Web page automatic operation processing method and device

Abstract

The invention discloses an automatic operation processing method and device for a Web page. The method comprises the following steps: acquiring structural aspect data and various types of data of the Web page; inputting page resource data of the Web page to be processed into the large model, and outputting multiple pieces of page element data in the Web page to be processed; the large model is obtained by training a pre-training model by utilizing historical page resource data and page element data of a plurality of Web pages in advance; inputting the operation intention description into the large model, and outputting an operation step sequence; the large model is further trained by using historical operation intention description input by a user and a user operation behavior sequence in advance; and automatically executing operation on the Web page to be processed according to the operation step sequence. According to the method, the understanding of the user intention in the Web page automatic operation can be improved, and the flexibility and adaptability of the Web page automatic operation are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for automatic operation processing of a Web page. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No admission is made that the description herein is prior art by inclusion in this section.

[0003] Web automation refers to the use of software tools or scripts to automatically complete a series of operations on Web pages without manual intervention. With the widespread use of the Internet, the demand for automation of Web page operations is growing. Traditional automation tools often rely on pre-set script rules to simulate user operations in web browsers to achieve automation. However, pre-set script rules can only operate strictly according to pre-defined rules. When encountering situations outside the rules (such as changes in page layout and the appearance of new elements), errors may occur, and there is a lack of flexible understanding and adaptability to complex page changes and user intentions. Summary of the invention

[0004] The embodiment of the present invention provides a method for processing a Web page automated operation, which is used to improve the understanding of user intentions in the Web page automated operation and enhance the flexibility and adaptability of the Web page automated operation. The method includes:

[0005] Acquire page resource data of the Web page to be processed; the page resource data includes data on the structure of the Web page;

[0006] Inputting the page resource data of the to-be-processed Web page into the trained large model, and outputting a plurality of page element data in the to-be-processed Web page; the large model pre-uses the historical page resource data and page element data of a plurality of Web pages to train the pre-trained model;

[0007] Receive the operation intention description input by the user;

[0008] Inputting the operation intention description into the large model and outputting the operation step sequence; the large model is also trained in advance using the historical: operation intention description input by the user and the user operation behavior sequence;

[0009] Automatically execute operations on the web page to be processed according to the sequence of operation steps.

[0010] In one embodiment, the large model is trained as follows:

[0011] Collect historical data: page resource data, page element data, user input operation intention description and user operation behavior sequence of multiple web pages to form a training set;

[0012] The data in the training set is used as prompt words to train the pre-trained model to learn the functions and operation logic of the web page and obtain a trained large model.

[0013] In one embodiment, the data in the training set is used as prompt words to train the pre-trained model to learn the functions and operation logic of the web page, including: the data in the training set interacts with the pre-trained model multiple times in the form of prompt words.

[0014] In one embodiment, the large model uses a natural language processing library to perform natural language intent analysis on the operation intent description.

[0015] In one embodiment, the operation step sequence includes a plurality of operation semantic tags and a page element mapped to each operation semantic tag; the operation semantic tag includes operation content;

[0016] According to the sequence of operation steps, the operation is automatically performed on the Web page to be processed, including: according to the page element mapped by each operation semantic tag, the operation is performed at the corresponding page element position on the Web page to be processed.

[0017] In one embodiment, after automatically performing the operation on the Web page to be processed according to the sequence of operation steps, the method further includes: obtaining the operation result and exception information; and training and updating the large model using the operation intention description, operation result and exception information.

[0018] The embodiment of the present invention further provides a Web page automatic operation processing device, which is used to improve the understanding of user intention in the Web page automatic operation and enhance the flexibility and adaptability of the Web page automatic operation. The device includes:

[0019] A page data acquisition module is used to acquire page resource data of the Web page to be processed; the page resource data includes data on the structure of the Web page;

[0020] The page perception and analysis module is used to input the page resource data of the web page to be processed into the trained large model, and output multiple page element data in the web page to be processed; the large model pre-uses the historical page resource data and page element data of multiple web pages to train the pre-trained model;

[0021] A user operation receiving module, used to receive an operation intention description input by a user;

[0022] An operation intention analysis module is used to input the operation intention description into the large model and output an operation step sequence; the large model is also trained in advance using the historical: operation intention descriptions input by users and user operation behavior sequences;

[0023] The automated operation execution module is used to automatically execute operations on the Web page to be processed according to the sequence of operation steps.

[0024] An embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned Web page automatic operation processing method when executing the computer program.

[0025] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned Web page automatic operation processing method is implemented.

[0026] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above-mentioned Web page automatic operation processing method is implemented.

[0027] In the embodiment of the present invention, the page resource data of the Web page to be processed is obtained; the page resource data includes the data of the Web page structure; the page resource data of the Web page to be processed is input into the trained big model, and multiple page element data in the Web page to be processed is output; the big model pre-uses the historical: page resource data and page element data of multiple Web pages to train the pre-trained model; receives the operation intention description input by the user; inputs the operation intention description into the big model, and outputs the operation step sequence; the big model also pre-uses the historical: operation intention description input by the user and the user operation behavior sequence for training; according to the operation step sequence, the operation is automatically performed on the Web page to be processed. In the embodiment of the present invention, a training big model is constructed, and the Web page to be processed and the user operation intention are parsed based on the big model. Compared with the existing technical solutions, the intelligent operation of the Web page is realized, the understanding of the user intention in the automatic operation of the Web page is improved, and the flexibility and adaptability of the automatic operation of the Web page are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0029] Figure 1 A schematic diagram of a process flow of a method for automatically operating a Web page according to an embodiment of the present invention;

[0030] Figure 2 A specific example diagram of a method for processing automatic operation of a Web page in an embodiment of the present invention;

[0031] Figure 3 A schematic diagram of a Web page automatic operation processing device according to an embodiment of the present invention;

[0032] Figure 4 A diagram showing a specific example of a Web page automatic operation processing device in an embodiment of the present invention;

[0033] Figure 5 Schematic diagram of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0035] The acquisition, transmission, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0036] Existing Web automation tools based on fixed rules need to traverse all elements of the current Web page in advance and operate according to certain rules. They cannot meet operations outside the rules and need to continuously improve the rules according to specific demand scenarios. The embodiment of the present invention is based on big model technology. By guiding the big model to learn the operation rules of Web pages, the automatic operation page can be operated more freely and flexibly.

[0037] Figure 1 FIG. 1 is a flow chart of a method for processing automatic operation of a Web page according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0038] Step 101: Obtain page resource data of a Web page to be processed; the page resource data includes data on the structure of the Web page;

[0039] Step 102: input the page resource data of the to-be-processed Web page into the trained large model, and output a plurality of page element data in the to-be-processed Web page; the large model is pre-trained by using the historical page resource data and page element data of a plurality of Web pages to train the pre-trained model;

[0040] Step 103: receiving the operation intention description input by the user;

[0041] Step 104: Input the operation intention description into the large model and output the operation step sequence; the large model is also trained in advance using the historical user input operation intention description and user operation behavior sequence;

[0042] Step 105: Automatically perform operations on the Web page to be processed according to the sequence of operation steps.

[0043] from Figure 1 It can be seen from the process shown that in the embodiment of the present invention, a large training model is constructed, and the Web pages to be processed and the user operation intentions are parsed based on the large model. Compared with the existing technical solutions, intelligent operation of Web pages is realized, the understanding of user intentions in the automated operation of Web pages is improved, and the flexibility and adaptability of the automated operation of Web pages are enhanced.

[0044] During implementation, the page resource data of the web page to be processed is obtained, including: using the browser developer tool to obtain the page resource data of the web page to be processed; or, obtaining it through programming languages ​​and related libraries, for example, first using a network request library (such as requests) to obtain the web page content, and then using BeautifulSoup to parse the obtained HTML content; or, using a crawler tool to obtain the page resource data.

[0045] The page resource data includes data on the structure of the Web page and various types of data on the Web page, such as page layout, element features, user operation behavior sequence and other information.

[0046] Among them, the structure of the page can be regarded as a form of resource organization. Just like the frame of a building, it is the basis for carrying other resources. For example, the hierarchical structure of HTML (such as the frame composed of tags such as , , etc.) can be regarded as a container-type resource for organizing other resources such as text and images. In terms of type data, data types such as text, images, links, and forms are all resources on the Web page. Text is an information resource, images are visual resources, links are navigation resources, and forms are interactive resources. Together, they constitute a rich and diverse set of Web page resources.

[0047] Afterwards, the page resource data of the Web page to be processed is input into the trained large model, and the data of multiple page elements in the Web page to be processed is output.

[0048] Page element data reflects the properties of the components that make up a web page. It is the basic unit that makes up a web page, including specific components such as text boxes, buttons, links, menus, etc. During the web page development and design process, these elements can be determined by code (such as HTML, CSS, and JavaScript) to determine their position in the web page (through coordinate positioning, layout method, etc.), size (such as pixel value or percentage of width and height), type (for example, a button may be a submit button, a reset button, etc., a text box may be a single-line text box, a multi-line text box, etc.) and many other attribute information.

[0049] In the embodiment of the present invention, the large model pre-uses historical: page resource data and page element data of multiple web pages to train the pre-trained model;

[0050] Figure 2 FIG. 1 is a specific example of a method for processing automatic operation of a Web page according to an embodiment of the present invention. Figure 2 As shown, the large model can be trained as follows:

[0051] Step 201: collect historical: page resource data, page element data, user input operation intention description and user operation behavior sequence of multiple web pages to form a training set;

[0052] Step 202: Use the data in the training set as prompt words to train the pre-trained model to learn the functions and operation logic of the web page, and obtain a trained large model.

[0053] During implementation, information such as page layout, element features, and user operation behavior sequences are used as prompts to induce the large model to learn the functions and operation logic of the Web page, so that it can understand the various element attributes, functions, and common operation logic of the Web page, including clicking buttons, filling out forms, scrolling pages, and more complex customized Web page logic operation steps.

[0054] In a preferred embodiment, using the data in the training set as prompt words to train the pre-trained model to learn the functions and operation logic of the web page can include: interacting the data in the training set with the pre-trained model multiple times in the form of prompt words.

[0055] For example, during the model learning process, the DOM elements and operation logic of the web page are interacted with the big model through prompting. After multiple interactions, the big model understands the operation logic of the DOM elements of the web page.

[0056] After the large model is trained, it is put into use. When faced with a new web page to be processed, the page is first fully perceived and analyzed. The large model is used to identify various elements in the page, such as text boxes, buttons, links, menus, etc., and determine their location, size, type and other attribute information. At the same time, the overall structure and layout pattern of the page are analyzed to provide a basis for subsequent operational decisions.

[0057] Afterwards, the operation intention description input by the user is received, for example, the operation intention description input by the user is received through a client interface or a user UI interface, which can be input in text form or in voice form.

[0058] The operation intention description is input into the large model, and the operation step sequence is output. The large model is also trained in advance using the historical user input operation intention description and user operation behavior sequence.

[0059] The natural language processing library is used in the large model to perform natural language intent analysis on the operation intent description.

[0060] Finally, according to the sequence of operation steps, the operation is automatically performed on the Web page to be processed.

[0061] For example, a user enters "log in to a website and view specific information". The large model parses the user input and generates a corresponding sequence of operation steps based on the page analysis results. For example, it determines to find the login button first, enter the user name and password, then click login, and then navigate to the specified information page.

[0062] In one embodiment, the operation step sequence includes a plurality of operation semantic tags and a page element mapped to each operation semantic tag; the operation semantic tag includes operation content;

[0063] Automatically performing operations on the Web page to be processed according to the sequence of operation steps may include: performing operations at corresponding page element positions on the Web page to be processed according to the page elements mapped by each operation semantic tag.

[0064] During implementation, the corresponding operations are automatically performed on the Web page according to the generated sequence of operation steps. By using technical means such as simulated mouse clicks and keyboard input, the system can accurately interact with page elements to achieve automated page browsing, data filling, task execution and other operations. In addition, during the operation, the system can dynamically adjust the operation strategy according to the real-time feedback and changes of the page, such as responding to pop-up windows, repositioning after page jumps, etc.

[0065] In one embodiment, after automatically performing the operation on the Web page to be processed according to the sequence of operation steps, the method may further include: obtaining the operation result and exception information; and training and updating the large model using the operation intention description, operation result and exception information.

[0066] Operation results include, for example, page status changes, data acquisition and storage, internal page jumps, external website jumps, pop-up boxes and prompts, etc. Page status changes include element attribute changes, element content updates, and page layout adjustments.

[0067] Then, the large model is trained and updated using the operation intention description, operation results and exception information.

[0068] During implementation, during the execution of automated operations, operation data and page feedback information are continuously collected, and the big model is used to further learn and optimize these data. For example, if a new element type or operation process is encountered on a certain page, the big model can learn and update its knowledge system so that it can handle similar pages more efficiently in the future. At the same time, based on the success rate and efficiency indicators of the operation, the generation algorithm of the operation step sequence is continuously optimized to improve the accuracy and speed of automated operations.

[0069] The present invention also provides a Web page automatic operation processing device, as described in the following embodiments. Since the principle of the device to solve the problem is similar to that of the Web page automatic operation processing method, the implementation of the device can refer to the implementation of the Web page automatic operation processing method, and the repeated parts will not be repeated.

[0070] Figure 3 FIG. 1 is a schematic diagram of a Web page automatic operation processing device according to an embodiment of the present invention. Figure 3 As shown, the device 300 includes:

[0071] The page data acquisition module 301 is used to acquire page resource data of the Web page to be processed; the page resource data includes data on the structure of the Web page and data of various types of data on the Web page;

[0072] The page perception and analysis module 302 is used to input the page resource data of the to-be-processed Web page into the trained large model, and output a plurality of page element data in the to-be-processed Web page; the large model is pre-trained using the historical page resource data and page element data of a plurality of Web pages to train the pre-trained model;

[0073] The user operation receiving module 303 is used to receive the operation intention description input by the user;

[0074] The operation intention analysis module 304 is used to input the operation intention description into the large model and output the operation step sequence; the large model is also trained in advance using the historical user input operation intention description and user operation behavior sequence;

[0075] The automated operation execution module 305 is used to automatically execute operations on the Web page to be processed according to the operation step sequence.

[0076] In one embodiment, the large model is trained as follows:

[0077] Collect historical data: page resource data, page element data, user input operation intention description and user operation behavior sequence of multiple web pages to form a training set;

[0078] The data in the training set is used as prompt words to train the pre-trained model to learn the functions and operation logic of the web page and obtain a trained large model.

[0079] In one embodiment, the data in the training set is used as prompt words to train the pre-trained model to learn the functions and operation logic of the web page, including:

[0080] The data in the training set interacts with the pre-trained model multiple times through prompt words.

[0081] In one embodiment, the large model uses a natural language processing library to perform natural language intent analysis on the operation intent description.

[0082] In one embodiment, the operation step sequence includes a plurality of operation semantic tags and a page element mapped to each operation semantic tag; the operation semantic tag includes operation content;

[0083] The automated operation execution module 305 is specifically used for:

[0084] According to the page element mapped by each operation semantic tag, the operation is performed at the corresponding page element position on the Web page to be processed.

[0085] In one embodiment, the apparatus 300 further includes:

[0086] The feedback and optimization module is used to: obtain the operation results and exception information after the automated operation execution module automatically executes the operation on the Web page to be processed according to the sequence of operation steps; and train and update the large model using the operation intention description, operation results and exception information.

[0087] Figure 4 FIG. 1 is a specific example diagram of a Web page automatic operation processing device according to an embodiment of the present invention. Figure 4As shown, the device includes a large model learning module, a page perception and analysis module, an operation intention analysis module, an automated operation execution module, and a feedback and optimization module. The large model is trained by means of web data inscription, and then data is processed in sequence by the page perception and analysis module, the operation intention analysis module, the automated operation execution module, and the feedback and optimization module.

[0088] In the embodiment of the present invention, the hardware platform can run on an ordinary server, which requires sufficient memory (such as 16GB and above), a high-speed CPU (such as multi-core i7 and above series) and a high-speed network connection.

[0089] Software platform: The big model uses a third-party big model on the cloud, uses the LangChain library to interact with the big model on the cloud, and uses Python language for tool development.

[0090] The page perception and analysis module, operation intention analysis module, automated operation execution module, large model training module, feedback and optimization module are developed in Python, and the browser automation operation is realized by combining the Selenium WebDriver library, the HTML page is analyzed by the BeautifulSoup library, and the natural language processing library (such as NLTK or SpaCy) is used for natural language intention analysis. The modules communicate and collaborate with each other through clearly defined interfaces and data structures. For example, the operation intention analysis module encapsulates the operation semantic labels and element mapping information into specific data objects and passes them to the automated operation execution module, and the automated operation execution module feeds back the operation results and abnormal information to the learning and optimization module in a unified format.

[0091] Figure 5 Schematic diagram of a computer device in an embodiment of the present invention. Figure 5 As shown, an embodiment of the present invention further provides a computer device 500, including a processor 501, a memory 502, and a computer program 503 stored in the memory 502 and executable on the processor 501, and when the processor 501 executes the computer program 503, the above-mentioned Web page automatic operation processing method is implemented.

[0092] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned Web page automatic operation processing method is implemented.

[0093] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above-mentioned Web page automatic operation processing method is implemented.

[0094] In the embodiment of the present invention, the page resource data of the Web page to be processed is obtained; the page resource data includes the data of the Web page structure; the page resource data of the Web page to be processed is input into the trained big model, and multiple page element data in the Web page to be processed is output; the big model pre-uses the historical: page resource data and page element data of multiple Web pages to train the pre-trained model; receives the operation intention description input by the user; inputs the operation intention description into the big model, and outputs the operation step sequence; the big model also pre-uses the historical: operation intention description input by the user and the user operation behavior sequence for training; according to the operation step sequence, the operation is automatically performed on the Web page to be processed. In the embodiment of the present invention, a training big model is constructed, and the Web page to be processed and the user operation intention are parsed based on the big model. Compared with the existing technical solutions, the intelligent operation of the Web page is realized, the understanding of the user intention in the automatic operation of the Web page is improved, and the flexibility and adaptability of the automatic operation of the Web page are improved.

[0095] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0096] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0097] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0098] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0099] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for processing automatic operations on a Web page, characterized in that: include: Obtain page resource data of the Web page to be processed; The page resource data includes data on the structure of the Web page; Inputting the page resource data of the to-be-processed Web page into the trained large model, and outputting a plurality of page element data in the to-be-processed Web page; the large model pre-uses the historical page resource data and page element data of a plurality of Web pages to train the pre-trained model; Receive the operation intention description input by the user; Inputting the operation intention description into the large model and outputting the operation step sequence; the large model is also trained in advance using the historical: operation intention description input by the user and the user operation behavior sequence; Automatically execute operations on the web page to be processed according to the sequence of operation steps.

2. The method according to claim 1, characterized in that The large model is trained as follows: Collect historical data: page resource data, page element data, user input operation intention description and user operation behavior sequence of multiple web pages to form a training set; The data in the training set is used as prompt words to train the pre-trained model to learn the functions and operation logic of the web page and obtain a trained large model.

3. The method according to claim 2, characterized in that Use the data in the training set as prompt words to train the pre-trained model to learn the functions and operation logic of the web page, including: The data in the training set interacts with the pre-trained model multiple times through prompt words.

4. The method according to claim 1, characterized in that The large model uses a natural language processing library to perform natural language intent analysis on the operation intent description.

5. The method according to claim 1, characterized in that The operation step sequence includes a plurality of operation semantic tags and a page element mapped by each operation semantic tag; the operation semantic tag includes operation content; Automatically perform operations on the web page to be processed according to the sequence of operation steps, including: According to the page element mapped by each operation semantic tag, the operation is performed at the corresponding page element position on the Web page to be processed.

6. The method according to claim 1, characterized in that After automatically performing operations on the web page to be processed according to the sequence of operation steps, it also includes: Obtain operation results and exception information; Use operation intention descriptions, operation results, and exception information to train and update the large model.

7. A Web page automatic operation processing device, characterized in that: include: A page data acquisition module is used to acquire page resource data of the Web page to be processed; The page resource data includes data on the structure of the Web page; The page perception and analysis module is used to input the page resource data of the web page to be processed into the trained large model, and output multiple page element data in the web page to be processed; the large model pre-uses the historical page resource data and page element data of multiple web pages to train the pre-trained model; A user operation receiving module, used to receive an operation intention description input by a user; An operation intention analysis module is used to input the operation intention description into the large model and output an operation step sequence; the large model is also trained in advance using the historical: operation intention descriptions input by users and user operation behavior sequences; The automated operation execution module is used to automatically execute operations on the Web page to be processed according to the sequence of operation steps.

8. The device according to claim 7, characterized in that The large model is trained as follows: Collect historical data: page resource data, page element data, user input operation intention description and user operation behavior sequence of multiple web pages to form a training set; The data in the training set is used as prompt words to train the pre-trained model to learn the functions and operation logic of the web page and obtain a trained large model.

9. The device according to claim 8, characterized in that Use the data in the training set as prompt words to train the pre-trained model to learn the functions and operation logic of the web page, including: The data in the training set interacts with the pre-trained model multiple times through prompt words.

10. The device according to claim 7, characterized in that The large model uses a natural language processing library to perform natural language intent analysis on the operation intent description.

11. The device according to claim 7, characterized in that The operation step sequence includes a plurality of operation semantic tags and a page element mapped by each operation semantic tag; the operation semantic tag includes operation content; The automated operation execution module is specifically used for: According to the page element mapped by each operation semantic tag, the operation is performed at the corresponding page element position on the Web page to be processed.

12. The device according to claim 7, characterized in that Also includes: The feedback and optimization module is used to: obtain the operation results and exception information after the automated operation execution module automatically executes the operation on the Web page to be processed according to the operation step sequence; and train and update the large model using the operation intention description, operation results and exception information.

13. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

15. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images