Intelligent low-code platform development system based on large model

Abstract

The invention provides an intelligent low-code platform development system based on a large model, and the system comprises a data processing and storage module which is used for collecting and processing business demand data; the low-code development module is used for visually arranging page layout and business logic to generate configuration metadata and supporting a user to operate based on a dialogue mode; the test module is used for generating, executing and feeding back a test case and ensuring the quality of the developed application program; and the artificial intelligence and reasoning module is used for providing capability based on machine learning and natural language processing and supporting intelligent data processing, content generation and reasoning. And the interaction and agent module is used for supporting a user to interact with the system based on a dialogue mode and realizing automatic task execution. According to the system provided by the invention, the code development efficiency and the application program quality can be remarkably improved, and meanwhile, the development cost is reduced.

Description

Technical Field

[0001] The present application relates to the field of computers, and specifically, to an intelligent low-code platform development system based on a large model. Background Art

[0002] With the rapid development of information technology, it has been widely used in all walks of life, and the demand for the construction of information systems in all walks of life has grown rapidly. The traditional system development process usually includes product managers designing products, outputting requirement documents and prototypes, and developers writing code to implement products. This process is highly dependent on the participation of developers and designers, and the development cycle is long, making it difficult to meet the needs of rapid launch.

[0003] In order to improve the efficiency of front-end development, many companies and teams have begun to adopt self-developed or open source low-code platforms. These platforms encapsulate common methods and components, allowing developers to build applications through simple drag and drop and configuration. However, this solution still has some problems. First, the quality of the generated code varies, and developers often need to further modify and debug it. When the generated code cannot be executed, the code needs to be rewritten, which may reduce development efficiency. Secondly, the configuration items of the low-code platform are not easy to understand, and the configuration threshold is still high, which is a challenge for non-professional developers. These problems limit the effectiveness of the low-code platform in actual applications.

[0004] Therefore, the present application provides an intelligent low-code platform development system based on a large model to solve one of the above-mentioned technical problems. Summary of the invention

[0005] The purpose of this application is to provide an intelligent low-code platform development system based on a large model, which can solve at least one of the technical problems mentioned above. The specific solution is as follows:

[0006] According to the specific implementation of the present application, in the first aspect, the present application provides an intelligent low-code platform development method based on a large model, including:

[0007] The data processing and storage module is used to collect and process business demand data; the low-code development module is used to visually orchestrate page layout and business logic to generate configuration metadata and support user operations based on conversations; the testing module is used to generate, execute and feedback test cases to ensure the quality of developed applications; the artificial intelligence and reasoning module is used to provide modules with capabilities based on machine learning and natural language processing, supporting intelligent data processing, content generation and reasoning; the interaction and agent module is used to support users to interact with the system based on conversations and realize automated task execution.

[0008] In one embodiment, the data processing and storage module includes: a business demand collection module, used to collect the business demand data; wherein the business demand data includes demand descriptions from different carriers, and the composition forms of the demand descriptions include demand documents, sample pages, pictures, audio and / or video files; a business demand data preprocessing module, used to preliminarily format the collected business demand data and establish management relationships by type classification; a business demand data storage module, used to store business demand data in the form of different data carriers such as text, pictures, videos, audio, etc., and maintain the association relationship between data; a business demand data analysis and extraction module, used to analyze business demand data based on machine learning algorithms and extract feature information such as entities, relationships, attributes, etc., to construct a business demand knowledge graph.

[0009] In one implementation, the requirement document includes pictures and document sections, and the business requirement data preprocessing module is further used to: extract the pictures from the requirement document, and establish a relationship between the pictures and the document sections.

[0010] In one embodiment, the low-code development module includes: a low-code page orchestration module, which allows users to create page layouts and interactive elements based on a visual interface and generate corresponding configuration metadata; a low-code business logic orchestration module, which is used to support users to orchestrate business logic based on a visual interface and generate corresponding business logic configuration metadata; a low-code configuration storage module, which is used to format and store the metadata generated by the low-code platform and provide a data source for the low-code configuration analysis and extraction module; a low-code configuration analysis and extraction module, which is used to analyze and extract the metadata generated by the low-code platform based on the machine learning module and provide data for the knowledge graph module.

[0011] In one embodiment, the test module includes: a test case generation module, which is used to generate test cases, test data and scripts for specific business needs; a test case execution module, which is used to provide an automated testing environment, automatically execute test tasks and feedback the results; a test feedback module, which is used to feedback and process the test results, and provide test data for the generative large model.

[0012] In one embodiment, the test case execution module is also used to: record the test data generated during the test process; wherein the test data is used for data analysis and supplementation of the knowledge graph, including test cases, scripts and execution results.

[0013] In one embodiment, the artificial intelligence and reasoning module includes: a machine learning module, which is used to provide capabilities based on machine learning, including convolutional neural networks, generative adversarial networks, three-dimensional convolutional neural networks, and / or natural language processing; a generative large model module, which is used to provide natural language processing, content generation and reasoning capabilities based on the transformer architecture after pre-training and fine-tuning; a knowledge graph module, which is used to build a structured semantic knowledge base using business demand data and low-code configuration data; an enhanced search module, which is used to improve the relevance and accuracy of retrieval results and update the knowledge graph in real time.

[0014] In one embodiment, the enhanced search module is based on a unified knowledge graph to improve retrieval efficiency and accuracy while maintaining the timeliness and relevance of answers.

[0015] In one embodiment, the interaction and agent module includes: a low-code platform agent, used to assist users in performing conversational operations and automated task execution; a manual annotation module, used to manually intervene in the data extracted by the business requirements analysis extraction module, the low-code configuration analysis extraction module, and the test feedback module.

[0016] In one embodiment, the low-code platform intelligent agent supports user conversational operations based on natural language processing, content generation and reasoning capabilities, and realizes automated task execution based on the platform interface.

[0017] Compared with the prior art, the above scheme of the embodiment of the present application has at least the following beneficial effects: through the collaborative work of the data processing and storage module, the low-code development module, the test module, the artificial intelligence and reasoning module, and the interaction and intelligent agent module, the system can efficiently collect and process business demand data, support visual arrangement of page layout and business logic, automatically generate and execute test cases, provide intelligent capabilities based on machine learning and natural language processing, and interact with users through dialogue to realize automated task execution. The system provided by this application can significantly improve code development efficiency and application quality, while reducing development costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A block diagram of a large-model-based intelligent low-code platform development system is shown;

[0019] Figure 2 A block diagram of a data processing and storage module is shown;

[0020] Figure 3 A block diagram of the composition of a low-code development module is shown;

[0021] Figure 4 A block diagram of the composition of a test module is shown;

[0022] Figure 5 A block diagram of the components of an artificial intelligence and reasoning module is shown;

[0023] Figure 6 A block diagram of the components of an interaction and agent module is shown;

[0024] Figure 7 A complete structural diagram of an intelligent low-code platform based on a large model is shown;

[0025] Figure 8 A business demand processing flow chart is shown;

[0026] Fig. 9 Another business requirement processing flow chart is shown;

[0027] Fig.10 A business requirement low-code testing flow chart is shown. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0029] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings, and "multiple" generally includes at least two.

[0030] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0031] It should be understood that although the terms first, second, third, etc. may be used to describe in the embodiments of the present application, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.

[0032] As used herein, the words "if" and "if" may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0033] It should also be noted that the term "includes", "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, so that a commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprising a ..." do not exclude the existence of other identical elements in the commodity or device including the elements.

[0034] It should be particularly noted that any symbols and / or numbers in the specification that are not marked in the accompanying drawings are not drawing marks.

[0035] The optional embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0036] The embodiment provided in this application is an embodiment of an intelligent low-code platform development method based on a big model.

[0037] Combine the following Figure 1 The embodiments of the present application are described in detail.

[0038] Figure 1 The following is a block diagram of a large model-based intelligent low-code platform development system. Figure 1 As shown, it includes a data processing and storage module 10, a low-code development module 11, a testing module 12, an artificial intelligence and reasoning module 13, and an interaction and agent module 14.

[0039] The data processing and storage module 10 is used to collect and process business demand data.

[0040] The low-code development module 11 is used to visually arrange page layout and business logic to generate configuration metadata and support user operations based on dialogue.

[0041] The testing module 12 is used to generate, execute and provide feedback on test cases to ensure the quality of the developed application.

[0042] The artificial intelligence and reasoning module 13 is used to provide the module with capabilities based on machine learning and natural language processing, supporting intelligent data processing, content generation and reasoning.

[0043] The interaction and agent module 14 is used to support users to interact with the system based on dialogue and realize automated task execution.

[0044] In the embodiment of the present application, through the collaborative work of the data processing and storage module 10, the low-code development module 11, the testing module 12, the artificial intelligence and reasoning module 13, and the interaction and intelligent agent module 14, the system can efficiently collect and process business demand data, support visual arrangement of page layout and business logic, automatically generate and execute test cases, provide intelligent capabilities based on machine learning and natural language processing, and interact with users through dialogue to achieve automated task execution. This integrated approach significantly improves development efficiency and application quality while reducing development costs.

[0045] Figure 2 A block diagram of the composition of a data processing and storage module 10 is shown.

[0046] In one embodiment, if Figure 2 As shown, the data processing and storage module 10 includes a business demand collection module 101 , a business demand data preprocessing module 102 , a business demand data storage module 103 , and a business demand data analysis and extraction module 104 .

[0047] The business requirement collection module 101 is used to collect business requirement data, wherein the business requirement data includes requirement descriptions from different carriers, and the composition forms of the requirement descriptions include requirement documents, sample pages, pictures, audio and / or video files.

[0048] The business demand data preprocessing module 102 is used to preliminarily format the collected business demand data and establish management relationships by type classification.

[0049] The business demand data storage module 103 is used to store business demand data in different data carrier forms such as text, picture, video, audio, etc., and maintain the association relationship between data.

[0050] The business requirement data analysis and extraction module 104 is used to analyze the business requirement data based on a machine learning algorithm and extract characteristic information such as entities, relationships, and attributes to construct a business requirement knowledge graph.

[0051] In the embodiment of the present application, the data processing and storage module 10 includes a business demand collection module 101, a preprocessing module, a storage module and a data analysis and extraction module, which can comprehensively collect demand descriptions from different carriers (such as documents, pictures, audio and video), and perform preliminary formatting, classification management, storage and analysis on these data to build a business demand knowledge graph. This not only ensures the integrity and consistency of the data, but also provides a solid data foundation for subsequent low-code development and testing, and improves the overall performance and accuracy of the system.

[0052] In this application, the requirements document may include pictures and document sections.

[0053] In one implementation, the business requirement data preprocessing module 102 is further used to extract images from the requirement document and to establish a relationship between the images and document sections.

[0054] In the embodiment of the present application, the business demand data preprocessing module 102 can extract images from the demand document and establish the relationship between the images and the document sections, which helps to more accurately understand the content of the demand document and improve the accuracy of demand analysis. In this way, the system can better identify and process key information in the demand document, thereby generating a more accurate knowledge graph and providing more reliable data support for subsequent development and testing.

[0055] Figure 3 A composition block diagram of a low-code development module 11 is shown.

[0056] In one embodiment, if Figure 3 As shown, the low-code development module 11 includes a low-code page orchestration module 111, a low-code business logic orchestration module 112, a low-code configuration storage module 113, and a low-code configuration analysis and extraction module 114.

[0057] The low-code page orchestration module 111 allows users to create page layouts and interactive elements based on a visual interface and generate corresponding configuration metadata.

[0058] The low-code business logic orchestration module 112 is used to support users in orchestrating business logic based on a visual interface and generate corresponding business logic configuration metadata.

[0059] The low-code configuration storage module 113 is used to format and store the metadata generated by the low-code platform, providing a data source for the low-code configuration analysis and extraction module 114.

[0060] The low-code configuration analysis and extraction module 114 is used to analyze and extract the metadata generated by the low-code platform based on the machine learning module 131, and provide data for the knowledge graph module 133.

[0061] In the embodiment of the present application, the low-code development module 11 includes a page layout module, a business logic layout module, a configuration storage module, and a configuration analysis and extraction module, allowing users to create page layouts and business logic through a visual interface and generate corresponding configuration metadata. These modules not only simplify the development process and improve development efficiency, but also support further analysis and extraction of the generated metadata, providing data support for the knowledge graph module 133, thereby realizing a more intelligent development process.

[0062] Figure 4 A block diagram of the components of a test module 12 is shown.

[0063] In one embodiment, if Figure 4 As shown, the test module 12 includes a test case generation module 121 , a test case execution module 122 , and a test feedback module 123 .

[0064] The test case generation module 121 is used to generate test cases, test data and scripts for specific business requirements.

[0065] The test case execution module 122 is used to provide an automated testing environment, automatically execute test tasks and feed back results.

[0066] The test feedback module 123 is used for feedback and processing of test results, and providing test data for the generative large model.

[0067] In one implementation, the test case execution module 122 is further configured to:

[0068] Record the test data generated during the test process.

[0069] Among them, test data is used for data analysis and knowledge graph supplement, including test cases, scripts and execution results.

[0070] In the embodiment of the present application, the test module 12 includes a test case generation module 121, a test case execution module 122 and a test feedback module 123, which can automatically generate test cases and test data for specific business needs, and execute test tasks in an automated test environment and feedback test results. This automated test process not only improves test efficiency, but also ensures the quality of the developed application, reduces the workload of manual testing, and speeds up the development cycle.

[0071] The test case execution module 122 not only executes the test task, but also records various data generated during the test, such as test cases, scripts, and execution results. These data are used for subsequent data analysis and supplementation of the knowledge graph, which helps to continuously optimize the test process and improve the test coverage, ensuring that the developed application has higher quality and reliability.

[0072] Figure 5 A block diagram of the composition of the artificial intelligence and reasoning module 13 is shown.

[0073] For example, Figure 5 As shown, the artificial intelligence and reasoning module 13 includes a machine learning module 131, a generative large model module 132, a knowledge graph module 133, and an enhanced search module 134.

[0074] The machine learning module 131 is used to provide capabilities based on machine learning foundations, including convolutional neural networks, generative adversarial networks, three-dimensional convolutional neural networks, and / or natural language processing.

[0075] The generative large model module 132 is used to provide natural language processing, content generation and reasoning capabilities based on the transformer architecture after pre-training and fine-tuning.

[0076] The knowledge graph module 133 is used to build a structured semantic knowledge base using business requirement data and low-code configuration data.

[0077] The enhanced search module 134 is used to improve the relevance and accuracy of the search results and update the knowledge graph in real time.

[0078] As a feasible implementation, the enhanced search module 134 is based on a unified knowledge graph to improve retrieval efficiency and accuracy while maintaining the timeliness and relevance of answers.

[0079] In the embodiment of the present application, the artificial intelligence and reasoning module 13 includes a machine learning module 131, a generative large model module 132, a knowledge graph module 133 and an enhanced search module 134, which provides capabilities based on machine learning and natural language processing, and supports intelligent data processing, content generation and reasoning. These modules work together to improve the intelligence level of the system, enhance the accuracy of data processing and reasoning, and provide users with more efficient and accurate services.

[0080] The interaction and agent module 14 includes a low-code platform agent 141 and a manual annotation module 142, which supports users to operate through dialogue and realize automated task execution. The low-code platform agent 141 can assist users in completing various development tasks, while the manual annotation module 142 further improves the accuracy of the data through manual intervention. This human-machine collaboration method not only improves the user experience, but also ensures the flexibility and accuracy of the system.

[0081] Figure 6 A block diagram of the components of the interaction and agent module 14 is shown.

[0082] For example, Figure 6As shown, the interaction and agent module 14 includes a low-code platform agent 141 and a manual annotation module 142.

[0083] The low-code platform agent 141 is used to assist users in performing conversational operations and automated task execution.

[0084] The manual annotation module 142 is used to manually intervene in the data extracted by the business requirements analysis and extraction module, the low-code configuration analysis and extraction module 114, and the test feedback module 123.

[0085] In one embodiment of the present application, the low-code platform agent 141 supports user conversational operations based on natural language processing, content generation and reasoning capabilities, and realizes automated task execution based on the platform interface.

[0086] In the embodiment of the present application, the enhanced search module 134 is based on a unified knowledge graph, which improves the retrieval efficiency and accuracy while maintaining the timeliness and relevance of the answers. This enables users to quickly find the information they need, improves the system's response speed and user satisfaction, and ensures the efficient operation of the system.

[0087] Among them, the low-code platform agent 141 supports user conversational operations based on natural language processing, content generation and reasoning capabilities, and realizes automated task execution through the platform interface. This agent not only improves the user's operational convenience, but also realizes highly automated task processing, significantly improves development efficiency and user experience, and reduces development difficulty and time cost.

[0088] Figure 7 A complete structural block diagram of an intelligent low-code platform based on a big model is shown.

[0089] For example, Figure 7As shown in the figure, an intelligent low-code platform based on a large model includes multiple modules that work together to achieve efficient business demand processing, development and testing. The business demand collection module is responsible for collecting business system demand descriptions from various carriers (such as demand documents, sample pages, pictures, videos, and audios) and uploading them to the platform. The business demand data processing module performs preliminary formatting and classification management on these data, such as extracting pictures from demand documents and establishing associations. The business demand data storage module stores business demand data in the form of different data carriers such as text, pictures, videos, and audios, and maintains the associations between the data. Among them, the business demand data analysis and extraction module uses machine learning algorithms (such as OCR, edge detection, convolutional neural network CNN, three-dimensional convolutional neural network 3D CNN, speech recognition, etc.) to analyze business demand data, extract feature information such as entities, relationships, and attributes, and provide data support for the construction of knowledge graphs. The low-code page orchestration module allows users to create interactive pages through a visual interface and generate page configuration metadata. The low-code business logic orchestration module supports users to orchestrate business logic through a visual interface and generate business logic configuration metadata. The low-code configuration storage module formats and stores the generated metadata, providing a data source for subsequent analysis and extraction. The low-code configuration analysis and extraction module uses the machine learning module to further analyze and extract the metadata generated by the low-code platform, providing data for the knowledge graph module. The test case generation module automatically generates test cases, test data, and scripts in a conversational manner through the low-code platform agent. The test case execution module provides an automated testing environment (such as Python runtime environment, Jmeter, etc.) and automatically calls and executes through the low-code platform agent. The test feedback module feeds back and processes the test results, provides test data for the generative large model, and improves the accuracy of its generated test cases.

[0090] In this application, the machine learning module provides various basic capabilities based on machine learning, such as convolutional neural networks (CNN), generative adversarial networks (GAN), three-dimensional convolutional neural networks (3D CNN) and natural language processing. The generative large model module is based on the Transformer architecture and provides natural language processing, content generation and reasoning capabilities for the intelligent agent through two stages of pre-training and fine-tuning. The knowledge graph module uses business demand data and low-code configuration data to build a structured semantic knowledge base to provide accuracy and relevance support for enhanced search. The manual annotation module manually intervenes in the data extracted by the business demand analysis extraction module, the low-code configuration analysis extraction module and the test feedback module to further improve the data accuracy. The enhanced search module improves the accuracy of the retrieval results through a unified and rich knowledge base, reduces misunderstandings or omissions, and introduces new information to the generative large model in real time to maintain the timeliness and accuracy of the answer. The low-code platform intelligent agent is autonomous, adaptable and interactive, and supports the conversational use and automated task execution of the low-code platform through the natural language processing, content generation and reasoning capabilities provided by the generative large model, as well as the platform real-time interface. Through the collaborative work of these modules, the entire system achieves efficient and high-quality business demand processing, development and testing.

[0091] Figure 8 A business demand processing flow chart is shown.

[0092] For example, Figure 8 As shown in the figure, when the business system puts forward new requirements, the requirements personnel will upload the written requirements documents, designed interaction diagrams, audio or video files of requirements communication or sample functions to the platform. The platform first uses text processing technology to extract chapters in the requirements documents and constructs the requirements function points. For example, the user management function module is constructed from the user management in the first chapter, and the user addition function point under the user management module is constructed from the new user in the first section of the first chapter. For other types of requirements files, such as the designed interaction diagrams, audio or video files, the platform will extract basic information such as file format, size and file name. By uploading batches, time and extracted chapter information, the platform establishes an association relationship between the requirements files and function points, and further establishes an association relationship between the function modules and function points and the business system. These association relationships, as well as the characteristic information of the requirements files, the relationship between the function modules and function points are stored in the relational database, and each type of requirements file is stored in the object storage.

[0093] Next, the platform uses text preprocessing techniques such as word segmentation (such as HMM, CRF), stop word removal, and stem extraction to preprocess the requirements document, and uses convolutional neural networks (CNN) and recurrent neural networks (RNN) to extract entity, relationship, attribute and other feature information of business requirements, such as user entities and their related attributes and relationships. Based on this extracted information, the knowledge graph of business requirements is initially constructed. For pictures, audio and video files, the platform uses OCR technology to recognize text information in pictures, automatic speech recognition (ASR) to transcribe audio into text, and CNN and RNN are used again to extract entity, relationship and attribute information. At the same time, the platform also uses CNN to extract interactive interface feature data (such as rectangles, buttons, tables, input boxes, etc.) in images, and uses a temporal convolutional network (3D CNN) to extract interactive operation feature data (such as clicks, pop-up dialog boxes, etc.) in demonstration videos. These feature information extracted from pictures, audio and video further supplements and improves the knowledge graph of business requirements. Finally, the knowledge graph of business requirement data is annotated and modified through manual annotation to further improve the accuracy of business requirement data. The entire process ensures the comprehensive collection, processing and storage of business requirements data, thereby providing a solid data foundation for subsequent low-code development and testing.

[0094] Fig. 9 Another business requirement processing flow chart is shown.

[0095] For example, Fig. 9 As shown in the figure, when developing business functions on an intelligent low-code platform based on a large model, first create function modules and function points, and then create low-code platform object metadata such as function objects, attributes, data tables, and data fields. Subsequently, create low-code platform interface metadata such as function services, function interfaces, input parameters, and output parameters. Next, create platform metadata for display interfaces such as function pages, page layouts, page forms, form attributes, attribute styles, and button operations. These configured metadata will be stored in the database and matched and associated with entities in the business requirements knowledge graph. Furthermore, these configuration metadata are added to the business requirements knowledge graph to accumulate data for subsequent business requirements development and improve the accuracy of the results generated by the large model.

[0096] During the entire configuration process, users can manually create these metadata through a visual interface. During the creation process, the low-code platform uses a multimodal generative large model to take user input data, business demand knowledge graphs, and various types of business demand files as model input data, and combines enhanced search to perform reasoning and give recommended information to help users create. In addition, users can also create conversationally through the low-code platform agent. For example, the user can tell the agent that the user management function needs to be created. The agent will infer the required function modules, function points, objects, attributes, and a series of low-code configuration data through the multimodal large model, and automatically call the creation through the real-time interface provided by the low-code platform, thereby completing the development of the user management function. As a feasible embodiment, after the start, function objects, attributes, data tables, and data fields can be created in sequence, and then function services, function interfaces, input parameters, and output parameters can be created. Next, create function modules, function points, and function pages, and then create page layouts, page forms, form attributes, attribute styles, and button operations. Store these low-code configuration metadata in the database, extract information associated with business demand entities, supplement them to the business demand knowledge graph, and finally end the entire process.

[0097] Fig.10 A business requirement low-code testing flow chart is shown.

[0098] For example, Fig.10As shown in the figure, when the business function test is performed on the intelligent low-code platform based on the big model, the user first selects the business function to be tested. The multimodal big model generates test cases and test data (including various types of test data such as text and pictures) through the business requirement knowledge graph and the multimodal business requirement file. The user reviews and adjusts the generated test cases and test data. Then, the low-code platform agent simulates the test operation through the platform real-time interface and performs the test task. For example, the agent obtains the business function request address and request parameters through the platform real-time interface, calls the big model to generate an automated test script, and executes the script by executing the plug-in to obtain the test results. The system records the test information, including various types of data such as the generated test cases, test scripts, and execution results. Then, machine learning is used to analyze and extract feature data from these test cases and test information. The extracted feature data is supplemented to the business requirement knowledge graph. Based on the test results and the knowledge graph containing the complete functions of business requirements, low-code configuration information, and test information, the big model generates repair suggestions or directly calls the platform interface to modify and adjust the configuration. During the entire testing process, users can manually create test cases through the visual interface. During the creation process, the low-code platform uses a multimodal generative large model to take user input data, business requirements knowledge graphs, and various types of business requirements files as model input data, and combines enhanced search to perform reasoning and provide recommended information to help users create. Users can also create conversationally through the low-code platform agent, and finally the user optimizes and confirms. After the test case is created, the low-code platform agent executes the Python test script by calling various plug-ins (such as Python execution plug-ins). The obtained test results, test cases, and data are processed by machine learning, and a complete business function knowledge graph is constructed together with the business requirements knowledge graph generated by the business requirements processing process and the business requirements configuration data generated by the business requirements low-code development process. The generative large model is assisted by a complete business function knowledge graph to provide more accurate repair suggestions for the results of functional testing, and optimizes and modifies the low-code configuration of business functions through the low-code agent, helping the team to quickly identify problems and repair them.

[0099] In the implementation of this application, at the beginning, the user selects the business function that needs to be tested, and the big model generates test cases and other information through the business demand knowledge graph and multimodal business demand files. The user optimizes and adjusts the generated test cases, and then the low-code platform agent builds a simulated test operation through the platform's real-time interface to perform the test task. Record the test information, use machine learning to analyze the test cases and test information to extract feature data, and add it to the business demand knowledge graph. The big model gives modification suggestions based on the test data and knowledge graph, and finally ends the whole process. The intelligent low-code platform based on the big model uses a variety of machine learning algorithms to perform multi-level processing on various types of business demand data, extract various types of feature data, and build a highly accurate business demand knowledge graph, providing an intelligent data foundation for the low-code development and testing of subsequent business functions. In the process of business function development, the multimodal generative big model combines knowledge graphs and enhanced search to perform reasoning and give recommended information to help users create or create conversationally through low-code platform agents, thereby improving development efficiency and quality. In the business function testing process, the multimodal generative large model combines knowledge graphs and enhanced search to generate test cases and test data. Users can create them in a conversational manner through a visual interface or a low-code platform agent, further improving the efficiency and quality of writing test cases and test data. The low-code platform agent combines the platform's real-time interface and plug-ins to automate the execution of test tasks and analyze test results, and generates repair suggestions to help the team quickly identify problems and repair them, thereby improving the efficiency and quality of business function testing.

[0100] The technical advantages of this application include high efficiency and high quality: starting from the source of system development, through multiple ways and multi-level processing of business demand data, combined with multimodal generative large models, knowledge graphs and enhanced search, it provides efficient and fast data support for low-code platform users. On the one hand, through the autonomy and task execution capabilities of low-code intelligent agents, the automation of using low-code to develop business functions is improved, thereby improving development efficiency. On the other hand, in business function testing, testing efficiency is improved by automatically creating and executing test cases and data. In addition, by constructing business requirements and low-code development configuration data into a complete knowledge graph, high-quality basic data is provided, the accuracy of generating business function configurations, test cases and test data is improved, and the autonomy and decision-making correctness of low-code platform intelligent agents are enhanced, thereby improving the overall quality of business function development.

[0101] Although operations are described in a particular order in the drawings, this should not be understood as requiring that the operations be performed in the particular order shown or in serial order, or that all shown operations be performed to achieve desired results. In certain circumstances, multitasking and parallel processing may be advantageous.

[0102] The methods, devices, equipment and storage media of the present application can be implemented using standard programming techniques, using rule-based logic or other logic to implement various method steps. It should also be noted that the words "device" and "module" used herein and in the claims are intended to include implementations using one or more lines of software code and / or hardware implementations and / or devices for receiving input.

[0103] Any steps, operations or procedures described herein may be performed or implemented using one or more hardware or software modules, either alone or in combination with other devices. In one embodiment, the software module is implemented using a computer program product including a computer-readable medium containing computer program code, which can be executed by a computer processor to perform any or all of the described steps, operations or procedures.

[0104] The foregoing description of the implementation of the present application has been given for the purpose of illustration and description. The foregoing description is not exhaustive nor is it intended to limit the present application to the exact form disclosed, and various variations and modifications may exist according to the above teachings or may be obtained from the practice of the present application. These embodiments are selected and described in order to illustrate the principles of the present application and its practical application, so that those skilled in the art can utilize the present application in various embodiments and various modifications suitable for the specific purpose contemplated.

[0105] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0106] It can be further understood that, unless otherwise specified, “connection” includes a direct connection without other components between the two, and also includes an indirect connection with other components between the two.

[0107] It is further understood that, although the operations are described in a specific order in the drawings in the embodiments of the present application, it should not be understood as requiring the operations to be performed in the specific order or serial order shown, or requiring the execution of all the operations shown to obtain the desired results. In certain environments, multitasking and parallel processing may be advantageous.

[0108] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to encompass any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the field of the present application that are not disclosed in the present application. The specification and embodiments are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following scope of rights.

[0109] It should be understood that the present application is not limited to the precise structures described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the scope of the appended claims.

[0110] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent low-code platform development system based on a large model, characterized in that: include: Data processing and storage module, used to collect and process business demand data; A low-code development module for visually choreographing page layout and business logic to generate configuration metadata and support user operations based on conversations; Testing module, used to generate, execute and provide feedback on test cases to ensure the quality of the developed applications; Artificial intelligence and reasoning module, which is used to provide the module with capabilities based on machine learning and natural language processing, and supports intelligent data processing, content generation and reasoning; The interaction and agent module is used to support users to interact with the system based on dialogue and realize automated task execution.

2. The system according to claim 1, characterized in that The data processing and storage module includes: A business requirement collection module, used to collect the business requirement data; wherein the business requirement data includes requirement descriptions from different carriers, and the composition forms of the requirement descriptions include requirement documents, sample pages, pictures, audio and / or video files; The business demand data preprocessing module is used to preliminarily format the collected business demand data and establish management relationships by type classification; The business demand data storage module is used to store business demand data in different data carriers such as text, pictures, videos, audios, etc., and maintain the association relationship between data; The business demand data analysis and extraction module is used to analyze business demand data based on machine learning algorithms and extract characteristic information such as entities, relationships, and attributes to build a business demand knowledge graph.

3. The system according to claim 2, characterized in that The demand document includes pictures and document sections, and the business demand data preprocessing module is also used for: The image is extracted from the requirement document, and a relationship between the image and the document section is established.

4. The system according to claim 1, characterized in that The low-code development module includes: The low-code page layout module allows users to create page layouts and interactive elements based on a visual interface and generate corresponding configuration metadata; The low-code business logic orchestration module is used to support users to orchestrate business logic based on a visual interface and generate corresponding business logic configuration metadata; The low-code configuration storage module is used to format and store the metadata generated by the low-code platform, providing a data source for the low-code configuration analysis and extraction module; The low-code configuration analysis and extraction module is used to analyze and extract metadata generated by the low-code platform based on the machine learning module, and provide data for the knowledge graph module.

5. The system according to claim 1, characterized in that The test module includes: Test case generation module, used to generate test cases, test data and scripts for specific business needs; The test case execution module is used to provide an automated testing environment, automatically execute test tasks and provide feedback on the results; The test feedback module is used to provide feedback and process test results, and provide test data for the generative large model.

6. The system according to claim 5, characterized in that The test case execution module is also used for: Record the test data generated during the test; Among them, the test data is used for data analysis and knowledge graph supplement, including test cases, scripts and execution results.

7. The system according to claim 1, characterized in that The artificial intelligence and reasoning module includes: A machine learning module for providing capabilities based on machine learning foundations, including convolutional neural networks, generative adversarial networks, three-dimensional convolutional neural networks, and / or natural language processing; Generative large model module, which is based on the transformer architecture and provides natural language processing, content generation and reasoning capabilities after pre-training and fine-tuning; The knowledge graph module is used to build a structured semantic knowledge base using business requirement data and low-code configuration data; Enhanced search module to improve the relevance and accuracy of retrieval results and update the knowledge graph in real time.

8. The system according to claim 7, characterized in that The enhanced search module is based on a unified knowledge graph, improving retrieval efficiency and accuracy while maintaining the timeliness and relevance of answers.

9. The system according to claim 1, characterized in that The interaction and agent module includes: Low-code platform agents are used to assist users in conversational operations and automated task execution; The manual labeling module is used to manually intervene in the data extracted by the business requirements analysis and extraction module, the low-code configuration analysis and extraction module, and the test feedback module.

10. The system according to claim 9, characterized in that The low-code platform intelligent agent supports user conversational operations based on natural language processing, content generation and reasoning capabilities, and realizes automated task execution based on the platform interface.

Images