A data full-link traceability method and system of a railway artificial intelligence data platform

Abstract

The application discloses a kind of data full-link traceability method and system of railway artificial intelligence data platform, it is related to railway data processing technical field, comprising: obtaining railway database design document and table structure information, railway database design document and table structure information are parsed, obtain structured mapping data, based on preset weighted keyword score, the business domain automatic classification of structured mapping data is carried out, and the business domain classification result is obtained;Ontology metadata is generated to the business domain classification result, based on ontology metadata and pre-defined relationship type table, automatic generation is carried out, and entity relationship graph is obtained, based on preset traceability specification, the ontology metadata in entity relationship graph is traced, and the first traceability item is obtained;Drill down knowledge of coding is obtained, based on the first traceability item and the drill down knowledge of coding, drill down query is carried out, and the next level traceability data is obtained, based on the next level traceability data, data full-link traceability is carried out, and problem root cause is determined.

Description

Technical Field

[0001] This invention relates to the field of railway data processing technology, specifically a method and system for full-link data traceability of a railway artificial intelligence data platform. Background Technology

[0002] The operation and management of railway transportation enterprises involves multiple business domains, including locomotive operation, crew management, material supply, equipment maintenance, engineering project construction, and safety management. Data from these domains is scattered across different business systems (such as the locomotive operation monitoring system LKJ, material management system, human resources management system, and safety management platform), resulting in heterogeneous data formats and inconsistent standards. When operational data anomalies occur, managers need to start with KPI indicators and trace the root cause layer by layer.

[0003] Existing data end-to-end tracing methods handle data modeling, query analysis, and user interaction in a fragmented manner, lacking coordination and integration between these stages, resulting in insufficient accuracy in locating data anomalies. Summary of the Invention

[0004] To address the shortcomings mentioned in the background section, the present invention aims to provide a data full-link traceability method and system for a railway artificial intelligence data platform.

[0005] Firstly, the objective of this invention can be achieved through the following technical solution: a method for full-link data traceability of a railway artificial intelligence data platform, the method comprising the following steps: Obtain railway database design documents and table structure information, parse the railway database design documents and table structure information to obtain structured mapping data, and automatically classify the structured mapping data into business domains based on preset weighted keyword scores to obtain business domain classification results; Generate ontology metadata from the business domain classification results. Based on the ontology metadata and the predefined relationship type table, automatically generate an entity relationship graph. Based on the preset traceability specifications, trace the ontology metadata in the entity relationship graph to obtain the first-level traceability item. Obtain the drill-down knowledge of the code, perform drill-down queries based on the first-level traceability item and the drill-down knowledge of the code to obtain the next-level traceability data, and perform full-link traceability of the data based on the next-level traceability data to determine the root cause of the problem.

[0006] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process of automatically classifying the structured mapping data based on a preset weighted keyword score into business domains, as follows: When the preset weighted keyword scoring includes matching table names with domain keywords and matching field text with domain keywords, the weight of the table name is greater than the weight of the text. The business domain classification results include: locomotive domain, personnel domain, equipment domain, materials domain, engineering project domain, safety domain, operation domain, production domain, performance domain, and system management domain.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the ontology metadata defines the types of entities managed by the platform, attribute structures, and classification systems. The types of entities include locomotives, equipment, personnel, engineering projects, and materials.

[0008] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the predefined relationship type table includes predefined graph relationships and ontology relationships.

[0009] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the entity relationship graph includes entity nodes and relationship edges, and the construction process includes: personnel node construction stage, locomotive node and driving relationship construction stage, material node construction stage, engineering project node and associated relationship construction stage, maintenance record node and maintenance relationship construction stage, scheduling relationship construction stage, contract node construction stage, supplier node construction stage, department node and subordinate relationship construction stage, safety event node construction stage, equipment node construction stage, and cross-domain influence chain construction stage.

[0010] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the cross-domain influence chain construction phase is based on preset rules, as follows: The system detects the inventory status field of a material entity. When the inventory status reaches a preset danger threshold, it automatically creates a material shortage impact relationship edge from the material entity to the relevant maintenance record entity.

[0011] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the drill-down query based on the first-level traceability item and the encoded drill-down knowledge, by maintaining a set of consumed dimensions after each drill-down and adding the currently used dimension to the consumed set, and when all dimensions have been consumed, the drill-down target automatically switches to a flat list mode.

[0012] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process of performing full-link data tracing based on the next-level tracing data, as follows: By managing metrics through the source trace stack and entity jump stack, we drill down to the hierarchical status and entity file page jump status, and return the source trace data to the front-end rendering, update the stack status and return the corresponding data to locate the root cause of the problem.

[0013] Secondly, in order to achieve the above objectives, this invention discloses a data end-to-end traceability system for a railway artificial intelligence data platform, comprising: The primary processing module is used to acquire railway database design documents and table structure information, parse the railway database design documents and table structure information to obtain structured mapping data, and automatically classify the structured mapping data into business domains based on preset weighted keyword scores to obtain business domain classification results. The secondary processing module is used to generate ontology metadata from the business domain classification results. Based on the ontology metadata and the predefined relationship type table, it automatically generates an entity relationship graph. Based on the preset traceability specifications, it traces the ontology metadata in the entity relationship graph to obtain the first-level traceability items. The source tracing output module is used to obtain the coded drill-down knowledge, perform drill-down queries based on the first-level source tracing items and the coded drill-down knowledge to obtain the next-level source tracing data, and perform full-link data tracing based on the next-level source tracing data to determine the root cause of the problem.

[0014] The beneficial effects of this invention are: This invention forms a complete railway artificial intelligence data platform, encompassing everything from automatic modeling of heterogeneous database design documents to unlimited-level drill-down queries and user-interactive tracing. Attached Figure Description To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system structure of the present invention; Figure 3 This is a schematic diagram of the homepage of the AI ​​intelligent management platform in an embodiment of the present invention; Figure 4 This is a schematic diagram of the holographic archive overview page in an embodiment of the present invention; Figure 5 This is a schematic diagram of the business insight page in an embodiment of the present invention; Figure 6 This is a schematic diagram of a one-item-one-file page in an embodiment of the present invention; Figure 7 This is a schematic diagram of the full data unlimited traceability indicator list page in an embodiment of the present invention; Figure 8 This is a schematic diagram of the AI-powered global intelligent data query page in an embodiment of the present invention; Figure 9 This is a schematic diagram of a one-person-one-file page in an embodiment of the present invention; Figure 10 This is a schematic diagram of a single page in an embodiment of the present invention; Figure 11 This is a schematic diagram of a single-machine-one-file page in an embodiment of the present invention; Figure 12 This is a schematic diagram of the intelligent traceability indicator selection page in an embodiment of the present invention; Figure 13 This is a schematic diagram of the complete traceability link of indicators in an embodiment of the present invention; Figure 14 This is a schematic diagram of the primary traceability drilling panel in an embodiment of the present invention; Figure 15 This is a schematic diagram of the secondary traceability locomotive file drilling page in an embodiment of the present invention; Figure 16 This is a schematic diagram of the drilling page of the three-level traceability locomotive operation record in an embodiment of the present invention; Figure 17 This is a schematic diagram of the four-level traceability operation details and breadcrumb navigation page in an embodiment of the present invention; Figure 18 This is a schematic diagram of the three-level tracing page after returning to the previous level in an embodiment of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1: like Figure 1 As shown, a method for full-link data traceability of a railway artificial intelligence data platform includes the following steps: S101: Obtain railway database design documents and table structure information, parse the railway database design documents and table structure information to obtain structured mapping data, and automatically classify the structured mapping data into business domains based on preset weighted keyword scores to obtain business domain classification results; In this embodiment, This method extracts table structure information from the database design documents of various business systems within railway enterprises, supporting multiple formats such as Word, Excel, PDF, and Markdown. The parsing process outputs structured table-field mapping data, including metadata such as table name, field name, field type, field description, and primary / foreign key relationships.

[0017] Input: 8 database design documents + 2 structured data files (taking a certain railway as an example, with a total of 258 tables and 2906 fields).

[0018] Output: Structured mapping data in `all_parsed.json` format.

[0019] The process of automatically classifying structured mapping data into business domains based on preset weighted keyword scores is as follows: When the preset weighted keyword scoring includes matching table names with domain keywords and matching field text with domain keywords, the weight of the table name is greater than the weight of the text. The business domain classification results include: locomotive domain, personnel domain, equipment domain, materials domain, engineering project domain, safety domain, operation domain, production domain, performance domain, and system management domain.

[0020] Define keyword rule sets for 10 business domains, each domain containing table-level keywords and field-level keywords. Perform the following scoring on each table: If the table name matches the domain keyword and scores 3 points, the table name has a high weight because it usually directly reflects the business meaning. If the field text (field name + description) matches the domain keyword and scores 1 point, the field has a low weight because there are many fields and there may be noise. The domain with the highest score is designated as the domain to which the table belongs. When the highest score is below the threshold, it is classified into the "System Management" domain.

[0021] The pre-set weighted keyword scoring is based on the following: Database table names are usually directly named using English terms (such as locomotive, crew, material). The business meaning carried by the table name is highly condensed and uniquely targeted, and a single match can determine the business domain with high confidence. However, field text (field name and field description) is numerous and contains a large number of cross-domain common words (such as "id", "name", "date", "status", etc.), making single field matching less distinguishable and prone to noise interference. Therefore, the table name matching weight is set to 3 times the field matching weight.

[0022] The specific scoring rules are as follows: 3 points are awarded for a table name matching a domain keyword, and 1 point is awarded for a field text matching a domain keyword. The domain with the highest score is designated as the domain to which the table belongs. When the highest score is 0, meaning that neither the table name nor the field text matches any domain keyword, the table is assigned to the system management domain.

[0023] When two or more business domains have the same score and are all the highest, the following decision rules determine the domain affiliation: The domain with the higher table name matching score is given priority. If the table name matching scores are still the same, the domain affiliation is determined according to a preset business domain priority order, from highest to lowest: Security Domain, Locomotive Domain, Personnel Domain, Materials Domain, Engineering Project Domain, Equipment Domain, Operations Domain, Production Domain, Performance Domain, and System Management Domain. Furthermore, system auxiliary tables whose names are prefixed with "sys_", "temp_", "backup_", or "log_" are not included in the domain classification and are directly assigned to the System Management Domain.

[0024] In one implementation, domain classification is performed on the table `locomotive_maintenance_record`: - Table name containing "locomotive" → vehicle field +3 points - Table name containing "maintenance" → equipment field +3 points - Field "fault_code" → vehicle field +1 point - Field "repair_date" → equipment field +1 point - Ultimately: Vehicle domain gets 4 points, Equipment domain gets 4 points. Based on the priority order of business domains, the locomotive domain takes precedence over the equipment domain and is therefore assigned to the locomotive domain. S102: Generate ontology metadata from the business domain classification results, automatically generate an entity relationship graph based on the ontology metadata and a predefined relationship type table, and trace the ontology metadata in the entity relationship graph based on the preset traceability specifications to obtain the first-level traceability items. Ontology metadata defines the types of entities, attribute structures, and classification systems managed by the platform. The types of entities include locomotives, equipment, personnel, engineering projects, and materials.

[0025] The ontology metadata is the internal data model of the railway artificial intelligence data platform, which defines the types of entities, attribute structures, and classification systems managed by the platform.

[0026] The graphical entity type system includes 20 types: crew, material, vehicle, equipment, project, maintenance, contract, supplier, department, safety, operations, finance, annual plan, budget indicator, procurement, acceptance, warranty, audit issue, supplier evaluation, and payment record. | Relationship | Source → Target | Tag | Weight | |------|---------|------|------| | drives | vehicle→crew | On-duty driver | 3 | | maintains | maintenance→vehicle | inspection | 5 | | uses | maintenance→material | accessories used | 4 | | Requires | Project → Material | Project Materials | 3 | | involves | project→vehicle | 4 | | supplies | contract→material | Contract supply | 5 | | procured_by | contract→supplier | Supplier signing | 5 | | dept_crew | department→crew | Department personnel | 3 | | safety_vehicle | safety→vehicle | The motorcycle involved in the accident | 5 | | material_shortage_impact | material→maintenance | Impact of material shortages | 4 | | vehicle_issue_impact | vehicle→safety | Locomotive malfunction impact | 5 | | ...and 14 other types| | | |; The ontology entity type system includes (17 types): locomotive, equipment, crew, project, material, contract, safety_alert, safety_rule, safety_incident, risk, training, violation, purchase_request, acceptance, inspection_plan, production_plan, dispatch_plan; 22 types of ontology relationships (ontology system): including co_drives (co-driver), operations (equipment service procedures), manages / participates (project management / participation), belongs_to (personnel assignment), generates (generate safety alarms), triggers (trigger procedure verification), monitors (monitor locomotive status), etc.

[0027] The relationship definition supports multiple sources and multiple targets (fromTypes array, toTypes array), enabling flexible many-to-many relationship modeling.

[0028] The entity relationship graph includes entity nodes and relationship edges. The construction process includes: personnel node construction stage, locomotive node and driving relationship construction stage, material node construction stage, engineering project node and related relationship construction stage, maintenance record node and maintenance relationship construction stage, scheduling relationship construction stage, contract node construction stage, supplier node construction stage, department node and subordinate relationship construction stage, safety event node construction stage, equipment node construction stage, and cross-domain impact chain construction stage.

[0029] The cross-domain influence chain construction phase is based on preset rules, as follows: The system detects the inventory status field of a material entity. When the inventory status reaches a preset danger threshold, it automatically creates a material shortage impact relationship edge from the material entity to the relevant maintenance record entity.

[0030] The specific generation process of the entity relationship graph is as follows: First, data sources from each business domain are extracted from the structured mapping data. Based on the entity type definitions in the ontology metadata, corresponding entity nodes of each data source record are created. Node attributes are mapped and filled from the field values ​​of the data source, and the node ID is generated by combining the entity type identifier and the local primary key value. Then, based on the relationship schema registered in the predefined relationship type table, the matching and construction of relationship edges are performed in each stage: For exact matching relationships (such as driving relationships, departmental affiliation relationships), the specified attribute values ​​of the source entity are matched with the primary key values ​​of the target entity; for fuzzy matching relationships (such as involvement relationships, material usage relationships), matching is achieved through substring inclusion detection between the attribute values ​​of the source entity and the attribute values ​​of the target entity. Each successfully matched relationship edge is deduplicated, assigned a predefined relationship type identifier, Chinese label, and weight value, and then written into the graph.

[0031] During the cross-domain impact chain construction phase, the system traverses all material entity nodes, checks their inventory status field values, and compares the inventory quantity with the safety stock threshold recorded for that material entity. When the inventory quantity is lower than the safety stock threshold, the system automatically queries maintenance record entities that have a "use spare parts" relationship with that material entity, and creates a relationship edge of type "material shortage impact" between the material entity and the maintenance record entity, with an edge weight set to 4. This relationship edge represents the cascading impact of material shortages on maintenance activities, providing a path basis for subsequent cross-domain tracing.

[0032] Entity ID and entity type are provided by the entity type definition and graph nodes of the ontology data construction subsystem.

[0033] In this embodiment, the edge weight is set to 4. The constructed graph is then loaded into the in-memory graph engine.

[0034] Persist entity relationships to a relational database, using a unified relational table structure: The unified relational table structure includes the following fields: from_type (source entity type, text type), from_id (source entity ID, text type), to_type (target entity type, text type), to_id (target entity ID, text type), relation_type (relation type identifier, text type), relation_label (relational Chinese label, text type), start_date (relation start date, text type), end_date (relation end date, text type), and metadata (JSON format extended metadata, text type). Indexes are created along three dimensions: the combination of from_type and from_id, the combination of to_type and to_id, and relation_type.

[0035] from_id TEXT, -- Source Entity ID to_type TEXT, -- Target entity type to_id TEXT, -- Target Entity ID relation_type TEXT, -- Relation type identifier relation_label TEXT, -- Chinese label for relation start_date TEXT, -- Relationship start date end_date TEXT, -- Relationship end date metadata TEXT -- Extended metadata in JSON format S103: Obtain the drill-down knowledge of the code, perform drill-down queries based on the first-level traceability item and the drill-down knowledge of the code to obtain the next-level traceability data, and perform full-link traceability of the data based on the next-level traceability data to determine the root cause of the problem.

[0036] Drill-down queries are performed based on the first-level traceability items and the encoded drill-down knowledge. After each drill-down, a set of consumed dimensions is maintained, and the currently used dimension is added to the consumed set. When all dimensions have been consumed, the drill-down target automatically switches to a flat list mode.

[0037] Based on the defined entity type, declare the traceability specifications, defining the dashboard, traceable fields, and relationship filtering rules to be displayed during the traceability process for that entity type. Extract primary KPI indicators from railway operation data sources to generate a list of traceability items.

[0038] Source traceability item data structure: The traceability item data structure is defined as follows: Each traceability item contains a unique identifier (traceKey, string type), indicator name (title, string type), value (value, numeric type), unit (unit, string type), multi-dimensional decomposition data (breakdowns, array type, each element contains the dimension name and the grouped data under that dimension), SQL calculation logic (sqlLogic, string type, recording the SQL statement for calculating the indicator), data source reference (dataSource, object type, containing the source system name, data table name, query time and number of records), and relationship graph (relations, object type, containing a list of parent indicators, sibling indicators and related indicators).

[0039] traceKey: string; / / Unique identifier title: string; / / Indicator name value: number; / / numerical value unit: string; / / unit breakdowns: DimensionBreakdown[; / / Multi-dimensional data decomposition sqlLogic: string; / / SQL calculation logic dataSource: DataSourceRef; / / Reference to the data source relations: TraceRelations; / / Relationship graph The declarative drill-down query subsystem builds an entity relationship graph and ontology metadata generated by the subsystem based on ontology data. Through declarative traceability specifications and coded drill-down identifiers, it enables infinite-level drill-down queries from top-level KPI indicators to bottom-level original records.

[0040] The following is an example of dimensional consumption: Layer 1: Initial State → Remaining Dimensions [locoType, ownership, route, department, status, model] Layer 2: Consume locoType → Remaining dimensions [ownership, route, department, status, model] Layer 3: Consume ownership → Remaining dimensions [route, department, status, model] Level 4: Consume route → Remaining dimensions [department, status, model] Level 5: Consume department → Remaining dimensions [status, model] Level 6: Consume status → Remaining dimensions [model] Level 7: Consume model → Remaining dimensions → Switch to flat list mode The multi-type drill-down engine routes to different query processors based on the prefix of the drill ID. In the query processor, the acquisition of entity attributes depends on the entity node data generated by the ontology data construction subsystem, and the query of related entities depends on the relationship traversal interface provided by the memory graph engine in step six.

[0041] Source tracing path caching and reuse The first-level traceability items and dimensional decomposition results are cached in memory.

[0042] - The cache key is a combination of `metric identifier + dimension`; - Subsequent drill-down operations will prioritize looking up from the cache to avoid duplicate queries; - The cache expiration policy is based on the data update timestamp.

[0043] The dual-stack interactive traceability subsystem serves as the user-level interface, presenting the query capabilities of the declarative drill-down query subsystem to the user through dual-stack state management and a visualization panel. In the traceability panel and entity archive page, all entity references are rendered as traceable links.

[0044] The process is as follows: 1. Parse the entity ID (format `type:localId`) to obtain the entity type and local ID; the entity type definition is provided by step three of the ontology data construction subsystem; 2. Push the target entity onto the entity jump stack using the DRILL operation; 3. Navigate to the entity profile page route ` / profile / {entityType} / {entityId}`; 4. The entity profile page automatically loads entity data of the corresponding type. Entity attribute data is provided by the memory graph engine of the ontology data construction subsystem.

[0045] Specifically, the present invention will be further illustrated below through embodiments: Taking a certain railway as an example, input 8 database design documents and 2 structured data files, which contain a total of 258 data tables and 2906 fields.

[0046] 1. Parse and output structured mapping data from 258 tables; 2. Domain classification results: Locomotive domain: 32 tables; Personnel domain: 28 tables; Materials domain: 45 tables; Engineering project domain: 36 tables; Safety domain: 22 tables; Equipment domain: 18 tables; Operations domain: 25 tables; Production domain: 15 tables; Performance domain: 12 tables; System management domain: 25 tables. 3. Generate 5 core ontology types (e.g., vehicle and gear), 20 graphical entity types, 17 ontology entity types, 25 graphical relationships, and 22 ontology relationships; 4. 12-stage construction: Generates approximately 500 entity nodes and approximately 2000 relationship edges; 5. Persist to the entity_relation table in the SQLite database.

[0047] To verify the technical effectiveness of this invention, five verification cases are provided below based on the actual operation results of the system: Case Study 1: Domain Classification Effectiveness Verification. Using eight database design documents and two structured data files (totaling 258 tables and 2906 fields) from a certain railway company as input, the domain classification process of the ontology data construction subsystem was executed. The classification results were as follows: Locomotive Domain: 32 tables (12.4%), Personnel Domain: 28 tables (10.9%), Materials Domain: 45 tables (17.4%), Engineering Project Domain: 36 tables (14.0%), Safety Domain: 22 tables (8.5%), Equipment Domain: 18 tables (7.0%), Operations Domain: 25 tables (9.7%), Production Domain: 15 tables (5.8%), Performance Domain: 12 tables (4.7%), System Management Domain: 25 tables (9.7%). After manual verification, the classification results for 249 out of the 258 tables were correct, resulting in a domain classification accuracy rate of 96.5%. Of the nine misclassified tables, six were classified into the locomotive domain after being manually verified because their table names contained keywords from both domains (e.g., locomotive_maintenance_record matched both the locomotive and equipment domains). This classification was corrected because the table names contained keywords from both domains.

[0048] Verification Case 2: Unlimited Level Drill-down Verification. Taking the "Daily Vehicle Kilometers" indicator as an example, the system performs 7 levels of drill-down: The first level displays two categories based on locomotive type: electric and diesel (see appendix). Figure 14 The drill ID is dailyMileage|locoType|{}; after clicking "Electric Locomotive" on the second level, group by assigned unit (see appendix). Figure 15 The [locoType] dimension has been consumed, with 5 dimensions remaining; layers 3 to 5 consume the ownership, route, and department dimensions respectively (see appendix). Figure 16 Once all six dimensions of the sixth layer have been consumed, the system automatically switches to a flat list mode, displaying all locomotive entities and their complete attributes (see appendix). Figure 17 Clicking on a locomotive at level 7 switches to the daily record leaf node, displaying the locomotive's daily operating data. Throughout the drill-down process, the dimension consumption tracking mechanism correctly maintains the set of consumed dimensions, the drill ID encoding / decoding at each level is error-free, and the caching mechanism effectively avoids duplicate queries, verifying the core technical effectiveness of the declarative drill-down query subsystem.

[0049] Verification Case 3: Verification of the traceability panel and dual-stack interaction. On the smart traceability page (see appendix)... Figure 12 The user first selects the KPI indicator to be traced. The right panel displays the complete traceability chain of that indicator, including the original data source, the number of data records and the query time, the calculation formula, the verification rules, and the final result (see appendix). Figure 13 After clicking the "Start Drilling Down" button, the traceability panel will slide out from the bottom (see appendix). Figure 14The panel displays a depth indicator ("Level 1 Source Tracing"), breadcrumb navigation, data source identifiers (showing the source system name, data table name, and number of records), SQL calculation logic, and a clickable details table. Clicking on a blue data row in the details table triggers a PUSH operation to enter Level 2 (see appendix). Figure 15 Clicking further will take you to level 3 and deeper levels (see appendix). Figure 16 After multi-level drilling, the crumb navigation displays the complete path in real time (see appendix). Figure 17 Users can click the history tab in the breadcrumbs to perform a JUMP_TO operation to jump to the corresponding level, or press the Backspace key to return level by level (see appendix). Figure 18 This verification case confirms the correctness of the PUSH, POP, and JUMP_TO operations of the traceability stack in the dual-stack interactive traceability subsystem, as well as the interactive functions of the bottom slide-out panel, capsule breadcrumb navigation, and keyboard shortcut operations (ESC to close, Backspace to go back).

[0050] Verification Case 4: Verification of "Five One Files" Entity Archives. The system generated independent archive pages for each of the five core entity categories. (See attached file for one item, one archive page) Figure 7 This displays basic information about materials, inbound and outbound records, inventory trend charts, and related contracts. The inventory trend charts are rendered by the ECharts component, and the data comes from the related query results of the in-memory graph engine. (See attached file for individual file details.) Figure 9 Displays basic personnel information, department, and skill level. Each item is listed on a separate page (see appendix). Figure 10 This displays basic information about the project, a list of related materials, and contract information. One file per machine (see appendix). Figure 11 The system displays the device's lifecycle status, maintenance records, and monitoring data. The panel rendering of the five profile pages is automatically assembled using a schema-driven mechanism, validating the effectiveness of the declarative traceability specification's TraceSpec registration and the panel's three-level fallback strategy.

[0051] Verification Case 5: Overall System Operation Verification. Based on the above four subsystems, the method of this invention was deployed as a railway artificial intelligence data platform and loaded and run in a browser. System homepage (see appendix) Figure 5 The dashboard displays an overall operational overview, including real-time data cards for 16 core operational metrics, a cost structure pie chart, a monthly profit trend chart, and a budget execution progress bar. All metric data is calculated in real-time from the browser-based SQLite database using a declarative drill-down query subsystem. (See attached holographic archive page.) Figure 6 The page displays category cards for eight business domains and a general search function, verifying the correctness of the domain classification results from the ontology data construction subsystem. (See attached image for the Smart Traceability page.) Figure 12 To be continued Figure 18 The entire interactive process, from indicator selection, drill-down panel pop-up, multi-level drill-down, breadcrumb navigation to keyboard return, was fully demonstrated, verifying the technical effectiveness of the three subsystems working together.

[0052] Example 2: To achieve the above objective, such as Figure 2 As shown, based on Embodiment 1, this invention discloses a data end-to-end traceability system for a railway artificial intelligence data platform, comprising: The primary processing module 11 is used to acquire railway database design documents and table structure information, parse the railway database design documents and table structure information to obtain structured mapping data, and automatically classify the structured mapping data into business domains based on preset weighted keyword scores to obtain business domain classification results. Secondary processing module 12 is used to generate ontology metadata from the business domain classification results, automatically generate entity relationship graphs based on ontology metadata and predefined relationship type tables, and trace the ontology metadata in the entity relationship graphs based on preset traceability specifications to obtain first-level traceability items. The traceability output module 13 is used to obtain the coded drill-down knowledge, perform drill-down queries based on the first-level traceability items and the coded drill-down knowledge to obtain the next-level traceability data, and perform full-link traceability of data based on the next-level traceability data to determine the root cause of the problem.

[0053] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0054] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0055] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0056] The foregoing has shown and described the basic principles, main features, and advantages of this disclosure. Those skilled in the art should understand that this disclosure is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this disclosure. Various changes and modifications can be made to this disclosure without departing from its spirit and scope, and all such changes and modifications fall within the scope of this disclosure as claimed.

Claims

1. A method for full-link data traceability of a railway artificial intelligence data platform, characterized in that, The method includes the following steps: Obtain railway database design documents and table structure information, parse the railway database design documents and table structure information to obtain structured mapping data, and automatically classify the structured mapping data into business domains based on preset weighted keyword scores to obtain business domain classification results; Generate ontology metadata from the business domain classification results. Based on the ontology metadata and the predefined relationship type table, automatically generate an entity relationship graph. Based on the preset traceability specifications, trace the ontology metadata in the entity relationship graph to obtain the first-level traceability item. Obtain the drill-down knowledge of the code, perform drill-down queries based on the first-level traceability item and the drill-down knowledge of the code to obtain the next-level traceability data, and perform full-link traceability of the data based on the next-level traceability data to determine the root cause of the problem.

2. The data full-link traceability method for a railway artificial intelligence data platform according to claim 1, characterized in that, The process of automatically classifying structured mapping data into business domains based on preset weighted keyword scores is as follows: When the preset weighted keyword scoring includes matching table names with domain keywords and matching field text with domain keywords, the weight of the table name is greater than the weight of the text. The business domain classification results include: locomotive domain, personnel domain, equipment domain, materials domain, engineering project domain, safety domain, operation domain, production domain, performance domain, and system management domain.

3. The data end-to-end traceability method for a railway artificial intelligence data platform according to claim 1, characterized in that, The ontology metadata defines the types of entities, attribute structures, and classification systems managed by the platform. The types of entities include locomotives, equipment, personnel, engineering projects, and materials.

4. The data full-link traceability method for a railway artificial intelligence data platform according to claim 1, characterized in that, The predefined relationship type table includes predefined graphical relationships and ontology relationships.

5. The data full-link traceability method for a railway artificial intelligence data platform according to claim 1, characterized in that, The entity relationship graph includes entity nodes and relationship edges. The construction process includes: personnel node construction stage, locomotive node and driving relationship construction stage, material node construction stage, engineering project node and related relationship construction stage, maintenance record node and maintenance relationship construction stage, scheduling relationship construction stage, contract node construction stage, supplier node construction stage, department node and subordinate relationship construction stage, safety event node construction stage, equipment node construction stage, and cross-domain influence chain construction stage.

6. The data full-link traceability method for a railway artificial intelligence data platform according to claim 5, characterized in that, The cross-domain influence chain construction phase is based on preset rules, as follows: The system detects the inventory status field of a material entity. When the inventory status reaches a preset danger threshold, it automatically creates a material shortage impact relationship edge from the material entity to the relevant maintenance record entity.

7. The data full-link traceability method for a railway artificial intelligence data platform according to claim 1, characterized in that, The drill-down query is based on the first-level traceability item and the encoded drill-down knowledge. After each drill-down, the set of consumed dimensions is maintained and the currently used dimension is added to the consumed set. When all dimensions have been consumed, the drill-down target automatically switches to a flat list mode.

8. The data full-link traceability method for a railway artificial intelligence data platform according to claim 1, characterized in that, The process of performing end-to-end data tracing based on the next-level tracing data is as follows: By managing metrics through the source trace stack and entity jump stack, we drill down to the hierarchical status and entity file page jump status, and return the source trace data to the front-end rendering, update the stack status and return the corresponding data to locate the root cause of the problem.

9. A data end-to-end traceability system for a railway artificial intelligence data platform, employing the data end-to-end traceability method for a railway artificial intelligence data platform as described in any one of claims 1 to 8, characterized in that, include: The primary processing module is used to acquire railway database design documents and table structure information, parse the railway database design documents and table structure information to obtain structured mapping data, and automatically classify the structured mapping data into business domains based on preset weighted keyword scores to obtain business domain classification results. The secondary processing module is used to generate ontology metadata from the business domain classification results. Based on the ontology metadata and the predefined relationship type table, it automatically generates an entity relationship graph. Based on the preset traceability specifications, it traces the ontology metadata in the entity relationship graph to obtain the first-level traceability items. The source tracing output module is used to obtain the coded drill-down knowledge, perform drill-down queries based on the first-level source tracing items and the coded drill-down knowledge to obtain the next-level source tracing data, and perform full-link data tracing based on the next-level source tracing data to determine the root cause of the problem.

10. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, The memory stores a computer program that can run on the processor. When the processor loads and executes the computer program, it employs a data full-link traceability method for a railway artificial intelligence data platform as described in any one of claims 1 to 8.

Images