Data query method, device, apparatus, medium and program product

By processing user query requests through a large model, performing semantic understanding and task decomposition, generating and executing structured query statements, the problem of low data query efficiency in existing technologies is solved, and an efficient and stable data query process is achieved.

CN123068461APending Publication Date: 2026-11-10INDUSTRIAL AND COMMERCIAL BANK OF CHINA +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611396036.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-09
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing technologies suffer from low efficiency in querying data metrics and insufficient timeliness in response, resulting in untimely responses when structured data changes to meet business needs.

Method used

The system receives user query requests through a large model, performs semantic understanding, task decomposition, and generates structured query statements using preset prompts, and finally executes the query in the database.

Benefits of technology

It achieves a highly automated, scalable, fault-tolerant, and controllable data query process, improving query efficiency and stability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123068461A_ABST
    Figure CN123068461A_ABST
Patent Text Reader

Abstract

This application provides a data query method, apparatus, device, storage medium, and program product, which can be applied to the fields of big data and artificial intelligence, specifically relating to the application of large models in fintech scenarios. The method includes: receiving a first query request from a user; obtaining semantic information based on the first query request and a first preset prompt word as input to a preset large model; obtaining a task sequence based on the semantic information and a second preset prompt word as input to the preset large model, the task sequence including N sub-tasks and a task execution order, where N is a positive integer; obtaining N sets of structured query statements based on the task sequence and a third preset prompt word as input to the preset large model; and executing the N sets of structured query statements according to the task execution order to query data in a preset database and obtain query results.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the fields of big data and artificial intelligence, specifically to the application of large models in fintech scenarios, and more specifically to a data query method, device, equipment, medium, and program product. Background Technology

[0002] Data metrics serve as a core reference for enterprise operational decisions, and the effectiveness of their query and analysis technologies directly impacts the quality of business decisions. Currently, the technology for querying data metrics primarily relies on manually developing structured query statements and then displaying and interacting with the retrieved data. However, existing technologies suffer from low efficiency and long delivery cycles due to the manual development of interface-based query statements, leading to insufficient responsiveness when structured data changes to business requirements.

[0003] Therefore, improving data query efficiency is a pressing technical issue that needs to be addressed when business changes occur. Summary of the Invention

[0004] In view of the above problems, embodiments of this application provide a data query method, apparatus, device, medium, and program product.

[0005] According to a first aspect of this application, a data query method is provided, the method comprising: receiving a first query request from a user; obtaining semantic information based on the first query request and a first preset prompt word as input to a preset large model; obtaining a task sequence based on the semantic information and a second preset prompt word as input to the preset large model, the task sequence including N subtasks and a task execution order, where N is a positive integer; obtaining N sets of structured query statements based on the task sequence and a third preset prompt word as input to the preset large model; and executing the N sets of structured query statements according to the task execution order to query data in a preset database and obtain query results.

[0006] According to an embodiment of this application, wherein the first query request is a multimodal input, after receiving the first query request from the user and before obtaining semantic information based on the first query request and the first preset prompt word as a preset large model input, the method further includes: converting the first query request into text form; and performing preprocessing on the text form of the first query request to obtain a standardized first query request.

[0007] According to an embodiment of this application, obtaining semantic information based on the first query request and the first preset prompt word as input to a preset large model includes: standardizing and mapping non-standard words in the first query request based on the standardized indicator names in a preset knowledge graph to obtain a second query request; and extracting entity information and constraint information based on the second query request and the first preset prompt word through a preset large model.

[0008] According to an embodiment of this application, obtaining a task sequence based on the semantic information and the second preset prompt word as input to a preset large model includes: obtaining N sub-tasks based on the entity information, the constraint information, and the second preset prompt word as input to a preset large model, wherein each sub-task includes corresponding context information; determining the execution priority of the N sub-tasks; and generating the task sequence based on the N sub-tasks and the execution priority.

[0009] According to an embodiment of this application, the step of obtaining N sets of structured query statements based on the task sequence and the third preset prompt word as input to the preset large model includes: for any subtask, obtaining the structured query statement based on the context information and the corresponding third preset prompt word as input to the preset large model.

[0010] According to an embodiment of this application, the step of executing the N sets of structured query statements to query data in a preset database according to the task execution order and obtaining query results includes: performing syntax verification and security review on the N sets of structured query statements; if the syntax verification and the security review are both passed, evaluating the execution cost of the N sets of structured query statements; and if the execution cost is greater than a preset cost threshold, selecting distributed execution; or if the execution cost is less than the preset cost threshold, selecting single-machine execution.

[0011] According to an embodiment of this application, before receiving the first query request from the user, the method further includes: constructing a user profile based on historical data generated from the user's historical query behavior, with the user's authorization; generating a recommendation list based on the user profile and a preset access control model; and pushing the recommendation list to the user interface for the user to select.

[0012] According to a second aspect of this application, a data query apparatus is provided, the apparatus comprising: a receiving module for receiving a first query request from a user; a first input module for obtaining semantic information based on the first query request and a first preset prompt word as input to a preset large model; a second input module for obtaining a task sequence based on the semantic information and the second preset prompt word as input to the preset large model, the task sequence including N subtasks and a task execution order, where N is a positive integer; a third input module for obtaining N sets of structured query statements based on the task sequence and the third preset prompt word as input to the preset large model; and a query module for executing the N sets of structured query statements according to the task execution order to query data in a preset database and obtain query results.

[0013] According to a third aspect of this application, an electronic device is provided, comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0014] According to a fourth aspect of this application, a computer-readable storage medium is also provided, on which a computer program or instructions are stored, wherein the computer program or instructions, when executed by a processor, implement the steps of the above-described method.

[0015] According to a fifth aspect of this application, a computer program product is also provided, comprising a computer program or instructions that, when executed by a processor, implement the steps of the above-described method.

[0016] To address the technical challenge of increasing the development efficiency of structured query statements, this application's embodiment first extracts semantic information from the query request by calling a large model. Then, it sorts the query tasks by calling the large model again, outputs structured statements, and finally, executes these structured statements to directly query data from the database. The beneficial effects are: 1. High degree of automation: Users only need to input the query request; they don't need to understand the database structure or query syntax. The system automatically completes the entire process of understanding, decomposing, and generating the query. 2. Strong scalability: By replacing or adjusting three preset prompts, it can adapt to different domains and databases without modifying the core process. 3. Good fault tolerance: Complex tasks are broken down into multiple subtasks to generate query statements separately; an error in a single subtask does not affect the retrying or repair of other subtasks. 4. Controllable response: Since the large model call task is clearly defined at each stage, timeout and retry strategies can be set separately, improving overall stability. Attached Figure Description

[0017] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0018] Figure 1 The illustrations depict application scenarios of data query methods, apparatuses, devices, media, and program products according to embodiments of this application.

[0019] Figure 2 A flowchart illustrating a data query method according to an embodiment of this application is shown schematically;

[0020] Figure 3 This illustration schematically shows the architecture of a data query method according to an embodiment of this application;

[0021] Figure 4 A schematic diagram illustrating the entire process of a data query method according to an embodiment of this application is provided.

[0022] Figure 5 This schematic diagram illustrates the structure of a data query apparatus according to an embodiment of the present application;

[0023] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a data query method according to an embodiment of this application. Detailed Implementation

[0024] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0025] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0026] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0027] In the technical solution of this application, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant laws, regulations, and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0028] In scenarios where personal information is used for automated decision-making, the methods, devices, and systems provided in this application all provide users with corresponding operation entry points for users to choose to agree to or reject the automated decision results; if the user chooses to reject, the process enters the expert decision-making process.

[0029] As used in this paper, the term "model" refers to a model that learns the relationship between inputs and outputs from training data, enabling it to generate corresponding outputs for a given input after training. Model generation can be based on machine learning techniques. Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs using multiple layers of processing units. A neural network model is an example of a deep learning-based model. In this paper, "model" may also be referred to as a "machine learning model," "learning model," "machine learning network," or "learning network," and these terms are used interchangeably.

[0030] It's important to note that the term "neural network" can refer to a machine learning network based on deep learning. A neural network processes input and provides corresponding output, typically consisting of an input layer, an output layer, and one or more hidden layers between them. Neural networks used in deep learning applications often include many hidden layers, increasing the network's depth. The layers of a neural network are connected sequentially, so that the output of the previous layer serves as the input to the next layer. The input layer receives the input to the neural network, while the output layer's output becomes the final output. Each layer of a neural network includes one or more nodes (also called processing nodes or neurons), each processing the input from the layer above.

[0031] It should be understood that machine learning generally includes three phases: training, testing, and application (also known as inference). In the training phase, a given model is trained using a large amount of training data, iteratively updating parameter values ​​until the model can consistently generate inferences that meet the expected goals from the training data. Through training, the model can be considered to have learned the relationship between inputs and outputs (also known as the input-output mapping) from the training data. The parameter values ​​of the trained model are determined. In the testing phase, test inputs are applied to the trained model to test whether it can provide the correct output, thus determining the model's performance. In the application phase, the model can be used to process actual inputs based on the trained parameter values ​​to determine the corresponding output.

[0032] In one or more embodiments described herein, the term "large model" can refer to a deep learning model with a large number of model parameters, which can include hundreds of millions, tens of billions, hundreds of billions, trillions, or even tens of trillions of model parameters. Large models can also be called foundational models or basic models. They are pre-trained using large-scale unlabeled corpora to produce pre-trained models with hundreds of millions of parameters. Such models can adapt to a wide range of downstream tasks and have good generalization ability, such as large language models and multimodal pre-trained models. It should be understood that in practical applications, large models only require a small number of samples to fine-tune the pre-trained model before being applied to different tasks. Large models can be widely used in natural language processing, computer vision, and other fields. Specifically, they can be applied to computer vision tasks such as visual question answering, image captioning, and image generation, as well as natural language processing tasks such as text-based sentiment classification, text summarization, and machine translation. Major application scenarios for large models can include digital assistants, intelligent robots, search, online education, office software, e-commerce, and intelligent design.

[0033] Figure 1 The diagram illustrates an application scenario of the data query method according to an embodiment of this application. For example... Figure 1 As shown, application scenario 100 according to an embodiment of this application may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables. For example, a user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send information, etc.

[0034] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be electronic devices such as smartphones, wearable devices, personal computers, intelligent voice interaction devices, smart home appliances, intelligent vehicles, in-vehicle terminals, aircraft, unmanned vending terminals, and extended reality devices. Extended reality devices can include virtual reality devices, augmented reality devices, and mixed reality devices. A client application for the target application can be installed and run on the terminal devices. This target application can include, but is not limited to, financial transaction applications, payment applications, shopping applications, web browser applications, search applications, instant messaging tools, email clients, and social media platform software (these are just examples). Furthermore, this application embodiment does not limit the form of the target application, and it can include, but is not limited to, applications, mini-programs, etc., installed on the terminal devices, and can also be in the form of web pages.

[0035] Server 105 can be a server providing various services, such as a backend management server supporting websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process received user requests and other data, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services such as cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks, and basic cloud computing services such as big data. The server can be the backend server of the aforementioned target application, used to provide backend services to the clients of the target application.

[0036] It should be noted that the data query method provided in this application embodiment can generally be executed by server 105 and / or terminal devices 101-103. Accordingly, the data query device provided in this application embodiment can generally be set in server 105 and / or terminal devices 101-103.

[0037] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0038] Figure 2 A flowchart illustrating a data query method according to an embodiment of this application is shown. Figure 2 As shown, the data query method 200 according to the embodiments of this application may include steps S210 to S250.

[0039] In step S210, a first query request from the user is received.

[0040] The first query request is the original input information initiated by the user to express the intent of data query, which can be in various forms such as text, voice, and images.

[0041] In step S220, semantic information is obtained based on the first query request and the first preset prompt word as input to the preset large model.

[0042] The first preset prompt is a pre-defined instruction template used to guide the large model to perform the "semantic understanding" task, such as "Please extract the core semantic information from the following query".

[0043] Semantic information refers to the structured expression of meaning obtained after understanding the user's query request, including the data indicators, condition constraints, and sorting requirements that the user wants to query.

[0044] In step S230, based on the semantic information and the second preset prompt word as input to the preset large model, a task sequence is obtained. The task sequence includes N sub-tasks and the task execution order, where N is a positive integer.

[0045] The second preset prompt is a pre-defined instruction template used to guide the large model to perform the "task decomposition and orchestration" task, such as "Please break down the following query requirements into executable sub-tasks and sort them".

[0046] Among them, the task sequence is an ordered set of multiple subtasks arranged in a specific execution order, and each subtask corresponds to a data processing step that can be executed independently.

[0047] In step S240, based on the task sequence and the third preset prompt word as input to the preset large model, N sets of structured query statements are obtained.

[0048] The third preset prompt is a pre-defined instruction template used to guide the large model to execute the task of "generating structured query statements", such as "Please generate a database-executable query statement based on the following task information".

[0049] Among them, N sets of structured query statements are query command texts that conform to database execution specifications and are generated for N subtasks respectively. Each set of statements can be run independently and retrieve data.

[0050] In step S250, according to the task execution order, the N sets of structured query statements are executed to query the data in the preset database and obtain the query results.

[0051] The pre-configured database refers to a pre-configured data storage system that stores business data; it can be a single-machine database, a distributed data lake, or another data warehouse.

[0052] Specifically, after receiving the user's first query request, the system calls the large model three times using different preset prompts to obtain semantic information, task sequence, and structured query statement in sequence, and finally executes the query statement to obtain the result.

[0053] In a typical scenario, consider a user entering the text "Please help me find the product with the highest sales last week" into an intelligent customer service system: This text is received as the first query request. Internally, a first preset prompt is stored: "You are a semantic understanding assistant, extracting key information such as the query metric, time range, and sorting criteria from the user's question." The first query request and the first preset prompt are combined and sent to the large model. The large model returns semantic information: "Metric = Sales, Time Range = Last Week, Sorting Requirement = Descending, Aggregation Method = Grouped by Product, Filtering Condition = Take Maximum Value." Next, a second preset prompt is used: "You are a task planning assistant, breaking down the following semantic information into executable data acquisition subtasks and providing the execution order." The large model returns a task sequence: Subtask 1 "Get the product list," Subtask 2 "Get the sales of each product last week," Subtask 3 "Sort and extract the highest value." Finally, a third preset prompt is used: "You are a code generation assistant, generating database query statements for each subtask." The large model generates three sets of structured query statements: the first set queries the product table to retrieve all product identifiers, the second set queries the sales table to summarize last week's sales by product, and the third set sorts the results and retrieves the first result. The three sets of statements are executed in sequence to obtain the query result "Product A, sales revenue 120,000 yuan," which is then returned to the user.

[0054] In the embodiments of this application, firstly, semantic information in the query request is extracted by calling a large model; then, the query task is sorted by calling the large model; secondly, a structured statement is output by calling the large model; and finally, data in the database can be directly queried by executing the structured statement. The beneficial effects are: 1. High degree of automation: Users only need to input the query request, without needing to understand the database structure or query syntax, and the entire process of understanding, decomposition, and generation is automatically completed. 2. Strong scalability: By replacing or adjusting three preset prompt words, it can adapt to different domains and different databases without modifying the core process. 3. Good fault tolerance: Complex tasks are broken down into multiple subtasks to generate query statements separately; an error in a single subtask does not affect the retry or repair of other subtasks. 4. Controllable response: Since the large model calling task is clear at each stage, timeout and retry strategies can be set separately, improving overall stability.

[0055] According to an embodiment of this application, wherein the first query request is a multimodal input, after receiving the first query request from the user and before obtaining semantic information based on the first query request and the first preset prompt word as a preset large model input, the method further includes: converting the first query request into text form; and performing preprocessing on the text form of the first query request to obtain a standardized first query request.

[0056] Multimodal input refers to user input that includes two or more information formats, such as voice plus text, voice plus image, text plus gesture, etc.

[0057] Among them, text form refers to the linear sequence composed of character encoding obtained after conversion, which is convenient for subsequent natural language processing.

[0058] Preprocessing refers to a set of operations that clean and standardize text-based query requests, including but not limited to removing invalid characters, correcting obvious typos, standardizing punctuation, and splitting long texts.

[0059] The standardized first query request refers to clean text that has been preprocessed, has a uniform format, reduced noise, and is suitable for inputting into a large model.

[0060] Specifically, the first query request is a multimodal input, and modality conversion (converting non-text modalities to text) and text preprocessing (standardizing and cleaning the converted text) are performed before inputting into the large model. Regardless of the original input modality, it is ultimately unified into standardized text before entering the subsequent large model processing flow.

[0061] In a typical scenario, a user speaks into their phone, "Check last month's return rate." This voice message is received as the first query request. First, the speech-to-text module is invoked to convert the speech into the text string "Check last month's return rate." Then, preprocessing is performed: a missing period is detected and automatically added; typos are checked; multiple consecutive spaces are compressed into one; finally, the standardized first query request "Check last month's return rate." is obtained.

[0062] In another typical scenario, a user takes a picture containing the handwritten text "Display items with inventory below the safety stock level" and uploads it. This image is received as the first query request. The image text extraction module (based on optical character recognition technology) is first invoked to locate and recognize the handwritten text in the image, converting it into the text "Display items with inventory below the safety stock level". Preprocessing stage: Due to potential errors in handwriting recognition, error correction is performed—the word "inventory" is compared with the business dictionary to confirm its accuracy; "below" is corrected to "below" based on the context; irrelevant spaces and line breaks are removed. The standardized text "Display items with inventory below the safety stock level" is obtained.

[0063] It is understood that the following beneficial effects can be achieved in the embodiments of this application: 1. User-friendly input method: Users can ask questions in the most natural way using voice or images, reducing the barrier to entry. 2. Improved robustness: The preprocessing step can eliminate noise in the original input, reducing errors in understanding large models. 3. Wide compatibility: After unifying to text, subsequent modules do not need to be developed separately for different modalities, reducing system complexity.

[0064] According to an embodiment of this application, obtaining semantic information based on the first query request and the first preset prompt word as input to a preset large model includes: standardizing and mapping non-standard words in the first query request based on the standardized indicator names in a preset knowledge graph to obtain a second query request; and extracting entity information and constraint information based on the second query request and the first preset prompt word through a preset large model.

[0065] Among them, knowledge graph refers to a knowledge base organized in a graph structure, where nodes represent business concepts (such as indicator names and data fields), and edges represent the relationships between concepts (such as synonyms, inclusion, computational dependencies, etc.).

[0066] Among them, the standardized indicator name refers to the unique and standardized indicator identifier that is predefined in the knowledge graph or indicator library. For example, "number of days of safe production" is the standardized name, while "number of safe days" and "number of accident-free days" are non-standard expressions.

[0067] Non-standard terminology refers to colloquialisms, abbreviations, alternative names, or incorrect expressions used by users in their queries that are not entirely consistent with the standardized indicator names.

[0068] Standardization mapping refers to the process of matching non-standard terms to corresponding standardized indicator names by querying a knowledge graph.

[0069] The second query request refers to the new query text obtained by replacing non-standard terms in the original query with standard names.

[0070] Entity information refers to the business objects extracted from the query, such as metric names, dimension names, and condition values.

[0071] Among them, constraint information refers to the limiting conditions extracted from the query, such as time range ("first quarter"), numerical conditions ("greater than 90%), comparison relationship ("highest" "lowest"), etc.

[0072] Specifically, the non-standard terms in the first query request are first standardized and mapped using a knowledge graph to obtain the second query request. Then, entity information and constraint information are extracted from the second query request using a large model. "Terminology standardization" and "semantic understanding" are decoupled. The knowledge graph is first responsible for terminology alignment, and then the large model is responsible for general semantic extraction. The two work together.

[0073] In a typical scenario, a user's first query is "Check the number of safe production days". The knowledge graph stores the standardized indicator name "safe production days" and records its synonyms, including "safe days", "accident-free days", and "continuous safe days". Using "safe days" as the query term, the knowledge graph is searched, and the standardized indicator name "safe production days" is found. Then, "safe days" in the first query request is replaced with "safe production days", resulting in the second query request "Check the number of safe production days". The second query request, along with the first preset prompt term, is input into the large model. The large model extracts the entity information: "safe production days", with the constraint information: no additional conditions.

[0074] In another typical scenario, a user queries "pass rate". The knowledge graph has two standardized metrics: "product pass rate" and "first-pass pass rate". Based on the context (e.g., the user's department is the production department) or the user's historical behavior (previously frequently querying product pass rate), the more relevant "product pass rate" is selected for mapping, resulting in a second query request: "Check the product pass rate". The larger model extracts the entity information: "product pass rate".

[0075] It is understood that the following beneficial effects can be achieved in the embodiments of this application: 1. Improved semantic understanding accuracy: The knowledge graph pre-stores domain terminology relationships, avoiding misunderstandings of spoken expressions by large models due to a lack of domain knowledge. 2. Reduced burden on large models: By standardizing terminology in advance, large models only need to process standard names, reducing the possibility of errors. 3. Easier maintenance: When business indicator names change or new aliases are added, only the knowledge graph needs to be updated, without retraining the large model.

[0076] According to an embodiment of this application, obtaining a task sequence based on the semantic information and the second preset prompt word as input to a preset large model includes: obtaining N sub-tasks based on the entity information, the constraint information, and the second preset prompt word as input to a preset large model, wherein each sub-task includes corresponding context information; determining the execution priority of the N sub-tasks; and generating the task sequence based on the N sub-tasks and the execution priority.

[0077] Contextual information refers to auxiliary data related to a subtask, such as the metadata definitions that the subtask depends on, the output results of other subtasks, and the structural description of the data table.

[0078] Among them, execution priority refers to the numerical value or rule that determines the order in which multiple subtasks are executed. Subtasks with higher priority are executed first, and subtasks with dependencies automatically follow the order.

[0079] Generating the task sequence refers to arranging N subtasks into an ordered list according to their execution priority, and possibly adding input and output definitions for each subtask.

[0080] Specifically, based on entity information, constraint information, and a second preset prompt, the large model is invoked to obtain N subtasks (each subtask comes with context information). Then, the execution priority of these subtasks is determined, and finally, a task sequence is generated. The large model not only outputs a list of subtasks but also outputs the context information required for each subtask, and ensures the correct execution order of tasks through a priority mechanism.

[0081] In a typical scenario, assume the entity information is "safety production days" and the constraint information is "grouped by construction unit, time range: first quarter". The main model outputs the following based on the second preset prompt ("Please break down the following query requirements into subtasks, each subtask indicating the required context information and giving its execution priority; the smaller the number, the earlier it is executed"): Subtask 1: Obtain the metadata definition of the standard indicator "safety production days" (context information: indicator code, calculation formula, dependent tables). Priority: 1. Subtask 2: Obtain the time boundary of the first quarter (context information: start date, end date). Priority: 2. Subtask 3: Obtain the list of construction units (context information: construction unit identifier, name). Priority: 2 (no dependency on Subtask 2, can be parallelized). Subtask 4: Calculate the safety production days for each construction unit (context information: depends on the output of Subtasks 1, 2, and 3). Priority: 3. After determining the priorities, the task sequence is generated: [Subtask 1, Subtask 2 + Subtask 3 in parallel, Subtask 4].

[0082] It is understood that the following beneficial effects can be achieved in the embodiments of this application: 1. Complex tasks can be handled: Complex queries are intelligently decomposed into multiple simple sub-tasks, reducing the difficulty of generating statements from a single large model. 2. High execution efficiency: Parallelizable tasks are identified by priority, allowing for concurrent execution and reducing overall response time. 3. Strong interpretability: The task sequence records the reasoning steps, facilitating debugging and optimization.

[0083] According to an embodiment of this application, the step of obtaining N sets of structured query statements based on the task sequence and the third preset prompt word as input to the preset large model includes: for any subtask, obtaining the structured query statement based on the context information and the corresponding third preset prompt word as input to the preset large model.

[0084] Structured query statements are query command texts generated for a single subtask that can be executed by the database.

[0085] Specifically, for each subtask in the task sequence, the preceding information accompanying the subtask and the third preset prompt are input into the large model to generate the structured query statement corresponding to the subtask.

[0086] In a typical scenario, suppose there is a subtask in the task sequence: "Get a list of products," with the context information: "The product table name is 'Product Main Table,' containing the fields 'Product Identifier' and 'Product Name.'" A third preset prompt (containing the message "You are a database query generation expert; based on the task description and context information, generate an accurate structured query statement") is concatenated with the information from this subtask and input into the main model. The main model returns the structured query statement: "Select Product Identifier and Product Name from the Product Main Table."

[0087] In another typical scenario, the subtask is: "Retrieve stores with sales exceeding 100,000 last week," with the context: "The sales table contains fields 'Store ID,' 'Sales Amount,' and 'Sales Date,' with the date field formatted as year-month-day." The main model generates the statement: "Select store IDs from the sales table where the sales date falls within the range of last week and the sales amount is greater than 100,000, and output the summarized store IDs."

[0088] In another typical scenario, the subtask is: "Get the average inventory turnover rate for each product category." The context information includes: "Inventory table structure," "Sales table structure," and "Product category mapping table." The large model generates statements involving multi-table joins and complex calculations, but these statements are still automatically generated by the large model based on the context.

[0089] It is understood that the following beneficial effects can be achieved in the embodiments of this application: 1. Modular generation: Each subtask generates query statements independently, and errors in a single statement do not affect the retrying of other subtasks. 2. Context awareness: The context information of the subtasks contains the necessary data structure descriptions, enabling the large model to generate accurate statements and reducing guesswork. 3. Strong scalability: When adding a new subtask type, only the corresponding context information needs to be provided, without modifying the generation logic.

[0090] According to an embodiment of this application, the step of executing the N sets of structured query statements to query data in a preset database according to the task execution order and obtaining query results includes: performing syntax verification and security review on the N sets of structured query statements; if the syntax verification and the security review are both passed, evaluating the execution cost of the N sets of structured query statements; and if the execution cost is greater than a preset cost threshold, selecting distributed execution; or if the execution cost is less than the preset cost threshold, selecting single-machine execution.

[0091] Syntax validation refers to the process of checking whether a structured query statement conforms to the database syntax rules, such as keyword spelling, bracket matching, and table name existence.

[0092] Security review refers to the process of checking whether structured query statements contain dangerous operations (such as deleting tables or modifying structures) or unauthorized access (querying unauthorized data).

[0093] Among them, execution overhead refers to the estimated computing resources required for the database to execute one or a set of query statements, which usually includes central processing unit time, memory usage, disk input / output, network transmission, etc.

[0094] Among them, the preset cost threshold refers to a pre-configured numerical limit used to distinguish between "small queries" and "large queries".

[0095] Distributed execution refers to breaking down a query task into multiple subtasks, distributing them to multiple computing nodes for parallel processing, and finally aggregating the results.

[0096] Single-machine execution refers to executing queries directly on a single database node without task splitting.

[0097] Specifically, each set of structured query statements is first subjected to syntax validation and security review. After passing the review, the execution cost is evaluated. If the cost is greater than a preset threshold, distributed execution is selected; otherwise, single-machine execution is selected.

[0098] In a typical scenario, a query is generated: "Select employee names from the employee table, where employee number equals 100. This statement is expected to access only a small amount of data, and its execution overhead is assessed as low (e.g., estimated time less than 100 milliseconds, memory usage less than 10 megabytes). The preset overhead threshold is medium (e.g., estimated time 1 second or memory usage 100 megabytes). Since the actual overhead is much lower than the threshold, single-machine execution is chosen, running the statement directly on the database node to quickly return the results."

[0099] In another typical scenario, a query was generated: "Summarize sales revenue by product from the annual sales details table, which contains over 100 million sales records." The execution overhead was assessed as high (estimated execution time exceeds 10 seconds, requiring scanning a large amount of data). This overhead exceeded a preset threshold, so the query was automatically converted into a distributed execution plan: the data was partitioned by product identifier hash across multiple computing nodes, each node performed parallel local summaries, and the results were finally merged.

[0100] In another typical scenario, a large model incorrectly generates a statement: "Clear the user table". Before execution, a security review detects that this statement is a high-risk operation (clearing data). Execution is immediately rejected, a security warning is returned, and the process does not proceed to the cost assessment stage. For example, if a user's permissions only allow querying data from department A, but the generated statement includes data from department B's table, the security review will also block it.

[0101] In another typical scenario, the generated statement contains a syntax error, such as "select from". The syntax check finds that the keywords are incomplete, so the statement is not executed. Instead, a retry mechanism is triggered (either the large model is called again to generate the statement or an error message is returned).

[0102] It is understood that the following beneficial effects can be achieved in the embodiments of this application: 1. High resource utilization: Small queries do not use a distributed framework, avoiding scheduling overhead; large queries are automatically parallelized, shortening response time. 2. Strong security: Dual protection of syntax verification and security review prevents invalid or malicious statements from harming the database. 3. Adaptive capability: No manual configuration of execution mode is required; the optimal method is automatically selected based on the actual query characteristics.

[0103] According to an embodiment of this application, before receiving the first query request from the user, the method further includes: constructing a user profile based on historical data generated from the user's historical query behavior, with the user's authorization; generating a recommendation list based on the user profile and a preset access control model; and pushing the recommendation list to the user interface for the user to select.

[0104] Historical query behavior refers to records generated by users in the past, including query time, query content, query frequency, and which recommended results were clicked.

[0105] User profiles refer to abstract descriptions of user characteristics and preferences, typically including labels across multiple dimensions, such as "users who frequently query safety production indicators" or "users whose preferences are displayed in charts."

[0106] Among them, the access control model refers to the role-based access control model (in Chinese: role-based permission control model), which grants different data access permissions according to the user's role (such as administrator, ordinary employee, department head).

[0107] The recommended list refers to a collection of metrics or query suggestions sorted by relevance, selected based on user profiles and permissions.

[0108] Specifically, with the user's authorization, a user profile is built based on the user's historical query behavior. Then, a personalized recommendation list is generated by combining a role-based access control model and actively pushed to the user interface for the user to choose from.

[0109] In a typical scenario, when a user logs in, their role is detected as "Safety Production Supervisor." First, the permissions for this role are determined: they can only query indicators such as "Safety Production Days," "Hazard Rectification Rate," and "Accident Statistics." The most frequently used queries for all users under this role are analyzed from the historical behavior database, revealing that "Safety Production Days" is the most frequent indicator. A recommended list is generated, in the order of "Safety Production Days," "Hazard Rectification Rate," and "Accident Statistics." This list is then pushed to the user interface homepage, allowing users to directly query the data without needing to type any natural language.

[0110] In another typical scenario, three dimensions are extracted based on user profiles: department (finance department), position (analyst), and recent behavior (frequently checked year-on-year and month-on-month comparisons). A role-based access control model provides a set of allowed metrics (budget execution rate, cost deviation rate, etc.). A ranking model combines the two, with the first recommendation being "month-on-month cost growth rate," the second being "year-on-year change in budget execution rate," and the third being "department expense details." After being pushed to users, they can quickly query by clicking the first recommendation.

[0111] It is understood that the following beneficial effects can be achieved in the embodiments of this application: 1. Improved query efficiency: Frequently used indicators are directly displayed, reducing input and even achieving "zero-input" queries. 2. Personalized experience: Different users see different recommendation lists, conforming to personal habits and job requirements. 3. Security and compliance: Recommendation results automatically follow the permission model, ensuring no unauthorized data is leaked. 4. Discovering new needs: Through popular trend recommendations, users may discover valuable indicators that were previously unnoticed.

[0112] Figure 3The diagram illustrates the architecture of a data query method according to an embodiment of this application.

[0113] like Figure 3 As shown, the architecture of this data query method includes an interaction layer, an intelligent processing layer, and a support layer. The interaction layer is used to interact with users and includes multimodal input parsing and result visualization. The intelligent processing layer is used to intelligently process data in conjunction with a large model and includes intent understanding, dynamic orchestration, and task execution. The intelligent processing layer also includes writing query statements, executing query predictions, and statistical recommendations. The support layer is used for data support and includes a knowledge base, a large model, metadata, and a data lake.

[0114] Figure 4 A schematic diagram illustrating the complete flowchart of a data query method according to an embodiment of this application is provided.

[0115] like Figure 4 As shown, users perform the following steps when querying data:

[0116] Step 1: User login recommendation module.

[0117] Step 2: The recommendation module recommends query metrics.

[0118] Step 3: The user makes a query using natural language.

[0119] Step 4: Input the parsing module to process the audio file.

[0120] Step 5: Send the processed speech text to the intent understanding module to identify the intent, and use the large model to identify the query indicator information.

[0121] Step 6: The large model returns information such as metrics and dimensions.

[0122] Step 7: Organize the tasks by calling the model to decompose and sort the tasks.

[0123] Step 8: The large model returns a task list.

[0124] Step 9: Execute the task by calling the large model to write the query statement through task scheduling.

[0125] Step 10: The large model returns the query statement.

[0126] Step 11: Execute the query statement to retrieve data from the data lake.

[0127] Step 12: The query results are returned.

[0128] Based on the above data query method, embodiments of this application also provide a data query device. The following will be combined with... Figure 5The device is described in detail.

[0129] Figure 5 A schematic block diagram of a data query apparatus according to an embodiment of this application is shown.

[0130] like Figure 5 As shown, the data query device 500 of this embodiment includes a receiving module 510, a first input module 520, a second input module 530, a third input module 540, and a query module 550.

[0131] The receiving module 510 is used to receive a first query request from the user. In one embodiment, the receiving module 510 can be used to perform step S210 described above, which will not be repeated here.

[0132] The first input module 520 is used to obtain semantic information based on the first query request and the first preset prompt word as preset large model input. In one embodiment, the first input module 520 can be used to execute step S220 described above, which will not be repeated here.

[0133] The second input module 530 is used to obtain a task sequence based on the semantic information and the second preset prompt word as input to a preset large model. The task sequence includes N sub-tasks and a task execution order, where N is a positive integer. In one embodiment, the second input module 530 can be used to execute step S230 described above, which will not be repeated here.

[0134] The third input module 540 is used to obtain N sets of structured query statements based on the task sequence and the third preset prompt words as input to the preset large model. In one embodiment, the third input module 540 can be used to execute step S240 described above, which will not be repeated here.

[0135] The query module 550 is used to execute the N sets of structured query statements to query data in a preset database according to the task execution order, and obtain query results. In one embodiment, the query module 550 can be used to execute step S250 described above, which will not be repeated here.

[0136] In the embodiments of this application, the semantic information of the query request is first extracted by calling the large model. Then, the query task is sorted by calling the large model. Next, the structured statement is output by calling the large model. Finally, the data in the database can be directly queried by executing the structured statement. The beneficial effects are: 1. High degree of automation: Users only need to input the query request and do not need to understand the database structure or query syntax; the system automatically completes the entire process of understanding, decomposition, and generation. 2. Strong scalability: By replacing or adjusting three preset prompt words, it can adapt to different fields and different databases without modifying the core process. 3. Good fault tolerance: Complex tasks are broken down into multiple sub-tasks to generate query statements separately; an error in a single sub-task does not affect the retry or repair of other sub-tasks. 4. Controllable response: Since the large model calling task is clear at each stage, timeout and retry strategies can be set separately, improving overall stability.

[0137] According to an embodiment of this application, wherein the first query request is a multimodal input, the apparatus further includes: a preprocessing module, configured to convert the first query request into text form; and to perform preprocessing on the text form of the first query request to obtain a standardized first query request.

[0138] According to an embodiment of this application, the first input module is specifically used to standardize and map the non-standard terms in the first query request based on the standardized indicator names in the preset knowledge graph to obtain a second query request; and to extract entity information and constraint information based on the second query request and the first preset prompt words through a preset large model.

[0139] According to an embodiment of this application, the second input module is specifically used to obtain N sub-tasks based on the entity information, the constraint information, and the second preset prompt word as input to a preset large model, wherein the sub-tasks include corresponding context information; determine the execution priority of the N sub-tasks; and generate the task sequence based on the N sub-tasks and the execution priority.

[0140] According to an embodiment of this application, the third input module obtains the structured query statement based on the context information and the corresponding third preset prompt word as input for any subtask using a preset large model.

[0141] According to an embodiment of this application, the query module is specifically used to perform syntax verification and security review on the N sets of structured query statements; if the syntax verification and the security review are both passed, evaluate the execution cost of the N sets of structured query statements; and if the execution cost is greater than a preset cost threshold, select distributed execution; or if the execution cost is less than the preset cost threshold, select single-machine execution.

[0142] According to an embodiment of this application, the device further includes: a personalized recommendation module, configured to construct a user profile based on historical data generated from the user's historical query behavior, with the user's authorization; generate a recommendation list based on the user profile and a preset access control model; and push the recommendation list to the user interface for the user to select.

[0143] According to embodiments of this application, any plurality of modules among the receiving module 510, the first input module 520, the second input module 530, the third input module 540, and the query module 550 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least part of the functionality of one or more of these modules can be combined with at least part of the functionality of other modules and implemented in one module. According to embodiments of this application, at least one of the receiving module 510, the first input module 520, the second input module 530, the third input module 540, and the query module 550 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array, a programmable logic array, a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit, or any other reasonable means of integrating or packaging the circuitry, or implemented in software, hardware, or firmware, or in any one of the three implementation methods, or in a suitable combination of any of them. Alternatively, at least one of the receiving module 510, the first input module 520, the second input module 530, the third input module 540, and the query module 550 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0144] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a data query method according to an embodiment of this application.

[0145] like Figure 6 As shown, an electronic device 1200 according to an embodiment of this application includes a processor 1201, which can perform various appropriate actions and processes according to a program stored in a read-only memory 1202 or a program loaded from a storage portion 1208 into a random access memory 1203. The processor 1201 may include, for example, a general-purpose microprocessor, an instruction set processor and / or an associated chipset and / or a dedicated microprocessor. The processor 1201 may also include onboard memory for caching purposes. The processor 1201 may include a single processing unit or multiple processing units for executing different steps of the method flow according to an embodiment of this application.

[0146] Random access memory 1203 stores various programs and data required for the operation of electronic device 1200. Processor 1201, read-only memory 1202, and random access memory 1203 are interconnected via bus 1204. Processor 1201 executes various steps of the method flow according to embodiments of this application by executing programs in read-only memory 1202 and / or random access memory 1203. It should be noted that the programs may also be stored in one or more memories other than read-only memory 1202 and random access memory 1203. Processor 1201 may also execute various steps of the method flow according to embodiments of this application by executing programs stored in said one or more memories.

[0147] According to embodiments of this application, the electronic device 1200 may further include an input / output interface 1205, which is also connected to the bus 1204. The electronic device 1200 may also include one or more of the following components connected to the input / output interface 1205: an input section 1206 including a keyboard, mouse, etc.; an output section 1207 including a cathode ray tube, liquid crystal display, etc., and a speaker, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card, such as a local area network card, modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the input / output interface 1205 as needed. A removable medium 1211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 1210 as needed so that computer programs read from it can be installed into the storage section 1208 as needed.

[0148] Embodiments of this application also provide a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0149] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory, portable compact disk read-only memory, optical storage devices, magnetic storage devices, or any suitable combination thereof. In embodiments of this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include the read-only memory 1202, and / or random access memory 1203, and / or one or more memories other than read-only memory 1202 and random access memory 1203 described above.

[0150] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to cause the computer system to implement the methods provided in the embodiments of this application.

[0151] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 1209, and / or installed from the removable medium 1211. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0152] In embodiments of this application, the computer program can be downloaded and installed from a network via communication section 1209, and / or installed from removable medium 1211. When the computer program is executed by processor 1201, it performs the functions defined in the system of this application embodiment. According to embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0153] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0154] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0155] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

Claims

1. A data query method, characterized in that, The method includes: Receive the first query request from the user; Based on the first query request and the first preset prompt words as input to the preset large model, semantic information is obtained; Based on the semantic information and the second preset prompt word as input to the preset large model, a task sequence is obtained. The task sequence includes N sub-tasks and the task execution order, where N is a positive integer. Based on the task sequence and the third preset prompt words as input to the preset large model, N sets of structured query statements are obtained; and According to the task execution order, the N sets of structured query statements are executed to query data in the preset database and obtain query results.

2. The method according to claim 1, characterized in that, in, The first query request is a multimodal input. After receiving the first query request from the user, and before obtaining semantic information based on the first query request and the first preset prompt words as input to the preset large model, the method further includes: Convert the first query request into text format; and The first query request in text form is preprocessed to obtain a standardized first query request.

3. The method according to claim 1, characterized in that, The semantic information obtained based on the first query request and the first preset prompt words as input to the preset large model includes: Based on the standardized indicator names in the preset knowledge graph, the non-standard terms in the first query request are standardized and mapped to obtain the second query request; and Based on the second query request and the first preset prompt, entity information and constraint information are extracted through a preset large model.

4. The method according to claim 3, characterized in that, The task sequence, obtained by using the semantic information and the second preset prompt word as input to the preset large model, includes: Based on the entity information, the constraint information, and the second preset prompt word as input to the preset large model, N sub-tasks are obtained, and each sub-task includes corresponding context information. Determine the execution priority of N subtasks; and The task sequence is generated based on the N sub-tasks and the execution priority.

5. The method according to claim 4, characterized in that, in, The process, based on the task sequence and the third preset prompt word as input to the preset large model, yields N sets of structured query statements, including: For any subtask, the structured query statement is obtained by using the context information and the corresponding third preset prompt word as input to the preset large model.

6. The method according to claim 1, characterized in that, The step involves executing the N sets of structured query statements according to the task execution order to query data in a preset database and obtain query results, including: Perform syntax validation and security audit on the N sets of structured query statements; If the syntax check and the security review both pass, evaluate the execution cost of the N sets of structured query statements; and If the execution overhead is greater than a preset overhead threshold, distributed execution is selected; or if the execution overhead is less than the preset overhead threshold, single-machine execution is selected.

7. The method according to any one of claims 1-6, characterized in that, Before receiving the first query request from the user, the method further includes: With user authorization, a user profile is constructed based on historical data generated from the user's past query behavior. Based on the user profile and the preset access control model, a recommendation list is generated; and The recommended list is pushed to the user interface for the user to select.

8. A data query device, characterized in that, The device includes: The receiving module is used to receive the first query request from the user. The first input module is used to obtain semantic information based on the first query request and the first preset prompt word as input to the preset large model. The second input module is used to obtain a task sequence based on the semantic information and the second preset prompt word as input to the preset large model. The task sequence includes N sub-tasks and the task execution order, where N is a positive integer. The third input module is used to obtain N sets of structured query statements based on the task sequence and the third preset prompt words as input to the preset large model; and The query module is used to execute the N sets of structured query statements according to the task execution order to query data in the preset database and obtain query results.

9. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 7.