Application code generation method, device and equipment based on low-code platform

Abstract

The invention provides an application program code generation method based on a low-code platform. The method can be applied to the technical field of artificial intelligence and software development. The method comprises the following steps: identifying an application programming interface in a development environment of a low-code platform; integrating the identified application programming interface to a low-code platform, and generating a configuration file; generating a recommendation table according to the configuration file and the feature information of the user of the low-code platform; selecting an application programming interface suitable for the low-code platform, the application program and the user according to the recommendation table; and according to the configuration file corresponding to the selected application programming interface, integrating the selected application programming interface into the application program to generate an application program code. The invention further provides an application program code generation device and equipment, a storage medium and a program product.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence and software development technology, and in particular to an application code generation method, device, electronic device, storage medium and program product based on a low-code platform. Background Art

[0002] With the acceleration of digital transformation, enterprises and developers have put forward higher requirements for the development efficiency and flexibility of applications. As an emerging development tool, low-code platform has emerged. It allows developers to quickly build applications through a visual interface and a small amount of code writing, significantly shortening the development cycle and lowering the technical threshold. In the application development process of low-code platform, integrating application programming interface has become a key means to enrich application functions. Application programming interface can provide a variety of professional services, such as payment processing, map navigation, data analysis, etc., avoiding the tedious process of developers developing these functions from scratch. However, in the face of massive application programming interface resources, how to efficiently identify, select and integrate application programming interfaces suitable for specific applications and user needs in low-code platform has become a difficult problem to be solved. Traditional application programming interface selection methods often rely on the experience of developers and manual review and analysis of application programming interface documents. This method is not only inefficient, but also easily affected by the limitations of personal knowledge of developers, resulting in the selected application programming interface may not be the optimal solution. In addition, the lack of comprehensive consideration of user characteristics, application context and low-code platform characteristics may cause the integrated application programming interface to have problems such as poor compatibility and poor user experience in actual use.

[0003] In recent years, with the development of artificial intelligence technology, some methods based on machine learning and deep learning have begun to be applied to the field of application programming interface recommendation. However, most existing methods focus on single-dimensional analysis, such as only considering the functional similarity of application programming interfaces or the historical usage behavior of users, and fail to fully integrate the complex correlation between the characteristics of application programming interfaces themselves, low-code platform information, application context, and multiple aspects of user characteristics. For example, some recommendation systems recommend new application programming interfaces only based on the frequency of users' past use of application programming interfaces, while ignoring the business scenarios of the current applications and the restrictions on application programming interface selection imposed by the technical architecture of the low-code platform. Summary of the invention

[0004] In view of the above problems, the embodiments of the present application provide an application code generation method, device, electronic device, medium and program product based on a low-code platform.

[0005] According to the first aspect of the present application, a method for generating application code based on a low-code platform is proposed, comprising the following steps: in the development environment of the low-code platform, identifying an application programming interface; integrating the identified application programming interface into the low-code platform, and generating a configuration file, wherein the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform; generating a recommendation table based on the configuration file and characteristic information of the user of the low-code platform, wherein the recommendation table includes recommendation information of the application programming interface; selecting an application programming interface suitable for the low-code platform, the application and the user according to the recommendation table; and integrating the selected application programming interface into the application according to the configuration file corresponding to the selected application programming interface to generate application code, wherein generating the recommendation table based on the configuration file and the characteristics of the user of the low-code platform comprises: using a hypergraph neural network model to process the configuration file and the user characteristics of the low-code platform to obtain the recommendation information of the application programming interface.

[0006] According to an embodiment of the present application, a hypergraph neural network model is used to process configuration files and user features of a low-code platform to obtain recommendation information of an application programming interface, specifically including: extracting application programming interface feature information related to the application programming interface from the configuration file; obtaining low-code platform feature information and context feature information in the application code where the application programming interface to be integrated is located from the low-code platform; obtaining feature information of users of the low-code platform, wherein the user feature information includes user attribute information and user behavior feature information; constructing an application programming interface hypergraph based on the application programming interface feature information, low-code platform feature information, context feature information, user attribute information and user behavior feature information, wherein the hyperedges in the application programming interface hypergraph reflect the association between at least two of the application programming interface feature information, low-code platform feature information, context feature information, user attribute information and user behavior feature information; and processing the application programming interface hypergraph using a hypergraph neural network model to generate recommendation information of the application programming interface.

[0007] According to an embodiment of the present application, an application programming interface hypergraph is constructed based on application programming interface feature information, low-code platform feature information, context feature information, user attribute information and user behavior feature information, specifically including: constructing an independent application programming interface node for each application programming interface to be recommended; constructing at least one virtual node for the low-code platform; constructing at least one environment node for the context environment in the application code where the application programming interface to be integrated is located; and classifying users based on user attribute information and user behavior feature information, and constructing an independent user node for each type of user.

[0008] According to an embodiment of the present application, an application programming interface hypergraph is constructed based on application programming interface feature information, low-code platform feature information, context feature information, user attribute information and user behavior feature information, and specifically includes: constructing a first hyperedge between at least two application programming interface nodes based on the application programming interface function attributes in the application programming interface feature information; constructing a second hyperedge between the application programming interface node and at least one of the user nodes and a virtual node based on the low-code platform feature information and the application programming interface feature information; constructing a third hyperedge between the application programming interface node and the environment node based on the context feature information and the application programming interface feature information; constructing a fourth hyperedge between the user node and the environment node based on the context feature information and the user behavior feature information; constructing a fifth hyperedge between the application programming interface node and the user node based on the application programming interface feature information, user attribute information and user behavior feature information; and constructing a sixth hyperedge between at least two user nodes based on the user attribute information and the user behavior feature information.

[0009] According to an embodiment of the present application, an application programming interface hypergraph is constructed based on application programming interface feature information, low-code platform feature information, context feature information, user attribute information and user behavior feature information, and specifically includes: integrating application programming interface nodes, virtual nodes, environment nodes and user nodes as well as the first hyperedge, the second hyperedge, the third hyperedge, the fourth hyperedge, the fifth hyperedge and the sixth hyperedge to form a complete application programming interface hypergraph structure, in which the hyperedges between each node can reflect the association relationship between at least two of the application programming interface feature information, low-code platform feature information, context feature information, user attribute information and user behavior feature information; and optimizing the application programming interface hypergraph structure to form an application programming interface hypergraph.

[0010] According to an embodiment of the present application, a hypergraph neural network model is used to process an application programming interface hypergraph to generate recommendation information of the application programming interface, specifically including: extracting and integrating each node feature from the application programming interface node, virtual node, environment node and user node to form an initial node feature vector for each node; performing information aggregation according to each hyperedge relationship in the application programming interface hypergraph; for each node, using the information obtained from the information aggregation to update the initial node feature vector to form an updated node feature vector; and using multi-layer hypergraph convolution and feature propagation of the hypergraph neural network model to process the updated node feature vector to generate recommendation information of the application programming interface.

[0011] According to an embodiment of the present application, a recommendation table is generated based on characteristic information of a configuration file and a user of a low-code platform, and specifically includes: generating a recommendation table based on recommendation degree information of an application programming interface, application programming interface characteristic information, low-code platform characteristic information and contextual characteristic information, wherein the recommendation table includes recommendation degree information in numerical form and application programming interface characteristic information, low-code platform characteristic information and contextual characteristic information in textual form.

[0012] According to an embodiment of the present application, an application programming interface suitable for a low-code platform, an application program and a user is selected according to a recommendation table, specifically including: organizing and encoding various information in the recommendation table to form a recommendation feature vector; and processing the recommendation feature vector using a recommendation model to output an application programming interface suitable for a low-code platform, an application program and a user, wherein the recommendation model includes at least one of a multilayer perceptron, a convolutional neural network or a recurrent neural network.

[0013] According to an embodiment of the present application, based on the configuration file corresponding to the selected application programming interface, the selected application programming interface is integrated into the application to generate application code, specifically including: obtaining and parsing the configuration file corresponding to the selected application programming interface; extracting key parameters closely related to the application programming interface integration from the parsed configuration file; in the application code, based on the request method and key parameter format requirements of the application programming interface, generating code for constructing requests, receiving and processing responses returned by the application programming interface; integrating the functions of the application programming interface into the business logic of the application; and generating complete application code including the application programming interface integration through the code generation function of the low-code platform.

[0014] According to the second aspect of the present application, an application code generation device based on a low-code platform is also proposed, including: an identification module, used to identify an application programming interface in the development environment of the low-code platform; an integration module, used to integrate the identified application programming interface into the low-code platform and generate a configuration file, wherein the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform; a recommendation table generation module, used to generate a recommendation table based on the configuration file and the characteristic information of the user of the low-code platform, wherein the recommendation table includes the recommendation degree information of the application programming interface, and generating the recommendation table based on the configuration file and the characteristics of the user of the low-code platform includes: using a hypergraph neural network model to process the configuration file and the user characteristics of the low-code platform to obtain the recommendation degree information of the application programming interface; a selection module, used to select an application programming interface suitable for the low-code platform, the application and the user according to the recommendation table; and a code generation module, used to integrate the selected application programming interface into the application according to the configuration file corresponding to the selected application programming interface to generate application code.

[0015] According to the third aspect of the present application, an electronic device is also proposed, comprising: one or more processors; 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 any one of the above methods.

[0016] According to a fourth aspect of the present application, a computer-readable storage medium is also proposed, on which a computer program or instruction is stored, and when the computer program or instruction is executed by a processor, the steps of any one of the above methods are implemented.

[0017] According to the fifth aspect of the present application, a computer program product is also proposed, including a computer program or instructions, which implement the steps of any one of the above methods when the computer program or instructions are executed by a processor. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above contents and other purposes, features and advantages of the present application will become more apparent through the following description of the embodiments of the present application with reference to the accompanying drawings, in which:

[0019] Figure 1 An application scenario diagram of an application code generation method based on a low-code platform according to an embodiment of the present application is schematically shown;

[0020] Figure 2 A flowchart of a method for generating application code based on a low-code platform according to an embodiment of the present application is schematically shown;

[0021] Figure 3A detailed flow chart schematically illustrates the steps of integrating an application programming interface and generating a configuration file in a method according to an embodiment of the present application;

[0022] Figure 4 A detailed flow chart of generating a recommendation table in a method according to an embodiment of the present application is schematically shown;

[0023] Figure 5 A detailed flow chart of generating recommendation degree information using a hypergraph neural network model in a method according to an embodiment of the present application is schematically shown;

[0024] Figure 6 A detailed flow chart of constructing an application programming interface hypergraph in a method according to an embodiment of the present application is schematically shown;

[0025] Figure 7 A schematic diagram showing the structure of an application programming interface hypergraph constructed according to the method of an embodiment of the present application is shown;

[0026] Figure 8 A further detailed flow chart of generating recommendation degree information by using a hypergraph neural network model in a method according to an embodiment of the present application is schematically shown;

[0027] Fig. 9 A detailed flow chart of the step of selecting an application programming interface according to a recommendation table in a method according to an embodiment of the present application is schematically shown;

[0028] Fig.10 A detailed flow chart schematically shows the steps of integrating a selected application programming interface into an application in a method according to an embodiment of the present application;

[0029] Fig.11 A block diagram schematically illustrates an application code generation device based on a low-code platform of the present application; and

[0030] Fig.12 A block diagram of an electronic device suitable for implementing an application code generation method based on a low-code platform according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0031] Below, embodiments of the present application will 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 the present application. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present application.

[0032] The terms used herein are only for describing specific embodiments and are not intended to limit the present application. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0033] In the technical solution of this application, the user information involved (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with relevant laws, regulations and standards, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation entrances are provided for users to choose to authorize or refuse. In the scenario of using personal information for automated decision-making, the methods, devices and systems provided in the embodiments of this application provide users with corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process is entered. The expression "automated decision-making" here refers to the activity of automatically analyzing and evaluating personal behavior habits, interests and hobbies or economic, health, credit status, etc. through computer programs, and making decisions. The expression "expert decision-making" here refers to the activity of making decisions by people who specialize in a certain field, have specialized experience, knowledge and skills and have reached a certain level of professionalism.

[0034] First, the technical terms recorded in this article are explained and illustrated as follows. Low-code platform: A rapid software or application development platform that allows users to create applications using a graphical interface and model-driven logic without traditional programming. Application Programming Interface (API): It is essentially a set of rules and protocols that define how different components in a software system interact with each other. These components can be different software applications, software libraries, or different modules in the system. Hypergraph neural network: A technology in the field of neural networks specifically used to process hypergraph structured data. Hypergraph is a generalized graph structure in which hyperedges can connect multiple nodes, which is different from traditional graphs where edges can only connect two nodes. Hypergraph neural networks can effectively use this complex structure of hypergraphs to learn high-order relationships in data. Its working principle is similar to that of traditional neural networks to process ordinary graph structured data, but it is optimized for the characteristics of hypergraphs. In a hypergraph neural network, the representation of a node is updated based on the information of the hyperedges connected to it and the other nodes connected to the hyperedges. For example, through the hypergraph convolution operation, a node can aggregate feature information from multiple adjacent nodes (connected by hyperedges), and this aggregation process takes into account the complex relationships contained in the hyperedges. This capability enables hypergraph neural networks to perform well in processing data with complex relationships, such as group relationships in social networks, multi-entity associations in knowledge graphs, and complex interactive relationships between application programming interfaces and users and platforms in the embodiments of the present application. Hypergraph: An extended form of a normal graph that allows edges (hyperedges) to connect any number of vertices (nodes). The edges of a normal graph can only connect two vertices, but the hypergraph breaks this limitation, allowing it to more flexibly represent complex many-to-many relationships. For example, in a knowledge graph scenario, a hyperedge can connect multiple related knowledge points (nodes), which may together constitute a knowledge topic. For another example, in a recommendation system, a hyperedge can connect a group of users (with similar purchasing behaviors or interests) and a group of products (these products often appear at the same time in the purchase history of these users), thereby better representing the complex association relationship between users and products. Hyperedge: The core element in a hypergraph, which is an edge connecting multiple nodes. The set of nodes connected by a hyperedge represents a combination of entities with common properties or mutual associations.

[0035] The embodiment of the present application provides an application code generation method based on a low-code platform, comprising the following steps: in the development environment of the low-code platform, identifying an application programming interface; integrating the identified application programming interface into the low-code platform, and generating a configuration file, wherein the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform; generating a recommendation table according to the configuration file and the characteristic information of the user of the low-code platform, wherein the recommendation table includes the recommendation degree information of the application programming interface; selecting an application programming interface suitable for the low-code platform, the application and the user according to the recommendation table; and integrating the selected application programming interface into the application according to the configuration file corresponding to the selected application programming interface to generate application code, wherein generating the recommendation table according to the configuration file and the characteristics of the user of the low-code platform includes: processing the configuration file and the user characteristics of the low-code platform using a hypergraph neural network model to obtain the recommendation degree information of the application programming interface. In the embodiment of the present application, the hypergraph neural network mines high-order relationships based on the comprehensive configuration file and user characteristics, accurately quantifies the recommendation degree, adapts the low-code platform, the application and the user needs, and improves the selection accuracy of the application programming interface. Moreover, it can improve development efficiency, reduce the cost of manually screening application programming interfaces, smoothly integrate according to configuration files, accelerate the development process, enhance application quality and user experience, ensure functional coordination and compatibility, and meet the personalized needs of users. In addition, it is conducive to subsequent maintenance and expansion, and configuration files help maintenance. The Supergraph neural network recommendation system is scalable and can be continuously optimized with changes in data, platforms, and application programming interfaces, adapt to business and technology updates, and extend the application life cycle.

[0036] Figure 1 An application scenario diagram of an application code generation method based on a low-code platform according to an embodiment of the present application is schematically shown.

[0037] like Figure 1As shown, the application scenario 100 according to this embodiment may include a usage scenario of application code generation based on a low-code platform. The network 104 is used to provide a medium for 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, wireless communication links or optical fiber cables, etc. The 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 messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, mailbox clients, social platform software, etc. (only as examples). The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices with display screens and supporting web browsing, including but not limited to smartphones, tablet computers, laptop portable computers, desktop computers, etc. The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device. It should be noted that the application code generation method based on the low-code platform provided in the embodiment of the present application can generally be executed by the server 105. Accordingly, the application code generation device based on the low-code platform provided in the embodiment of the present application can generally be set in the server 105. The application code generation method based on the low-code platform provided in the embodiment of the present application can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the application code generation isolation device based on the low-code platform provided in the embodiment of the present application can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0038] The following will be based on Figure 1 The scene described by Figure 2~Figure 10 The application code generation method based on the low-code platform according to the embodiment of the present application is described in detail.

[0039] Reference Figure 2 The application code generation method based on the low-code platform provided in an embodiment of the present application may include steps S10~S50.

[0040] In step S10, in the development environment of the low-code platform, an application programming interface is identified.

[0041] In step S20, the identified application programming interface is integrated into the low-code platform, and a configuration file is generated, wherein the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform.

[0042] In step S30, a recommendation table is generated based on the configuration file and characteristic information of users of the low-code platform, wherein the recommendation table includes recommendation degree information of the application programming interface.

[0043] In step S40, an application programming interface suitable for the low-code platform, application, and user is selected according to the recommendation table.

[0044] In step S50, the selected application programming interface is integrated into the application program according to the configuration file corresponding to the selected application programming interface to generate application program code.

[0045] In an embodiment of the present application, in step S30, generating a recommendation table based on the configuration file and the user characteristics of the low-code platform includes: using a hypergraph neural network model to process the configuration file and the user characteristics of the low-code platform to obtain the recommendation information of the application programming interface.

[0046] The method provided by the embodiment of the present application has at least one of the following beneficial effects: (1) By combining the various parameters and rules of the application programming interface itself in the configuration file and the characteristic information of the low-code platform user to generate a recommendation table, multiple factors are fully considered. It no longer relies solely on a single dimension (for example, only focusing on the application programming interface function or the user's historical usage frequency, etc.) to select the application programming interface, but comprehensively integrates the technical characteristics of the application programming interface, the low-code platform environment and the personalized characteristics of the user, so that the recommended application programming interface is more compatible with the specific low-code platform, application and user. (2) The hypergraph neural network model is used to process relevant information to obtain the recommendation information of the application programming interface. The model can effectively capture the complex high-order relationships between the application programming interface, users and platforms. Compared with traditional methods, it can dig out the deep-level associations hidden in the data, thereby more accurately quantifying the recommendation degree of each application programming interface for the current application scenario. (3) By automatically generating a recommendation table containing recommendation information, developers can directly filter based on the recommendation table, greatly reducing the workload of manually checking and evaluating application programming interfaces one by one, shortening the time spent in the application programming interface selection stage, and accelerating the entire application development process. (4) After selecting the application programming interface, the corresponding configuration file is used to integrate it into the application. Since the configuration file has detailed records of various parameters and rules for integration, the integration process is smoother and more standardized. It reduces integration failures or repeated debugging problems caused by incorrect parameter configuration and unclear rules, further improves the efficiency of integrating application programming interfaces into application code, and enables applications to be developed and put into use faster. (5) Consider the characteristic information of low-code platform users to filter application programming interfaces, so that the application programming interfaces finally integrated into the application are more in line with the usage habits and expectations of the target user group. (6) The generated configuration file records in detail the parameters and rules for integrating the application programming interface with the low-code platform, which not only serves as a guide during the initial integration, but also provides a clear reference for developers during the subsequent maintenance of the application. When the API needs to be upgraded, replaced, or the platform needs to be updated, the key information of the API integration can be quickly understood based on the configuration file, which facilitates accurate adjustment and maintenance, and reduces the maintenance difficulty caused by system changes. (7) The recommendation system constructed by the hypergraph neural network model has a certain degree of scalability. As more user behavior data is accumulated during the use of the application, platform characteristics change, and new APIs emerge, the model can be continuously updated and learned to continuously optimize the recommendation results of the API.This enables more scientific selection and integration of applications when expanding new functions and connecting to new application programming interfaces in the future, ensuring that the applications can flexibly adapt to the needs of business development and technological updates, thereby extending the life cycle of the applications.

[0047] Continue to refer to Figure 2 In some exemplary embodiments, in step S10, first, an analysis is performed based on the application development goals and functional requirements set by the low-code platform. For example, the various functional modules required by the application at the business logic level are analyzed. For example, if an e-commerce application is built, application programming interfaces corresponding to payment functions, order management functions, product display and search functions, etc. are required; if it is a social application, application programming interfaces for user authentication, messaging, friend management, etc. are required. Then, with the help of the application programming interface library equipped by the low-code platform itself or the docking interface with the application programming interface market, a list of preliminary matching application programming interfaces is screened out according to the predetermined functional requirements. When screening, not only is the functional description of the application programming interface considered to be consistent, but also its compatibility information is preliminarily evaluated, including whether the supported operating system, programming language version, and data format are compatible with the low-code platform. At the same time, factors such as the developer reputation, community activity, and version update frequency of the application programming interface are referred to ensure that the identified application programming interface has stability, reliability, and sustainable development, laying a solid foundation for subsequent integration and application. Exemplarily, a document review of the application programming interface can be performed to achieve the identification of the application programming interface. Specifically, the application programming interface has detailed documentation, including the functional description of the interface, the request method, the format and requirements of the request parameters, the data structure of the response, etc. The steps of identifying the interface can be performed by reading these documents. For example, for a map application programming interface, the documentation will explain how to obtain detailed information about a location on the map by sending a request with latitude and longitude parameters. Exemplarily, the interface testing tool can be used for identification. For example, the application programming interface can be tested using the interface testing tool. The request can be constructed according to the document description, sent to the application programming interface endpoint, and the response result can be observed. This helps to verify whether the interface functions as expected, and to understand the performance characteristics of the interface, such as response time, data accuracy, etc. For example, when testing a payment application programming interface, a simulated payment request is sent to check whether the returned payment status code is correct.

[0048] Figure 3 The detailed flow chart of the steps of integrating the application programming interface and generating the configuration file in the method according to the embodiment of the present application is schematically shown. Figure 2 and Figure 3 , the above step S20 may include sub-steps S201~S204.

[0049] In sub-step S201, connection information may be configured. For example, in a low-code development platform, connection information of an application programming interface may be configured according to the requirements of the platform. This may include a uniform resource locator of an application programming interface endpoint, authentication information (e.g., an application programming interface key), and the like.

[0050] In sub-step S202, data mapping and conversion can be performed. For example, analyze the difference between the data structure returned by the application programming interface and the internal data processing mechanism of the low-code platform application. If the data format returned by the application programming interface is in JSON format, and the application expects to display the data in a specific table format, the data mapping conversion is achieved by using the data conversion tool provided by the low-code platform or writing customized script code. Parse the JSON data through the script code, extract key data fields such as name, value, timestamp, etc. according to the application requirements, and reorganize and assign values ​​according to the row and column logic of the table.

[0051] In sub-step S203, event-driven and callback mechanisms can be configured. For example, for an application programming interface with event-driven or callback functions, set the event processing logic and callback receiving endpoint that match it in the low-code platform. Determine the specific uniform resource locator address for receiving callback information in the low-code platform, and register the callback endpoint in the event management module of the platform so that it can effectively monitor real-time event information from the application programming interface. When the application programming interface triggers an event and sends data through the callback uniform resource locator, the low-code platform responds according to the pre-written event handler.

[0052] In an embodiment of the present application, by executing step S30, the functionality of the application programming interface can be embedded into an application built by a low-code development platform, enabling the application to call the services provided by the application programming interface to implement extended functionality; a data channel can be established between the low-code platform and the application programming interface to ensure that data can be accurately and efficiently transmitted between the two, including converting the data in the application into a format that meets the requirements of the application programming interface request and sending it out, and correctly processing the data returned by the application programming interface; the application programming interface can also work in conjunction with other components of the low-code development platform to ensure the consistency and stability of the entire application system.

[0053] Continue to refer to Figure 3 , in sub-step S204, a configuration file is generated.

[0054] Exemplarily, after completing the connection information configuration, data mapping and conversion settings, and event-driven and callback mechanism settings between the application programming interface and the low-code platform, first, key information in these integration processes can be collected. This includes the application programming interface endpoint uniform resource locator and authentication information obtained from the connection information configuration link; data format conversion rules determined from the data mapping and conversion link, such as the mapping relationship between JSON data and platform-specific data structures (such as tables, objects, etc.), data type conversion rules (such as string to number, date format conversion, etc.); and event types and corresponding callback endpoint uniform resource locators, event processing logic (such as display or storage logic corresponding to message push events) recorded from the event-driven and callback mechanism link. Then, the collected information is classified and organized, and grouped according to different functional modules or parameter types, so that a structured configuration file can be generated later. For example, all connection-related parameters are grouped together, data processing-related parameters are grouped together, event processing-related parameters are grouped together, etc. Then, according to the regulations of the low-code platform or the settings of the developer, a suitable configuration file format is selected. These formats are well structured and readable, which is convenient for machine parsing and manual viewing. Then, according to the selected format, a pre-designed configuration file template is applied. The template defines the location, naming conventions, and hierarchical structure of each parameter in the file. Fill the sorted parameter information into the corresponding position according to the template requirements to generate the preliminary configuration file content.

[0055] Optionally, after generating the preliminary configuration file, perform a verification operation. Check whether the syntax of the file is correct and verify whether the parameter values ​​in the file meet the expected range and format requirements. Optimize and adjust the configuration file based on the verification results. If syntax errors or parameter errors are found, correct them in time; if it is found that some parameter settings may affect performance or compatibility, optimize according to the best practices or rules of thumb of the platform.

[0056] In this embodiment, the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform. Exemplarily, the various parameters and rules include but are not limited to the following parameters and rules: (1) connection-related parameters and rules, such as application programming endpoint uniform resource locator, communication protocol, authentication information, request timeout setting, maximum number of concurrent requests, etc.; (2) data format-related parameters and rules, such as request data format, response data format, data mapping rules, data validation rules, etc.; (3) event processing-related parameters and rules, such as event type definition, callback uniform resource locator configuration, event processing logic, etc.

[0057] Figure 4The detailed flow chart of generating a recommendation table in the method according to the embodiment of the present application is schematically shown. In the embodiment of the present application, step S30 may include sub-steps S31 to S32.

[0058] In sub-step S31, the hypergraph neural network model is used to process the configuration files and user characteristics of the low-code platform to obtain the recommendation information of the application programming interface.

[0059] Figure 5 The detailed flow chart of generating recommendation degree information by using a hypergraph neural network model in the method according to the embodiment of the present application is schematically shown. In the embodiment of the present application, step S31 may include sub-steps S311 to S315.

[0060] In sub-step S311, application programming interface feature information related to the application programming interface is extracted from the configuration file. Exemplarily, the application programming interface feature information can be extracted and sorted from the configuration file. Specifically, the application programming interface feature information extracted from the configuration file, such as the functional attributes, interface specifications, performance indicators, security requirements, and version information of the application programming interface. Then, these feature information are classified and sorted so that the relevant information can be clearly used when determining nodes and constructing hyperedges in the future.

[0061] In sub-step S312, low-code platform feature information and context feature information in the application code where the application programming interface to be integrated is located are obtained from the low-code platform. Exemplarily, the obtained low-code platform feature information is analyzed, including but not limited to the technical architecture adopted by the platform (for example, based on cloud computing, microservice architecture or other architectures), the supported programming languages ​​or script types, the built-in components and functional modules provided, the scalability of the platform, and the ability to integrate with external systems, etc., to clarify how these features are associated with the application programming interface and the user, and to prepare for the construction of the hypergraph. Exemplarily, for the context feature information in the application code where the application programming interface to be integrated is located, factors such as the business scenario of the current application (e-commerce transaction scenario, enterprise office scenario or social interaction scenario, etc.), the time pattern when using the application programming interface (for example, some application programming interfaces are frequently used during the daytime on weekdays, while some are frequently used at night), the device environment (mobile terminal, desktop terminal or specific IoT device, etc.) and the geographical location (whether there are regional restrictions or different usage patterns for different regions, etc.) are considered to analyze how these context information affects the relationship between the application programming interface, the user and the platform.

[0062] In sub-step S313, characteristic information of users of the low-code platform is obtained, and the characteristic information of users includes user attribute information and user behavior characteristic information. Exemplarily, user attribute information and user behavior characteristic information are obtained, analyzed and integrated. For example, for user attribute information, basic information of users is sorted out, such as age, gender, occupation, industry, educational background, etc., and possible preference differences for different application programming interfaces among user groups with different attributes are analyzed. At the same time, in-depth mining of user behavior characteristic information is carried out, including the user's history of using application programming interfaces (application programming interfaces used, order of use, frequency of use, etc.), the dwell time on each application programming interface, the operation behavior of the application programming interface (such as whether complex configurations have been performed, whether specific function requests are frequently initiated, etc.), and the evaluation feedback of the application programming interface (scoring, leaving messages, evaluations, etc.), so as to find the connection between user behavior and other types of characteristics.

[0063] In sub-step S314, an application programming interface hypergraph is constructed based on the application programming interface characteristic information, low-code platform characteristic information, contextual feature information, user attribute information and user behavior characteristic information, wherein the hyperedges in the application programming interface hypergraph reflect the association relationship between at least two of the application programming interface characteristic information, low-code platform characteristic information, contextual feature information, user attribute information and user behavior characteristic information.

[0064] In sub-step S315, the application programming interface hypergraph is processed using a hypergraph neural network model to generate recommendation information of the application programming interface.

[0065] In this embodiment, a hypergraph is constructed by integrating multi-source feature information, comprehensively integrating the application programming interface itself, the low-code platform, context and user characteristics, breaking down information silos, making the recommendation basis more sufficient and comprehensive, and improving the accuracy of application programming interface recommendations.

[0066] Figure 6 A detailed flow chart of constructing an application programming interface hypergraph in a method according to an embodiment of the present application is schematically shown. Figure 7 The schematic diagram of the structure of the application programming interface hypergraph constructed according to the method of the embodiment of the present application is shown schematically. Figure 6 and Figure 7 In an embodiment of the present application, step S314 may include sub-steps S3141~S3152.

[0067] In sub-step S3141, an independent application programming interface node 71 is constructed for each application programming interface to be recommended. Exemplarily, each application programming interface to be recommended is included in the application programming interface hypergraph as an independent application programming interface node. The initial attribute of each application programming interface node is set to its corresponding application programming interface feature information, that is, in some exemplary embodiments, the initial feature vector of the application programming interface node can be constructed based on the features related to the application programming interface functional attributes, the performance related features obtained from the configuration file, the version information of the application programming interface, the developer information and other features.

[0068] In sub-step S3142, at least one virtual node 72 is constructed for the low-code platform. In sub-step S3143, at least one environment node 73 is constructed for the context environment in the application code where the application programming interface to be integrated is located. Exemplarily, for the virtual node representing the low-code platform, the platform feature information is extracted, such as the technical architecture adopted by the platform, the external integration methods supported, the number of active users of the platform, etc., and the information is appropriately processed to form a feature vector of the virtual node. For the environment node, the context feature information is extracted, such as the business scenario of the application, the device type, geographic location, time information when using the application programming interface, etc., and the context information is encoded or quantized to form a feature vector of the environment node.

[0069] In sub-step S3144, users are classified according to user attribute information and user behavior characteristic information, and an independent user node 74 is constructed for each type of user. Exemplarily, users are classified according to user attribute information and user behavior characteristic information. Users are divided into different categories using a suitable classification algorithm or based on a pre-set classification rule. For example, users can be classified based on their occupational type and the main functional field of using the application programming interface. Users with similar attribute information and behavior characteristic information can be clustered together to form different categories using a cluster analysis algorithm. After completing the user classification, an independent user node is constructed for each type of user. The attributes of each user node are set to a comprehensive representation of the typical characteristic information of the user of that type. For example, for a certain type of e-commerce industry-related user node, its attributes may include the average age range of the user of that type, the type of device mainly used (e.g., mainly mobile phones), the application programming interface function tags used frequently (e.g., order management, payment, product recommendation, etc.), and the average frequency and success rate of using these application programming interfaces. By constructing user nodes in this way, it is possible to more accurately reflect the complex relationship between different types of users and application programming interfaces and low-code platforms when subsequently constructing the application programming interface hypergraph, providing a strong data foundation for calculating the application programming interface recommendation degree using the hypergraph neural network model.

[0070] In this embodiment, by constructing corresponding nodes for different elements, the role of each part in the system is clearly presented. Nodes are constructed for application programming interfaces, low-code platforms, contexts, and users respectively, making the association between the parties more intuitive and conducive to subsequent analysis. Nodes are constructed for user classification to accurately reflect user differences and commonalities. Overall, it is easy to integrate multivariate information, lay a good foundation for subsequent operations such as using hypergraphs to mine high-order relationships and accurately recommend application programming interfaces, and improve application development efficiency and quality.

[0071] In sub-step S3145, a first hyperedge 710 is constructed between at least two application programming interface nodes according to the application programming interface function attributes in the application programming interface feature information. Exemplarily, first, the application programming interface function attributes corresponding to each application programming interface node can be analyzed. Including detailed classification of application programming interface functions, such as data processing class (such as data filtering, sorting, aggregation and other functions), visualization class (such as generating charts, graphic display, etc.), communication class (such as message sending, receiving, network request, etc.), storage class (such as data storage, reading, caching, etc.). Then, a functional similarity metric is established, for example, a similarity algorithm based on text description can be used to analyze the functional description document of the application programming interface. If the cosine similarity of the functional description documents of the two application programming interfaces in the vector space is high, it indicates that they have certain functional similarities. After analysis, it may be found that the cosine similarity of their functional description vectors reaches a predetermined similarity threshold, and it can be preliminarily judged that they have a strong functional association. For example, in addition to similarity, the functional complementary relationship between application programming interfaces can also be focused on. For example, an image upload application programming interface and an image compression application programming interface have different functions, but they have a sequential and complementary role in the image processing process. The uploaded image may need to be compressed for better storage or transmission. In this case, the first hyperedge should also be constructed. Then, for at least two application programming interface nodes that are determined to be functionally related (similar or complementary), a first hyperedge 710 is constructed to connect them. For example, the weight of the hyperedge reflects the closeness of the functional association between the application programming interfaces.

[0072] In sub-step S3146, a second hyperedge 720 is constructed between at least one of the application programming interface node and the user node and the virtual node according to the low-code platform feature information and the application programming interface feature information. Exemplarily, after extracting the low-code platform feature information, the adaptation of the application programming interface feature information to the platform features is analyzed, including but not limited to: considering the compliance of the application programming interface with the platform security mechanism, such as whether the application programming interface supports the encryption algorithm required by the platform for data transmission, and if it does, it scores higher in security adaptability. Evaluate whether the response time and concurrent processing capability of the application programming interface meet the platform requirements. If the average response time of the application programming interface is within the acceptable range of the platform and the concurrent processing capability matches the platform, it scores higher in performance adaptability. For application programming interface nodes or user nodes with good adaptability and virtual nodes representing low-code platforms, a second hyperedge 720 is constructed. For example, the hyperedge weight reflects the degree of adaptation of the application programming interface or user to the platform.

[0073] In sub-step S3147, a third hyperedge 730 is constructed between the application programming interface node and the environment node according to the context feature information and the application programming interface feature information. Exemplarily, the association between the application programming interface feature information and the context is analyzed. For example, in an e-commerce transaction scenario, the payment application programming interface and the order management application programming interface are used more frequently; in a mobile device environment, some lightweight application programming interfaces may be more commonly used. For example, the hyperedge weight reflects the strength of the association between the application programming interface and the context, for example, the weight calculation can be determined based on the frequency of occurrence or importance of the context factor.

[0074] In sub-step S3148, a fourth hyperedge 740 is constructed between the user node and the environment node based on the context feature information and the user behavior feature information. Exemplarily, the behavioral differences of users in different context environments are analyzed in combination with the context feature information. For example, in a social entertainment scenario, users are more inclined to share the experience of using certain entertainment application programming interfaces, and the frequency of use is higher; in an enterprise office scenario, users' operation behaviors on office application programming interfaces during the daytime on weekdays are more professional and regular, such as frequent document editing, data processing and other operations. For example, the hyperedge weight reflects the degree of correlation between user behavior and context. For example, the weight calculation can be determined based on the frequency or intensity of the user's behavior in a specific context.

[0075] In sub-step S3149, a fifth hyperedge 750 is constructed between the application programming interface node and the user node according to the application programming interface feature information, the user attribute information and the user behavior feature information. Exemplarily, from the aspect of the application programming interface feature information, analyze whether the function of the application programming interface meets the needs implied by the user attributes and the preferences reflected by the user behavior. From the user behavior feature information, analyze the user's frequency of use of different application programming interfaces, satisfaction with use, etc. For example, if a user frequently uses an image editing application programming interface and gives a high evaluation, it means that there is a strong association between the user and the application programming interface. At the same time, considering the user attribute information, for example, users whose occupation is a designer may have a higher demand and usage tendency for professional design application programming interfaces. For application programming interface nodes and user nodes that are clearly associated, the fifth hyperedge 750 is constructed. For example, the hyperedge weight reflects the degree of association between the application programming interface and the user.

[0076] In sub-step S3150, a sixth hyperedge 760 is constructed between at least two user nodes based on the user attribute information and the user behavior feature information. Exemplarily, the similarity between different users is analyzed based on the user attribute information and the user behavior feature information. For at least two user nodes with similarity, a sixth hyperedge 760 is constructed. The hyperedge weight reflects the degree of similarity between users. For example, the weight calculation can be determined comprehensively based on the attribute similarity score and the behavior similarity score.

[0077] In this embodiment, by constructing multiple hyperedges, the complex relationships between nodes are accurately portrayed, such as multi-dimensional relationships such as function, adaptation, environment, and user. The first hyperedge clarifies the connection between application programming interfaces, which is conducive to functional collaboration considerations; each hyperedge jointly helps the hypergraph to fully present the interaction of multiple information. This provides a detailed structure for subsequent hypergraph neural network analysis, making the recommended application programming interface more accurate, thereby improving the rationality of application programming interface selection in low-code platform application development and optimizing the overall performance and adaptability of the application.

[0078] In sub-step S3151, the application programming interface node 71, the virtual node 72, the environment node 73 and the user node 74 as well as the first hyperedge 710, the second hyperedge 720, the third hyperedge 730, the fourth hyperedge 740, the fifth hyperedge 750 and the sixth hyperedge 760 are integrated to form a complete application programming interface hypergraph structure. In the application programming interface hypergraph structure, the hyperedges between each node can reflect the association relationship between at least two of the application programming interface feature information, low-code platform feature information, context feature information, user attribute information and user behavior feature information.

[0079] In sub-step S3152, the application programming interface hypergraph structure is optimized to form an application programming interface hypergraph 700. Exemplarily, in this optimization step, redundant hyperedges or isolated nodes can be removed to simplify the hypergraph structure while ensuring its effective expression of various relationships. At the same time, the weights of the hyperedges and the attribute representations of the nodes can be dynamically adjusted and optimized based on feedback data in actual applications (such as subsequent recommendation effect evaluation, etc.), so that the application programming interface hypergraph can more accurately reflect the actual association situation and provide a better data basis for using the hypergraph neural network model to calculate the recommendation degree.

[0080] In this embodiment, it is possible to effectively integrate multivariate information, integrate and optimize each node and hyperedge, and form a complete hypergraph structure. Through hyperedge association, the complex relationship between different feature information is clearly presented, laying a solid foundation for the subsequent use of hypergraphs for in-depth analysis. The optimized hypergraph is more reasonable, which helps the hypergraph neural network to accurately mine information, improve the accuracy of hypergraph-based application programming interface recommendations, promote better adaptation to business scenarios and user needs in low-code platform application development, and improve overall development efficiency and application quality.

[0081] Figure 8 A further detailed flow chart of generating recommendation degree information using a hypergraph neural network model in a method according to an embodiment of the present application is schematically shown. In an embodiment of the present application, step S315 may include sub-steps S3153 to S3156.

[0082] Combined with reference Figures 5 to 8 In sub-step S3153, each node feature is extracted and integrated from the application programming interface node, virtual node, environment node and user node to form an initial node feature vector for each node. In sub-step S3154, information aggregation is performed according to each hyperedge relationship in the application programming interface hypergraph. In sub-step S3155, for each node, the initial node feature vector is updated using the information obtained by information aggregation to form an updated node feature vector. In sub-step S3156, the updated node feature vector is processed using the multi-layer hypergraph convolution and feature propagation of the hypergraph neural network model to generate the recommendation information of the application programming interface.

[0083] In this embodiment, by extracting and integrating node features to form an initial vector, updating the vector based on hyperedge relationship aggregation information, and then processing through multi-layer hypergraph convolution and feature propagation, the complex associations contained in the hypergraph can be fully explored. Thus, the recommendation degree of the application programming interface can be accurately quantified and the subjectivity of the recommendation can be reduced. At the same time, the recommendation efficiency is improved, helping applications to better adapt to platforms, users and application scenarios, and optimizing the overall development and usage experience. In some exemplary embodiments, the constructed application programming interface hypergraph is used as the input data structure of the hypergraph neural network. For example, the node features in the hypergraph can be preprocessed, such as normalizing the numerical features to facilitate the numerical calculation and comparison in the subsequent model training; for text features, unique hot encoding and other methods can be used for conversion so that they can be effectively recognized and processed by the model. The appropriate hypergraph neural network architecture can be selected according to the task requirements and data characteristics, such as the spatial domain hypergraph convolution network or the spectral domain hypergraph convolution network. Determine the basic parameters of the model, such as the number of network layers, the number of neurons in each layer, the initial learning rate, the selection of the activation function, etc. Random initialization can be used to assign small random values ​​to the weights of each layer in the network, laying the foundation for the subsequent training process, so that the model can gradually adjust the weights during training to learn the data pattern in the hypergraph. In the training stage of the hypergraph neural network, the training objectives and loss functions can be defined. For example, the training goal is set to accurately predict the recommendation information of the application programming interface. The actual use of the application programming interface by users can be used as supervision information. If there is a lack of clear annotation data such as user ratings, unsupervised learning can also be used to train based on the distribution law and association relationship of the data, such as by making similar application programming interfaces closer in the vector representation space. Optimize. Select a suitable loss function according to the training goal. In addition, a loss function that is more suitable for the recommendation scenario can also be selected according to the specific situation. It can better learn the relationship between the relative recommendation order between application programming interfaces and prompt the model to output the recommendation result that conforms to the actual preference order of the user. In the calculation process of the hypergraph neural network, a hypergraph convolution operation is performed. Taking the spatial hypergraph convolution as an example, the first stage of information aggregation is to collect information from other nodes associated with the hyperedges connected to it, and fuse the feature information of neighboring nodes into its own node representation according to certain aggregation rules (such as weighted average, etc.). This process enables the node to obtain its local context information in the hypergraph structure. Then, the information update stage is carried out. The node uses the aggregated information combined with its original features to update its own node representation through nonlinear transformations such as activation functions, realize feature propagation and update, and enable the node representation to continuously fuse high-order relationship information in the hypergraph. After multiple layers of hypergraph convolution operations, the node representation can gradually capture the complex associations and mutual influences between multiple factors such as application programming interfaces, users, platforms, and contexts.In the hypergraph convolution process, the node features learned at different levels are fused. For example, the original feature representations of the application programming interface, users, etc. learned by the shallow network are concatenated and weighted summed with the abstract feature representations based on high-order relationships learned by the deep network, so that the final node representation can comprehensively reflect multiple feature information and enrich the description of elements such as the application programming interface and users. According to the defined loss function, the difference between the model prediction result and the actual target is calculated, and then the error is back-propagated from the output layer to each layer of the network through the back-propagation algorithm. The weight parameters of the hypergraph neural network are updated according to a certain optimization algorithm, so that the model gradually reduces the loss and learns more accurate hypergraph data patterns and recommendation prediction capabilities. After completing the training of the hypergraph neural network, for each application programming interface node corresponding to the application programming interface to be recommended, the recommended degree is calculated using the trained model. The recommendation degree can be obtained by performing specific operations on the final representation vector of the application programming interface node in the hypergraph, such as using a simple linear combination method to weight the key elements in the node representation vector (such as dimensions related to functional importance, user preference matching, platform adaptability, etc.) according to pre-set weights to obtain a value representing the recommendation degree; or a more complex method is used, such as further processing the node representation vector through a small multi-layer perceptron to output the recommendation degree score. Exemplarily, the value range of the recommendation degree can be set according to the actual application scenario, for example, set in the interval [0, 1], where 0 means not recommended, 1 means strongly recommended, and the intermediate values ​​represent different degrees of recommendation, which is convenient for users to intuitively understand and compare the recommendations of different application programming interfaces. Optionally, a suitable evaluation index can be used to measure the accuracy and effectiveness of the recommendation information generated by the hypergraph neural network model. According to the results of the evaluation index, if it is found that the model performance does not meet expectations, the hypergraph neural network can be optimized.

[0084] Exemplarily, in sub-step S3154, information aggregation operation is performed according to the hyperedge relationship in the application programming interface hypergraph. For example, for an application programming interface node, feature information of adjacent application programming interface nodes is aggregated through the first hyperedge connected to other application programming interface nodes. For example, weighted aggregation is performed according to the hyperedge weight. For another example, for a user node, context environment feature information is aggregated through the fourth hyperedge (between the user node and the environment node); feature information of similar users is aggregated through the sixth hyperedge (between the user node and other user nodes).

[0085] Exemplarily, in sub-step S3155, node feature update is performed. For example, after completing information aggregation, each node is updated using the aggregated information combined with its own original features. A nonlinear activation function is used to transform the aggregated features to enhance the expression ability of the model.

[0086] Exemplarily, in sub-step S3156, after multi-layer hypergraph convolution and feature propagation, the application programming interface node obtains the final feature vector, which contains rich information about the application programming interface itself and information related to the user, platform, and context. The recommendation degree can be calculated using a method based on vector similarity. Optionally, a multi-layer perceptron can be used to calculate the recommendation degree. The final feature vector of the application programming interface node is input into a pre-trained multi-layer perceptron, which is calculated through multiple hidden layers and finally outputs a recommendation score.

[0087] Return to reference Figure 4 In sub-step S32, a recommendation table is generated based on the recommendation information of the application programming interface, the application programming interface feature information, the low-code platform feature information and the context feature information, wherein the recommendation table includes the recommendation information in numerical form and the application programming interface feature information, the low-code platform feature information and the context feature information in text form. Exemplarily, in addition to the recommendation information of the application programming interface that has been acquired, other information of each application programming interface is also collected. This information can be extracted from the configuration files, application programming interface documents or platform records involved in the previous steps. Integrate user and platform related feature information. In order to make the recommendation table more referenceable, some feature information summaries related to the low-code platform, applications and users can also be selectively added. This information helps to show the adaptability of the application programming interface to the overall environment from a more macro perspective.

[0088] Table 1 is an example of a recommendation table. See Table 1. The header of the recommendation table can contain multiple fields to clearly display various aspects of the application programming interface and its recommendation status. For example, you can set fields such as "application programming interface name", "recommendation score", "function description", "classification", "developer", "version number", "adaptation platform type", and "adaptation application field". According to the fields designed in the header, fill in the corresponding information of each application programming interface row by row. Fill the calculated recommendation information into the "recommendation score" field. For example, the value range is between 0 and 1, and the higher the value, the more recommended it is. The other fields are filled in sequentially according to the information of the corresponding application programming interface collected and sorted before, so that each row fully presents the comprehensive situation of an application programming interface and the degree of recommendation.

[0089] Table 1 Recommended table example

[0090]

[0091] For example, all application programming interfaces can be sorted in descending order according to the recommendation scores, and the application programming interface with the highest recommendation score can be placed at the front of the recommendation list, so that users can quickly browse and filter out the application programming interfaces that are given priority. In this way, the generated recommendation list can be presented to users in an intuitive and organized manner, helping them understand the overall situation of each application programming interface and the relative recommendation level.

[0092] Fig. 9 The detailed flow chart of the step of selecting an application programming interface according to the recommendation table in the method according to the embodiment of the present application is schematically shown. Fig. 9 , step S40 may include sub-steps S401~S402.

[0093] In sub-step S401, various information in the recommendation table is sorted and encoded to form a recommendation feature vector. Exemplarily, the data in the recommendation table can be sorted and encoded and converted into a format suitable for deep learning model input. For example, for classified data, the numerical data is normalized by using a unique hot encoding or embedded encoding method, and then all feature vectors are concatenated as the input vector of the model. In sub-step S402, the recommendation feature vector is processed by a recommendation model to output an application programming interface suitable for the low-code platform, the application and the user, wherein the recommendation model includes at least one of a multi-layer perceptron, a convolutional neural network or a recurrent neural network. In this embodiment, the feature vector is formed by sorting and encoding, and is processed with the help of a recommendation model including a multi-layer perceptron. On the one hand, multiple factors can be considered more scientifically and comprehensively, avoiding the limitations of manual screening and improving the accuracy of the selected application programming interface. On the other hand, the deep learning network model is used to enhance the ability to process complex information, so that the output application programming interface is more in line with the low-code platform, application and user needs, helping to improve the quality of application development and user experience.

[0094] Fig.10 The detailed flow chart of the steps of integrating the selected application programming interface into the application program in the method according to the embodiment of the present application is schematically shown. Fig.10 , step S50 may include sub-steps S501~S505.

[0095] In sub-step S501, the configuration file corresponding to the selected application programming interface is obtained and parsed. For example, the configuration file corresponding to the selected application programming interface is obtained, the configuration file is opened using the corresponding file reading mechanism, and is parsed according to its specific format, and the various parameters and rules therein are extracted and converted into a data structure that can be recognized and processed by the program to facilitate subsequent operations. In sub-step S502, key parameters closely related to the integration of the application programming interface are extracted from the parsed configuration file. For example, the key parameters include the endpoint uniform resource locator of the application programming interface, authentication related information, the request method supported by the application programming interface and the corresponding request parameter format requirements, the response data format returned by the application programming interface, etc.

[0096] Optionally, before sub-step S503, the development environment can also be configured. For example, according to the requirements of the application programming interface and the characteristics of the low-code platform, ensure that the development environment where the application is located meets the integration conditions. For different programming languages ​​​​or development frameworks, configure the relevant dependent library versions to ensure smooth interaction with the selected application programming interface. Optionally, before sub-step S503, network connections and permissions can also be set. For example, configure network connection-related settings in the low-code platform to ensure that the application can access the server where the endpoint uniform resource locator of the application programming interface is located. According to the authentication method of the application programming interface, set the corresponding permission configuration in the security module of the low-code platform.

[0097] In sub-step S503, in the code of the application, according to the request method and key parameter format requirements of the application programming interface, the code for constructing the request, receiving and processing the response returned by the application programming interface is generated. In this sub-step, the application programming interface call logic is implemented. For example, in the code of the application, according to the request method and parameter format requirements of the application programming interface, write code to construct the request. To handle the asynchronous and synchronous situations of the request, for some time-consuming application programming interface requests, an asynchronous request method can be used, and an asynchronous programming mechanism can be used in the code to avoid blocking the main thread of the application and improve the user experience; for simple and fast application programming interface requests, a synchronous request method can be used to send and receive request responses in sequence to ensure clear logic and concise code. For another example, write code to receive and process the response returned by the application programming interface.

[0098] In sub-step S504, the function of the application programming interface is integrated into the business logic of the application. In this sub-step, the business logic integration is realized. The function of the application programming interface is integrated into the original business logic of the application, making it an organic part of the overall function of the application. For example, in an e-commerce application, after integrating the payment application programming interface, in the business logic process of the user placing an order, the code logic for calling the payment application programming interface to perform the payment operation is added.

[0099] In sub-step S505, the code generation function of the low-code platform is used to generate complete application code including the integration of the application programming interface. In this step, after the integration of the application programming interface and the integration of business logic are completed, the code generation function of the low-code platform is used to generate complete application code including the integration of the application programming interface. Optionally, the generated code can be optimized, such as removing redundant code, optimizing resource utilization, performing performance testing, and making targeted adjustments to the code based on the test results, so that the application after the integration of the application programming interface can run efficiently and stably to meet the actual needs of users.

[0100] In this embodiment, the configuration file can be accurately obtained and parsed, and key parameters can be extracted to ensure integration accuracy. The corresponding code is generated according to the request method and parameter requirements, and the interaction with the application programming interface is smoother. The application programming interface function is integrated into the business logic to improve the application function. With the help of the low-code platform code generation function, the complete application code with application programming interface integration can be efficiently produced, which improves the overall application development efficiency, ensures the stable operation of the integrated application, and better meets business and user needs.

[0101] It should be noted that although Figure 2~Figure 10 The steps in the flowchart are shown in sequence as indicated by the arrows, but the steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of the steps, and the steps can be executed in other orders. Figure 2~Figure 10 At least part of the steps in the flowchart shown may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0102] Based on the above-mentioned method for generating application code based on a low-code platform, the embodiment of the present application also provides an application code generating device based on a low-code platform. Fig.11The device is described in detail.

[0103] Fig.11 The following is a block diagram of a low-code platform-based application code generation device of the present application. Fig.11 As shown, an embodiment of the present application provides an application code generation device 1100 based on a low-code platform, including: an identification module 1110, an integration module 1120, a recommendation table generation module 1130, a selection module 1140 and a code generation module 1150. The identification module 1110 is used to identify the application programming interface in the development environment of the low-code platform. In some embodiments, the identification module 1110 can be used to execute step S10 and its sub-steps described above, which are not repeated here. The integration module 1120 is used to integrate the identified application programming interface into the low-code platform and generate a configuration file, wherein the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform. In some embodiments, the integration module 1120 can be used to execute step S20 and its sub-steps described above, which are not repeated here. The recommendation table generation module 1130 is used to generate a recommendation table according to the characteristic information of the configuration file and the user of the low-code platform, wherein the recommendation table includes the recommendation degree information of the application programming interface, and the generation of the recommendation table according to the characteristics of the configuration file and the user of the low-code platform includes: using a hypergraph neural network model to process the configuration file and the user characteristics of the low-code platform to obtain the recommendation degree information of the application programming interface. In some embodiments, the recommendation table generation module 1130 can be used to perform the step S30 and its sub-steps described above, which are not repeated here. The selection module 1140 is used to select an application programming interface suitable for the low-code platform, the application and the user according to the recommendation table; in some embodiments, the selection module 1140 can be used to perform the step S40 and its sub-steps described above, which are not repeated here. The code generation module 1150 is used to integrate the selected application programming interface into the application according to the configuration file corresponding to the selected application programming interface to generate the application code. In some embodiments, the code generation module 1150 can be used to perform the step S50 and its sub-steps described above, which are not repeated here.

[0104] According to an embodiment of the present application, any multiple modules among the identification module 1110, the integration module 1120, the recommendation table generation module 1130, the selection module 1140 and the code generation module 1150 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present application, at least one of the identification module 1110, the integration module 1120, the recommendation table generation module 1130, the selection module 1140 and the code generation module 1150 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of the identification module 1110, the integration module 1120, the recommendation table generation module 1130, the selection module 1140 and the code generation module 1150 may be at least partially implemented as a computer program module, which may perform a corresponding function when executed.

[0105] Fig.12 A block diagram of an electronic device suitable for implementing an application code generation method based on a low-code platform according to an embodiment of the present application is schematically shown.

[0106] like Fig.12As shown, the electronic device 1200 according to the embodiment of the present application includes a processor 1201, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1202 or the program loaded from the storage part 1208 to the random access memory (RAM) 1203. The processor 1201 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a dedicated microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 1201 may also include an on-board memory for cache purposes. The processor 1201 may include a single processing unit or multiple processing units for performing different actions of the method flow according to the embodiment of the present application. In the RAM 1203, various programs and data required for the operation of the electronic device 1200 are stored. The processor 1201, the ROM 1202 and the RAM 1203 are connected to each other through a bus 1204. The processor 1201 performs various operations of the method flow according to the embodiment of the present application by executing the programs in the ROM 1202 and / or the RAM 1203. It should be noted that the program may also be stored in one or more memories other than ROM 1202 and RAM 1203. The processor 1201 may also perform various operations of the method flow according to the embodiment of the present application by executing the program stored in the one or more memories. According to an embodiment of the present application, the electronic device 1200 may also include an input / output (I / O) 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 (I / O) interface 1205: an input part 1206 including a keyboard, a mouse, etc.; an output part 1207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage part 1208 including a hard disk, etc.; and a communication part 1209 including a network interface card such as a LAN card, a modem, etc. The communication part 1209 performs communication processing via a network such as the Internet. The driver 1210 is also connected to the input / output (I / O) interface 1205 as needed. A removable medium 1211, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like, is mounted on the drive 1210 as needed, so that a computer program read therefrom is installed into the storage section 1208 as needed.

[0107] The present application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or it may exist alone and not be assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present application is implemented. According to an embodiment of the present application, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present application, the computer-readable storage medium may include the ROM 1202 and / or RAM 1203 described above and / or one or more memories other than ROM 1202 and RAM 1203.

[0108] The embodiment of the present application also includes a computer program product, which includes a computer program, which includes a program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the application code generation method based on the low-code platform provided in the embodiment of the present application. When the computer program is executed by the processor 1201, the above-mentioned functions defined in the system / device of the embodiment of the present application are executed. According to the embodiment of the present application, the system, device, module, unit, etc. described above can be implemented by a computer program module. In one embodiment, the computer program can rely on tangible storage media such as optical storage devices and magnetic storage devices. In another embodiment, the computer program can also be transmitted and distributed in the form of a signal on a network medium, and downloaded and installed through the communication part 1209, and / or installed from a removable medium 1211. The program code contained in the computer program can be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1209, and / or installed from the removable medium 1211. When the computer program is executed by the processor 1201, the above functions defined in the system of the embodiment of the present application are executed. According to the embodiment of the present application, the system, device, apparatus, module, unit, etc. described above can be implemented by a computer program module.

[0109] According to the embodiments of the present application, the program code for executing the computer program provided by the embodiments of the present application can be written in any combination of one or more programming languages. Specifically, these computer programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server.

[0110] It will be appreciated by those skilled in the art that the features described in the various embodiments of the present application may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments of the present application may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present application.

Claims

1. A method for generating application code based on a low-code platform, characterized in that: The following steps are involved: Identify application programming interfaces in the development environment of low-code platforms; Integrate the identified application programming interface into the low-code platform and generate a configuration file, wherein the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform; Generate a recommendation table according to the configuration file and characteristic information of users of the low-code platform, wherein the recommendation table includes recommendation degree information of the application programming interface; According to the recommendation table, selecting an application programming interface suitable for the low-code platform, the application, and the user; and Integrate the selected application programming interface into the application program according to the configuration file corresponding to the selected application programming interface to generate the application program code, Among them, generating a recommendation table based on the configuration file and the user characteristics of the low-code platform includes: using a hypergraph neural network model to process the configuration file and the user characteristics of the low-code platform to obtain the recommendation information of the application programming interface.

2. The method according to claim 1, characterized in that The using of the hypergraph neural network model to process the configuration file and the user features of the low-code platform to obtain the recommendation information of the application programming interface specifically includes: extracting application programming interface characteristic information related to the application programming interface from the configuration file; Obtaining low-code platform feature information and context feature information in the application code where the application programming interface to be integrated is located from the low-code platform; Acquire characteristic information of users of the low-code platform, wherein the characteristic information of users includes user attribute information and user behavior characteristic information; Constructing an application programming interface hypergraph based on the application programming interface characteristic information, the low-code platform characteristic information, the contextual characteristic information, the user attribute information, and the user behavior characteristic information, wherein the hyperedges in the application programming interface hypergraph reflect the association relationship between at least two of the application programming interface characteristic information, the low-code platform characteristic information, the contextual characteristic information, the user attribute information, and the user behavior characteristic information; and The application programming interface hypergraph is processed using a hypergraph neural network model to generate recommendation information of the application programming interface.

3. The method according to claim 2, characterized in that The constructing of an application programming interface hypergraph according to the application programming interface characteristic information, the low-code platform characteristic information, the context characteristic information, the user attribute information and the user behavior characteristic information specifically includes: Constructing a separate API node for each API to be recommended; Building at least one virtual node for the low-code platform; constructing at least one environment node for a context environment in the application code in which the application programming interface to be integrated is located; and The user attribute information and the user behavior feature information are used to classify the users, and an independent user node is constructed for each type of user.

4. The method according to claim 3, characterized in that The constructing of an application programming interface hypergraph according to the application programming interface characteristic information, the low-code platform characteristic information, the context characteristic information, the user attribute information and the user behavior characteristic information further specifically includes: constructing a first hyperedge between at least two application programming interface nodes according to the application programming interface function attributes in the application programming interface feature information; constructing a second hyperedge between the virtual node and at least one of the application programming interface node and the user node based on the low-code platform characteristic information and the application programming interface characteristic information; constructing a third hyperedge between the application programming interface node and the environment node based on the context feature information and the application programming interface feature information; constructing a fourth hyperedge between the user node and the environment node according to the context feature information and the user behavior feature information; constructing a fifth hyperedge between the application programming interface node and the user node according to the application programming interface characteristic information, the user attribute information and the user behavior characteristic information; and A sixth hyperedge is constructed between at least two of the user nodes according to the user attribute information and the user behavior feature information.

5. The method according to claim 4, characterized in that The constructing of an application programming interface hypergraph according to the application programming interface characteristic information, the low-code platform characteristic information, the context characteristic information, the user attribute information and the user behavior characteristic information further specifically includes: Integrate the application programming interface node, the virtual node, the environment node, the user node, and the first hyperedge, the second hyperedge, the third hyperedge, the fourth hyperedge, the fifth hyperedge, and the sixth hyperedge to form a complete application programming interface hypergraph structure, in which the hyperedges between the nodes can reflect the association relationship between at least two of the application programming interface feature information, the low-code platform feature information, the context feature information, the user attribute information, and the user behavior feature information; and The application programming interface hypergraph structure is optimized to form the application programming interface hypergraph.

6. The method according to claim 4 or 5, characterized in that: The step of processing the application programming interface hypergraph using a hypergraph neural network model to generate recommendation information of the application programming interface specifically includes: extracting and integrating node features from the application programming interface node, the virtual node, the environment node, and the user node to form an initial node feature vector for each node; Aggregating information according to each hyperedge relationship in the application programming interface hypergraph; For each node, using the information obtained by information aggregation to update the initial node feature vector to form an updated node feature vector; and The updated node feature vector is processed using multi-layer hypergraph convolution and feature propagation of the hypergraph neural network model to generate recommendation degree information of the application programming interface.

7. The method according to claim 6, characterized in that The generating of a recommendation table according to the configuration file and the characteristic information of the user of the low-code platform specifically includes: A recommendation table is generated based on the recommendation degree information of the application programming interface, the application programming interface feature information, the low-code platform feature information and the context feature information, wherein the recommendation table includes the recommendation degree information in numerical form and the application programming interface feature information, the low-code platform feature information and the context feature information in text form.

8. The method according to claim 7, characterized in that The selecting, according to the recommendation table, an application programming interface suitable for the low-code platform, the application, and the user specifically includes: Arranging and encoding various information in the recommendation table to form a recommendation feature vector; and The recommendation feature vector is processed using a recommendation model to output an application programming interface suitable for the low-code platform, the application, and the user, wherein the recommendation model includes at least one of a multilayer perceptron, a convolutional neural network, or a recurrent neural network.

9. The method according to claim 7 or 8, characterized in that: The step of integrating the selected application programming interface into the application program according to the configuration file corresponding to the selected application programming interface to generate the application program code specifically includes: Obtain and parse the configuration file corresponding to the selected application programming interface; Extract key parameters closely related to API integration from the parsed configuration files; In the application code, according to the request method and key parameter format requirements of the application programming interface, generate codes for constructing requests, receiving and processing responses returned by the application programming interface; Integrate the functionality of the application programming interface into the business logic of the application; and Generate complete application code including application programming interface integration through the code generation function of the low-code platform.

10. An application code generation device based on a low-code platform, characterized in that: include: An identification module for identifying application programming interfaces in the development environment of a low-code platform; An integration module, used to integrate the identified application programming interface into the low-code platform and generate a configuration file, wherein the configuration file records various parameters and rules for integrating the application programming interface with the low-code platform; A recommendation table generation module, for generating a recommendation table according to the configuration file and the characteristic information of the user of the low-code platform, wherein the recommendation table includes the recommendation degree information of the application programming interface, and the generating of the recommendation table according to the configuration file and the characteristics of the user of the low-code platform includes: using a hypergraph neural network model to process the configuration file and the user characteristics of the low-code platform to obtain the recommendation degree information of the application programming interface; a selection module, configured to select an application programming interface suitable for the low-code platform, the application, and the user according to the recommendation table; and The code generation module is used to integrate the selected application programming interface into the application program according to the configuration file corresponding to the selected application programming interface, so as to generate the application program code.

11. An electronic device, comprising: one or more processors; a memory for storing one or more computer programs, It is characterized in 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 9.

12. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

13. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

Images