Method, system, device and medium for generating a business query language

Abstract

The application relates to the field of natural language processing and large-scale language models, and proposes a business query language generation method, system, device and medium, the method comprising the following steps: receiving a natural language query sentence input by a user, preprocessing the natural language query sentence, and converting the natural language query sentence into structured information; calling a business entity compression structure service, obtaining a compression structure of a corresponding business entity based on the structured information, optimizing the compression structure, and integrating the compression structure into a large model; obtaining user permission information, extracting permission restrictions in the user permission information, and inputting the permission restrictions into the large model; based on the optimized and integrated compression structure and the permission restrictions, generating a query sentence of a business query language through the large model, querying according to the query sentence of the business query language, and displaying a query result. Through multi-layer conversion and compression structure optimization, and in combination with permission management, efficient conversion from a natural language to a business query language is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of natural language processing and large-scale language models, in particular to a method, system, device and medium for business query language generation. BACKGROUND

[0002] With the popularity of data-driven business decisions in modern enterprises, it is essential to quickly and accurately extract data from databases. However, existing natural language to structured query language (SQL) conversion techniques have the following shortcomings in dealing with complex queries and permission management: Complex query processing is insufficient: existing methods are difficult to handle complex natural language queries, especially those involving multi-table association and complex conditions. Lack of permission management: methods that directly convert SQL statements often lack effective management of user permissions, which may lead to data leakage and security problems. Low efficiency: existing methods often have low efficiency when dealing with large-scale databases and complex queries.

[0003] With the rapid development of information technology, business query language (BQL) as a query tool for business users greatly reduces the threshold for non-technical users to access databases through a visual interface and simplified query operations. Existing natural language conversion BQL conversion methods rely heavily on fixed templates or rules, making it difficult to handle complex and variable natural language problems, resulting in low conversion accuracy and efficiency. SUMMARY

[0004] Therefore, the present application provides a method, system, device and medium for business query language generation, which solves the accuracy and security problems of large models in generating business query language.

[0005] To achieve the above purpose, the present application provides a method for generating business query language, comprising:

[0006] receiving a natural language query sentence input by a user, preprocessing the natural language query sentence, and converting it into structured information;

[0007] calling a business entity compression structure service, obtaining the compression structure of the corresponding business entity based on the structured information, optimizing the compression structure, and integrating it into a large model;

[0008] obtaining user permission information, extracting permission restrictions from the user permission information and inputting them into a large model;

[0009] based on the optimized and integrated compression structure and the permission restrictions, generating a query sentence of the business query language through a large model, querying and displaying the query results according to the query sentence of the business query language.

[0010] In some embodiments, the structured information includes a business entity ID and information.

[0011] In some embodiments, the step of obtaining a compressed structure of a corresponding business entity based on the structured information comprises:

[0012] retrieving a corresponding table structure in a database according to the business entity ID and identifying key attributes in the table;

[0013] extracting the table structure and key attributes from the database as data of the business entity;

[0014] transforming the data of the business entity and extracting key attributes of each entity, and constructing a compressed structure of the business entity based on the key attributes.

[0015] In some embodiments, the step of optimizing the compressed structure and integrating it into a large model comprises:

[0016] performing data cleaning, feature dimensionality reduction and encoding, and standardization processing on the compressed structure;

[0017] building an input pipeline to convert the standardized compressed structure into a required format of the large model and integrate it according to the input requirements of the large model;

[0018] using the integrated compressed structure as input to train or infer the large model.

[0019] In some embodiments, the step of receiving a natural language query statement input by a user, preprocessing the natural language query statement, and converting it into structured information comprises:

[0020] parsing the natural language query statement input by the user;

[0021] According to the parsed natural language query statement, identifying the user's query intent and mapping the user's query intent to a pre-set fixed content template;

[0022] identifying and extracting entities, attributes and their relationships in the query as structured information.

[0023] In some embodiments, the step of generating a query statement of a business query language by a large model based on the optimized and integrated compressed structure and the permission limit comprises:

[0024] retrieving business entities and attributes from the optimized and integrated compressed structure;

[0025] determining the user's permission range according to the permission limit as a constraint condition for generating the query statement, and filtering inaccessible fields according to the constraint condition;

[0026] Form a complete query statement of the business query language according to the filtered fields and the constraint conditions.

[0027] In some embodiments, the method further comprises: providing context information for the large model through a preset fixed content template when the large model generates the query statement of the business query language.

[0028] The present application proposes a system for generating a business query language, comprising:

[0029] An acquisition unit configured to receive a natural language query statement input by a user, pre-process the natural language query statement, and convert the natural language query statement into structured information;

[0030] A compression unit configured to call a business entity compression structure service, obtain a compression structure of a corresponding business entity based on the structured information, and optimize and integrate the compression structure into a large model;

[0031] A permission unit configured to obtain user permission information, extract permission restrictions in the user permission information, and input the permission restrictions into the large model;

[0032] A query unit configured to generate a query statement of a business query language through the large model based on the optimized and integrated compression structure and the permission restrictions, perform a query according to the query statement of the business query language, and display a query result.

[0033] The present application proposes a computer device, comprising:

[0034] At least one processor; and a memory storing a computer program executable on the processor, the processor executing the program to perform the steps of the method for generating a business query language.

[0035] The present application proposes a computer-readable storage medium storing a computer program, the computer program being executed by a processor to perform the steps of the method for generating a business query language.

[0036] The present application has at least the following beneficial technical effects:

[0037] The application provides a method, system, device and medium for generating a business query language, the method comprising: receiving a natural language query sentence input by a user, preprocessing the natural language query sentence and converting it into structured information; calling a business entity compression structure service, obtaining a compression structure of a corresponding business entity based on the structured information, optimizing the compression structure and integrating it into a large model; obtaining user permission information, extracting permission restrictions in the user permission information and inputting them into the large model; based on the optimized and integrated compression structure and the permission restrictions, generating a query sentence of the business query language through the large model, querying and displaying the query results according to the query sentence of the business query language.

[0038] The application realizes efficient and secure conversion from natural language to business query language through multi-layer conversion and compression structure optimization combined with permission management. The semantics and context of natural language problems can be more accurately understood, thereby improving the conversion accuracy. The construction of the compression structure reduces the amount of data for subsequent processing, making the conversion process more efficient. Accurate BQL query results can be obtained through simple natural language problems without the need to master complex query syntax. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other embodiments can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0040] Figure 1 A method flow chart for generating a business query language provided by the present application;

[0041] Figure 2 A system module diagram for generating a business query language provided by the present application;

[0042] Figure 3 A flow chart of an embodiment of the method for generating a business query language provided by the present application;

[0043] Figure 4 A flow chart of an embodiment of the method for generating a business query language provided by the present application according to the compression structure to generate BQL;

[0044] Figure 5 A structural schematic diagram of an embodiment of a computer device provided by the present application;

[0045] Figure 6 A structural schematic diagram of an embodiment of a computer readable storage medium provided by the present application. DETAILED DESCRIPTION

[0046] In order to make the objects, technical solutions and advantages of the present application clearer, the embodiments of the present application are further described in detail below with reference to the drawings.

[0047] It should be noted that all the expressions of "first" and "second" in the embodiments of the present application are used to distinguish two same name non-same entities or non-same parameters. It can be seen that "first" and "second" are only for the convenience of description, and should not be understood as a limitation of the embodiments of the present application. The subsequent embodiments will not be described one by one.

[0048] The present application proposes a method for generating business query language, please refer to Figure 1 、 Figure 3 and Figure 4 , including:

[0049] S1: receiving a natural language query sentence input by a user, preprocessing the natural language query sentence, and converting it into structured information;

[0050] S2: calling a business entity compression structure service, obtaining a compression structure of a corresponding business entity based on the structured information, optimizing the compression structure, and integrating it into a large model;

[0051] S3: obtaining user permission information, extracting permission restrictions in the user permission information and inputting them into the large model;

[0052] S4: based on the optimized and integrated compression structure and the permission restrictions, generating a query sentence of the business query language through the large model, querying and displaying the query results according to the query sentence of the business query language.

[0053] Further, the present application includes the following steps:

[0054] Multi-layer natural language processing: converting the natural language query input by the user into structured information including business entity ID and information by calling external REST API;

[0055] Business entity compression structure acquisition: obtaining and processing compression structure information of business entities, which is optimized to help improve query performance and efficiency;

[0056] Processing and managing the user's permission information to ensure that the generated query sentence meets the permission control requirements and guarantees data security;

[0057] Providing context information for the large model through fixed content templates to ensure consistency and flexibility of conversion.

[0058] Large model generates BQL (Business Query Language) query: using the powerful natural language processing capability of the large model, structured information and user query are generated into BQL query statements.

[0059] Query result processing and display: the query results are processed and displayed by various methods, including converting the results into string form for user understanding and use.

[0060] The present application realizes efficient and secure natural language to business query language conversion through multi-layer conversion and compression structure optimization combined with permission management. It can more accurately understand the semantics and context of natural language problems, thereby improving conversion accuracy. The construction of the compression structure reduces the data volume of subsequent processing, making the conversion process more efficient. Accurate BQL query results can be obtained through simple natural language questions without the need to master complex query syntax.

[0061] Large model selects large language models (LLMs) with powerful natural language processing (NLP) capabilities, such as GPT, BERT, etc. The query statements of the business query language are queried through the BQL engine and the visual interface. Input the query statement on the visual interface, and the BQL engine will automatically convert the query conditions into SQL statements and execute in the database.

[0062] In some embodiments, please refer to Figure 1 , Figure 3 and Figure 4 , the structured information includes business entity ID and information.

[0063] In the generation process of the business query language, structured information includes business entity ID and information, which is crucial for accurately locating and querying business data. Business entity ID is usually used as the primary key or unique identifier of the data table, while information may include multiple attributes or fields related to the business entity.

[0064] In some embodiments, please refer to Figure 1 , Figure 3 and Figure 4 , the step of obtaining the compression structure of the corresponding business entity based on the structured information includes:

[0065] Retrieving the corresponding table structure in the database according to the business entity ID and identifying the key attributes in the table;

[0066] Extracting the table structure and key attributes from the database as the data of the business entity;

[0067] Convert the data of the business entity and extract the key attributes of each entity, and construct the compression structure of the business entity based on the key attributes.

[0068] In the business query language, the storage and retrieval of large amounts of data can be greatly reduced by using compression structures. Compression structures based on structured information acquisition usually contain rich business logic and data relationships, which help generate more accurate and efficient query statements to meet complex query requirements. The multi-level conversion and index optimization features in compression structures can improve the flexibility of queries, allowing the system to quickly adjust query strategies according to different query conditions and requirements, improving the accuracy and efficiency of queries.

[0069] Compression structures have a unified data format and storage specification, which helps simplify data management and maintenance workflows. As businesses develop and change, more data types and query logic need to be supported. Compression structures based on structured information acquisition have good scalability and upgradeability, making it easy to adapt to these changes.

[0070] In some embodiments, please refer to Figure 1 , Figure 3 and Figure 4 , the step of optimizing the compression structure and integrating it into the large model includes:

[0071] Data cleaning and feature dimensionality reduction and encoding are performed on the compression structure, and standardization processing is performed;

[0072] An input pipeline is constructed to convert the standardized compression structure into the required format of the large model and integrate it according to the input requirements of the large model;

[0073] The integrated compression structure is used as input to train or infer the large model.

[0074] Structured information processing is one of the innovations of the present invention, and its core is to obtain the compression structure information of business entities and use it to optimize the process of converting natural language into business query language. This is achieved through the following steps:

[0075] Receive structured information: Obtain structured information from the NLP conversion module, including business entity ID and information.

[0076] Call compression structure service: Use the business entity compression structure service (such as BECompressEntity service) to obtain the compression structure information of the corresponding business entity.

[0077] Optimize compression structure: Optimize the obtained compression structure information to improve query performance and efficiency.

[0078] Integrate compression structure information: Integrate the optimized compression structure information into the input of the large model to generate BQL queries.

[0079] Compression structure optimization is a key step to ensure efficient querying. By analyzing and optimizing the compression structure, query performance can be significantly improved. Integrating the optimized compression structure information into the input of the large model ensures that the generated BQL query is accurate and efficient.

[0080] In some embodiments, referring to Figure 1 , Figure 3 and Figure 4 , the step of receiving a natural language query sentence from a user, preprocessing the natural language query sentence, and converting it into structured information includes:

[0081] Parsing the natural language query sentence input by the user;

[0082] According to the parsed natural language query sentence, identifying the user's query intent, and mapping the user's query intent to a pre-set fixed content template;

[0083] Identify and extract entities, attributes and their relationships in the query as structured information.

[0084] Through preprocessing steps such as word segmentation, part-of-speech tagging, and syntax analysis, the user's query intent and context can be more accurately understood, which helps the system more accurately parse the user's question and reduce misunderstandings and ambiguities.

[0085] Converting natural language queries into structured information makes query requests more explicit and specific, which helps the system accurately locate the required data or resources and improves query accuracy.

[0086] In some embodiments, referring to Figure 1 , Figure 3 and Figure 4 , the step of generating a query statement of a business query language through a large model based on the optimized and integrated compression structure and the permission limit includes:

[0087] Retrieving business entities and attributes from the optimized and integrated compression structure;

[0088] According to the permission limit, determine the user's permission range as a constraint condition when generating a query statement, and filter inaccessible fields according to the constraint condition;

[0089] Form a complete query statement of the business query language according to the filtered fields and constraint conditions.

[0090] The optimized and integrated compression structure reduces data redundancy and improves data processing efficiency. This allows the large model to access and process relevant data more quickly when generating a query statement, thereby improving query efficiency.

[0091] The authority restriction ensures that only query requests that meet security specifications and data privacy policies are processed. This helps reduce invalid queries and potential security risks, improving the accuracy and security of queries.

[0092] The large model has strong natural language processing and knowledge reasoning capabilities, enabling accurate understanding of user intent and generation of corresponding query statements. Combined with the compression structure and authority restriction, the large model can generate more accurate and efficient query statements. Through the large model query statement generation method, the system can adapt to different query scenarios and needs. Whether it is simple data retrieval or complex business analysis, the system can generate corresponding query statements based on user input and authority restrictions.

[0093] In some embodiments, referring to Figure 1 , Figure 3 and Figure 4 , the method further comprises: when generating a query statement of the business query language by the large model, providing context information for the large model through a preset fixed content template.

[0094] The preset fixed content template helps the model more accurately understand the user's query intent. The keywords, sentence structures and logical relationships in the template can guide the model to generate replies that better meet user expectations. After receiving user input, the model can quickly match with the preset template. This matching process is usually much faster than parsing and understanding natural language from scratch, so it can significantly improve processing efficiency and response speed. The preset fixed content template can be flexibly extended and modified according to business needs. When business scenarios change, only the template needs to be updated, without the need for large-scale retraining of the model. This enhances the adaptability and flexibility of the model.

[0095] The present application proposes a system for generating business query language, referring to Figure 2 , comprising:

[0096] The acquisition unit 100 is configured to receive user input natural language query statements, preprocess the natural language query statements, and convert them into structured information;

[0097] The compression unit 200 is configured to call a business entity compression structure service, obtain a compression structure of a corresponding business entity based on the structured information, optimize the compression structure, and integrate it into the large model;

[0098] The authority unit 300 is configured to obtain user authority information, extract authority restrictions from the user authority information, and input them into the large model;

[0099] The query unit 400 is configured to generate a query statement of a business query language through a large model based on the optimized and integrated compression structure and the permission limit, and query and display a query result according to the query statement of the business query language.

[0100] The system architecture of the application further comprises:

[0101] A user input module is configured to receive a query input by a user in a natural language.

[0102] An NLP conversion module is configured to convert the natural language into structured information.

[0103] A structured information processing module is configured to process the structured information and obtain a business entity compression structure.

[0104] A permission management module is configured to process and manage user permission information.

[0105] A large model generation module is configured to generate a BQL query statement by using a large model.

[0106] A query execution module is configured to execute the generated BQL query statement.

[0107] A result processing module is configured to process and display a query result.

[0108] In the application, when the accuracy of the LLM in converting the natural language into the BQL is not high enough, the natural language query is efficiently and safely converted into the business query language through a multi-layer conversion and optimization process, especially in the improvement of processing complex queries and managing permissions. The application belongs to the field of intelligent large models and solves the accuracy and security problems of the large model in generating the business query language. The system can automatically generate the business query language according to the user input or the preset condition without manually writing complex query statements. This greatly saves time and improves the efficiency of business processing. The system can deeply understand the query intention and demand of the user by using the natural language processing technology and the machine learning algorithm, and generate accurate query statements. This helps to reduce query errors and misunderstandings and improve the accuracy of the query result.

[0109] Based on the same inventive concept, according to another aspect of the application, as Figure 5 shown, the embodiment of the application also provides a computer device 30, which comprises a processor 310 and a memory 320, the memory 320 stores a computer program 321 which can run on the processor, and the processor 310 executes the steps of the method as above when executing the program.

[0110] Based on the same inventive concept, according to another aspect of the application, as Figure 6As shown, the embodiment of the present application further provides a computer readable storage medium 40, which stores a computer program 410 executed by a processor to perform the method as above.

[0111] The embodiment of the present application can also include a corresponding computer device. The computer device includes a memory, at least one processor, and a computer program stored on the memory and executable on the processor, and the processor executes the program to perform any of the above methods.

[0112] The memory is a non-volatile computer readable storage medium, which can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules in the embodiments of the present application. The processor executes the non-volatile software programs, instructions and modules stored in the memory to perform various functional applications and data processing of the device, i.e. to implement the above method.

[0113] The memory can include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required by a function; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage device. In the embodiments, the memory can optionally include a memory remotely arranged with respect to the processor, and these remote memories can be connected to the local module through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0114] Finally, it should be noted that those skilled in the art can understand that all or part of the above-mentioned embodiment methods can be implemented by a computer program to instruct related hardware to complete, and the program can be stored in a computer readable storage medium. When the program is executed, it can include the flow of the above-mentioned embodiments of each method. The storage medium of the program can be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc. The above-mentioned embodiments of the computer program can achieve the same or similar effects as the corresponding any of the above-mentioned method embodiments.

[0115] Those of skill would further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the disclosure herein can be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans can implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present embodiments.

[0116] The above are only exemplary embodiments of the present embodiments, but it should be noted that various changes and modifications can be made without departing from the scope of the present embodiments defined by the claims. The functions, steps and / or actions of the method claims described above need not be performed in any particular order. Unless explicitly stated otherwise, the ordinal use of terms such as first, second, etc., in the description above does not reflect an ordinal use of the terms in the description of the embodiments. Furthermore, although the elements of the present embodiments can be described or claimed in individual forms, unless expressly limited to individual form, multiple elements can be substituted for a single element of an embodiment. The embodiments of the present embodiments disclosed above can be implemented in the form of software, firmware, hardware, or any combination thereof.

[0117] It should be understood that, as used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising", or "includes" and / or "including" when used herein, specify the presence of stated features, integers, steps, operations, elements, and / or components but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0118] Those of ordinary skill in the art will appreciate that the above discussions related to various embodiments are merely illustrative and not intended to suggest restrictive implementations of the present embodiments. Implementations of the present embodiments disclosed above can be implemented in any of the various forms of software, firmware, hardware, or a combination thereof. The various embodiments disclosed above can be combined or combined in any manner to provide other embodiments of the present embodiments. Therefore, the above embodiments should not be construed as limiting the scope of the present embodiments, but merely as illustrative.

Claims

1. A method of generating a business query language, characterized by, Comprise: Receiving a natural language query sentence input by a user, preprocessing the natural language query sentence, and converting it into structured information; The structured information includes business entity ID and information; Call the business entity compression structure service, obtain the compression structure of the corresponding business entity based on the structured information, wherein the corresponding table structure is retrieved in the database according to the business entity ID, and the key attributes in the table are identified; Extract the table structure and key attributes from the database as the data of the business entity; Convert the data of the business entity and extract the key attributes of each entity, and based on the key attributes, the compression structure of the business entity is formed, and the compression structure is optimized and integrated into a large model; Obtain user permission information, extract the permission restrictions in the user permission information and input them into the large model; Based on the optimized and integrated compression structure and the permission restrictions, generate a query sentence of a business query language through a large model, query and display the query results according to the query sentence of the business query language.

2. The method of generating a business query language according to claim 1, wherein, The step of optimizing the compression structure and integrating it into the large model comprises: Data cleaning, feature dimensionality reduction and coding are performed on the compression structure, and standardization processing is performed; Build an input pipeline to convert the standardized compression structure into the required format of the large model and integrate it according to the input requirements of the large model; Use the integrated compression structure as input to train or infer the large model.

3. The method of generating a business query language according to claim 1, wherein, The step of receiving a natural language query sentence input by a user, preprocessing the natural language query sentence, and converting it into structured information comprises: Parse the natural language query sentence input by the user; According to the parsed natural language query sentence, identify the query intent of the user, map the query intent of the user to a preset fixed content template, and identify and extract the entities, attributes and their relationships in the query as structured information. The step of generating a query sentence of a business query language through a large model based on the optimized and integrated compression structure and the permission restrictions comprises:

4. The method of generating a business query language according to claim 3, wherein, Retrieve the business entity and attribute from the optimized and integrated compression structure; Determine the user's permission range according to the permission restrictions, and use it as a constraint condition when generating a query sentence, and filter the inaccessible fields according to the constraint condition; Form a complete query sentence of the business query language according to the filtered fields and constraint conditions. Also includes:

5. The method of generating a business query language according to claim 4, wherein, When generating a query sentence of a business query language through a large model, provide context information for the large model through a preset fixed content template. Comprise:

6. A system for generating a business query language, characterized by An acquisition unit configured to receive a natural language query sentence input by a user, preprocess the natural language query sentence, and convert it into structured information; The structured information includes business entity ID and information, ​ The compression unit is configured to invoke a business entity compression structure service, obtain a compression structure of a corresponding business entity based on the structured information, retrieve a corresponding table structure in a database according to the business entity ID, and identify key attributes in the table; extract the table structure and the key attributes from the database as data of the business entity; convert the data of the business entity and extract key attributes of each entity, construct a compression structure of the business entity based on the key attributes, optimize the compression structure, and integrate the compression structure into a large model; The permission unit is configured to obtain user permission information, extract permission restrictions in the user permission information, and input the permission restrictions into the large model; The query unit is configured to generate a query statement of a business query language through the large model based on the optimized and integrated compression structure and the permission restrictions, query according to the query statement of the business query language, and display a query result.

7. A computer device, comprising: at least one processor; and a memory storing a computer program executable on the processor, characterized in that the processor executes the program to execute the steps of the method for generating the business query language according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium comprising: The computer program is executed by the processor to execute the steps of the method for generating the business query language according to any one of claims 1 to 5.

Images