Qwen-VL-Max-based UI style code intelligent walkthrough formatting system and method

Through the UI style code intelligent check formatting system based on Qwen-VL-Max, using deep learning and natural language processing technology, the problems of low efficiency and insufficient accuracy of UI style code processing in the existing technology are solved, and efficient Accurate UI style processing and adaptive formatting rules to adapt to complex front-end technical architectures.

CN119065671APending Publication Date: 2024-12-03BEIYIN FINANCIAL TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411289331.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-14
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Existing UI style code checking and formatting methods are inefficient, insufficient accuracy, and poor adaptability. They cannot fully detect complex UI style problems. They lack flexibility and low degree of intelligence, making it difficult to deal with large-scale code bases and diversified front-ends. Technical architecture.

Method used

Using a UI-style code intelligent walk-through formatting system based on Qwen-VL-Max, including data input, preprocessing, feature extraction, Qwen-VL-Max model analysis, rule formulation and formatting execution, through deep learning and natural language processing Technology extracts key features and combines computer vision and rules engines to achieve intelligent formatting and walk-through reports.

Benefits of technology

It realizes efficient and accurate UI style processing, and can adaptively generate and adjust formatting rules according to user needs and code characteristics, improves code quality and user experience, and adapts to complex and diverse front-end technical architectures.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119065671A_ABST
    Figure CN119065671A_ABST
Patent Text Reader

Abstract

The invention discloses a Qwen-VL-Max-based UI style code intelligent walkthrough formatting system and a Qwen-VL-Max-based UI style code intelligent walkthrough formatting method, and the system comprises a data input module which is used for receiving a UI style code to be processed; the preprocessing module is used for carrying out preliminary cleaning, conversion and standardization operation on the received codes; the feature extraction module is used for extracting key features from the preprocessed codes; according to the Qwen-VL-Max model module, a Qwen-VL-Max model is used for carrying out deep analysis on the input UI style codes; the formatting execution module is used for formatting the UI style codes; and the result output and feedback module is used for outputting the formatted UI style code to the user. According to the method, a set of complete flow from data input, preprocessing, feature extraction and model analysis to rule making, formatting execution and result output feedback is constructed, and efficient and accurate UI style processing is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of software development, and specifically is a UI style code intelligent walkthrough formatting system and method based on Qwen-VL-Max. Background Art

[0002] In today's software development field, especially in front-end development, the design and implementation of the user interface (UI) is crucial. A good UI can not only provide an excellent user experience, but also improve the usability and competitiveness of the software. With the continuous development of front-end technology, UI style codes are becoming more and more complex and diverse. In order to ensure the quality, readability and maintainability of the code, it is essential to effectively check and format the UI style code.

[0003] At present, common UI style code review and formatting methods mainly include manual inspection and the use of traditional code formatting tools. Manual inspection takes a lot of time and manpower, and is prone to omissions and inconsistencies. Traditional code formatting tools usually format based on fixed rules and patterns, lacking in-depth understanding and intelligent processing capabilities for complex UI style codes.

[0004] The current technology has the following disadvantages:

[0005] Inefficiency: Manual review and some automated tools are slow when processing large-scale front-end codes and are unable to cope with large-scale code bases, which seriously affects development progress.

[0006] Lack of accuracy: Existing tools have difficulty accurately identifying various complex UI style issues, especially those related to visual perception and design aesthetics, and are prone to miss some potential problems, especially style issues related to specific business logic.

[0007] Poor adaptability: Unable to flexibly respond to changing design requirements and diverse front-end technology architectures.

[0008] Lack of comprehensiveness: It can only detect some common problems, and some deep-seated UI style issues that are closely related to user experience are easily overlooked.

[0009] Lack of flexibility: Cannot adapt well to the specific needs and style requirements of different projects.

[0010] Low intelligence: Unable to perform precise formatting based on context and semantics. Summary of the invention

[0011] In view of the above problems, the present invention is proposed to provide a UI style code intelligent walkthrough formatting system and method based on Qwen-VL-Max that overcomes the above problems or at least partially solves the above problems.

[0012] To achieve the above object, the present invention adopts the following technical solutions:

[0013] A UI style code intelligent walkthrough formatting system based on Qwen-VL-Max, the system comprising:

[0014] A data input module, used for receiving UI style codes to be processed;

[0015] The preprocessing module is used to perform preliminary cleaning, conversion and normalization operations on the received code;

[0016] Feature extraction module, used to extract key features from the preprocessed code;

[0017] Qwen-VL-Max model module, which performs in-depth analysis on the input UI style code through the Qwen-VL-Max model;

[0018] Formatting execution module, used to format UI style code;

[0019] The result output and feedback module is used to output the formatted UI style code to the user and provide a walkthrough report and feedback information.

[0020] Optionally, the system further includes a rule formulation and adjustment module for defining and adjusting formatting rules, and the rule formulation and adjustment module includes:

[0021] The architecture based on rule engine supports the definition and management of complex rules;

[0022] Use a visual rule editing interface to intuitively create and modify rules;

[0023] Combine with database or configuration file to store and read rule information.

[0024] Optionally, the data input module includes:

[0025] Input code files in various data formats and upload them;

[0026] Use the Gzip compression algorithm to compress uploaded files.

[0027] Optionally, the preprocessing module includes:

[0028] Use lexical analysis tools and syntax analysis tools to parse the code;

[0029] Use image processing technology to pre-process image data;

[0030] Use natural language processing technology to process text data.

[0031] Optionally, the feature extraction module includes:

[0032] Extract image features through deep learning convolutional neural network;

[0033] Processing sequence data through recurrent neural networks or long short-term memory networks;

[0034] Combine the bag-of-words model of natural language processing and TF-IDF technology to extract text features. Optional, Qwen-VL-Max model modules include:

[0035] It uses a deep neural network architecture and integrates natural language processing and computer vision technology; it uses large-scale pre-training and fine-tuning strategies to adapt to UI style analysis tasks. Optional formatting execution modules include:

[0036] Use code refactoring and optimization techniques to adjust the structure and modify the style of the original code; combine with the version control system to ensure the traceability and security of formatting operations;

[0037] Integrate with your code editor or IDE to provide real-time formatting support.

[0038] Optionally, the result output and feedback module includes:

[0039] Use data visualization techniques to present walkthrough results and statistical information;

[0040] Integrate with the data analysis platform to analyze the inspection results and feedback data;

[0041] Use the difference comparison algorithm to show the changes of the code before and after formatting.

[0042] The present invention also provides a method for intelligently checking and formatting UI style codes based on Qwen-VL-Max, the method being a system for intelligently checking and formatting UI style codes based on Qwen-VL-Max as described in any of the above, the method comprising:

[0043] Enter UI style code and UI specifications;

[0044] Preprocess the input data;

[0045] Extract key features from the preprocessed data;

[0046] The extracted features were analyzed using the Qwen-VL-Max model;

[0047] Formulate or adjust rules based on model analysis results;

[0048] Optimize the rules;

[0049] Perform formatting operations according to optimized rules;

[0050] Output formatted results and detailed reports.

[0051] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0052] This paper innovatively applies the Qwen-VL-Max model to the inspection and formatting of UI style codes, constructs a complete set of processes from data input, preprocessing, feature extraction, model analysis to rule formulation, formatting execution and result output feedback, and realizes efficient and accurate UI style processing. It can also adaptively generate and adjust formatting rules according to user needs and code characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A schematic diagram of a UI style code intelligent walkthrough and formatting system based on Qwen-VL-Max provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0055] Embodiment 1:

[0056] See also Figure 1 This embodiment provides a UI style code intelligent walkthrough formatting system based on Qwen-VL-Max, the system comprising:

[0057] The data input module is used to receive the UI style code to be processed and the UI style related data to be processed from various sources. It provides the system with a variety of data sources, making it convenient for users to input different forms of UI style data into the system for processing. It ensures the integrity and accuracy of the data and lays the foundation for subsequent processing steps.

[0058] The data input module includes:

[0059] Enter code files in various data formats, such as common code file formats (such as .html, .css, .js), image formats (such as .png, .jpg), and text formats (such as .txt, .docx, etc.), and upload the code files.

[0060] Use the web interface, file upload, or integration with the version control system to obtain data, and use the Gzip compression algorithm to compress the uploaded files to reduce transmission bandwidth.

[0061] The algorithms used are:

[0062] For the network interface, the HTTP protocol related algorithms may be used for data transmission and reception.

[0063] When uploading files, use file reading and parsing algorithms to ensure accurate data content.

[0064] When integrating with a code management system, follow the API rules and data extraction algorithms of the corresponding system.

[0065] The preprocessing module is used to perform preliminary cleaning, conversion and normalization operations on the received code, and preliminary cleaning, conversion and normalization operations on the input data. The UI style specifications specified by the user are parsed, sorted and transmitted to the rule formulation and adjustment module; improve the quality and consistency of the data, and reduce the impact of noise and abnormal data on subsequent processing. Convert data of different formats and styles into a unified and easy-to-process form to prepare for feature extraction.

[0066] The preprocessing modules include:

[0067] Use lexical analysis tools (such as Flex) and syntax analysis tools (such as Bison) to parse the code;

[0068] Image processing technology (such as OpenCV) is used to preprocess the image data, including cropping, scaling, grayscale and other operations.

[0069] The text data is processed using natural language processing techniques (such as word segmentation and part-of-speech tagging).

[0070] The algorithms used are:

[0071] Whitespace processing algorithm to remove extra spaces, tabs, and newlines.

[0072] Code standardization algorithm converts different styles of code (such as indentation method, variable naming) into a unified standard.

[0073] Image denoising algorithm reduces noise interference in images.

[0074] Text cleaning algorithm to remove special characters and correct spelling errors.

[0075] The feature extraction module is used to extract key features from the preprocessed code. It uses deep learning technology and natural language processing methods to extract key features from the preprocessed code, including style attributes, element structure, code logic, etc. It converts complex data into feature vectors that can be understood and processed by the model, highlights the key information in the data, reduces the complexity of the data, improves the learning efficiency and accuracy of the model, and provides strong support for subsequent model analysis and decision-making.

[0076] The feature extraction module includes:

[0077] Image features are extracted through deep learning convolutional neural network.

[0078] Process sequence data through recurrent neural networks or long short-term memory networks.

[0079] Combine the bag-of-words model of natural language processing and TF-IDF technology to extract text features.

[0080] The algorithms used are:

[0081] Convolution operation algorithm extracts features such as shape, color, and texture from images.

[0082] Recursive algorithms are used for information transfer and memory in RNNs and LSTMs.

[0083] Feature selection and dimensionality reduction algorithms, such as principal component analysis (PCA) and linear discriminant analysis (LDA), are used to reduce feature dimensions and improve computational efficiency and model performance.

[0084] The Qwen-VL-Max model module performs in-depth analysis on the input UI style code through the Qwen-VL-Max model, applies the Qwen-VL-Max model to deeply understand the input UI style code and related descriptions, extracts key features and semantic information, identifies different UI elements, style attributes and the relationships between them, and compares them with predefined standards and specifications to detect potential style inconsistencies or parts that do not conform to best practices.

[0085] Qwen-VL-Max model modules include:

[0086] It uses a deep neural network architecture and integrates natural language processing and computer vision technologies.

[0087] Apply large-scale pre-training and fine-tuning strategies to adapt to UI style analysis tasks.

[0088] The algorithms used are:

[0089] The encoding-decoding mechanism based on the Transformer architecture can effectively process long sequences of text and image data.

[0090] Use the attention mechanism to focus on and analyze the input UI style related codes, images and descriptions.

[0091] Multi-layer neural networks are used for feature extraction and pattern recognition to discover potential rules and problems from complex data.

[0092] The rule formulation and adjustment module is used to define and adjust formatting rules, and can be flexibly adjusted according to actual conditions. The system can also automatically optimize the rules based on the analysis results of the model.

[0093] The rule making and adjustment module includes:

[0094] The rule engine-based architecture supports the definition and management of complex rules.

[0095] Use a visual rule editing interface to intuitively create and modify rules.

[0096] Combine with database or configuration file to store and read rule information.

[0097] The algorithms used are:

[0098] Use the classification and clustering algorithms in machine learning to summarize and classify common UI style issues, providing a reference for rule making.

[0099] A heuristic search algorithm is used to quickly find the optimal or nearly optimal rule combination when adjusting the rules.

[0100] The formatting execution module is used to format the UI style code; according to the established rules and the analysis results of the model, the UI style code is actually formatted, and the UI style code that does not conform to the rules is corrected and optimized to conform to the preset specifications.

[0101] The formatting execution module includes:

[0102] Use code refactoring and optimization techniques to adjust the structure and modify the style of the original code.

[0103] Combined with the version control system, the traceability and security of formatting operations are guaranteed.

[0104] Integrate with your code editor or IDE to provide real-time formatting support.

[0105] The algorithms used are:

[0106] Use a greedy algorithm or a dynamic programming algorithm to select the best one among multiple possible formatting schemes.

[0107] Use code template matching and replacement algorithms to quickly apply unified style standards.

[0108] The result output and feedback module is used to output the formatted UI style code to the user, provide inspection reports and feedback information, and analyze the trends of multiple inspection results to help users understand the improvement of UI style quality.

[0109] The result output and feedback modules include:

[0110] Use data visualization techniques to present walkthrough results and statistical information.

[0111] Integrate with the data analysis platform to analyze the inspection results and feedback data.

[0112] Use the difference comparison algorithm to clearly show the changes in the code before and after formatting.

[0113] This embodiment innovatively applies the Qwen-VL-Max model to the inspection and formatting of UI style codes, constructs a complete system flow from data input, preprocessing, feature extraction, model analysis to rule formulation, formatting execution and result output feedback, and realizes efficient and accurate UI style processing. It can also adaptively generate and adjust formatting rules according to user needs and code characteristics.

[0114] Embodiment 2:

[0115] Embodiment 2 discloses a method for intelligently checking and formatting UI style codes based on Qwen-VL-Max. The method is based on the above-mentioned system for intelligently checking and formatting UI style codes based on Qwen-VL-Max. The method comprises the following steps:

[0116] S1. Input UI style code and UI specification.

[0117] S2. Preprocess the input data.

[0118] S3. Extract key features from the preprocessed data.

[0119] S4. Use the Qwen-VL-Max model to analyze the extracted features.

[0120] S5. Formulate or adjust rules based on the model analysis results.

[0121] S6. Optimize the rules.

[0122] S7. Perform formatting operations according to the optimized rules.

[0123] S8. Output formatted results and detailed reports.

[0124] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. A UI style code intelligent walkthrough formatting system based on Qwen-VL-Max, characterized in that: The system comprises: A data input module, used for receiving UI style codes to be processed; The preprocessing module is used to perform preliminary cleaning, conversion and normalization operations on the received code; Feature extraction module, used to extract key features from the preprocessed code; Qwen-VL-Max model module, which performs in-depth analysis on the input UI style code through the Qwen-VL-Max model; Formatting execution module, used to format UI style code; The result output and feedback module is used to output the formatted UI style code to the user and provide a walkthrough report and feedback information.

2. A UI style code intelligent walkthrough formatting system based on Qwen-VL-Max as claimed in claim 1, characterized in that: The system further includes a rule formulation and adjustment module for defining and adjusting formatting rules, wherein the rule formulation and adjustment module includes: The architecture based on rule engine supports the definition and management of complex rules; Use a visual rule editing interface to intuitively create and modify rules; Combine with database or configuration file to store and read rule information.

3. The UI style code intelligent walkthrough formatting system based on Qwen-VL-Max as claimed in claim 1, characterized in that: The data input module includes: Input code files in various data formats and upload them; Use the Gzip compression algorithm to compress uploaded files.

4. The UI style code intelligent walkthrough formatting system based on Qwen-VL-Max as claimed in claim 1, characterized in that: The preprocessing modules include: Use lexical analysis tools and syntax analysis tools to parse the code; Use image processing technology to pre-process image data; Use natural language processing technology to process text data.

5. The UI style code intelligent walkthrough formatting system based on Qwen-VL-Max as claimed in claim 1, characterized in that: The feature extraction module includes: Extract image features through deep learning convolutional neural network; Processing sequence data through recurrent neural networks or long short-term memory networks; Combine the bag-of-words model of natural language processing and TF-IDF technology to extract text features.

6. The UI style code intelligent walkthrough formatting system based on Qwen-VL-Max as claimed in claim 1, characterized in that: Qwen-VL-Max model modules include: It uses a deep neural network architecture and integrates natural language processing and computer vision technologies; Apply large-scale pre-training and fine-tuning strategies to adapt to UI style analysis tasks.

7. The UI style code intelligent walkthrough formatting system based on Qwen-VL-Max as claimed in claim 1, characterized in that: The formatting execution module includes: Use code refactoring and optimization techniques to adjust the structure and modify the style of the original code; Combined with the version control system, the traceability and security of formatting operations are guaranteed; Integrate with your code editor or IDE to provide real-time formatting support.

8. The UI style code intelligent walkthrough formatting system based on Qwen-VL-Max as claimed in claim 1, characterized in that: The result output and feedback modules include: Use data visualization techniques to present walkthrough results and statistical information; Integrate with the data analysis platform to analyze the inspection results and feedback data; Use the difference comparison algorithm to show the changes of the code before and after formatting.

9. A method for intelligently checking and formatting UI style codes based on Qwen-VL-Max, the method being based on a system for intelligently checking and formatting UI style codes based on Qwen-VL-Max according to any one of claims 1 to 8, the method comprising: Enter UI style code and UI specifications; Preprocess the input data; Extract key features from the preprocessed data; The extracted features were analyzed using the Qwen-VL-Max model; Formulate or adjust rules based on model analysis results; Optimize the rules; Perform formatting operations according to optimized rules; Output formatted results and detailed reports.