A data processing method, apparatus, device, medium, and product

Abstract

This application provides a data processing method, apparatus, device, medium, and product, relating to the field of data processing. The specific technical solution involves: acquiring metadata information of an aircraft in a target flight dimension, prompt word templates for the aircraft, and fusion information templates for the aircraft from multiple heterogeneous data sources. The templates containing metadata information are input into the processing model corresponding to the aircraft in the target flight dimension, outputting processing text for each target parameter. Based on the fusion information templates, the target parameters, processing texts, and corresponding visualization information are integrated. This method has two advantages: First, the hierarchical relationships in the templates are fixed, providing clear analysis instructions for the subsequent model and avoiding broad, unfounded generation tasks, thus improving the accuracy of data processing. Second, the use of a combination of graphics, text, and data in the presentation enhances the user experience.

Description

Technical Field

[0001] This application relates to the field of data processing, and more particularly to a data processing method, apparatus, device, medium, and product. Background Technology

[0002] As a technology-intensive industry, the aviation sector's operational decisions rely on accurate data analysis reports. For example, operational decisions regarding aircraft depend on the accuracy of the data analysis information available for the aircraft.

[0003] In traditional technologies, aircraft data analysis relies on human experience, which limits the accuracy and efficiency of aircraft data analysis. Summary of the Invention

[0004] This application provides a data processing method, apparatus, device, medium, and product that can improve the accuracy and efficiency of data processing.

[0005] Firstly, this application provides a data processing method, which includes: acquiring metadata information of an aircraft in a target flight dimension, a prompt word template for the aircraft, and a fusion information template for the aircraft from multiple heterogeneous data sources. The target flight dimension includes a flight operation status dimension, a route network analysis dimension, or a flight resource allocation dimension; the metadata information reflects the aircraft's operation status in the target flight dimension; the prompt word template includes multiple semantic tags with preset structures, which are tags written using natural language. By acquiring metadata information from multiple data sources and automatically integrating the information, data silos are broken down, information unification is achieved, and a data foundation is provided for subsequent comprehensive analysis. Using preset prompt word templates and fusion information templates ensures that the generated visualized status information conforms to aviation professional standards in terms of framework, terminology, and depth, avoiding inconsistent quality and formatting caused by analysts' discretionary decisions.

[0006] This preprocessing method, based on metadata type matching, parameterizes multiple metadata elements to obtain multiple target parameters, which are mapped to semantic tags. This transforms the original descriptive metadata into target parameters with clear business definitions, achieving data-to-knowledge conversion. This benefits subsequent large-scale models in understanding knowledge and improves the accuracy of their output. Furthermore, this metadata type matching preprocessing method enhances data quality and usability, further improving the quality of subsequent model outputs.

[0007] Based on the mapping relationship and each target parameter, the semantic tags in the prompt word template are filled to obtain the filled template; the filled template includes multiple target parameters with a preset structure. Specific target parameters are then filled into the prompt word template to obtain the filled template. Because the contextual relationships in the filled template are fixed, it provides clear analysis instructions for the subsequent large model, avoiding the model from performing broad and unfounded generation tasks and improving the accuracy of data processing.

[0008] The process involves inputting a template into the processing model corresponding to the target flight dimension of the aircraft, and outputting processing text for each target parameter. The processing text includes at least one of the following: content generation text for the target parameter, data analysis text, or verification and review text. By inputting the template into the corresponding processing model and outputting the processing text for the target parameters, the process replaces a large amount of repetitive work that operators would otherwise have to manually consult manuals, compare standards, and write reports. This automates the analysis process and improves the standardization and efficiency of data processing.

[0009] Based on a fusion information template, this system integrates various target parameters, processed texts, and corresponding visualization information to obtain and display visualized status information of the aircraft in the target flight dimension. The visualized status information includes at least one of the following: flight operation status visualization, route network analysis visualization, or flight resource allocation visualization. By integrating various target parameters, processed texts, and their corresponding visualization information, this system combines data, insight, and visual perspectives to obtain visualized status information. Presenting this information using a combination of graphics, text, and data helps reduce the cognitive load on operators and improves their user experience.

[0010] In one possible implementation of the first aspect, a preprocessing method based on metadata type matching involves parameterizing multiple metadata entries to obtain multiple target parameters. This includes: sequentially performing syntax-level, logic-level, and association-level validations on the multiple metadata entries, and using the data that passes these validations as the target parameters. The metadata type matching preprocessing method includes syntax-level, logic-level, and association-level validations. Specifically, the syntax-level validation targets the format and structure of the metadata; the logic-level validation targets the inherent business logic and relationships within the metadata; and the association-level validation targets the metadata, related data, and the consistency between the metadata and related data.

[0011] This solution employs three main methods: First, it performs syntax-level validation on the metadata, filtering out erroneous data and ensuring that fields conform to format requirements, thus providing a data foundation for subsequent processing. Second, it performs logic-level validation on the metadata, ensuring the correctness of business logic and improving the accuracy of subsequent data processing. Third, it performs association-level validation on the metadata, enabling cross-system data comparison. This cross-system comparison identifies inconsistencies between systems, thereby improving the accuracy of data processing.

[0012] In one possible implementation of the first aspect, the method further includes: obtaining raw data of the aircraft in the target flight dimension from a heterogeneous data source, wherein the raw data is unstructured; and extracting parameters from the raw data based on a parameter extraction method configured for the target flight dimension to obtain the target parameters.

[0013] This solution has two main aspects. First, it obtains raw data from heterogeneous data sources. Unstructured data contains rich details, preserving the potential for in-depth analysis. Second, it transforms the raw data into target parameters through parameter extraction, realizing the transformation from data to knowledge. Furthermore, by transforming the data to obtain target parameters, it reduces the data dimensionality, which helps improve the efficiency of subsequent data processing.

[0014] In one possible implementation of the first aspect, the semantic tags include semantic placeholders and preset placeholder tags; the target parameters include text type parameters and rich media type parameters; based on the mapping relationship and each target parameter, the semantic tags in the prompt word template are filled to obtain a filled template, specifically including: based on the mapping relationship and text type parameters, the semantic placeholders in the prompt word template are filled to obtain a first template; based on the mapping relationship and rich media type parameters, the preset placeholder tags in the prompt word template are filled to obtain a second template; the first template and the second template are merged to obtain the filled template.

[0015] Firstly, this solution separates the text and rich media parameter processing logic, decoupling the processing logic. Text processing and visualization processing can use different technology stacks and algorithms, while also enabling parallel processing of text and rich media parameters, thus improving data processing efficiency. Secondly, the first template reflects the text filling result, and the second template reflects the rich media filling result. Combining the first and second templates yields the filling template, providing a data foundation for the organic integration of subsequent text analysis and visualization.

[0016] In one possible implementation of the first aspect, a fill template is input into a processing model corresponding to the aircraft for the target flight dimension, and the processing text for each target parameter is output. Specifically, this includes: obtaining the analysis type of the aircraft for the target flight dimension, and a processing model matching the analysis type. The analysis type includes at least one of content generation type, data analysis type, or verification and review type. The fill template is then input into the processing model, and the processing text for each target parameter is output.

[0017] This solution has three main advantages: First, it acquires the processing type that matches the analysis type, identifies the analysis type, and matches the processing model accordingly, achieving precise matching of task models. Second, different processing models have different processing preferences, which can meet different data processing needs, thereby improving the overall accuracy of data processing. Third, it provides multiple processing models for operators to select and configure, improving the user experience.

[0018] In another possible implementation of the first aspect, a filling template is input into a processing model, and the output is processing text for each target parameter. This includes: obtaining a template analysis scenario configured for the filling template, and template samples matching the template analysis scenario. The template analysis scenario includes aircraft technology scenarios, anomaly handling scenarios, or trend determination scenarios. Among the target parameters in the filling template, the target parameters matching the template analysis scenario are retained as reserved parameters. Multiple sample parameters with template structures in the template samples are integrated with the retained parameters with preset structures to obtain an adjusted template. The adjusted template is input into the processing model, and the output is processing text for each target parameter.

[0019] This solution has three main aspects: First, it automatically focuses on the most relevant parameters to retain based on different scenarios, simplifying the target parameters to ensure data processing accuracy while improving efficiency. Second, it matches and obtains template samples, providing a professional template structure for subsequent data processing, which helps improve accuracy. Third, it adjusts the template to include both professional structure and specific data, ensuring the quality of the processed text output by the subsequent processing model.

[0020] Secondly, a data processing device is provided, which includes: an acquisition module for acquiring metadata information of an aircraft in the target flight dimension, prompt word templates for the aircraft, and fusion information templates for the aircraft from multiple heterogeneous data sources; The target flight dimension includes flight operation status dimension, route network analysis dimension, or flight resource allocation dimension; metadata information reflects the aircraft's operation in the target flight dimension; the prompt word template includes multiple semantic tags with preset structures, which are tags written in natural language; The parameterization processing module is used for preprocessing based on information type matching of metadata information. It performs parameterization processing on multiple metadata in the metadata information to obtain multiple target parameters. The target parameters have a mapping relationship with semantic tags. The template filling module is used to fill the semantic tags in the prompt word template based on the mapping relationship and each target parameter to obtain the filled template; the filled template includes multiple target parameters with preset structures; The model processing module is used to input the filling template into the processing model corresponding to the aircraft for the target flight dimension, and output the processing text for each target parameter; the processing text includes at least one of the following: content generation text for the target parameter, data analysis text, or verification and review text; The information integration module is used to integrate various target parameters, various processed texts, and the visualization information corresponding to each target parameter based on the fusion information template, so as to obtain and display the visualization status information of the aircraft in the target flight dimension; the visualization status information includes at least one of flight operation status visualization information, route network analysis visualization information, or flight resource configuration visualization information.

[0021] Thirdly, embodiments of this application provide a data processing apparatus, the method comprising: a memory and at least one processor. The memory is communicatively connected to the processor. The memory is used to store computer program code, the computer program code including computer instructions. When the processor executes the computer instructions, it causes the electronic device to perform the method as described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer instructions. When these computer instructions are executed by a processor, they are used to implement the method as described in the first aspect and any possible implementation thereof.

[0023] Fifthly, embodiments of this application provide a computer program product that, when run on a computer or executed by the computer's processor, implements the method described in the first aspect and any possible design thereof. The computer may be the data processing device described in the third aspect and any possible implementation thereof.

[0024] Understandably, the beneficial effects achieved by the data processing apparatus of the second aspect, the data processing device of the third aspect, the computer-readable storage medium of the fourth aspect, and the computer program product of the fifth aspect provided above can be referred to as the beneficial effects of the first aspect and any possible implementation thereof, which will not be repeated here. Attached Figure Description

[0025] Figure 1This is a schematic diagram of the structure of a data processing system provided in an embodiment of this application; Figure 2 A flowchart illustrating a data processing method provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a data processing device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a data processing device provided in an embodiment of this application. Detailed Implementation

[0026] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this embodiment, unless otherwise stated, "a plurality of" means two or more.

[0027] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0028] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0029] As a technology-intensive industry, the aviation sector's operational decisions rely on accurate data analysis reports. For example, operational decisions regarding aircraft depend on the accuracy of the data analysis information available for the aircraft.

[0030] In traditional technologies, aircraft data analysis information relies on human experience to compile. When humans lack experience in compiling such information, the accuracy of the aircraft data analysis information is relatively low.

[0031] For example, the current methods of generating aircraft data analysis information have the following problems: First, the traditional data collection-table-experience analysis model suffers from problems such as time-consuming data extraction, fixed analysis dimensions, and long update cycles, making it difficult to meet the requirements of real-time data processing. Studies have shown that, under this traditional model, the average analysis cycle for aircraft data is as long as 3-5 working days, with data extraction accounting for more than 60% of the total time.

[0032] Second, the data analysis results from aircraft are presented in a simplistic format, lacking visual representation. Studies have shown that purely tabular reports are more than 40% less efficient at extracting user information than visual reports.

[0033] Third, the data analysis of aircraft is limited by the professional level of the analysts, and there is a risk of subjective bias. Traditional analysis methods struggle to achieve deep cross-dimensional correlation mining, and the consistency of analytical conclusions is difficult to guarantee, resulting in low accuracy. Statistics show that the difference in analytical conclusions among different analysts on the same dataset can reach as high as 35%.

[0034] Fourth, due to the distinct seasonality of the aviation industry and the poor scalability of existing systems, adding new data analysis report types requires full-code development, making it impossible to quickly respond to changes in business needs.

[0035] In this embodiment, by inputting a fill template into the corresponding processing model and outputting the processing text corresponding to the target parameters, a large amount of repetitive work that operators manually do by consulting manuals, comparing standards, and writing reports is replaced, thus automating the analysis process and improving data processing efficiency and accuracy. Based on this, by integrating each target parameter, each processing text, and the corresponding visualization information, multi-faceted information from data, insight, and visual perspectives is obtained to obtain visualized status information. This information is displayed using a combination of graphics, text, and data, which helps reduce the cognitive load on operators regarding professional information and improves the user experience.

[0036] Specifically, in this embodiment of the application, regarding the first point mentioned above, information integration is automatically completed by obtaining metadata information from multiple data sources, eliminating the need for manual data collection, reducing the time consumed by data collection, and thus improving data processing efficiency.

[0037] In this embodiment of the application, regarding the second point mentioned above, each target parameter, each processed text, and the corresponding visualization information of each target parameter are integrated. This combines information from multiple perspectives, including data, insight, and visual angles, to obtain visualization status information. By displaying this information using a combination of graphics, text, and data, it helps reduce the cognitive load of operators on professional information and improves the user experience for operators.

[0038] In this embodiment, regarding the third point mentioned above, the target parameters in the filling template are processed using the processing model corresponding to the aircraft to obtain the processed text of the target parameters. This method does not rely on human experience, ensuring the objectivity of the analysis. Since the model parameters in the processing model are fixed, the analysis results for the same dataset can also remain consistent. Furthermore, since the processed text includes at least one of content generation text, data analysis text, or verification and review text, multi-dimensional text is obtained. Moreover, by integrating each target parameter, each processed text, and the corresponding visualization information of each target parameter, multi-angle information from data, insight, and visual perspectives is combined to obtain visualized status information, achieving cross-dimensional analysis and improving the accuracy of data processing.

[0039] In this embodiment of the application, regarding the fourth point mentioned above, by configuring the prompt word template and the integrated information template for flight respectively, the configured template can adapt to business needs. At the same time, by configuring the template, there is no need to carry out full code development, which saves code development time and improves the response speed to business needs.

[0040] The data processing method provided in this application embodiment can be applied to, for example... Figure 1 The data processing system shown includes a data processing device 101, a model interface 102, and a data storage system 103.

[0041] The data processing device 101 is used to acquire metadata information of the aircraft in the target flight dimension, prompt word templates for the aircraft, and fused information templates for the aircraft from multiple heterogeneous data sources. The target flight dimension includes flight operation status dimension, route network analysis dimension, or flight resource allocation dimension; the metadata information reflects the aircraft's operation in the target flight dimension; the prompt word template includes multiple semantic tags with preset structures, which are tags written using natural language.

[0042] The data processing device 101 is used for a preprocessing method based on information type matching of metadata information. It performs parameterization processing on multiple metadata in the metadata information to obtain multiple target parameters. The target parameters have a mapping relationship with semantic tags.

[0043] The data processing device 101 is used to fill in the semantic tags in the prompt word template based on the mapping relationship and each target parameter to obtain a filled template; the filled template includes multiple target parameters with a preset structure; The data processing device 101 is used to input the filling template into the processing model corresponding to the aircraft for the target flight dimension through the model interface 102, and output the processing text for each target parameter; the processing text includes at least one of the following: content generation text for the target parameter, data analysis text, or verification and audit text.

[0044] The data processing device 101 is used to integrate various target parameters, various processed texts, and the visualization information corresponding to each target parameter based on the fusion information template, to obtain and display the visualization status information of the aircraft in the target flight dimension; the visualization status information includes at least one of flight operation status visualization information, route network analysis visualization information, or flight resource configuration visualization information.

[0045] The data processing device 101 can be a server, which can be implemented using a standalone server or a server cluster composed of multiple servers.

[0046] Model interface 102 can be an interface used to call and process the model.

[0047] The data storage system 103 can be a processor or server with data storage capabilities. It can be a single server or a server cluster consisting of multiple servers. The data storage system 103 can be used to store metadata information of the aircraft in the target flight dimension, aircraft cue word templates, and aircraft fusion information templates, etc.

[0048] In one embodiment, such as Figure 2 As shown, a data processing method is provided that can be applied to... Figure 1 The data processing device 101 in the middle, or applied to a cloud computing platform, edge computing device, chip, or device with computing capabilities, etc. This application embodiment does not limit the specific form of the device executing the method; taking the method applied to a server as an example, the method specifically includes: S201, obtains metadata information of the aircraft in the target flight dimension, prompt word templates for the aircraft, and fusion information templates for the aircraft from multiple heterogeneous data sources.

[0049] Metadata information of the aircraft is obtained from multiple heterogeneous data sources in the dimensions of flight operation status, route network analysis, or flight resource allocation. Prompt word templates corresponding to multiple semantic tags with preset structures are obtained for the aircraft. Fusion information templates for the aircraft are also obtained to facilitate subsequent preprocessing of metadata information.

[0050] Heterogeneous data sources can be a collection of multiple data sources with different technical formats, protocol standards, storage methods, and access interfaces. Each heterogeneous data source is generated and stored independently.

[0051] An aircraft can refer to any device that flies within or outside the atmosphere. Aircraft can include flying vehicles such as aircraft, spacecraft, and rockets. Aircraft can include fixed-wing aircraft, helicopters, etc.

[0052] The target flight dimension can represent a specific framework or perspective for analyzing and evaluating aircraft operations from different business viewpoints and management levels. Each dimension focuses on a different set of issues.

[0053] The target flight dimension can include the flight operation status dimension, the route network analysis dimension, or the flight resource allocation dimension.

[0054] Among them, the flight operation status dimension can focus on the real-time or near-real-time operation status, efficiency, and safety level of flights within a specific time window.

[0055] The dimensions of route network analysis can include the structural characteristics and operational efficiency of the network as a whole composed of multiple routes.

[0056] Flight resource allocation can focus on the allocation, utilization, and optimization of core resources such as fleet, crew, and flight slots.

[0057] Metadata information can represent structured summary information extracted from heterogeneous data sources, describing the operational characteristics and status of an aircraft in a specific dimension. It can be understood that metadata information can be the raw data of the aircraft after preliminary screening and standardization.

[0058] The cue word template can be a predefined, semantically tagged natural language analysis framework for a specific aircraft type and target flight dimension. Firstly, the semantic tags in the cue word template can represent data injection points. Secondly, the cue word template includes a professional analysis framework and evaluation criteria. Thirdly, the same cue word template is applicable to the analysis of similar flights. Fourthly, all content in the cue word template is configurable, allowing for adjustments to the template complexity based on the analysis depth.

[0059] The fused information template can be a natural language analysis framework predefined for specific aircraft types and target flight dimensions, featuring parameter areas, text areas, and visual information areas. Each area is configurable, allowing for corresponding area configurations to adapt to changing business needs.

[0060] Semantic tags can be parameterized placeholders within a prompt word template. Written in natural language, semantic tags improve readability and provide accurate references for operators.

[0061] Multiple semantic tags with predefined structures can represent a certain structural relationship between them. For example, semantic tag A and semantic tag B belong to the same chapter structure. Another example is that semantic tag C belongs to a parent chapter structure, and semantic tag D belongs to a child chapter structure of that parent chapter structure.

[0062] For example, the server can obtain metadata information of the aircraft in the flight operation status dimension (summary data such as time, fuel consumption, and altitude profile of all recent flights on the route) from heterogeneous data sources, as well as prompt word templates for the aircraft (wide-body aircraft route performance evaluation templates). Based on this, a series of subsequent processing can be performed based on the metadata information and prompt word templates to obtain the visualized status information of the aircraft in the flight operation status dimension (on-time performance trend chart, fuel consumption distribution box plot, altitude profile comparison chart, and relevant analysis text generated by the large model).

[0063] For example, the server can define the preset structure corresponding to the semantic tags through a declarative template language, and define the structural hierarchy (parent level, child level, etc.) of each semantic tag through the declarative template language.

[0064] For example, obtaining metadata information about the aircraft in the target flight dimension from multiple heterogeneous data sources can specifically include: using an adapter pattern to uniformly access multiple heterogeneous data sources, supporting various data source types such as relational databases, time-series databases, and API interfaces, with data source configurations supporting hot updates. After accessing multiple heterogeneous data sources, the metadata information about the aircraft in the target flight dimension can be obtained from these multiple heterogeneous data sources.

[0065] In this embodiment, firstly, by acquiring metadata information of the aircraft in the target flight dimension from multiple heterogeneous data sources, information integration is automatically completed, breaking down data silos and achieving information unification, thus providing a data foundation for subsequent comprehensive analysis. Secondly, by using preset prompt word templates and fused information templates, it is ensured that the generated visualized status information conforms to aviation professional standards in terms of framework, terminology, and depth, avoiding inconsistent quality and formatting caused by analysts' free interpretation.

[0066] S202, a preprocessing method based on information type matching of metadata information, performs parameterization processing on multiple metadata in the metadata information to obtain multiple target parameters.

[0067] After obtaining the metadata information, the preprocessing method based on the information type matching of the metadata information is used to parameterize multiple metadata in the metadata information to obtain multiple target parameters. There is a mapping relationship between the target parameters and semantic tags, so that the target parameters can be filled in later according to the mapping relationship.

[0068] In some embodiments, the preprocessing method may be a series of data cleaning, transformation, calculation and verification operations that are automatically selected and executed based on the source, format, quality characteristics and business purpose of the metadata.

[0069] Metadata can be raw, descriptive information.

[0070] Information type can represent the type of metadata information. For example, metadata information can specifically be storage metadata, management metadata, descriptive metadata, etc.

[0071] Target parameters can be standardized, quantified feature variables obtained after preprocessing and directly used for business analysis and model input. Target parameters have clear business definitions; for example, a target parameter might be takeoff distance, not the ACARS field F012. Target parameters have data or categorical values, such as 15 seconds or mild turbulence. Target parameters have complete context, including necessary metadata (units, time range, calculation method, etc.). The quality of target parameters is traceable; for example, target parameters can include quality scores.

[0072] The mapping relationship can represent the one-to-one correspondence between the target parameter and the semantic tags in the prompt word template.

[0073] In one possible implementation, a preprocessing method based on information type matching of metadata information is used to parameterize multiple metadata elements to obtain multiple target parameters. Specifically, this may include: constructing a parameter computation pipeline based on a distributed computing framework, which supports priority scheduling and resource isolation. The server then uses the computation pipeline to parameterize multiple metadata elements based on information type matching of metadata information to obtain multiple target parameters.

[0074] After obtaining the metadata information, in order to improve the data quality of the metadata information, in one possible implementation, S202 above, based on the information type matching preprocessing method of the metadata information, parameterizes multiple metadata in the metadata information to obtain multiple target parameters, including: performing syntax layer verification, logic layer verification, and association layer verification on multiple metadata in the metadata information in sequence, and using the data that passes the syntax layer verification, logic layer verification, and association layer verification as target parameters; the information type matching preprocessing method based on the metadata information includes syntax layer verification, logic layer verification, and association layer verification. Specifically, the verification object of the syntax layer verification is the format and structure of the metadata; the verification object of the logic layer verification is the internal business logic and relationships of the metadata; and the verification object of the association layer verification is the metadata and related data that are associated with the metadata, as well as the consistency between the metadata and related data.

[0075] Syntax layer validation involves performing technical, superficial verification on metadata regarding its basic format, structure, type, and value range. It ensures the data is syntactically correct. The object of syntax layer validation is the format and structure of the metadata. For example, syntax layer validation might verify that a height value is less than 40,000 feet.

[0076] Logical layer validation verifies the consistency, rationality, and compliance of business logic within and between metadata. It ensures that the data is logically correct at the business semantic level. The validation object of logical layer validation can be the inherent business logic and relationships within the metadata. For example, logical layer validation could verify that the actual takeoff time must be later than the pushback time and earlier than the landing time.

[0077] Correlation layer verification involves cross-referencing current metadata with relevant external data sources, historical data, and benchmark data to verify their consistency and consistency. This ensures that data is mutually corroborative within a global context. The verification objects of correlation layer verification can be metadata, related data that are associated with the metadata, and the consistency between metadata and related data. For example, correlation layer verification can verify the difference between the Quick Access Recorder (QAR) (metadata) and radar monitoring speed (related data) (the difference should be less than 10 sections).

[0078] Related data refers to other datasets that have business, spatiotemporal, or logical relationships with the metadata currently being validated. These datasets provide the context and benchmark for validation.

[0079] Specifically, after obtaining the metadata information, the metadata information may include at least one of the following: stored metadata, administrative metadata, or descriptive metadata. For stored metadata, since stored metadata changes infrequently, multiple metadata items within the stored metadata need to be validated at the syntax layer, logic layer, and association layer, respectively. The data that passes the syntax layer, logic layer, and association layer validations is used as the target parameter.

[0080] Specifically, based on the syntax layer validation, logic layer validation, and association layer validation of the metadata information, if the data in the metadata information fails the aforementioned validation, the data that fails the aforementioned validation can be classified according to the type of validation that failed and the corresponding validation question. The classification of the questions is then sent to the operators for manual review.

[0081] For management metadata, which changes frequently, in order to preserve data processing efficiency, we can perform syntax-level validation on multiple metadata items in the management metadata. Once the system calculates that the remaining resources are greater than the resource threshold, we can then perform logic-level validation and association-level validation on the multiple metadata items in the management metadata. The data that passes the syntax-level validation, logic-level validation, and association-level validation will be used as the target parameters.

[0082] For descriptive metadata, which helps users understand, find and utilize data content, multiple metadata in the stored metadata can be validated at the syntax layer, logic layer and association layer respectively during the data processing stage. The data that passes the syntax layer validation, logic layer validation and association layer validation is used as the target parameter.

[0083] In this embodiment, firstly, syntax-level validation is performed on the metadata in the metadata information to filter out metadata with incorrect formats and ensure that the fields of the metadata meet the formal requirements, providing a data foundation for subsequent processing. Secondly, logic-level validation is performed on the metadata in the metadata information to ensure the correctness of the business logic, which helps to improve the accuracy of subsequent data processing. Thirdly, association-level validation is performed on the metadata in the metadata information to realize cross-system data comparison. Through cross-system comparison, inconsistencies between systems can be found, thereby improving the accuracy of data processing.

[0084] Heterogeneous data sources contain not only structured metadata but also unstructured data. To improve the comprehensiveness of subsequent data processing, one possible implementation includes: obtaining the raw data of the aircraft in the target flight dimension from the heterogeneous data source; this raw data is unstructured. Based on the parameter extraction method configured for the target flight dimension, parameters are extracted from the raw data to obtain the target parameters.

[0085] Raw data can be unstructured flight data collected directly from heterogeneous data sources without standardization. For example, raw data could include weather reports, NOTAMs, radar data messages, etc.

[0086] The parameter extraction method can be a systematic approach designed and implemented to extract, transform, and construct effective features from raw data, targeting the target flight dimension.

[0087] Specifically, the server retrieves raw aircraft data for the target flight dimension from heterogeneous data sources, aggregating multi-source data to avoid the incompleteness issues of a single data source. It also preserves the original information; unstructured data contains rich details, allowing for potential in-depth analysis later. For example, non-standard descriptions in the crew's free-text report may contain important safety clues. After obtaining the raw data, the server can extract parameters based on the parameter extraction method configured for the target flight dimension, obtaining the target parameters. This achieves the transformation from data to knowledge, that is, converting raw data into features with business meaning.

[0088] In this embodiment, firstly, raw data is obtained from heterogeneous data sources. Unstructured data contains rich details, preserving the possibility for in-depth analysis. Secondly, the raw data is transformed into target parameters through parameter extraction, realizing the transformation from data to knowledge. Furthermore, transforming the data to obtain target parameters reduces the data dimensionality, which helps improve the efficiency of subsequent data processing.

[0089] In this embodiment, firstly, the original descriptive metadata information is transformed into target parameters with clear business definitions, realizing the transformation from data to knowledge. This is beneficial for subsequent large-scale models to understand knowledge and improve the accuracy of the output of large-scale models. Secondly, based on the information type matching preprocessing method of metadata information, multiple metadata in the metadata information are parameterized, improving data quality and usability, and further improving the data quality of subsequent model output.

[0090] S203. Based on the mapping relationship and each target parameter, the semantic tags in the prompt word template are filled to obtain the filled template.

[0091] After obtaining the mapping relationship and each target parameter, the target parameters can be filled into the semantic tags corresponding to the target parameters in the prompt word template based on the mapping relationship and each target parameter, resulting in multiple target parameters with preset structures.

[0092] In some embodiments, the filling template can be a complete set of prompts that is data-complete and can be directly input into a large model, obtained by replacing the target parameters with the semantic labels of the prompt word template according to the mapping relationship. The filling template can be a template with data completeness; the template obtained after replacing all the semantic labels in the prompt word template can be considered a filling template. The filling template can also have grammatical correctness, ensuring the fluency of natural language in the resulting filling template after the prompt word template is filled. The filling template has complete context, ensuring the necessary analytical background information in the resulting filling template after the prompt word template is filled. This improves the readability of the filling template and helps improve the user experience for operators.

[0093] It is understandable that the template includes multiple target parameters with preset structures, which can be interpreted as analysis prompts for the subsequent processing model. The processing model can then process these prompts and output processed text tailored to the target parameters.

[0094] The preset structure can represent the hierarchical relationship between target parameters. For example, target parameter 1 is the parent level of target parameter 2. Target parameter 3 and target parameter 4 belong to the same level.

[0095] Semantic tags include semantic placeholders and preset placeholder tags. Target parameters include text type parameters and rich media type parameters. To fill in different target parameters, in one possible implementation, step S203 above, based on the mapping relationship and each target parameter, fills in the semantic tags in the prompt word template to obtain a filled template. Specifically, this includes: filling in the semantic placeholders in the prompt word template based on the mapping relationship and text type parameters to obtain a first template; filling in the preset placeholder tags in the prompt word template based on the mapping relationship and rich media type parameters to obtain a second template; and merging the first and second templates to obtain the filled template.

[0096] Semantic placeholders can be variable markers used in prompt word templates to embed text type analysis content, employing semantic naming in natural language. For example, semantic placeholders could be "maximum altitude of a certain aircraft," "number of flight cycles of a certain aircraft," or "weather impact analysis." The names of semantic placeholders themselves express the business meaning, and compared to machine language or programming languages, they are more readable, facilitating real-time configuration by operators according to business needs.

[0097] Preset placeholder labels can be special markers in the prompt template used to reference rich media visualizations, pointing to the configured location of specific charts, images, or other non-text elements.

[0098] Text parameters are processed using parameter extraction methods to obtain structured data parameters that can be directly used for text analysis. These parameters include numerical results, classification labels, and evaluation conclusions.

[0099] Rich media parameters contain configuration information for visualization elements, describing how to generate or reference non-text content such as charts, images, and maps.

[0100] The first template fills the semantic placeholders of the prompt word template with text type parameters, resulting in a pure text analysis framework that contains all text analysis content but has not yet been associated with visualization elements.

[0101] The second template converts rich media type parameters into visual element references or generation instructions, resulting in a visual content configuration framework that indicates where and what kind of visual content should be placed in the report.

[0102] Specifically, after obtaining multiple target parameters, the server can categorize them into text-type parameters and rich media-type parameters. The server can also categorize multiple semantic tags in the prompt word template into semantic placeholders and preset placeholder tags. Based on the mapping relationship and text-type parameters, the server fills in the semantic placeholders in the prompt word template to obtain the first template. Based on the mapping relationship and rich media-type parameters, the server fills in the preset placeholder tags in the prompt word template to obtain the second template. It can be understood that the above two filling processes can be processed in parallel, thereby improving filling efficiency. The first and second templates are merged to obtain the filled template.

[0103] In this embodiment, firstly, the text and rich media parameter processing logics are separated, decoupling the processing logic. Text processing and visualization processing can use different technology stacks and algorithms, while also enabling subsequent parallel processing of text and rich media parameters, thus improving data processing efficiency. Secondly, the first template reflects the text filling result, and the second template reflects the rich media filling result. Combining the first and second templates yields a filling template, providing a data foundation for the organic integration of subsequent text analysis and visualization.

[0104] In this embodiment, firstly, specific target parameters are filled into the prompt word template to obtain a filled template. Since the contextual relationships in the filled template are fixed, it provides clear analysis instructions for subsequent large-scale models, avoiding the model from performing broad and unfounded generation tasks. Secondly, operators do not need to write complex prompt words from scratch for each analysis task; they only need to maintain the mapping relationship between parameters and semantic tags. Based on the mapping relationship and each target parameter, the filled template is automatically generated, saving prompt word writing time and improving data processing efficiency and accuracy.

[0105] S204 inputs the filling template into the processing model corresponding to the aircraft for the target flight dimension, and outputs the processing text for each target parameter.

[0106] After obtaining the filling template, the filling template can be input into the processing model corresponding to the aircraft for the target flight dimension, and the output will be at least one of the following: content generation text, data analysis text, or verification and audit text for the target parameters.

[0107] In some embodiments, the processing model may be a large language model specifically optimized or fine-tuned for a particular target flight dimension and aircraft type, possessing domain expertise and the ability to understand aviation terminology and business logic. The processing model may include a content generation model, a data analysis model, or a verification and auditing model.

[0108] Text processing can involve natural language analysis, interpretation, or evaluation generated by the processing model for each target parameter.

[0109] Content generation can automatically generate descriptive and narrative content based on target parameters, transforming data into easily understandable text reports. For example, if the target parameter is a 15-minute departure delay, the generated content could read, "Flight XX, departing from point YY from location A, is 15 minutes late."

[0110] Data analysis texts provide in-depth analysis, interpretation, and insight mining of target parameters, revealing the business implications and impacts behind the data. For example, a target parameter could be a 5% increase in fuel consumption compared to the plan. The data analysis text could be something like: "Analysis of factors influencing increased fuel consumption includes: 1. An 8-minute wait in the air due to air traffic control during the departure phase, resulting in an additional fuel consumption of approximately 200 kg; 2. Encountering headwinds during the cruise phase, with an average wind speed 15 knots higher than forecast; 3. Actual payload increased by 2 tons compared to the plan. We recommend optimizing departure time and payload forecast accuracy."

[0111] The verification and audit text performs compliance checks, logical verification, and anomaly flagging on the target parameters to ensure data rationality and business compliance. For example, the target parameter could be the crew's cumulative flight time of X hours this month. The verification and audit text could be: "The crew's cumulative flight time this month is X hours, approaching the monthly limit of Y hours stipulated by the Civil Aviation Administration. Subsequent scheduling needs to be monitored to ensure compliance. Current compliance status: Compliant but requiring warning."

[0112] To ensure the comprehensiveness of the generated text, in one possible implementation, S204 above inputs the filling template into the processing model corresponding to the aircraft for the target flight dimension, and outputs the processing text for each target parameter. Specifically, this includes: inputting the filling template into the processing model corresponding to the aircraft for the target flight dimension, and outputting the generated text, data analysis text, and verification and review text for each target parameter.

[0113] For example, after the fill template is input into the processing model corresponding to the target flight dimension of the aircraft, the server processes the parameters in the fill template through the distributed engine and the processing model, and outputs the processing text for each target parameter. The distributed engine supports load balancing and failover mechanisms.

[0114] To achieve configurability of the generated text, in one possible implementation, S204 above involves inputting the fill template into the processing model corresponding to the aircraft for the target flight dimension, and outputting the processing text for each target parameter. Specifically, this includes: obtaining the analysis type of the aircraft for the target flight dimension, and the processing model matching the analysis type. The analysis type includes at least one of content generation type, data analysis type, or verification and review type. The fill template is then input into the processing model, and the processing text for each target parameter is output.

[0115] After obtaining the fill template, the analysis type of the aircraft for the target flight dimension can be acquired, along with a processing model matching the analysis type. The analysis type includes at least one of content generation, data analysis, or verification / audit types. The fill template is input into the processing model, which outputs processing text for each target parameter, allowing for subsequent integration based on the processing text.

[0116] The analysis types are categorized into different levels of depth and focus based on the operator's identification. Each type corresponds to a different cognitive processing method and output goal.

[0117] Content generation type can represent a light analysis type that primarily aims at objective description and narrative summary. The corresponding processing model focuses on transforming data into easily understandable natural language descriptions.

[0118] Data analysis type can refer to a professional analysis type that aims at in-depth insights, pattern discovery, and root cause analysis. The corresponding processing model focuses on revealing the business meaning and inherent laws behind the data.

[0119] The verification and audit type can represent a type of supervisory analysis aimed at compliance checks, quality verification, and risk identification. The processing model corresponding to this type focuses on ensuring the accuracy, compliance, and security of the data.

[0120] Specifically, the server can obtain the operator's operator identifier and multiple candidate types of the aircraft for the target flight dimension. The server can then select the candidate type that matches the operator identifier from among the multiple candidate types as the analysis type.

[0121] In this embodiment, firstly, the processing type matching the analysis type is obtained, the analysis type is identified, and the processing model is matched, achieving accurate matching of the task model. Secondly, different processing models have different processing preferences, which can meet different data processing needs, thereby improving the overall accuracy of data processing. Thirdly, multiple processing models are provided for operators to select and configure, improving the user experience.

[0122] After obtaining the processing model, the filled template can be further optimized to improve the output accuracy of the processing model. In one example, the filled template is input into the processing model, and the output is the processing text for each target parameter. This includes: obtaining the template analysis scenario configured for the filled template, and template samples matching the template analysis scenario. The template analysis scenarios include aircraft technology scenarios, anomaly handling scenarios, or trend determination scenarios. Among the target parameters in the filled template, the target parameters that match the template analysis scenario are retained as parameters. Multiple sample parameters with template structures in the template samples are integrated with the retained parameters with preset structures to obtain the adjusted template. The adjusted template is input into the processing model, and the output is the processing text for each target parameter.

[0123] The template analysis scenario is a specific analytical context predefined based on the report's purpose, audience characteristics, and decision-making needs. The template analysis scenario determines the focus, depth, and professional emphasis of the analysis.

[0124] Aircraft technology scenarios can be specialized technical analysis scenarios focusing on aircraft system performance, technical status, and maintenance requirements. These scenarios are geared towards technical personnel such as aircraft mechanics, performance engineers, and manufacturers.

[0125] Anomaly handling scenarios can be emergency analysis and response scenarios for operational anomalies, security incidents, and emergencies. This scenario is geared towards positions in security departments, operations control, and risk management.

[0126] Trend identification scenarios can be macro-analysis scenarios focusing on long-term development, cyclical changes, and future forecasts. This scenario is geared towards decision-makers in strategic planning, market analysis, fleet management, and other related fields.

[0127] Template samples are high-quality analysis report examples pre-designed for specific analysis scenarios. They include a complete structure, professional analytical logic, and best practice expressions.

[0128] The parameters to be retained are those that have analytical value in the current analysis scenario and need to be retained and analyzed in depth. These parameters are directly related to the core issues of the scenario.

[0129] A template structure can be the logical organizational framework and content arrangement pattern embodied in a template sample. Specifically, the template structure can be the hierarchical relationship between sample parameters and other parameters. For example, sample parameter a and target parameter b are at the same hierarchical level. Or, sample parameter c is at a higher hierarchical level than sample parameter d.

[0130] Sample parameters can be example data parameters used in template samples.

[0131] The adjustment template can be a structured analysis framework that can be directly input into an AI model by intelligently integrating the target parameters of actual flights based on specific analysis scenarios and high-quality sample templates.

[0132] Specifically, after obtaining the template analysis scenario, the server can retrieve template samples that match the scenario from the sample library. After obtaining the template samples, the server can select the target parameters that match the scenario from the target parameters used to fill in the template, and automatically focus on the most relevant parameters based on different scenarios. This simplifies the target parameters, ensuring data processing accuracy while improving data processing efficiency.

[0133] Specifically, after obtaining the retained parameters, multiple sample parameters with template structures in the template sample can be combined with retained parameters with preset structures to obtain the adjustment template. The adjustment template contains a professional template structure, the most relevant retained parameters, multiple sample parameters, and preset structures, which helps to perform accurate data processing based on the adjustment template in the future.

[0134] In this embodiment, firstly, the most relevant parameters are automatically focused and retained based on different scenarios, and the target parameters are simplified, ensuring data processing accuracy while improving data processing efficiency. Secondly, template samples are matched and obtained, providing a professional template structure for subsequent data processing, which helps improve the accuracy of data processing. Thirdly, the template is adjusted to include both professional structure and specific data, ensuring the quality of the processed text output by the subsequent processing model.

[0135] In this embodiment, firstly, by inputting a fill template into the corresponding processing model and outputting the processing text corresponding to the target parameters, a large amount of repetitive work that operators manually do by consulting manuals, comparing standards, and writing reports is replaced, thus automating the analysis process and improving data processing efficiency. Secondly, the processing text output by the processing model includes content-generated text, data analysis text, or verification and review text for the target parameters, ensuring the comprehensiveness of data in subsequent visualization status information, providing operators with comprehensive data references, and improving user experience.

[0136] S205, based on the fusion information template, integrates various target parameters, various processed texts, and the visualization information corresponding to each target parameter to obtain and display the visualization status information of the aircraft in the target flight dimension.

[0137] After obtaining the target parameters and processing text, based on the fusion information template, the target parameters, processing text, and corresponding visualization information can be integrated to obtain and display at least one of the following: flight operation status visualization information, route network analysis visualization information, or flight resource configuration visualization information for the aircraft in the target flight dimension.

[0138] In some embodiments, visualization information can be a form of information in which target parameters are encoded and presented using visual elements such as graphics, charts, and maps. Visualization information can be used to enhance the understandability of data and pattern recognition.

[0139] Flight operation status visualization information can be a visual presentation that shows the spatiotemporal trajectory, performance parameters, and status changes of a single flight throughout its entire operation.

[0140] Route network analysis visualization information can be a visual presentation that shows the network structure, traffic distribution, and efficiency characteristics of multiple routes.

[0141] Flight resource allocation visualization information can be a visual presentation of the allocation, utilization, and optimization status of core resources such as fleet, crew, and time slots.

[0142] Visualized status information can be a final integrated, interactive, and comprehensive information presentation that organically combines target parameters, processing text, and visual information into a unified interface, providing a complete understanding of the business status.

[0143] In one possible implementation, the server can integrate each target parameter, each processed text, and the corresponding visualization information into a fused information template to obtain and display the visualized status information of the aircraft in the target flight dimension.

[0144] For example, after obtaining the visual status information, the server can display it. More specifically, after obtaining the visual status information, the server can build aviation-specific chart components based on the visualization framework, supporting flexible configuration and reuse for professional-level visualization needs. The visual status information is then displayed based on these aviation-specific chart components.

[0145] In the aviation-specific icon component, interactive events are handled uniformly, achieving standardized processing of chart interaction events and supporting event management and state synchronization for various interactive operations. The aviation-specific chart component also features adaptive layout, supports multi-device adaptation, and centralized management of visual styles.

[0146] In this embodiment, each target parameter, each processed text, and the corresponding visualization information of each target parameter are integrated. This combines information from multiple perspectives, including data, insight, and visual perspectives, to obtain visualized status information. By displaying this information using a combination of graphics, text, and data, it helps reduce the cognitive load of operators on professional information and improves the user experience for operators.

[0147] In this embodiment, metadata information of the aircraft in the target flight dimension, prompt word templates for the aircraft, and fused information templates for the aircraft are obtained from multiple heterogeneous data sources. The target flight dimension includes flight operation status, route network analysis, or flight resource allocation. The metadata information reflects the aircraft's operational status in the target flight dimension. The prompt word templates include multiple semantic tags with preset structures, which are tags written in natural language. By obtaining metadata information from multiple data sources and automatically integrating the information, data silos are broken down, information unification is achieved, and a data foundation is provided for subsequent comprehensive analysis. Using preset prompt word templates and fused information templates ensures that the generated visualized status information conforms to aviation professional standards in terms of framework, terminology, and depth, avoiding inconsistent quality and formatting caused by analysts' discretionary decisions.

[0148] This preprocessing method, based on metadata type matching, parameterizes multiple metadata elements to obtain multiple target parameters, which are mapped to semantic tags. This transforms the original descriptive metadata into target parameters with clear business definitions, achieving data-to-knowledge conversion. This benefits subsequent large-scale models in understanding knowledge and improves the accuracy of their output. Furthermore, this metadata type matching preprocessing method enhances data quality and usability, further improving the quality of subsequent model outputs.

[0149] Based on the mapping relationship and each target parameter, the semantic tags in the prompt word template are filled to obtain the filled template; the filled template includes multiple target parameters with a preset structure. Specific target parameters are then filled into the prompt word template to obtain the filled template. Because the contextual relationships in the filled template are fixed, it provides clear analysis instructions for the subsequent large model, avoiding the model from performing broad and unfounded generation tasks and improving the accuracy of data processing.

[0150] The process involves inputting a template into the processing model corresponding to the target flight dimension of the aircraft, and outputting processing text for each target parameter. The processing text includes at least one of the following: content generation text for the target parameter, data analysis text, or verification and review text. By inputting the template into the corresponding processing model and outputting the processing text for the target parameters, the process replaces a large amount of repetitive work that operators would otherwise have to manually consult manuals, compare standards, and write reports, thus automating the analysis process and improving data processing efficiency.

[0151] Based on a fusion information template, this system integrates various target parameters, processed texts, and corresponding visualization information to obtain and display visualized status information of the aircraft in the target flight dimension. The visualized status information includes at least one of the following: flight operation status visualization, route network analysis visualization, or flight resource allocation visualization. By integrating various target parameters, processed texts, and their corresponding visualization information, this system combines data, insight, and visual perspectives to obtain visualized status information. Presenting this information using a combination of graphics, text, and data helps reduce the cognitive load on operators and improves their user experience.

[0152] Figure 3 This is a schematic diagram of the structure of a data processing device provided in an embodiment of this application. Figure 3 As shown, the data processing device includes: an acquisition module 301, a parameterization processing module 302, a template filling module 303, a model processing module 304, and an information integration module 305.

[0153] The acquisition module 301 is used to acquire metadata information of the aircraft in the target flight dimension, prompt word templates for the aircraft, and fusion information templates for the aircraft from multiple heterogeneous data sources; The target flight dimension includes flight operation status dimension, route network analysis dimension, or flight resource allocation dimension; metadata information reflects the aircraft's operation in the target flight dimension; the prompt word template includes multiple semantic tags with preset structures, which are tags written in natural language; The parameterization processing module 302 is used for a preprocessing method based on information type matching of metadata information. It performs parameterization processing on multiple metadata in the metadata information to obtain multiple target parameters. The target parameters have a mapping relationship with semantic tags. The template filling module 303 is used to fill the semantic tags in the prompt word template based on the mapping relationship and each target parameter to obtain a filled template; the filled template includes multiple target parameters with preset structures; The model processing module 304 is used to input the filling template into the processing model corresponding to the aircraft for the target flight dimension, and output the processing text for each target parameter; the processing text includes at least one of the following: content generation text for the target parameter, data analysis text, or verification and review text; The information integration module 305 is used to integrate various target parameters, various processed texts, and the visualization information corresponding to each target parameter based on the fusion information template, so as to obtain and display the visualization status information of the aircraft in the target flight dimension; the visualization status information includes at least one of flight operation status visualization information, route network analysis visualization information, or flight resource configuration visualization information.

[0154] In other embodiments, the preprocessing method for matching the information type of metadata information includes syntax layer verification, logic layer verification, and association layer verification; the parameterization processing module 302 is also used to perform syntax layer verification, logic layer verification, and association layer verification on multiple metadata in the metadata information, and use the data that passes the syntax layer verification, logic layer verification, and association layer verification as the target parameter. The validation objects of the syntax layer include the format and structure of the metadata; the validation objects of the logic layer include the business logic and relationships inherent in the metadata; and the validation objects of the association layer include the metadata and the associated data that are related to the metadata, as well as the consistency between the metadata and the associated data.

[0155] In other embodiments, the parameterization processing module 302 is further configured to obtain raw data of the aircraft in the target flight dimension from multiple heterogeneous data sources, wherein the raw data is unstructured data; and to extract parameters from the raw data based on the parameter extraction method configured for the target flight dimension to obtain target parameters.

[0156] In other embodiments, the semantic tags include semantic placeholders and preset placeholder tags; the target parameters include text type parameters and rich media type parameters; the template filling module 303 is further configured to fill the semantic placeholders in the prompt word template based on the mapping relationship and the text type parameters to obtain a first template; fill the preset placeholder tags in the prompt word template based on the mapping relationship and the rich media type parameters to obtain a second template; and merge the first template and the second template to obtain a filled template.

[0157] In other embodiments, the model processing module 304 is further configured to obtain the analysis type of the aircraft for the target flight dimension, and the processing model matching the analysis type, wherein the analysis type includes at least one of content generation type, data analysis type or verification and auditing type; input the filling template into the processing model, and output the processing text for each target parameter.

[0158] In other embodiments, the model processing module 304 is further configured to acquire a template analysis scenario configured for the filling template, and template samples matching the template analysis scenario. The template analysis scenario includes an aircraft technology scenario, an anomaly handling scenario, or a trend determination scenario. Among the target parameters in the filling template, the target parameters matching the template analysis scenario are retained as reserved parameters. Multiple sample parameters with template structures in the template samples are integrated with the reserved parameters with preset structures to obtain an adjustment template. The adjustment template is input into the processing model, and the processing text for each target parameter is output.

[0159] The data processing apparatus provided in this application embodiment can execute the methods shown in the above method embodiments. Its implementation principle and beneficial effects can be referred to the relevant descriptions in the method embodiments, and will not be repeated here.

[0160] Figure 4 This is a schematic diagram of the structure of a data processing device provided in an embodiment of this application. Figure 4 As shown, the data processing device includes: a memory 401, a transceiver 402, and at least one processor 403.

[0161] The transceiver 402 is used to interact with other devices to send and receive data. For example, in this embodiment, the transceiver 402 can specifically be used to send and receive metadata information.

[0162] The memory 401 is used to store computer program code, which includes computer instructions. These computer instructions run in the data processing device described above to implement the method shown in the above method embodiments. For example, the memory may include high-speed random access memory (RAM), and may also include non-volatile memory (NVM), such as at least one disk storage device, and may also be a USB flash drive, portable hard drive, read-only memory, disk, or optical disc, etc.

[0163] Processor 403 can be a general-purpose processor, including a Central Processing Unit (CPU), a network processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. Processor 403 can also be other general-purpose processors. The general-purpose processor can be a microprocessor or any conventional processor.

[0164] The memory 401, transceiver 402, and processor 403 are communicatively connected. For example, the memory 401 and transceiver 402 can be connected to the processor 403 via a system bus to complete mutual communication. The system bus can be a peripheral component interconnect (PCI) bus, an extended industry standard architecture (EISA) bus, an industry standard architecture (ISA) bus, etc. The system bus can be divided into address bus, data bus, control bus, etc. For ease of representation, only one thick line is used in the figure, but this does not mean that there is only one bus or one type of bus.

[0165] Optionally, the memory 401 can be either standalone or integrated with the processor 403. When the memory 401 is set up independently, it is connected to the processor 403 via a system bus.

[0166] This application also provides a chip for executing instructions, which is used to execute the data processing method described in the above embodiments.

[0167] This application also provides a computer-readable storage medium storing computer instructions. When these computer instructions are executed by a processor, they are used to implement the technical solution of the data processing method described in the above embodiments. Specifically, when the computer instructions are executed by a processor, the data processing device can execute the technical solution of the data processing method described in the above embodiments.

[0168] This application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium, and when the at least one processor executes the computer program, it can implement the technical solution of the data processing method in the above embodiments.

[0169] The aforementioned computer-readable storage media can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The computer-readable storage media can be any available medium accessible to a general-purpose or special-purpose computer.

[0170] An exemplary computer-readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the computer-readable storage medium can also be a component of the processor. The processor and the computer-readable storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the computer-readable storage medium can exist as discrete components in an electronic control unit or main control device; this application does not limit this.

[0171] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0172] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to implement the solution of this embodiment according to actual needs.

[0173] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.

[0174] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application.

[0175] It should be understood that the steps of the method disclosed in the embodiments of this application can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in the processor.

[0176] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A data processing method, characterized in that, The method includes: Metadata information of the aircraft in the target flight dimension, prompt word templates for the aircraft, and fusion information templates for the aircraft are obtained from multiple heterogeneous data sources. The target flight dimension includes a flight operation status dimension, a route network analysis dimension, or a flight resource allocation dimension; the metadata information reflects the aircraft's operation in the target flight dimension; the prompt word template includes multiple semantic tags with preset structures, and the semantic tags are tags written using natural language. Based on the information type matching preprocessing method of the metadata information, multiple metadata in the metadata information are parameterized to obtain multiple target parameters, and the target parameters have a mapping relationship with the semantic tags; Based on the mapping relationship and each of the target parameters, the semantic tags in the prompt word template are filled to obtain a filling template; the filling template includes multiple target parameters having the preset structure. The filling template is input into the processing model corresponding to the aircraft for the target flight dimension, and the processing text for each target parameter is output; the processing text includes at least one of the following: content generation text for the target parameter, data analysis text, or verification and review text; Based on the fusion information template, the target parameters, the processed text, and the visualization information corresponding to each target parameter are integrated to obtain and display the visualization status information of the aircraft in the target flight dimension; the visualization status information includes at least one of flight operation status visualization information, route network analysis visualization information, or flight resource configuration visualization information.

2. The method according to claim 1, characterized in that, The preprocessing methods for matching the information type of the metadata information include syntax layer verification, logic layer verification, and association layer verification; The preprocessing method based on information type matching of the metadata information parameterizes multiple metadata elements in the metadata information to obtain multiple target parameters, including: For multiple metadata in the metadata information, perform the syntax layer verification, the logic layer verification, and the association layer verification respectively, and use the data that passes the syntax layer verification, the logic layer verification, and the association layer verification as the target parameter; The verification objects of the syntax layer include the format and structure of the metadata; the verification objects of the logic layer include the inherent business logic and relationships of the metadata; and the verification objects of the association layer include the metadata, associated data that are related to the metadata, and the consistency between the metadata and the associated data.

3. The method according to claim 2, characterized in that, The method further includes: The raw data of the aircraft in the target flight dimension is obtained from the multiple heterogeneous data sources, and the raw data is unstructured data; Based on the parameter extraction method configured for the target flight dimension, the original data is used to extract parameters to obtain the target parameters.

4. The method according to claim 1, characterized in that, The semantic tags include semantic placeholders and preset placeholder tags; the target parameters include text type parameters and rich media type parameters. The step of filling the semantic tags in the prompt word template based on the mapping relationship and each of the target parameters to obtain a filled template includes: Based on the mapping relationship and the text type parameter, the semantic placeholders in the prompt word template are filled to obtain the first template; Based on the mapping relationship and the rich media type parameters, the preset placeholder tags in the prompt word template are filled to obtain the second template; The first template and the second template are merged to obtain the filling template.

5. The method according to claim 1, characterized in that, The step of inputting the filling template into the processing model corresponding to the aircraft for the target flight dimension and outputting the processing text for each of the target parameters includes: Obtain the analysis type of the aircraft for the target flight dimension, and the processing model matching the analysis type, wherein the analysis type includes at least one of content generation type, data analysis type, or verification and review type; The filling template is input into the processing model, and the processed text for each of the target parameters is output.

6. The method according to claim 5, characterized in that, The step of inputting the filling template into the processing model and outputting the processed text for each of the target parameters includes: Obtain template analysis scenarios configured for the filling template, and template samples matched with the template analysis scenarios. The template analysis scenarios include aircraft technology scenarios, anomaly handling scenarios, or trend determination scenarios. Among the target parameters in the filling template, the target parameter that matches the analysis scenario of the template is retained as the reserved parameter; The template sample is integrated with multiple sample parameters that have the template structure and the reserved parameters that have the preset structure to obtain the adjustment template; The adjustment template is input into the processing model, and the processing text for each of the target parameters is output.

7. A data processing apparatus, characterized in that, The device includes: The acquisition module is used to acquire metadata information of the aircraft in the target flight dimension, prompt word templates for the aircraft, and fusion information templates for the aircraft from multiple heterogeneous data sources; The target flight dimension includes a flight operation status dimension, a route network analysis dimension, or a flight resource allocation dimension; the metadata information reflects the aircraft's operation in the target flight dimension; the prompt word template includes multiple semantic tags with preset structures, and the semantic tags are tags written using natural language. The parameterization processing module is used to perform parameterization processing on multiple metadata in the metadata information based on the information type matching preprocessing method of the metadata information to obtain multiple target parameters, and the target parameters have a mapping relationship with the semantic tags; A template filling module is used to fill the semantic tags in the prompt word template based on the mapping relationship and each of the target parameters to obtain a filled template; the filled template includes multiple target parameters having the preset structure; The model processing module is used to input the filling template into the processing model corresponding to the aircraft for the target flight dimension, and output the processing text for each target parameter; the processing text includes at least one of the following: content generation text for the target parameter, data analysis text, or verification and review text. The information integration module is used to integrate each of the target parameters, each of the processed texts, and the visualization information corresponding to each of the target parameters based on the fusion information template, so as to obtain and display the visualization status information of the aircraft in the target flight dimension; the visualization status information includes at least one of flight operation status visualization information, route network analysis visualization information, or flight resource configuration visualization information.

8. A data processing device, characterized in that, include: A memory and at least one processor; the memory is communicatively connected to the processor; the memory is used to store computer program code, the computer program code including computer instructions; when the processor executes the computer instructions, the data processing device causes the data processing device to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, are used to implement the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, When the computer program product is run on a computer / executed by the computer's processor, it implements the method as described in any one of claims 1-6.

Images