Data service publishing method, system and equipment of data platform, and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and particularly provides a data service publishing method, system and device for a data-in-data table and a storage medium, and the method comprises the steps: generating a service demand document based on the logical relationship between business data of the data-in-data table and user interaction data through employing a large language model; generating a data service interface file based on a pre-constructed template according to the service demand document; the service demand document comprises a data service type, a data service object and a data service target. The data service interface can be automatically generated, the automation degree is improved, and the development time of the service interface is shortened.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and specifically relates to a data service publishing method, system, device and storage medium for a data middle station. Background Art

[0002] In the current data-driven era, enterprises rely on data platforms to integrate data from different sources to support business decisions. However, the data service publishing process often faces many challenges. Traditional methods mainly rely on manual demand collection, which is not only inefficient but also prone to misunderstandings, resulting in the data services finally released not meeting actual needs.

[0003] With the continuous advancement of artificial intelligence technology, especially the development of deep learning and natural language processing, many companies have begun to explore the use of AI to improve the efficiency and accuracy of demand analysis. Existing solutions usually focus on using AI for information extraction, pattern recognition, etc., but they are still insufficient in terms of how to systematically guide users to express their needs. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a data service publishing method, system, device and storage medium of a data middle platform to solve the above-mentioned technical problems.

[0005] In a first aspect, the present invention provides a data service publishing method of a data middle station, comprising: Generate service requirement documents using the big language model based on the logical relationship between business data and user interaction data in the data center; Generate data service interface files based on pre-built templates according to service requirement documents; The service requirement document includes data service type, data service object and data service target.

[0006] In an optional implementation, a service requirement document is generated using a large language model based on the logical relationship between business data and user interaction data of the data center, including: Standardize the business data of the data center to build a data dictionary containing the logical relationships of the data tables; Decomposing the generated service requirement document into multiple subtasks, and generating a chain prompt based on the multiple subtasks and user interaction data; The chained prompt is input into a pre-trained large language model, which generates a service requirement document based on a data dictionary.

[0007] In an optional implementation, the business data of the data center is standardized to construct a data dictionary containing the logical relationship of the data tables, including: Classify the business data of the data center according to the preset data types; Establishing data standards, assigning identifiers to business data based on the data standards, and generating associations between data tables based on the identifiers; A data dictionary is generated according to the association between data types and data tables, and the data dictionary includes the meaning of the fields of each table, the data type, and the association with other tables.

[0008] In an optional implementation, generating a service requirement document is decomposed into a plurality of subtasks, and a chain prompt is generated based on the plurality of subtasks and user interaction data, including: Set the initial prompt as an open-ended question, and the initial goal is to obtain the type of data service the user needs; Acquire user feedback data on the initial prompt, extract key information from the feedback data, and generate a subsequent prompt based on the key information; Use reinforcement learning algorithms to analyze user input and feedback in real time, and adjust the prompt's questioning method based on the analysis results; According to the key information of the user feedback data corresponding to the prompt, the task of generating the service requirement document is decomposed into multiple subtasks, each subtask corresponds to a prompt; Arrange each prompt in the logical order of the subtasks to form an orderly chain; Set the instruction, background information or constraint information of a single prompt to get a chained prompt.

[0009] In an optional implementation, the chained prompt is input into a pre-trained large language model, and the large language model generates a service requirement document based on a data dictionary, including: Use named entity recognition technology to extract key entities from key information; Use relationship extraction technology to determine the relationship between key entities based on the data dictionary; Extract data source information corresponding to key entities; Generate business SQL statements based on the determined key entities and the relationships between them; Retrieving target data from a corresponding data source according to the data source information by using the business SQL statement; Organize the acquired target data into a structured service requirement document.

[0010] In an optional implementation, generating a data service interface file based on a pre-built template according to a service requirement document includes: Construct a data service interface template that complies with the RESTful specification, wherein the data service interface template includes an interface endpoint template, a request parameter template, and a response format template; Extracting key parameters from the service requirement document using a natural language processing model; Fill the key parameters into the data service interface template; Use programming language to integrate the filled data service interface template into an interface code file, and generate an interface description file that complies with the OpenAPI specification.

[0011] In an optional embodiment, the method further comprises: Confirm that the interface description file has passed manual review and the interface code file has passed testing, and publish the data service interface file to the data center.

[0012] In a second aspect, the present invention provides a data service publishing system for a data middle station, comprising: The demand acquisition module is used to generate service demand documents based on the logical relationship between business data and user interaction data of the data center using a large language model; A file generation module is used to generate a data service interface file based on a pre-built template according to a service requirement document; The service requirement document includes data service type, data service object and data service target.

[0013] In a third aspect, a device is provided, comprising: Storage device, used to store data service publishing program of data center; A processor is used to implement the steps of the data service publishing method of the data middle station as provided in the first aspect when executing the data service publishing program of the data middle station.

[0014] In a fourth aspect, a computer-readable storage medium is provided, on which a data service publishing program of a data middle station is stored. When the data service publishing program of the data middle station is executed by a processor, the steps of the data service publishing method of the data middle station provided in the first aspect are implemented.

[0015] The beneficial effect of the present invention lies in that the data service publishing method, system, device and storage medium of the data middle platform provided by the present invention generate service requirement documents in the form of dialogue, and use artificial intelligence models to automatically generate data service interfaces based on the requirement documents and pre-built templates, thereby improving the degree of automation and shortening the development time of the service interface.

[0016] In addition, the invention has a reliable design principle, a simple structure and a very broad application prospect. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 is a schematic flow chart of a method according to an embodiment of the present invention.

[0019] Figure 2 It is a schematic principle diagram of a method according to an embodiment of the present invention.

[0020] Figure 3 is a schematic block diagram of a system according to an embodiment of the present invention.

[0021] Figure 4 A schematic diagram of the structure of a device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0024] The key terms appearing in the present invention are explained below.

[0025] Prompt means "prompt word" in Chinese. In the AI ​​big model, the main function of Prompt is to give the AI ​​model the context of input information and parameter information of the input model.

[0026] Large Language Model (LLM) refers to a deep learning model trained with a large amount of text data, which enables the model to generate natural language text or understand the meaning of language text. These models can provide in-depth knowledge and language production on various topics by training on huge data sets. The core idea is to learn the patterns and structures of natural language through large-scale unsupervised training, simulating the human language cognition and generation process to a certain extent.

[0027] Natural Language Processing (NLP) is an important research direction in the field of artificial intelligence. It integrates knowledge from multiple disciplines such as linguistics, computer science, machine learning, mathematics, cognitive psychology, etc. It is an interdisciplinary subject that integrates computer science, artificial intelligence and linguistics. It includes two main aspects: natural language understanding and natural language generation. The research content includes multiple levels such as characters, words, phrases, sentences, paragraphs and chapters. It is a bridge for communication between machine language and human language. It aims to enable machines to understand, interpret and generate human language, realize effective communication between humans and machines, and enable computers to perform tasks such as language translation, sentiment analysis, and text summarization.

[0028] The data service publishing method of the data middle platform provided in the embodiment of the present invention is executed by a computer device, and accordingly, the data service publishing system of the data middle platform runs in the computer device.

[0029] Figure 1 is a schematic flow chart of a method according to an embodiment of the present invention. Figure 1 The execution subject may be a data service publishing system of a data middle platform. According to different requirements, the order of the steps in the flowchart may be changed, and some may be omitted.

[0030] like Figure 1 As shown, the method includes: S1, using the big language model based on the logical relationship between the business data of the data center and the user interaction data, to generate a service requirement document, wherein the service requirement document includes the data service type, data service object and data service goal.

[0031] Use the generate_service_requirement_doc function to get the generated service requirement document by sending a prompt containing business logic and user interaction data to the GPT-3.5-turbo model.

[0032] S2, generates data service interface files based on pre-built templates according to the service requirement document.

[0033] Use the generate_data_service_interface_file function to read the pre-built template file and fill in the placeholders in the template according to the content in the service requirement document.

[0034] Finally, write the filled content into the new data service interface file.

[0035] In an embodiment of the present invention, based on step S1, a possible embodiment is given below to illustrate its specific implementation scheme in a non-limiting manner.

[0036] S101. Standardize the business data of the data center to build a data dictionary containing the logical relationships between data tables.

[0037] (1) Classify the business data of the data center according to the preset data type: First, clarify the preset data types, for example, they can be divided into structured data (such as tabular data in relational databases), semi-structured data (such as data in JSON and XML formats), and unstructured data (such as text files, images, audio, etc.).

[0038] For structured data, it is further subdivided according to business areas, such as customer data, product data, order data, financial data, etc.

[0039] For semi-structured and unstructured data, classify them according to their source and purpose. For example, user feedback text is classified as customer-related data, and product images are classified as product-related data.

[0040] (2) Establish data standards, assign identifiers to business data based on the data standards, and generate associations between data tables based on the identifiers: Establish data format standards, such as standardizing the date format to "YYYY - MM - DD" and the telephone number format to "1 [3 - 9]\d {9}".

[0041] Specify data encoding rules, such as using UTF-8 encoding, and adopt a unified encoding system for specific business codes.

[0042] Determine data quality standards, including data accuracy, completeness, consistency, etc. For example, stipulate that the customer name cannot be empty and the order amount must be a positive number.

[0043] Assign a unique identifier to each business data entity, such as customer, product, order, etc. For example, for a customer table, use the prefix "CUST_" followed by an 8-digit serial number as the customer ID; for a product table, use the prefix "PROD_" followed by a 6-digit alphanumeric combination as the product ID.

[0044] Ensure the uniqueness and stability of identifiers throughout the data platform to avoid duplication or conflict.

[0045] Determine the association method between data tables based on business logic and data sharing requirements. For example, the order table and customer table are associated through customer ID, and the order table and product table are associated through product ID.

[0046] Use foreign key constraints to establish and maintain these relationships to ensure data consistency and integrity. For example, add customer ID and product ID as foreign keys in the order table to reference the corresponding primary keys in the customer table and product table respectively.

[0047] (3) Generate a data dictionary based on the relationship between data types and data tables. The data dictionary includes the meaning of each table's fields, data types, and relationships with other tables: Create a data dictionary document that uses a table to record the information of each table in detail.

[0048] For each table, record basic information such as its table name, brief description, and business area.

[0049] For each field in the table, record information such as the field name, field meaning, data type (such as integer, string, date, etc.), whether it is required, and default value.

[0050] Clarify the relationship between each table and other tables, record the associated fields and the type of association (such as one-to-one, one-to-many, many-to-many). For example, the customer table and the order table are in a one-to-many relationship, associated through the customer ID; the order table and the product table are also in a one-to-many relationship, associated through the product ID.

[0051] S102. Decomposing the generated service requirement document into multiple subtasks, and generating a chain prompt based on the multiple subtasks and user interaction data; (1) Set the initial prompt as an open-ended question, and the initial goal is to obtain the type of data service required by the user: Design a clear, concise and guiding initial prompt, such as: "What type of data service do you want to get from the data center? For example, query, analysis, statistics or other types of data services?" Make sure the question stimulates thought and provides valuable feedback without limiting the range of responses.

[0052] (2) Obtaining user feedback data on the initial prompt, extracting key information from the feedback data, and generating a subsequent prompt based on the key information: Get feedback data: Receive the user's response to the initial prompt through the user interface (such as a web form, chat window, etc.).

[0053] Extract key information: Use natural language processing techniques, such as lexical analysis and syntactic analysis, to extract key information from user feedback. For example, if the user answers "I want to analyze the sales data of the past year", then extract key information such as "analysis", "past year", and "sales data".

[0054] Generate follow-up prompts: Generate more targeted follow-up prompts based on the extracted key information. For example, based on the above answer, you can generate "What aspects of sales data do you want to analyze? Is it sales volume, sales volume, or sales trend, etc.?" (3) Use reinforcement learning algorithms to analyze user input and feedback in real time, and adjust the prompt's questioning method based on the analysis results: Build a reinforcement learning model whose input is the user's input and feedback, and whose output is a suggestion for adjusting the way of asking questions.

[0055] Define a reward function to reward users based on the quality of their feedback and how close they are to the goal. For example, if users provide more detailed and clear information about their needs, give them a higher reward; if user feedback is vague or deviates from the goal, give them a lower reward.

[0056] According to the output of the reinforcement learning model, adjust the questioning method of subsequent prompts, such as changing the wording of the question, adding guiding information, etc., to improve the quality and effectiveness of user feedback.

[0057] (4) Based on the key information of the user feedback data corresponding to the prompt, the task of generating the service requirement document is decomposed into multiple subtasks, each of which corresponds to a prompt: Further analyze the key information from user feedback to determine the specific subtasks required to generate the service requirement document. For example, if the user requirement involves the analysis of sales data, the subtasks may include determining the time range for the analysis, selecting the analysis indicators, and determining the data source.

[0058] Design a corresponding prompt for each subtask, and ensure that each prompt guides the user to provide the information required to complete the subtask. For example, for the subtask of determining the time range for analysis, the prompt can be "What is the time range of the sales data you want to analyze? Is it the past month, a quarter, or a year?" (5) Arrange the prompts in the logical order of the subtasks to form an orderly chain: Sort each prompt according to the logical dependencies between the subtasks. For example, first determine the time range for analysis, then select the analysis indicators, and finally determine the data source.

[0059] Make sure each prompt in the chain naturally leads to the next prompt, forming a coherent, step-by-step questioning process.

[0060] (6) Set the instructions, background information or constraint information of a single prompt to obtain a chain prompt: For each prompt, make it clear what the user needs to answer. For example, "Please specify the specific indicators of the sales data you want to analyze, such as sales volume, sales volume, profit margin, etc."

[0061] Add background information or constraint information to each prompt as needed to help users better understand the question and provide accurate answers. For example, "Due to storage limitations of the data center, only sales data from the past three years can be provided. Please select the time range you want to analyze within this range."

[0062] S103. Input the chained prompt into a pre-trained large language model, and the large language model generates a service requirement document based on the data dictionary.

[0063] (1) Use named entity recognition technology to extract key entities from key information: The user feedback data corresponding to the chain prompt is input into the named entity recognition model, which can be a pre-trained model based on deep learning, such as BERT, GPT, etc.

[0064] The model identifies key entities in the text, such as data type (sales data, customer data, etc.), time (past year, last month, etc.), business object (product, order, etc.), operation (analysis, query, etc.), etc.

[0065] The identified key entities are classified and labeled for subsequent processing.

[0066] (2) Use relationship extraction technology to determine the relationship between key entities based on the data dictionary: The extracted key entities are matched with the data dictionary, and the relationship between entities is determined using relationship extraction technology. For example, the relationship between the "order" and "product" entities is determined based on the relationship between the order table and the product table in the data dictionary.

[0067] Analyze the semantic relationship between entities in the text, such as "analyze sales data" indicates the functional relationship between the "analysis" operation and "sales data".

[0068] Create a relationship diagram between key entities to clearly show the connections between them.

[0069] (3) Extract data source information corresponding to key entities: Determine the data source corresponding to the key entity based on the information of each table recorded in the data dictionary. For example, if the key entity is "sales data", the data dictionary shows that it is stored in the sales order table, and the database where the table is located is the data source.

[0070] Record the detailed information of the data source, such as database type (MySQL, Oracle, etc.), server address, port number, database name, table name, etc.

[0071] (4) Generate business SQL statements based on the determined key entities and the relationships between key entities: Construct SQL statements based on key entities and their relationships, as well as data source information. For example, if a user needs to query sales data for the past year, involving key entities such as "sales order table" and "time", the generated SQL statement is as follows: SELECT * FROM sales_order_table WHERE order_date>= '2023 - 01 - 01'AND order_date<= '2023 - 12 - 31'; According to specific needs, add necessary filtering conditions, aggregate functions, connection operations, etc. to ensure that the SQL statement can accurately obtain the required data.

[0072] (5) Using the business SQL statement, according to the data source information, retrieve the target data from the corresponding data source: Use a database connection tool (such as Python's pymysql library for connecting to a MySQL database) to establish a connection with the database based on the data source information.

[0073] Execute the generated SQL statement to retrieve the target data from the database.

[0074] Handle possible abnormal situations, such as database connection failure, SQL statement execution error, etc., and provide corresponding error prompts and handling measures.

[0075] (6) Organize the acquired target data into a structured service requirements document: The retrieved target data is cleaned and preprocessed to ensure the accuracy and consistency of the data.

[0076] According to requirements, organize the data into specific structures, such as tables, JSON formats, etc.

[0077] In the service requirements document, describe in detail the source, processing process, structure and meaning of the data, as well as the correspondence with user needs.

[0078] In an embodiment of the present invention, based on step S2, a possible embodiment is given below to illustrate its specific implementation scheme in a non-limiting manner.

[0079] S201. Construct a data service interface template that complies with the RESTful specification, wherein the data service interface template includes an interface endpoint template, a request parameter template, and a response format template.

[0080] Interface endpoint template: Determine the basic URL structure of the interface, such as / api / v1 as the version prefix, followed by the specific resource path. Define endpoints corresponding to different HTTP methods (GET, POST, PUT, DELETE, etc.) for each resource. For example, for an interface to obtain sales data, the endpoint of the GET method can be / api / v1 / sales_data; for an interface to create a new order, the endpoint of the POST method can be / api / v1 / orders. Define a basic description of the endpoint in the template, including a description of the function, the general format of the request and response, etc.

[0081] Request parameter template: Define the type (such as string, integer, date, etc.), name, required or not, default value, etc. of the request parameter according to different endpoints and business requirements. For complex request parameters, such as objects with multiple fields, you can define them in JSON format. For example, the request parameters for creating an order may include order amount, order date, customer ID, etc. The template can be defined as: { "order_amount": { "type": "number", "required": true, "description": "Order amount"}, "order_date": { "type": "string", "format": "YYYY- MM - DD", "required": true, "description": "Order date"}, "customer_id": { "type": "string", "required": true, "description": "Customer ID"}}.

[0082] Response format template: Define the status code, data structure, and format of the response according to different endpoints and business requirements. For a successful response, define the returned data fields and their types and meanings. For example, a successful response to obtain sales data may contain fields such as total sales and sales details. The template is as follows: { "status": { "type": "string", "description": "Response status, such as 'success' or 'error'"}, "total_sales": { "type": "number", "description": "Total sales"}, "sales_details": { "type": "array", "items": { "type": "object", "properties": { "product_id": { "type": "string", "description": "Product ID"},"quantity": { "type": "number", "description": "Sales quantity"}, "price": { "type": "number", "description": "Product unit price"}}}, "description": "Sales details"}}.

[0083] For error responses, define common error codes and error message formats. For example: { "status": "error", "error_code": { "type": "string", "description":"Error code, such as '400' means request parameter error"}, "error_message": { "type": "string", "description": "Error details"}}.

[0084] S202. Extract key parameters from the service requirement document using a natural language processing model.

[0085] Select a pre-trained BERT model and configure and optimize it appropriately according to the characteristics of the service requirements document and the task requirements. Fine-tune the model using annotated sample data to improve the model's ability to extract key parameters. The annotated sample data includes the text of the service requirements document and the location and type annotations of the key parameters therein.

[0086] Input the service requirement document into the BERT model. The model analyzes the text and identifies key parameters, such as interface endpoints, request parameters, response formats, and other related information. The extracted key parameters are sorted and classified to ensure the accuracy and completeness of the parameters.

[0087] S203. Fill the key parameters into the data service interface template.

[0088] (1) Interface endpoint filling: Fill in the interface endpoint template based on the information about the interface endpoint in the extracted key parameters, such as resource path, HTTP method, etc.

[0089] Ensure that the populated interface endpoints comply with RESTful specifications and business requirements.

[0090] (2) Request parameter filling: Fill the extracted request parameter information, including parameter name, type, required or not, default value, etc., into the request parameter template.

[0091] Check whether the filled request parameters are complete, reasonable, and match the functions of the interface endpoint.

[0092] (3) Response format filling: Fill in the response format template based on the extracted key parameters about the response format, such as status code, data field and its type.

[0093] Ensure that the populated response format accurately reflects the response requirements in the service requirements document.

[0094] S204. Integrate the populated data service interface template into an interface code file using a programming language, and generate an interface description file that complies with the OpenAPI specification.

[0095] According to the project requirements and team technology stack, choose the appropriate programming language and framework to implement the data service interface. For example, use Python and FastAPI framework, Java and Spring Boot framework, etc.

[0096] Write the interface code using the selected programming language and framework based on the filled interface endpoint template, request parameter template, and response format template.

[0097] Implement the logic of the interface in the code, including receiving requests, verifying request parameters, calling business logic processing functions, returning responses, etc. Ensure the readability, maintainability and security of the interface code.

[0098] Use the tools or libraries provided by the framework to generate an interface description file (usually in YAML or JSON format) that complies with the OpenAPI specification based on the interface code.

[0099] In the description file, define the interface path, method, request parameters, response format and other information in detail so that other developers can understand and use the interface.

[0100] Verify the generated interface description file to ensure that it complies with the requirements of the OpenAPI specification.

[0101] Please refer to Figure 2 , the intelligent data service publishing method comprises the following steps: The first step is to establish the data service model: Because the data in the data center is complex, the standardized business data distinction and the logical relationship between the tables require the establishment of a data service standard management mechanism, so that the standardized data service can be combined with the user needs based on the prompt chain to establish a demand model. This module will systematically standardize the business data tables and mapping relationships, and classify them by directory and theme. This facilitates the subsequent call of large models.

[0102] The second step is to establish demand and provide real-time feedback mechanism: This invention uses a pre-trained large language model and its powerful language understanding and generation capabilities to design a chain prompt mechanism. This mechanism guides users to gradually describe their data service needs through a series of questions. The initial questions are designed as open questions, and the subsequent questions are gradually refined according to the user's answers, for example: Initial question: What type of data services do you want to provide? Another question: For which specific business modules do you want to provide data services? In-depth question: What kind of data result information do you hope to obtain for this business module? … In the process of interacting with users, reinforcement learning algorithms are used to analyze user input and feedback in real time, continuously enhancing the accuracy of data requirements.

[0103] The third step is to confirm and generate requirements: After the user input is completed, the information extraction technology is used to analyze the input and extract key information, such as data source, functional requirements, user roles, final business SQL, etc. This process involves a variety of NLP technologies, including named entity recognition, relationship extraction, etc., to ensure that the information obtained is comprehensive and accurate. The generated structured requirements document will be presented to the user in a clear format for confirmation to ensure that the final requirements are fully understood.

[0104] Step 4: Connect AI Agent with standard data service products: After the requirements are confirmed, the structured requirements document is automatically generated through AI Agent to generate a data service API that complies with the RESTful specification. The system uses a template method, combined with the API design pattern of existing standard service products, to generate an API document containing necessary call parameters and format instructions.

[0105] Step 5: Review and Release: The generated data service API will be officially released after being reviewed by R&D personnel. The review content includes interface connectivity testing, security testing, performance evaluation, etc. to ensure that the generated service meets corporate standards. After passing the review, the API will be automatically released to the data center for subsequent use and call.

[0106] The present invention combines the prompt chain with AI Agent and the data middle platform to provide a flexible and efficient data service publishing solution. Its main beneficial effects include: Improved user experience: The data service API is released in the form of dialogue in the data center, which greatly improves the user experience.

[0107] Reduced error rate: Real-time feedback and dynamic adjustment mechanisms enhance the accuracy of demand confirmation and significantly reduce the error rate after release.

[0108] Automation and efficiency improvement: Through automation, APIs are generated and reviewed based on demand and standard template information, which shortens the time for service release, improves availability, and enables enterprises to respond quickly to market changes.

[0109] Decision support: This method not only improves the efficiency of data service publishing, but also provides a reliable data service foundation for enterprise decision support, helping digital transformation.

[0110] In some embodiments, the data service publishing system of the data middle station may include multiple functional modules composed of computer program segments. The computer program of each program segment in the data service publishing system of the data middle station may be stored in the memory of the computer device and executed by at least one processor to execute (see Figure 1 Description) The function of publishing data services in the data center.

[0111] In this embodiment, the data service publishing system of the data middle station can be divided into multiple functional modules according to the functions it performs, such as Figure 3 As shown. The functional modules of the system may include: a demand acquisition module and a file generation module. The module referred to in the present invention refers to a series of computer program segments that can be executed by at least one processor and can complete fixed functions, which are stored in a memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0112] The demand acquisition module is used to generate service demand documents based on the logical relationship between business data and user interaction data of the data center using a large language model; A file generation module is used to generate a data service interface file based on a pre-built template according to a service requirement document; The service requirement document includes data service type, data service object and data service target.

[0113] Figure 4 The data service publishing method of the data middle station provided for the embodiment of the present application can be applied to the device. It can be understood by those skilled in the art that the device structure involved in the embodiment of the present invention does not constitute a limitation on the device, and the device may include more or less components than shown, or combine certain components, or arrange different components. In an embodiment of the present invention, the device includes but is not limited to a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described and / or required herein.

[0114] The device 400 may include: a processor 410, a memory 420 and a communication unit 430. These components communicate via one or more buses. Those skilled in the art will appreciate that the server structure shown in the figure does not limit the present invention, and it may be a bus structure or a star structure, and may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently.

[0115] The memory 420 may be used to store the execution instructions of the processor 410, and the memory 420 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. When the execution instructions in the memory 420 are executed by the processor 410, the device 400 is enabled to perform some or all of the steps in the following method embodiments.

[0116] The processor 410 is the control center of the storage device, and uses various interfaces and lines to connect various parts of the entire electronic device. It runs or executes software programs and / or modules stored in the memory 420, and calls data stored in the memory to perform various functions of the electronic device and / or process data. The processor can be composed of an integrated circuit (IC), for example, it can be composed of a single packaged IC, or it can be composed of a plurality of packaged ICs with the same or different functions. For example, the processor 410 can include only a central processing unit (CPU). In an embodiment of the present invention, the CPU can be a single computing core or multiple computing cores.

[0117] The communication unit 430 is used to establish a communication channel so that the storage device can communicate with other devices, receive user data sent by other devices or send user data to other devices.

[0118] The present invention also provides a computer storage medium, wherein the computer storage medium may store a program, and when the program is executed, the program may include some or all of the steps in each embodiment provided by the present invention. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0119] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes, including several instructions for enabling a computer device (which can be a personal computer, a server, or a second device, a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.

[0120] In this specification, the same or similar parts between the various embodiments can be referred to each other. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method embodiment.

[0121] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of systems or modules, which can be electrical, mechanical or other forms.

[0122] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0123] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0124] Although the present invention has been described in detail with reference to the accompanying drawings and in combination with the preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, a person of ordinary skill in the art may make various equivalent modifications or substitutions to the embodiments of the present invention, and these modifications or substitutions shall be within the scope of the present invention. Any person of ordinary skill in the art may easily think of changes or substitutions within the technical scope disclosed by the present invention, and these shall be within the scope of protection of the present invention.

Claims

1. A data service publishing method for a data middle station, characterized in that: include: Generate service requirement documents using the big language model based on the logical relationship between business data and user interaction data in the data center; Generate data service interface files based on pre-built templates according to service requirement documents; The service requirement document includes data service type, data service object and data service target.

2. The method according to claim 1, characterized in that The big language model is used to generate service requirement documents based on the logical relationship between business data and user interaction data in the data center, including: Standardize the business data of the data center to build a data dictionary containing the logical relationships of the data tables; Decomposing the generated service requirement document into multiple subtasks, and generating a chain prompt based on the multiple subtasks and user interaction data; The chained prompt is input into a pre-trained large language model, which generates a service requirement document based on a data dictionary.

3. The method according to claim 2, characterized in that Standardize the business data of the data center to build a data dictionary containing the logical relationships of the data tables, including: Classify the business data of the data center according to the preset data types; Establishing data standards, assigning identifiers to business data based on the data standards, and generating associations between data tables based on the identifiers; A data dictionary is generated according to the association between data types and data tables, and the data dictionary includes the meaning of the fields of each table, the data type, and the association with other tables.

4. The method according to claim 2, characterized in that: Decompose the generated service requirement document into multiple subtasks, and generate a chain prompt based on the multiple subtasks and user interaction data, including: Set the initial prompt as an open-ended question, and the initial goal is to obtain the type of data service the user needs; Acquire user feedback data on the initial prompt, extract key information from the feedback data, and generate a subsequent prompt based on the key information; Use reinforcement learning algorithms to analyze user input and feedback in real time, and adjust the prompt's questioning method based on the analysis results; According to the key information of the user feedback data corresponding to the prompt, the task of generating the service requirement document is decomposed into multiple subtasks, each subtask corresponds to a prompt; Arrange each prompt in the logical order of the subtasks to form an orderly chain; Set the instruction, background information or constraint information of a single prompt to get a chained prompt.

5. The method according to claim 2, characterized in that: The chained prompt is input into a pre-trained large language model, and the large language model generates a service requirement document based on the data dictionary, including: Use named entity recognition technology to extract key entities from key information; Use relationship extraction technology to determine the relationship between key entities based on the data dictionary; Extract data source information corresponding to key entities; Generate business SQL statements based on the determined key entities and the relationships between them; Retrieving target data from a corresponding data source according to the data source information by using the business SQL statement; Organize the acquired target data into a structured service requirement document.

6. The method according to claim 1, characterized in that Generate data service interface files based on pre-built templates according to service requirement documents, including: Construct a data service interface template that complies with the RESTful specification, wherein the data service interface template includes an interface endpoint template, a request parameter template, and a response format template; Extracting key parameters from the service requirement document using a natural language processing model; Fill the key parameters into the data service interface template; Use programming language to integrate the filled data service interface template into an interface code file, and generate an interface description file that complies with the OpenAPI specification.

7. The method according to claim 6, characterized in that The method further comprises: Confirm that the interface description file has passed manual review and the interface code file has passed testing, and publish the data service interface file to the data center.

8. A data service publishing system for a data middle station, characterized in that: include: The demand acquisition module is used to generate service demand documents based on the logical relationship between business data and user interaction data of the data center using a large language model; A file generation module is used to generate a data service interface file based on a pre-built template according to a service requirement document; The service requirement document includes data service type, data service object and data service target.

9. A data service publishing device for a data middle station, characterized in that: include: Storage device, used to store data service publishing program of data center; A processor, used to implement the steps of the data service publishing method of the data middle station as described in any one of claims 1-7 when executing the data service publishing program of the data middle station.

10. A computer-readable storage medium storing a computer program, characterized in that: The readable storage medium stores a data service publishing program of the data middle station, and when the data service publishing program of the data middle station is executed by the processor, the steps of the data service publishing method of the data middle station as described in any one of claims 1-7 are implemented.

Images