module can be found in the relevant description in step S101 above.

[0173] The scheduling and location module is used to perform scheduling rule matching and entity library location on material codes based on scheduling rule information configured for intelligent agents, in order to determine the material category entity library corresponding to the material code. Specifically, the material master data is associated with the material category entity library through the material code, and the material master data is also associated with a parameter library that provides material parameter information. This module can perform scheduling rule matching and entity library location in the manner described in steps S102 and S201 to S206.

[0174] The data source acquisition module is used to retrieve candidate material data sources related to the material access request from the material master data, parameter library, and corresponding material category entity library based on the material code. The specific data acquisition method for this module can be found in the relevant description in step S103 above.

[0175] The data source determination module is used to determine the target data source based on the candidate material data sources. When there are multiple candidate material data sources, the module determines the target data source from among them according to a preset conflict priority. For specific details regarding the handling of data source identifiers, field integrity information, update time information, and conflict priority, please refer to the relevant descriptions in the aforementioned candidate material data source arbitration embodiment.

[0176] The information generation module is used to extract field data corresponding to the material access request based on the target data source, generate target material information, and return the target material information to the intelligent agent. In some implementations, the information generation module can also generate output records associated with the target material information to record the data source, priority source, or verification status of the target material information.

[0177] The attribute acquisition module is used to respond to an attribute activation request submitted by the agent and acquire the material attribute data to be verified carried in the attribute activation request. The material attribute data to be verified is the attribute data generated by the agent based on the target material information, and is requested to be written to or updated in the material data. This module can perform the relevant processing in the aforementioned step S106.

[0178] The attribute verification module invokes the attribute verification rules configured corresponding to the material category entity library, performing cross-consistency verification on at least two related basic attributes in the attribute data of the material to be verified. The attribute verification module can extract at least two of the basic attributes such as brand, model, code prefix, material category, interface type, and specification parameters, and compare the actual combination relationship with the constraint relationship in the attribute verification rules.

[0179] The data processing module is used to perform corresponding processing on the material attribute data to be verified based on the verification results of cross-consistency checks. For example, when the cross-consistency check passes, the data processing module can write the material attribute data to be verified into the corresponding material category entity library, or mark it as data that can be used to update the corresponding material category entity library; when the cross-consistency check fails, the data processing module can generate a verification record and write the material attribute data to be verified into the pending review queue. For the specific processes of anomaly control, manual confirmation, and rule or attribute definition updates, please refer to the relevant descriptions in the aforementioned anomaly control embodiments.

[0180] In some implementations, the processes of receiving raw material documents, feature sniffing, extracting structured parameters, detecting conflicts, and writing data can be carried out collaboratively by the data source acquisition module, the data source determination module, and the data processing module. Alternatively, they can be separated into independent document receiving modules, feature sniffing modules, parameter extraction modules, conflict detection modules, and writing modules, depending on the actual deployment needs.

[0181] It should be noted that the above modules can be implemented through software programs, hardware circuits, or a combination of both. The division between the modules is only for illustrating the technical solution of this application and does not constitute a limitation on the actual software architecture or hardware structure. In actual implementation, multiple modules can be integrated into one processing module, and a module can be divided into multiple sub-modules. As long as the corresponding steps in the foregoing method embodiments can be implemented, they should all be understood as implementations of this application.

[0182] This application provides a schematic block diagram of a computer device. The computer device 800 can be an enterprise local server, a cloud server, a privately deployed data management device, a material information management platform server, a research and development selection system server, or other electronic devices capable of executing computer programs. The material information management method is applied to the material information management platform, which is used to interact with intelligent agents, material master data, parameter libraries, and material category entity libraries.

[0183] See Figure 3The computer device 800 includes a processor 802, a memory, and a network interface 805 connected via a system bus 801. The memory may include a non-volatile storage medium 803 and internal memory 804.

[0184] The non-volatile storage medium 803 can store an operating system 8031 ​​and a computer program 8032. The computer program 8032 includes program instructions that, when executed, cause the processor 802 to perform a material information management method. The processor 802 provides computing and control capabilities to support the operation of the entire computer device 800. Specifically, upon receiving a material access request, the processor can obtain the material code and locate the entity library based on scheduling rule information. Subsequently, the processor obtains candidate material data sources from the material master data, parameter library, and corresponding material category entity library, and determines the target data source according to a preset conflict priority. After generating the target material information and returning it to the agent, the processor can also respond to an attribute activation request, obtain the attribute data of the material to be verified, and call the attribute verification rules to perform cross-consistency verification, and then process the attribute data of the material to be verified according to the verification result.

[0185] The internal memory 804 provides an environment for the operation of the computer program 8032 in the non-volatile storage medium 803. When the computer program 8032 is executed by the processor 802, the processor 802 can execute a material information management method.

[0186] The network interface 805 is used for network communication with other devices, such as intelligent agents, user terminals, devices containing material master data, devices containing parameter databases, and devices containing material category entity databases.

[0187] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the embodiments of this application, and does not constitute a limitation on the computer device 800 to which the embodiments of this application are applied. The specific computer device 800 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0188] It should be understood that, in the embodiments of this application, the processor 802 may be a central processing unit, a graphics processor, a digital signal processor, an application-specific integrated circuit, a field-programmable gate array, or other programmable logic devices. The memory may include volatile memory and / or non-volatile memory, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, or other storage devices that can be used to store computer programs and data.

[0189] Since the computer device implements the steps in the aforementioned method embodiments by executing computer programs, it can achieve the same or corresponding technical effects as the aforementioned method embodiments, which will not be repeated here.

[0190] This application also provides a storage medium storing a computer program, which includes program instructions. When executed by a processor, the program instructions cause the processor to perform the material information management method in any of the foregoing embodiments.

[0191] For example, when the program instructions are executed by the processor, the processor can receive material access requests, obtain material codes, perform scheduling rule matching and entity library location, obtain candidate material data sources, determine target data sources, generate and return target material information, obtain material attribute data to be verified, perform cross-consistency checks, and process data based on the check results. For the specific implementation of each step, please refer to the description in the foregoing method embodiments.

[0192] The storage medium can be any medium capable of storing program code, such as read-only memory, random access memory, disk, optical disk, flash memory, portable hard drive, solid-state drive, server storage media, or other non-transitory computer-readable storage media. Program instructions can be stored in the storage medium in the form of software, firmware, or other forms that can be read and executed by a processor. After the processor executes the program instructions, it can implement the aforementioned material information management method, thereby subjecting the intelligent agent to scheduling, arbitration, and verification mechanisms when accessing, generating, or requesting effective material attribute data, thus improving the reliability of the material data retrieval and writing process.

[0193] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0194] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0195] The steps in the methods of this application embodiment can be adjusted, merged, or deleted according to actual needs. The units in the apparatus of this application embodiment can be merged, divided, or deleted according to actual needs. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0196] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0197] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A material information management method, characterized in that, include: Receive a material access request from the intelligent agent and obtain the material code carried in the material access request; Based on the scheduling rule information configured for the intelligent agent, the material code is matched with the scheduling rule and located in the entity library to determine the material category entity library corresponding to the material code. The material master data is associated with the material category entity library through the material code, and the material master data is also associated with a parameter library for providing material parameter information. Based on the material code, candidate material data sources related to the material access request are obtained from the material master data, the parameter library, and the corresponding material category entity library; The target data source is determined based on the candidate material data source, wherein when there are multiple candidate material data sources, the target data source is determined from the multiple candidate material data sources according to a preset conflict priority; Based on the target data source, extract the field data corresponding to the material access request, generate target material information, and return the target material information to the agent; In response to receiving the attribute activation request submitted by the intelligent agent, the material attribute data to be verified carried in the attribute activation request is obtained. The material attribute data to be verified is the attribute data generated by the intelligent agent based on the target material information and requested to be written to or updated in the material data. Invoke the attribute verification rules configured corresponding to the material category entity library, and perform cross-consistency verification on at least two related basic attributes in the attribute data of the material to be verified; Based on the verification results of the cross-consistency check, the corresponding processing is performed on the material attribute data to be verified.

2. The method according to claim 1, characterized in that, Before receiving a material access request from the agent, the following is also included: Obtain material master data, and write the material code, material category and material model in the material master data into the index data area, write the material attributes in the material master data into the feature data area, and write the historical records, change records or usage records associated with the material master data into the archive data area. Based on the correspondence between the material code and the material category, a routing mapping relationship from the material code to the material category entity library is generated; Write the routing mapping relationship into the scheduling rule information; When locating the entity library based on the scheduling rule information, the category identification field is extracted from the material code, and the corresponding entity library path or entity library identifier is queried in the scheduling rule information based on the category identification field. The corresponding material category entity library is located according to the entity library path or entity library identifier.

3. The method according to claim 1, characterized in that, A target data source is determined based on the candidate material data sources. When there are multiple candidate material data sources, the target data source is determined from the multiple candidate material data sources according to a preset conflict priority. Field data corresponding to the material access request is extracted from the target data source, and target material information is generated, including: Obtain the data source identifier, field integrity information, and update time information of the candidate material data source respectively; When there are multiple candidate material data sources, the data level of each candidate material data source is determined based on the data source identifier, and the target data source is determined according to the preset conflict priority of master data priority, parameter library supplementation, and entity library expansion, combined with field integrity information and update time information. When the target field is missing in the material master data that is identified as the target data source, supplementary data corresponding to the target field is obtained from the parameter library or the corresponding material category entity library. The supplementary data is then merged with the field of the material master data to generate the target material information. Generate an output record associated with the target material information. The output record includes at least one of the following: the data source corresponding to the target material information, the conflict priority source identifier of the target data source, and the verification status of the target material information. After performing cross-consistency verification, write the verification result of the cross-consistency verification into the output record or the verification record associated with the output record.

4. The method according to claim 1, characterized in that, Perform cross-consistency checks on at least two related basic attributes in the material attribute data to be verified, and based on the results of the cross-consistency checks, perform corresponding actions on the material attribute data to be verified, including: Extract at least two basic attributes from the material attribute data to be verified. The basic attributes include at least two of the following: brand, model, code prefix, material category, interface type, and specification parameters. Based on the attribute verification rules, determine the constraint relationship between at least two basic attributes, and compare the actual combination relationship of at least two basic attributes with the constraint relationship to obtain the verification result of cross-consistency verification; If the cross-consistency check passes, the material attribute data to be verified is written into the corresponding material category entity library, or the material attribute data to be verified is marked as data that can be used to update the corresponding material category entity library; If the cross-consistency check fails, a verification record is generated to characterize the abnormal fields, abnormal attribute combinations, and corresponding attribute verification rules, and the material attribute data to be verified is written to the audit queue.

5. The method according to claim 4, characterized in that, If the cross-consistency check fails, an anomaly control mechanism will be triggered on the material attribute data to be verified. The anomaly control mechanism includes: In response to receiving a manual confirmation result for the verification record, the material attribute data to be verified is corrected, rejected, or kept pending review based on the manual confirmation result. Record user correction entries for the attribute data of the material to be verified; When the cumulative number of similar correction items under the same material category reaches a preset threshold, generate candidate items for attribute verification rule updates or attribute definition updates. After the attribute verification rule update candidate or attribute definition update candidate is approved, the attribute verification rule or attribute definition in the corresponding material category entity library is updated.

6. The method according to claim 1, characterized in that, When the material access request is a substitute material query request, the field data corresponding to the material access request is extracted based on the target data source, and target material information is generated and returned to the intelligent agent, including: Obtain the brand, model, and specification parameters of the material to be queried based on the target data source; Based on the corresponding material category entity library, determine the candidate alternative materials corresponding to the material to be queried; Call the alternative material rules and parameter mapping templates in the material category entity library to perform parameter alignment between the material to be queried and the candidate alternative materials; Calculate the compatibility score of the candidate alternative material relative to the material to be queried based on the parameter alignment results; A list of alternative materials is generated based on the compatibility score, and the list of alternative materials is returned to the agent as at least part of the target material information.

7. The method according to claim 1, characterized in that, Also includes: Receive original material documents; Feature sniffing is performed on the original material documents to obtain document features, including high-frequency technical terms obtained from text parsing of the original material documents, and structural features detected from the original material documents; Based on the matching results between document features and category-parameter-template mapping relationships, the material category and document type corresponding to the original material document are determined. The corresponding parameter template is called according to the material category and document type to perform structured parameter extraction on the original material document and obtain the structured parameters. Based on the attribute verification rules or parameter compliance rules configured in the corresponding material category entity library, perform conflict pre-checks on structured parameters; When a structured parameter passes the conflict pre-check, the structured parameter is written into the corresponding material category entity library, and the original material document is saved as a reference in the corresponding material category entity library; when a structured parameter fails the conflict pre-check, the fields that failed the pre-check and the corresponding conflict rules are written into the pending review directory, and the abnormal control mechanism for structured parameters is triggered. The structured parameters written serve as the data source for candidate materials, used to respond to subsequent material access requests.

8. A material information management device, characterized in that, include: The request receiving module is used to receive material access requests from the intelligent agent and obtain the material code carried in the material access request; The scheduling and positioning module is used to perform scheduling rule matching and entity library positioning on the material code based on the scheduling rule information configured for the intelligent agent, so as to determine the material category entity library corresponding to the material code. The material master data is associated with the material category entity library through the material code, and the material master data is also associated with a parameter library for providing material parameter information. The data source acquisition module is used to acquire, based on the material code, candidate material data sources related to the material access request from the material master data, the parameter library, and the corresponding material category entity library; A data source determination module is used to determine a target data source based on the candidate material data sources, wherein when there are multiple candidate material data sources, the target data source is determined from the multiple candidate material data sources according to a preset conflict priority; The information generation module is used to extract field data corresponding to the material access request based on the target data source, generate target material information, and return the target material information to the intelligent agent; The attribute acquisition module is used to, in response to receiving an attribute activation request submitted by the intelligent agent, acquire the material attribute data to be verified carried in the attribute activation request, wherein the material attribute data to be verified is attribute data generated by the intelligent agent based on the target material information and requests to write or update the material data; The attribute verification module is used to invoke the attribute verification rules configured corresponding to the material category entity library, and perform cross-consistency verification on at least two related basic attributes in the attribute data of the material to be verified; and The data processing module is used to perform corresponding processing on the material attribute data to be verified based on the verification results of the cross-consistency check.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the material information management method as described in any one of claims 1-7.

10. A storage medium, characterized in that, The storage medium stores a computer program, which includes program instructions that, when executed by a processor, cause the processor to perform the material information management method as described in any one of claims 1-7.

Images