Page generation method and apparatus

Abstract

The one or more embodiments of the specification provide a page generation method and device. Through the method, target data can be acquired first, wherein the target data is used to describe the page content of the page to be generated. Then, feature information is determined based on the target data, wherein the feature information is used to represent the theme feature of the target data. Next, target style information matching the feature information in the style specification library is determined, and the target style information is used to describe the page style of the page to be generated. Finally, the target page is generated based on the target data and the target style information.

Description

Technical Field

[0001] This specification relates to one or more embodiments in the field of computer technology, and more particularly to a page generation method and apparatus. Background Technology

[0002] With the rapid development of large language model technology, the generation of user pages based on user data using large language models is gradually moving from the technology verification stage to real-world scenarios.

[0003] Currently, when users generate user pages using a large language model, they need to directly input their user data into the large language model, which will then generate the user page based on that data.

[0004] However, when generating user pages using the large language model, it adopts a generic visual style. This means that regardless of whether the user uploads serious financial data, lively sports data, or concise billing data, the large language model uses the same default visual style for rendering. This results in the page style of the user pages generated by the large language model failing to match the user data.

[0005] Therefore, ensuring that the page style of the user pages generated by the large language model matches the user data has become a pressing technical problem that needs to be solved. Summary of the Invention

[0006] This specification provides one or more embodiments of a page generation method and apparatus to match the page style of the generated user page with the user data.

[0007] In a first aspect, one or more embodiments of this specification provide a page generation method, the method comprising: first, acquiring target data, the target data being used to describe the page content of the page to be generated; then, determining feature information based on the target data, the feature information being used to characterize the thematic features of the target data; second, determining target style information in a style specification library that matches the feature information, the target style information being used to describe the page style of the page to be generated; and finally, generating a target page based on the target data and the target style information.

[0008] In one possible implementation, before generating the target page based on the target data and target style information, the method further includes: obtaining initial requirement information, which describes the page generation requirements of the page to be generated; determining whether the initial requirement information meets the completeness condition; if the initial requirement information does not meet the completeness condition, determining the target keywords in the initial requirement information; determining at least one historical requirement document in the requirement document library that matches the target keywords; and generating complete requirement information based on the initial requirement information and at least one historical requirement document, which describes the page structure of the page to be generated.

[0009] In one possible implementation, determining at least one historical requirement document in the requirement document library that matches the target keyword includes: determining multiple historical requirement documents in the requirement document library that match the target keyword; and determining at least one historical requirement document among the multiple historical requirement documents whose similarity to the target keyword is greater than a preset threshold.

[0010] In one possible implementation, the completeness condition includes at least one of the following: determining whether the initial requirements information contains a page type, determining whether the initial requirements information contains a page functional module, determining whether the initial requirements information contains a page interaction flow, and determining whether the initial requirements information contains a page functional boundary; at least one historical requirements document includes at least one of the following: page type, page functional module, page interaction flow, and page functional boundary; and complete requirements information includes at least one of the following: page type, page functional module, page interaction flow, and page functional boundary.

[0011] One possible implementation involves generating a target page based on target data and target style information, including: constructing fusion prompts from the target data, complete requirement information, and target style information; and calling a large language model to generate the target page based on the fusion prompts.

[0012] In one possible implementation, before generating the target page based on the target data and target style information, the method further includes: obtaining a first image, which is an image containing a first reference page; determining the component information of each component in the first reference page; determining the hierarchical relationship between each component; and generating page layout information based on the component information of each component and the hierarchical relationship between each component, wherein the page layout information is used to describe the page structure of the page to be generated.

[0013] In one possible implementation, determining the component information of each component in the first reference page includes: calling a visual component detection model to identify the first reference page in the first image, and obtaining the component information of each component in the first reference page. The component information includes component category, component position information, component text, component color, and component geometric attributes.

[0014] In one possible implementation, determining the hierarchical relationship between components includes: determining the inclusion relationship between components based on the component location information of each component; and determining the hierarchical relationship between components based on the inclusion relationship between each component.

[0015] In one possible implementation, the first reference page includes a first component and a second component; the hierarchical relationship between the components is determined based on the inclusion relationship between them, including: when the first component includes the second component, the first component is determined as the parent and the second component is determined as the child; when the second component includes the first component, the second component is determined as the parent and the first component is determined as the child; when the first component and the second component partially overlap or do not overlap, the first component and the second component are determined as siblings.

[0016] One possible implementation involves generating a target page based on target data and target style information, including: constructing a fusion prompt word from the target data, page layout information, and target style information; and calling a large language model to generate the target page based on the fusion prompt word.

[0017] In one possible implementation, the method further includes: acquiring a second image, which is an image containing a second reference page; performing visual analysis on the second reference page to obtain target style information; receiving feature information associated with the target style information input by the user; and storing the mapping relationship between the feature information and the target style information in a style specification library.

[0018] One possible implementation involves determining the target style information in the style specification library that matches the feature information, including: determining multiple design style information in the style specification library that matches the feature information; and determining the target style information with the highest similarity to the feature information among the multiple design style information.

[0019] In one possible implementation, determining the target style information in the style specification library that matches the feature information includes: determining multiple design style information in the style specification library that matches the feature information; displaying multiple design style information; receiving a confirmation instruction from the user, the confirmation instruction being used to indicate that the target style information among the multiple design style information is selected by the user; and determining the selected target style information among the multiple design style information.

[0020] In one possible implementation, the feature information includes at least one keyword; determining the feature information based on the target data includes: performing semantic recognition on the target data to obtain at least one keyword.

[0021] In one possible implementation, the target style information includes at least one of color specifications, font specifications, spacing specifications, rounded corner specifications, shadow specifications, icon specifications, or button specifications, and the font specifications include at least one of font size, font weight, line height, and character spacing.

[0022] One possible implementation involves generating a target page based on target data and target style information, including: constructing fusion prompts from the target data and target style information; and calling a large language model to generate the target page based on the fusion prompts.

[0023] Secondly, one or more embodiments of this specification also provide a page generation apparatus, the apparatus comprising: an acquisition module for acquiring target data, the target data being used to describe the page content of the page to be generated; a determination module for determining feature information based on the target data, the feature information being used to characterize the thematic features of the target data; a matching module for determining target style information in a style specification library that matches the feature information, the target style information being used to describe the page style of the page to be generated; and a generation module for generating a target page based on the target data and the target style information.

[0024] In one possible implementation, the page generation device may also include a requirement expansion module.

[0025] In one possible implementation, a requirement expansion module is used to obtain initial requirement information, which describes the page generation requirements of the page to be generated; determine whether the initial requirement information meets the completeness condition; if the initial requirement information does not meet the completeness condition, determine the target keywords in the initial requirement information; determine at least one historical requirement document in the requirement document library that matches the target keywords; and generate complete requirement information based on the initial requirement information and at least one historical requirement document, which describes the page structure of the page to be generated.

[0026] In one possible implementation, the requirement expansion module is specifically used to identify multiple historical requirement documents in the requirement document library that match the target keyword; and to identify at least one historical requirement document among the multiple historical requirement documents whose similarity to the target keyword is greater than a preset threshold.

[0027] In one possible implementation, the completeness condition includes at least one of the following: determining whether the initial requirements information contains a page type, determining whether the initial requirements information contains a page functional module, determining whether the initial requirements information contains a page interaction flow, and determining whether the initial requirements information contains a page functional boundary; at least one historical requirements document includes at least one of the following: page type, page functional module, page interaction flow, and page functional boundary; and complete requirements information includes at least one of the following: page type, page functional module, page interaction flow, and page functional boundary.

[0028] In one possible implementation, the generation module is specifically used to construct fusion prompt words from the target data, complete requirement information, and target style information; and to call the large language model to generate the target page based on the fusion prompt words.

[0029] In one possible implementation, the page generation device may further include a component identification module. The component identification module is used to acquire a first image, which is an image containing a first reference page; determine the component information of each component in the first reference page; determine the hierarchical relationship between the components; and generate page layout information based on the component information and the hierarchical relationship between the components, the page layout information being used to describe the page structure of the page to be generated.

[0030] In one possible implementation, the component recognition module is specifically used to call the visual component detection model to recognize the first reference page in the first image, and obtain the component information of each component in the first reference page. The component information includes component category, component position information, component text, component color and component geometric attributes.

[0031] In one possible implementation, the component identification module is specifically used to determine the inclusion relationship between components based on the component location information of each component; and to determine the hierarchical relationship between components based on the inclusion relationship between each component.

[0032] In one possible implementation, the first reference page includes a first component and a second component. The component identification module is specifically configured to: determine the first component as the parent and the second component as the child when the first component contains the second component; determine the second component as the parent and the first component as the child when the second component contains the first component; and determine the first component and the second component as siblings when the first component and the second component partially overlap or do not overlap.

[0033] In one possible implementation, the component recognition module is specifically used to construct fusion prompts from target data, page layout information, and target style information; and to call a large language model to generate the target page based on the fusion prompts.

[0034] In one possible implementation, the page generation device may further include a style extraction module, which is used to acquire a second image, the second image being an image containing a second reference page; perform visual analysis on the second reference page to obtain target style information; receive feature information associated with the target style information input by the user; and store the mapping relationship between the feature information and the target style information in a style specification library.

[0035] In one possible implementation, the matching module is specifically used to determine multiple design style information in the style specification library that match the feature information; and to determine the target style information with the highest similarity to the feature information among the multiple design style information.

[0036] In one possible implementation, the matching module is specifically used to determine multiple design style information that match the feature information in the style specification library; display the multiple design style information; receive a confirmation instruction input by the user, the confirmation instruction being used to indicate that the target style information among the multiple design style information is selected by the user; and determine the selected target style information among the multiple design style information.

[0037] In one possible implementation, the feature information includes at least one keyword. A determining module is specifically used to perform semantic recognition on the target data to obtain at least one keyword.

[0038] In one possible implementation, the target style information includes at least one of color specifications, font specifications, spacing specifications, rounded corner specifications, shadow specifications, icon specifications, or button specifications, and the font specifications include at least one of font size, font weight, line height, and character spacing.

[0039] In one possible implementation, the generation module is specifically used to construct fusion prompts from the target data and target style information; and to call a large language model to generate the target page based on the fusion prompts.

[0040] Thirdly, one or more embodiments of this specification also provide an electronic device, which includes a memory and a processor; the memory is used to store a computer program product; the processor is used to execute the computer program product stored in the memory, and when the computer program product is executed, it implements the method of the first aspect described above.

[0041] Fourthly, one or more embodiments of this specification also provide a computer-readable storage medium storing computer program instructions that, when executed, implement the method described in the first aspect.

[0042] In summary, in one or more embodiments of this specification, a page generation method stores multiple different types of design style information in a style specification library, each corresponding to different feature information. After obtaining the target data, feature information can be determined based on the target data to ascertain its thematic characteristics. Then, from the multiple different types of design style information in the style specification library, the target style information matching the feature information of the target data can be identified, thereby finding a design style suitable for the target data. Therefore, since the generated target page uses target style information suitable for the target data, the page style of the generated target page matches the target data, thus realizing the generation of pages with different visual styles based on different types of user data. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of one or more embodiments of this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of one or more embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A schematic diagram of a scenario provided for one or more embodiments of this specification; Figure 2 This specification provides another scenario illustration for one or more embodiments; Figure 3 A flowchart illustrating a page generation method provided for one or more embodiments of this specification; Figure 4 A structural block diagram of a page generation apparatus provided for one or more embodiments of this specification; Figure 5 This is a structural block diagram of an electronic device provided for one or more embodiments of this specification. Detailed Implementation

[0045] The present specification provides one or more embodiments in further detail below with reference to the accompanying drawings and examples. Through these descriptions, the features and advantages of one or more embodiments of the present specification will become clearer and more apparent.

[0046] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments. Although various aspects of embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless specifically indicated otherwise.

[0047] Furthermore, the technical features described below in one or more embodiments of this specification may be combined with each other as long as they do not conflict with each other.

[0048] The data acquisition and collection actions involved in one or more embodiments of this specification are all performed after authorization by the user, the object, or after full authorization by all parties.

[0049] To facilitate understanding, the application scenarios of the technical solutions provided in one or more embodiments of this specification will be described below.

[0050] With the rapid development of large language model technology, the generation of user pages based on user data using large language models is gradually moving from the technology verification stage to real-world scenarios.

[0051] Currently, when users generate user pages using a large language model, they need to directly input their user data into the large language model, which will then generate the user page based on that data.

[0052] However, when generating user pages using the large language model, it adopts a generic visual style. This means that regardless of whether the user uploads serious financial data, lively sports data, or concise billing data, the large language model uses the same default visual style for rendering. This results in the page style of the user pages generated by the large language model failing to match the user data.

[0053] Therefore, ensuring that the page style of the user pages generated by the large language model matches the user data has become a pressing technical problem that needs to be solved.

[0054] To address the aforementioned technical problems, one or more embodiments of this specification provide a page generation method. The method includes: first, acquiring target data, which describes the page content of the page to be generated; then, determining feature information based on the target data, which characterizes the thematic features of the target data; second, determining target style information in a style specification library that matches the feature information, which describes the page style of the page to be generated; and finally, generating a target page based on the target data and the target style information.

[0055] In one or more embodiments of this specification, a page generation method stores multiple different types of design style information in a style specification library, each corresponding to different feature information. After obtaining target data, feature information can be determined based on the target data to ascertain its thematic characteristics. Then, target style information matching the feature information of the target data can be identified from the multiple different types of design style information in the style specification library, thereby finding a design style suitable for the target data. Therefore, since the generated target page uses target style information suitable for the target data, the page style of the generated target page matches the target data, thus realizing the generation of pages with different visual styles based on different types of user data.

[0056] For ease of understanding, the page generation method provided in one or more embodiments of this specification will be described in detail below with reference to the accompanying drawings.

[0057] Please see Figure 1 As shown, Figure 1 This is a schematic diagram of a scenario provided for one or more embodiments of this specification.

[0058] exist Figure 1In the example shown, the user can create design style information S1 and feature information X1 in the page generation system 100, establish a mapping relationship between feature information X1 and design style information S1, and then store the mapping relationship between feature information X1 and design style information S1 in the style specification library 200.

[0059] Users can manually input design style information S1 and feature information X1 directly in the page generation system 100, or they can generate design style information S1 and feature information X1 in the page generation system 100 through other means.

[0060] Then, using the same approach of creating design style information S1 and feature information X1, the user continuously creates other feature information and design style information in the page generation system 100, and stores the mapping relationship between the created feature information and design style information in the style specification library 200.

[0061] Finally, the style specification library 200 stores various types of design style information, and each design style information corresponds to characteristic information.

[0062] For example, the style specification library 200 stores the mapping relationship between feature information X1 and design style information S1, the mapping relationship between feature information X2 and design style information S2, and the mapping relationship between feature information X3 and design style information S3, etc.

[0063] The following example illustrates the feature information X1 and the design style information S1.

[0064] For example, feature information X1 includes the following keyword: "stock".

[0065] Design style information S1 includes the following specifications: color specifications, font specifications, spacing specifications, rounded corner specifications, shadow specifications, icon specifications, button specifications, and other element specifications.

[0066] Specifically, the color specifications for design style information S1 include: "Primary blue: 3A50ED, used for key buttons, important data, and selected states; Light blue: 1FB6FF, used for the background of secondary function blocks; Dark purple: 261E5A, used for main title text; ......." Specifically, the font specifications for design style information S1 include: "Main title: 18~20px, bold, dark purple; Subtitle: 16px, bold, black; Body text: 14px, regular, black; ...... Specifically, the spacing specifications for design style information S1 include: "Page margins: 24~32px on the left and right; Card inner margins: 16~20px; Vertical spacing between elements: 16~24px; ......". Specifically, the rounded corner specifications for design style information S1 include: "Rounded corners for large cards: 16px; rounded corners for medium elements: 12px; rounded corners for small buttons: 8px; ... Specifically, the shadow specifications for design style information S1 include: "Card shadow: light, ...; Floating element shadow: medium, ...". Specifically, the icon specifications for design style information S1 include: "Navigation icon size: 24×24px; Function icon size: 20×20px; Small icon size: 16×16px." Specifically, the button specifications for design style information S1 include: "Main button: blue background, white text, height approximately 40~44px; Secondary button: white background, blue border and text; Icon button: round or square, size approximately 40px." Specifically, other element specifications for design style information S1 include: "Progress bar height: 6~8px; Card spacing: 16px; Label style: Lightweight rounded corners, inner margin approximately 8px 12px; Table row height: approximately 48~56px." It should be noted that the examples of feature information X1 and design style information S1 in the above examples do not constitute a limitation on the scope of protection of this invention.

[0067] Furthermore, the structure of other design style information is similar to that of design style information S1. However, the specific content of other design style information differs from that of design style information S1, so that each piece of design style information corresponds to a unique visual style.

[0068] pass Figure 1 As shown in the example, users can create feature information and design style information through the page generation system 100 and store the mapping relationship between the two in the style specification library 200.

[0069] Please see Figure 2 As shown, Figure 2 This is a schematic diagram of another scenario provided for one or more embodiments of this specification.

[0070] exist Figure 2 In the example shown, after the style specification library 200 is established, the user can use the page generation system 100, the style specification library 200, and the large language model 300 to generate pages.

[0071] The following steps will explain... Figure 2 The execution process of generating page U1 in the middle: Step 1: The user inputs the target data D1 into the page generation system 100.

[0072] For example, the target data D1 can be stock data, sports data, news information, game data, to-do list information, membership information, or billing data, etc.

[0073] Typically, users expect the page generation system 100 to generate a page U1 that is compatible with the target data D1, meaning that the design style of the generated page U1 needs to match the target data D1.

[0074] Step 2: After the page generation system 100 receives the target data D1 input by the user, the page generation system 100 will determine the feature information Y1 based on the target data D1.

[0075] For example, suppose the target data D1 includes the following: "Company A's stock is currently valued at XXX yuan, its trading volume has been steadily increasing over the past year, and its stock price has generally been on an upward trend, ... Company B's stock is currently valued at YYY yuan, ... The page generation system 100 analyzes the semantic content of the target data D1 and finds one or more keywords from the target data D1.

[0076] For example, the feature information Y1 found by the page generation system 100 from the target data D1 includes the following keyword: "stock".

[0077] Step 3: After the page generation system 100 determines the feature information Y1, the page generation system 100 will send the feature information Y1 to the style specification library 200.

[0078] Step 4: Style specification library 200 will determine the design style information S1 that matches the feature information Y1.

[0079] For example, suppose feature information X1 in style specification library 200 includes the keyword "stock", feature information X2 in style specification library 200 includes the keyword "bill", and feature information X3 in style specification library 200 includes the keyword "health".

[0080] Among them, the keyword included in feature information Y1 is "stock", and the keyword included in feature information X1 is also "stock". Feature information Y1 is the same as feature information X1.

[0081] Furthermore, feature information X1 in style specification library 200 is associated with design style information S1, so feature information Y1 matches design style information S1.

[0082] Feature information X2 includes the keyword "bill". Feature information Y1 is different from feature information X2, so feature information Y1 does not match design style information S2.

[0083] Feature information X3 includes the keyword "health". Feature information Y1 is different from feature information X3, so feature information Y1 does not match design style information S3.

[0084] Step 5: Style specification library 200 sends design style information S1 that matches feature information Y1 to page generation system 100.

[0085] Step 6: After the page generation system 100 receives the design style information S1 sent by the style specification library 200, the page generation system 100 will send the target data D1 and the design style information S1 to the large language model 300.

[0086] Step 7: The large language model 300 generates page U1 based on the target data D1 and design style information S1.

[0087] Step 8: The large language model 300 sends the generated page U1 to the page generation system 100.

[0088] Combination Figure 1 and Figure 2 As shown in the example, the style specification library 200 stores multiple different types of design style information, each corresponding to different feature information. After the page generation system 100 obtains the target data D1, it can determine the feature information Y1 based on the target data D1 to understand the theme characteristics of the target data D1. Then, it can determine the design style information S1 that matches the feature information Y1 of the target data D1 from the multiple different types of design style information in the style specification library 200, thereby finding a design style that is suitable for the target data D1. Therefore, since the generated page U1 uses the design style information S1 that matches the target data D1, the page style of the generated page U1 matches the target data D1, thus realizing the generation of pages with different visual styles based on different types of user data.

[0089] The following describes embodiments of the page generation method provided in one or more embodiments of this specification.

[0090] See Figure 3 , Figure 3 This is a flowchart illustrating a page generation method provided in one or more embodiments of this specification. This method can be applied to electronic devices, such as computers, servers, or server clusters. Figure 3 As shown, the method may include the following steps.

[0091] Step S102: Obtain target data.

[0092] The target data is used to describe the page content of the page to be generated.

[0093] For example, please refer to... Figure 2 As shown, the user inputs target data D1 into the page generation system 100. Target data D1 can be stock data, sports data, news information, game data, to-do task information, membership information, or billing data, etc.

[0094] Typically, users expect the page generation system 100 to generate a page U1 that is compatible with the target data D1, meaning that the design style of the generated page U1 needs to match the target data D1.

[0095] Step S104: Determine feature information based on target data.

[0096] Among them, feature information is used to characterize the thematic features of the target data.

[0097] For example, please refer to... Figure 2 As shown, after the page generation system 100 receives the target data D1 input by the user, the page generation system 100 will determine the feature information Y1 based on the target data D1.

[0098] Suppose that the target data D1 includes the following: "Company A's stock is currently valued at XXX yuan, its trading volume has been steadily increasing over the past year, and its stock price has generally shown an upward trend, ... Company B's stock is currently valued at YYY yuan, ... The page generation system 100 analyzes the semantic content of the target data D1 and finds one or more keywords from the target data D1. For example, the feature information Y1 found by the page generation system 100 from the target data D1 includes the following keyword: "stock".

[0099] Step S106: Determine the target style information in the style specification library that matches the feature information.

[0100] The target style information describes the page style of the page to be generated. The target style information includes at least one of the following: color specifications, font specifications, spacing specifications, rounded corner specifications, shadow specifications, icon specifications, or button specifications. The font specifications include at least one of the following: font size, font weight, line height, and character spacing.

[0101] For example, please refer to... Figure 2 As shown, after the page generation system 100 determines the feature information Y1, it sends the feature information Y1 to the style specification library 200. The style specification library 200 then determines the design style information S1 that matches the feature information Y1. The style specification library 200 then sends the design style information S1 that matches the feature information Y1 to the page generation system 100.

[0102] Step S108: Generate the target page based on the target data and target style information.

[0103] For example, please refer to... Figure 2 As shown, after the page generation system 100 receives the design style information S1 sent by the style specification library 200, the page generation system 100 sends the target data D1 and the design style information S1 to the large language model 300. The large language model 300 generates page U1 based on the target data D1 and the design style information S1. The large language model 300 then sends the generated page U1 to the page generation system 100.

[0104] In some embodiments, step S108 may further include the following steps: constructing fusion prompt words from target data and target style information; and calling a large language model to generate a target page based on the fusion prompt words.

[0105] After obtaining the fused prompt words, it is necessary to assemble the fused prompt words together with the front-end framework code, system constraints, and output protocols into a target prompt word that can be understood by the large language model. The target prompt word is then input into the large language model, which will generate the target page based on the target prompt word.

[0106] In one or more embodiments of this specification, a page generation method stores multiple different types of design style information in a style specification library, each corresponding to different feature information. After obtaining target data, feature information can be determined based on the target data to ascertain its thematic characteristics. Then, target style information matching the feature information of the target data can be identified from the multiple different types of design style information in the style specification library, thereby finding a design style suitable for the target data. Therefore, since the generated target page uses target style information suitable for the target data, the page style of the generated target page matches the target data, thus realizing the generation of pages with different visual styles based on different types of user data.

[0107] In some embodiments, Figure 3 The page generation method shown may further include the following steps: obtaining initial requirement information, which describes the page generation requirements of the page to be generated; determining whether the initial requirement information meets the completeness condition; if the initial requirement information does not meet the completeness condition, determining the target keywords in the initial requirement information; determining at least one historical requirement document in the requirement document library that matches the target keywords; and generating complete requirement information based on the initial requirement information and at least one historical requirement document, which describes the page structure of the page to be generated.

[0108] After obtaining the initial requirements information input by the user, it is necessary to determine whether the initial requirements information meets the completeness criteria to identify any missing key information. If it is determined that some key information is missing, target keywords can be found in the initial requirements information, and at least one historical requirements document matching the target keywords can be retrieved from the requirements document library. Finally, complete requirements information is generated based on the initial requirements information and at least one historical requirements document. This expands the initial requirements information based on at least one historical requirements document, ensuring that the complete requirements information includes all key information describing the page structure. Consequently, the large language model can clearly understand how to generate the page structure of the target page based on the complete requirements information.

[0109] In some embodiments, completeness conditions include at least one of the following: determining whether the initial requirements information contains a page type, determining whether the initial requirements information contains page functional modules, determining whether the initial requirements information contains page interaction flows, and determining whether the initial requirements information contains page functional boundaries. At least one historical requirements document includes at least one of the following: page type, page functional modules, page interaction flows, and page functional boundaries. Complete requirements information includes at least one of the following: page type, page functional modules, page interaction flows, and page functional boundaries.

[0110] The page type can be a list, details page, form, dashboard, onboarding page, or empty page, and the page type determines the basic framework of the target page. The core functional modules determine the main information blocks and component sets carried by the target page. The main interaction flow determines the user's operation path on the target page. Functional boundaries can explicitly restrict the free play of the large language model and suppress functional illusions.

[0111] For example, suppose the user inputs the following initial request information into the page generation system: "Create a to-do list page".

[0112] Then, the page generation system will determine whether the initial requirement information meets the completeness requirements. If the initial requirement information is determined to be incomplete, the page generation system will identify the target keywords from the initial requirement information.

[0113] If the target keyword identified by the page generation system in the initial requirements information is "to-do list", then the page generation system will identify multiple historical requirements documents in the requirements document library that match the target keyword "to-do list".

[0114] Since the requirements document library stores product requirements documents that users have used in the past, all the historical requirements documents retrieved from the requirements document library are product requirements documents that meet the user's requirements.

[0115] Then, identify at least one historical requirement document from among multiple historical requirement documents that has a similarity greater than a preset threshold to the target keyword "to-do list".

[0116] Finally, the page generation system generates complete requirements information based on the initial requirements information and at least one historical requirements document.

[0117] The following example illustrates the structure and content of complete requirements information.

[0118] The structure of complete requirements information includes: "1. Requirements overview includes a statement of core issues and main functional requirements."

[0119] 2. Feature details include user roles and usage scenarios, user operation flow, component breakdown, data field requirements, interaction specifications, information architecture, responsive constraints, and copywriting strategies.

[0120] 3. Additional considerations include accessibility requirements and brand consistency requirements.

[0121] 4. Hard constraints include constraint 1 and constraint 2. Constraint 1 includes not adding functional modules that have not been expressed by the user or do not conform to normal logic. Constraint 2 also includes not modifying fields, components or interactions that have been explicitly specified by the user.

[0122] Complete requirements information includes: 1. Requirements Overview: 1.1. Core Problem Statement: Users need a simple and efficient task management tool to quickly record, organize, and complete daily tasks, and solve the problems of task forgetting and prioritization confusion.

[0123] 1.2. Main functional requirements: task creation, deletion, modification and query, status switching, category management and priority marking.

[0124] 2. Core Function List: 2.1. User Roles and Scenarios: Individual users manage daily tasks in work / life scenarios.

[0125] 2.2. User Flowchart: Enter the page → View the task list → Add a new task → Edit / mark as complete → Delete completed tasks.

[0126] 2.3. Detailed Description of Functional Modules: Task input box: Fixed at the top, supports adding shortcut keys; Task list: Displays all tasks, with checkboxes to toggle completion status; Task items: Display title, priority label (high / medium / low), and creation time; Operation menu: Edit, Delete, Set priority; Filter: Switches between three states: All / Incomplete / Completed; Statistics: The top displays the number of items yet to be completed.

[0127] 2.4. Data Requirements: Task ID, Title, Status, Priority, and Creation Time.

[0128] 2.5. Interaction Design Principles: Drag and drop sorting is supported, a delete button is displayed when sliding, a highlight animation is added when a task is completed, and a friendly prompt is provided for empty states.

[0129] 2.6. Information Architecture: Flat single-level list, sorted in descending order of creation time.

[0130] 2.7. Responsive design requirements: Single column layout on mobile devices, centered with a maximum width of 800px on desktop devices.

[0131] 2.8. Content Strategy: Concise prompts, such as "What do you want to accomplish today?"; double confirmation before deletion.

[0132] 3. Other key considerations: 3.1. Accessibility requirements: Keyboard navigation is supported, checkboxes are focusable, and color contrast meets WCAG 2.1 AA standards.

[0133] 3.2. Brand Consistency Guidelines: Employ a refreshing color scheme (primarily blue / green), rounded corners, and subtle animations to enhance the sense of enjoyment. In some embodiments, Figure 3 Step S108 may specifically include the following steps: constructing fusion prompt words from target data, complete requirement information and target style information; calling the large language model to generate the target page based on the fusion prompt words.

[0134] After the page generation system acquires the target data, complete requirement information, and target style information, it constructs a fusion prompt word from these three dimensions. Then, it calls the large language model to generate the target page based on this fusion prompt word. The page generation system organically integrates these three dimensions into a unified prompt word input to the large language model. The target data provides the specific content to be displayed on the page, such as the amount, username, and category list. The complete requirement information provides the page's functional structure and interaction logic, such as the modules included and user actions. The target style information provides the page's visual style, such as color scheme, font specifications, and spacing specifications. These three elements work synergistically, each fulfilling its function, to jointly constrain the generation behavior of the large language model.

[0135] In some embodiments, Figure 3The page generation method shown may further include the following steps: obtaining a first image, which is an image containing a first reference page; determining the component information of each component in the first reference page; determining the hierarchical relationship between each component; and generating page layout information based on the component information of each component and the hierarchical relationship between each component, wherein the page layout information is used to describe the page structure of the page to be generated.

[0136] After the page generation system obtains the first image, it determines the component information of each component in the first reference page within the first image, as well as the hierarchical relationship between each component. Based on the component information and the hierarchical relationship between each component, it can generate page layout information, thereby obtaining the page structure used to describe the page to be generated. This allows the large language model to clearly know how to generate the page structure of the target page based on the page layout information.

[0137] The visual component detection model in the page generation system can accurately extract structured information such as type, coordinates, size, color, and text of each component in the first reference page of the first image, thereby solving the problem of low recognition accuracy.

[0138] The visual component detection model can employ the Co-DETR model, which is a DETR model trained based on collaborative hybrid assignment. The core of the Co-DETR model is a deep learning model architecture designed for object detection tasks. In this scheme, a visual component detection model is trained based on the Co-DETR model, and this model is used to identify various components in a page screenshot.

[0139] To ensure the model's recognition accuracy and generalization ability, we trained the model according to the following process: First, prepare the training data.

[0140] A large number of real page screenshots needed to be collected. After data cleaning, deduplication, and filtering, some high-quality screenshots were retained. Subsequently, professional annotators meticulously annotated the page components in each screenshot, including the component's bounding box position (i.e., the component's coordinates in the screenshot) and the component's category (such as button, text, input box, etc.).

[0141] Then, the model is trained.

[0142] The model was trained using the Co-DETR architecture. During training, various data augmentation techniques (such as random pruning, color perturbation, and rotation) were employed to improve the model's generalization ability. The backbone network used a Swin-Large architecture with an initial learning rate of 1e-4. The number of samples selected in a single training iteration was 16, and the training was iterated for 120 epochs.

[0143] Finally, identify component properties.

[0144] After the model detects the location and category of the component, it further extracts the detailed attributes of the component: for the text content in the component, OCR technology is used to recognize the text in the component area; for the color of the component and the text, image processing technology is used to segment the component area into pixels, and the color value is extracted and calculated from the segmented pixels; for the rounded corner pixel value of the component, image processing technology is also used to analyze the pixels in the component area and calculate the rounded corner radius of the component.

[0145] After the above training process, the final visual component detection model has the following capabilities: identifying component type, locating component position, extracting text content, recognizing color information, obtaining geometric attributes, and outputting confidence scores.

[0146] Among these features, component type identification is used to accurately identify the category of each UI component in the screenshot, such as buttons, text, input boxes, labels, icons, card containers, etc.

[0147] Locate the component positions and output the bounding box coordinates [x1, y1, x2, y2] of each component in the screenshot with pixel-level precision.

[0148] Extract text content; for text-type components, identify and output the specific text content they contain.

[0149] Recognize color information and output the text color value for text components. This supports scenarios where a single component contains multiple colors.

[0150] Obtain geometric properties, identify and output the geometric properties of the component, such as the corner radius pixel value.

[0151] Output confidence scores, providing a confidence score for each detection result for reference by downstream processes.

[0152] Through the training described above, this model fundamentally solves the two core bottlenecks of general multimodal models in page-aware scenarios: inaccurate positioning and color recognition. The structured detection results output by the model can be directly used as a visual benchmark in the downstream generation process, providing precise layout and style constraints for subsequent page code generation, thereby ensuring a high degree of consistency between the generated page and the original design in terms of position, size, color, and hierarchy.

[0153] In some embodiments, determining the component information of each component in the first reference page may include the following steps: calling a visual component detection model to identify the first reference page in the first image, and obtaining the component information of each component in the first reference page, wherein the component information includes component category, component location information, component text, component color and component geometric attributes.

[0154] Among them, the visual component detection model is trained based on large-scale user interface annotation data and is specifically designed to identify various components in the user interface. It can accurately extract structured component information from a reference page image.

[0155] Specifically, this visual component detection model can identify the category of each component, such as whether the component belongs to the category of button, text, input box, label, icon or card container; it can determine the precise position of the component on the page and output it in the form of bounding box coordinates; it can extract the text content contained in the component; it can identify the color value of the text and supports scenarios where a single component contains multiple colors; it can also obtain the geometric attributes of the component, such as the rounded corner pixel value.

[0156] In some embodiments, determining the hierarchical relationship between components includes: determining the inclusion relationship between components based on the component location information of each component; and determining the hierarchical relationship between components based on the inclusion relationship between each component.

[0157] The component location information can be pixel coordinates.

[0158] In some embodiments, the first reference page includes a first component and a second component; the hierarchical relationship between the components is determined based on the inclusion relationship between the components, including: when the first component includes the second component, the first component is determined as the parent and the second component is determined as the child; when the second component includes the first component, the second component is determined as the parent and the first component is determined as the child; when the first component and the second component partially overlap or do not overlap, the first component and the second component are determined as siblings.

[0159] In some embodiments, Figure 3 Step S108 may specifically include the following steps: constructing fusion prompt words from target data, page layout information, and target style information; and calling a large language model to generate a target page based on the fusion prompt words.

[0160] After the page generation system acquires the target data, page layout information, and target style information, it can organically integrate these three dimensions to form a unified prompt input for the large language model. The target data provides the specific content to be displayed on the page, such as the username, transaction amount, and category list. The page layout information provides the precise position, size, and hierarchy of each component on the page, such as the coordinates of a component, its size, and its parent container. The target style information provides the visual style of the page, such as the color scheme, font specifications, and spacing specifications. These three elements work together, each fulfilling its function, to constrain the generation behavior of the large language model.

[0161] For example, suppose the first image uploaded by the user is a screenshot of a monthly bill, which contains three components: an overall card component t1, a title text component t2, and a date text component t3.

[0162] First, after the page generation system receives the screenshot of the monthly bill, the component detection model within the page generation system will identify the three components mentioned above in the screenshot of the monthly bill and determine the component information of the three components.

[0163] Among them, the component category of the overall card component t1 is "card container", and the component position information is bounding box [0, 0, 400, 500].

[0164] The title text component t2 has a component category of "title text" and a component position information of bounding box [20, 20, 380, 60].

[0165] The date text component t3 has a component category of "date text" and a component position information of bounding box [300, 20, 380, 60].

[0166] Then, after the page generation system obtains the component information of each component, the page generation system will determine the inclusion relationship between the three components based on the component position information of the three components.

[0167] For example, the page generation system compares the bounding box [0, 0, 400, 500] of the overall card component t1 with the bounding box [20, 20, 380, 60] of the title text component t2, determining that the overall card component t1 completely contains the title text component t2. Similarly, the overall card component t1 completely contains the date text component t3. However, there is no containment relationship between the title text component t2 and the date text component t3, and both of them have the overall card component t1 as their parent.

[0168] That is, the inclusion relationship between the three components is as follows: the overall card component t1 completely includes the title text component t2 and the date text component t3, and there is no inclusion relationship between the title text component t2 and the date text component t3.

[0169] Secondly, the page generation system determines the hierarchical relationship between components based on the above inclusion relationship.

[0170] Since the overall card component t1 completely contains the title text component t2 and the date text component t3, the overall card component t1 is the parent of the title text component t2 and the date text component t3, and the title text component t2 and the date text component t3 are the children of the overall card component t1.

[0171] Since there is no containment relationship between the title text component t2 and the date text component t3, the title text component t2 and the date text component t3 are at the same level.

[0172] Finally, the page generation system generates page layout information based on the component information of the three components and the hierarchical relationship between them.

[0173] The page layout information describes the structure of the page to be generated, specifying that it should contain an overall card component t1, a title text component t2, and a date text component t3, with the title text component t2 and date text component t3 located inside the overall card component t1. Furthermore, the page layout information can also describe the position, text, color, and geometric properties of each component.

[0174] In some embodiments, Figure 3 The page generation method shown may further include the following steps: obtaining a second image, which is an image containing a second reference page; performing visual analysis on the second reference page to obtain target style information; receiving feature information associated with the target style information input by the user; and storing the mapping relationship between feature information and target style information in a style specification library.

[0175] After the page generation system obtains the second image, it can invoke a multimodal model to perform visual analysis of the second reference page in the second image across multiple dimensions to obtain the target style information of the second reference page. Then, the page generation system receives feature information input by the user that is associated with the target style information and stores the mapping relationship between feature information and target style information in a style specification library. This allows users to obtain target style information by uploading images to the page generation system, thus avoiding the need for users to manually write target style information and making it more convenient to obtain target style information.

[0176] In some embodiments, Figure 3Step S106 may specifically include the following steps: determining multiple design style information in the style specification library that matches the feature information; determining the target style information with the highest similarity to the feature information among the multiple design style information.

[0177] In this process, after identifying multiple design style information matching the feature information in the style specification library, the page generation system calculates the similarity between the keywords and feature information of each design style information, selecting the design style information with the highest similarity as the target style information. Therefore, the page generation system completes the matching automatically, eliminating the need for manual user intervention and reducing user steps and time costs. Furthermore, during the matching process, the page generation system automatically filters based on similarity, ultimately selecting the target style information that best matches the theme characteristics of the target data, thus ensuring a high degree of consistency between the visual style of the generated target page and the theme characteristics of the target data.

[0178] For example, assuming the target data uploaded by the user is stock financial report data, the page generation system determines the characteristic information based on the stock financial report data as the keyword "stock".

[0179] The page generation system matched three design style information in the style specification library, namely design style information S1, design style information S2 and design style information S3.

[0180] In the style specification library, the keywords corresponding to design style information S1 include "stocks, finance, investment and financial statements", the keywords corresponding to design style information S2 include "stock investment", and the keywords corresponding to design style information S3 include "equity funds".

[0181] After the page generation system finds three design style information that match the keyword "stock" in the style specification library, it also needs to calculate the similarity between the keyword "stock" in the target data and the keywords corresponding to the three design style information.

[0182] Suppose that the similarity between the keyword "stock" in the target data calculated by the page generation system and the keyword "stock, finance, investment and financial statements" corresponding to the design style information S1 is 0.95, the similarity between the keyword "stock" in the target data calculated by the page generation system and the keyword "stock investment" corresponding to the design style information S2 is 0.8, and the similarity between the keyword "stock" in the target data calculated by the page generation system and the keyword "equity fund" corresponding to the design style information S3 is 0.7.

[0183] Then, the page generation system found that the highest similarity among the three similarity scores was 0.95 for design style information S1, indicating that design style information S1 is the most matched with the theme features of the target data, and thus design style information S1 was identified as the target style information.

[0184] In some embodiments, Figure 3 Step S106 may specifically include the following steps: determining multiple design style information that matches the feature information in the style specification library; displaying multiple design style information; receiving a confirmation instruction from the user, the confirmation instruction being used to indicate that the target style information among the multiple design style information is selected by the user; and determining the selected target style information among the multiple design style information.

[0185] In this process, after identifying multiple design styles in the style specification library that match the feature information, the page generation system displays these styles to allow the user to select the one they deem suitable. The selected style is then designated as the target style. Regardless of the user's choices within the style specification library, the chosen style always matches the feature information, providing the user with a choice and thus fulfilling their personalized needs.

[0186] For example, suppose the target data uploaded by the user is stock market trend data, the page generation system determines the feature information as the keyword "stock" based on the target data.

[0187] The page generation system matched three design style information in the style specification library. These three design style information are design style information S1, design style information S2 and design style information S3.

[0188] The keywords for design style information S1 include: "stocks, finance, investment, professionalism"; the keywords for design style information S2 include: "stocks, technology, AI, data"; and the keywords for design style information S3 include: "stocks, games, concept stocks".

[0189] Then, the page generation system displays three design style options on the presentation interface, waiting for the user to choose.

[0190] After seeing the three design style options on the page generation system's display interface, the user selected design style option S2.

[0191] At this point, the page generation system will receive a confirmation command from the user. Based on this confirmation command, the page generation system can determine that the user has selected design style information S2.

[0192] In some embodiments, the feature information includes at least one keyword; determining the feature information based on the target data may include the following steps: performing semantic recognition on the target data to obtain at least one keyword.

[0193] After acquiring the target data, the page generation system performs semantic recognition to obtain at least one keyword, thus acquiring feature information. In subsequent steps, the page generation system can then determine the target style information that matches the feature information from the style specification library. Therefore, users only need to provide the target data, and the page generation system can automatically complete the entire process from data understanding to keyword extraction and style matching, further reducing the operational threshold for users.

[0194] See Figure 4 , Figure 4 This is a structural block diagram of a page generation apparatus provided for one or more embodiments of this specification. Figure 4 As shown, the page generation device 700 may include: an acquisition module 701, a determination module 702, a matching module 703, a generation module 704, a requirement expansion module 705, a component identification module 706, and a style extraction module 707.

[0195] The acquisition module 701 is used to acquire target data, which describes the page content of the page to be generated.

[0196] The determination module 702 is used to determine feature information based on the target data. The feature information is used to characterize the thematic features of the target data.

[0197] The matching module 703 is used to determine the target style information in the style specification library that matches the feature information. The target style information is used to describe the page style of the page to be generated.

[0198] Module 704 is used to generate a target page based on target data and target style information.

[0199] In some embodiments, the page generation apparatus 700 may further include a requirement expansion module 705.

[0200] In some embodiments, the requirement expansion module 705 is configured to: obtain initial requirement information, which describes the page generation requirements of the page to be generated; determine whether the initial requirement information meets the completeness requirements; if the initial requirement information does not meet the completeness requirements, determine the target keywords in the initial requirement information; determine at least one historical requirement document in the requirement document library that matches the target keywords; and generate complete requirement information based on the initial requirement information and at least one historical requirement document, which describes the page structure of the page to be generated.

[0201] In some embodiments, the requirement expansion module 705 is specifically used to determine multiple historical requirement documents in the requirement document library that match the target keyword; and to determine at least one historical requirement document among the multiple historical requirement documents whose similarity to the target keyword is greater than a preset threshold.

[0202] In some embodiments, the completeness conditions include: determining whether the initial requirements information contains a page type, determining whether the initial requirements information contains a page functional module, determining whether the initial requirements information contains a page interaction flow, and determining whether the initial requirements information contains a page functional boundary; at least one historical requirements document includes a page type, a page functional module, a page interaction flow, and a page functional boundary; and complete requirements information includes a page type, a page functional module, a page interaction flow, and a page functional boundary.

[0203] In some embodiments, the generation module 704 is specifically used to generate a target page based on target data, complete requirement information, and target style information.

[0204] In some embodiments, the page generation apparatus 700 may further include a component identification module 706.

[0205] In some embodiments, the component identification module 706 is configured to acquire a first image, the first image being an image containing a first reference page; determine the component information of each component in the first reference page; determine the hierarchical relationship between each component; and generate page layout information based on the component information of each component and the hierarchical relationship between each component, the page layout information being used to describe the page structure of the page to be generated.

[0206] In some embodiments, component information includes component category, component location information, component text, component color, and component geometric attributes.

[0207] In some embodiments, the component identification module 706 is specifically used to determine the inclusion relationship between components based on the component location information of each component; and to determine the hierarchical relationship between components based on the inclusion relationship between each component.

[0208] In some embodiments, the first reference page includes a first component and a second component.

[0209] In some embodiments, the component identification module 706 is specifically configured to: determine the first component as the parent and the second component as the child when the first component contains the second component; determine the second component as the parent and the first component as the child when the second component contains the first component; and determine the first component and the second component as siblings when the first component and the second component partially overlap or do not overlap.

[0210] In some embodiments, the component identification module 706 is specifically used to generate a target page based on target data, page layout information, and target style information.

[0211] In some embodiments, the page generation apparatus 700 may further include a style extraction module 707.

[0212] In some embodiments, the style extraction module 707 is used to acquire a second image, which is an image containing a second reference page; perform visual analysis on the second reference page to obtain target style information; receive feature information associated with the target style information input by the user; and store the mapping relationship between feature information and target style information in a style specification library.

[0213] In some embodiments, the matching module 703 is specifically used to determine multiple design style information in the style specification library that matches the feature information; and to determine the target style information with the highest similarity to the feature information among the multiple design style information.

[0214] In some embodiments, the matching module 703 is specifically used to determine multiple design style information that match the feature information in the style specification library; display the multiple design style information; receive a confirmation instruction input by the user, the confirmation instruction being used to indicate that the target style information among the multiple design style information is selected by the user; and determine the selected target style information among the multiple design style information.

[0215] In some embodiments, the feature information includes at least one keyword.

[0216] In some embodiments, the determining module 702 is specifically used to perform semantic recognition on the target data to obtain at least one keyword.

[0217] In some embodiments, the target style information includes at least one of color specifications, font specifications, spacing specifications, rounded corner specifications, shadow specifications, icon specifications, or button specifications, and the font specifications include at least one of font size, font weight, line height, and character spacing.

[0218] See Figure 5 , Figure 5 This is a structural block diagram of an electronic device provided for one or more embodiments of this specification. Figure 5 As shown, the electronic device 800 may include a processor 801 and a memory 802; the memory 802 may be coupled to the processor 801. It is worth noting that... Figure 5 This is an example; other types of structures can also be used to supplement or replace this structure to achieve telecommunications functions or other functions.

[0219] In one possible implementation, the functionality of the page generation device 700 can be integrated into the processor 801. The processor 801 can be configured to perform the following operations: acquire target data, which describes the page content of the page to be generated; determine feature information based on the target data, which characterizes the thematic features of the target data; determine target style information in a style specification library that matches the feature information, which describes the page style of the page to be generated; and generate a target page based on the target data and the target style information.

[0220] In another possible implementation, the page generation device 700 can be configured separately from the processor 801, for example, the page generation device 700 can be configured as a chip connected to the processor 801.

[0221] Furthermore, in some alternative implementations, the electronic device 800 may also include: a communication module, an input unit, an audio processor, a display, a power supply, etc. It is worth noting that the electronic device 800 is not necessarily required to include these components. Figure 5 All components shown; in addition, the electronic device 800 may also include Figure 5 For components not shown, please refer to existing technologies.

[0222] In some alternative implementations, the processor 801, sometimes also referred to as a controller or operating control, may include a microprocessor or other processor device and / or logic device, which receives input and controls the operation of various components of the electronic device 800.

[0223] The memory 802 may be, for example, one or more of a cache, flash memory, hard drive, removable medium, volatile memory, non-volatile memory, or other suitable devices. It may store the information related to the page generation apparatus 700, and may also store a program for executing that information. The processor 801 may execute the program stored in the memory 802 to perform information storage or processing, etc.

[0224] An input unit can provide input to the processor 801. This input unit may be, for example, a button or touch input device. A power supply can be used to provide power to the electronic device 800. A display can be used to display images and text, etc. This display may be, for example, an LCD display, but is not limited to this.

[0225] Memory 802 can be a solid-state memory, such as read-only memory (ROM), random access memory (RAM), SIM card, etc. It can also be a memory that retains information even when power is off, can be selectively erased, and contains more data; examples of this type of memory are sometimes referred to as EPROM, etc. Memory 802 can also be some other type of device. Memory 802 includes buffer memory (sometimes referred to as a buffer). Memory 802 may include an application / function storage unit for storing application programs and function programs or processes for executing the operation of electronic device 800 via processor 801.

[0226] The memory 802 may also include a data storage section for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage section of the memory 802 may include various drivers for the computer device for communication functions and / or for performing other functions of the computer device (such as messaging applications, address book applications, etc.).

[0227] The communication module is a transmitter / receiver that sends and receives signals via an antenna. The communication module (transmitter / receiver) is coupled to the processor 801 to provide input signals and receive output signals, which is the same as in a conventional mobile communication terminal.

[0228] Based on different communication technologies, multiple communication modules can be configured in the same computer device, such as cellular network modules, Bluetooth modules, and / or wireless LAN modules. The communication module (transmitter / receiver) is also coupled to a speaker and microphone via an audio processor to provide audio output through the speaker and receive audio input from the microphone, thereby enabling typical telecommunications functions. The audio processor can include any suitable buffer, decoder, amplifier, etc. Additionally, the audio processor is coupled to processor 801, enabling on-device recording via the microphone and on-device playback of stored sound via the speaker.

[0229] One or more embodiments of this specification also provide a computer-readable storage medium capable of implementing all steps of the page generation method in the above embodiments. The computer-readable storage medium stores a computer program that, when executed by a processor, implements all steps of the page generation method in the above embodiments. For example, when the processor executes the computer program, it implements the following steps: acquiring target data, the target data being used to describe the page content of the page to be generated; determining feature information based on the target data, the feature information being used to characterize the thematic features of the target data; determining target style information in a style specification library that matches the feature information, the target style information being used to describe the page style of the page to be generated; and generating a target page based on the target data and the target style information.

[0230] While this specification provides method operation steps as shown in the embodiments or flowcharts through one or more embodiments, more or fewer operation steps may be included based on conventional or non-inventive labor. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only possible execution order. In actual device or client product execution, the method may be executed sequentially as shown in the embodiments or drawings, or in parallel (e.g., in a parallel processor or multi-threaded processing environment).

[0231] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, apparatus (systems), or computer program products. Therefore, the embodiments of this specification can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, one or more embodiments of this specification can 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.

[0232] This specification describes one or more embodiments of a method, apparatus (system), and computer program product according to one or more embodiments of this specification with reference to flowchart illustrations and / or block diagrams. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0233] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0234] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0235] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device and system embodiments are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0236] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Those skilled in the art will understand the specific meaning of the above terms in one or more embodiments of this specification, depending on the specific circumstances.

[0237] It should be noted that, unless otherwise specified, one or more embodiments and features thereof in this specification can be combined with each other. This specification is not limited to any single aspect, nor to any single embodiment, nor to any combination and / or substitution of such aspects and / or embodiments. Furthermore, each aspect and / or embodiment of one or more embodiments of this specification can be used alone or in combination with one or more other aspects and / or embodiments thereof.

[0238] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of one or more embodiments of this specification, and are not intended to limit them. Although one or more embodiments of this specification have been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of one or more embodiments of this specification, and they should all be covered within the scope of the claims and the specification of one or more embodiments of this specification.

[0239] The foregoing description of one or more embodiments of this specification has been provided in conjunction with optional implementation methods. However, these embodiments are merely exemplary and serve only an illustrative purpose. Based on this, various substitutions and modifications can be made to one or more embodiments of this specification, all of which fall within the protection scope of one or more embodiments of this specification.

Claims

1. A page generation method, characterized in that, The method includes: Obtain target data, which describes the page content of the page to be generated; Feature information is determined based on the target data, and the feature information is used to characterize the thematic features of the target data; Determine the target style information in the style specification library that matches the feature information, the target style information being used to describe the page style of the page to be generated; A target page is generated based on the target data and the target style information.

2. The method as described in claim 1, characterized in that, Before generating the target page based on the target data and the target style information, the method further includes: Obtain initial requirement information, which describes the page generation requirements of the page to be generated; Determine whether the initial requirement information meets the completeness requirements; If the initial requirement information does not meet the completeness condition, the target keywords in the initial requirement information are determined. Identify at least one historical requirement document in the requirement document library that matches the target keyword; Complete requirement information is generated based on the initial requirement information and the at least one historical requirement document. The complete requirement information is used to describe the page structure of the page to be generated.

3. The method as described in claim 2, characterized in that, Identify at least one historical requirement document in the requirement document library that matches the target keyword, including: Identify multiple historical requirement documents in the requirement document library that match the target keyword; Among the plurality of historical demand documents, at least one historical demand document is identified whose similarity to the target keyword is greater than a preset threshold.

4. The method as described in claim 2, characterized in that: The completeness condition includes at least one of the following: determining whether the initial requirement information contains a page type, determining whether the initial requirement information contains a page functional module, determining whether the initial requirement information contains a page interaction flow, and determining whether the initial requirement information contains a page functional boundary. The at least one historical requirement document includes at least one of the following: the page type, the page functional modules, the page interaction flow, and the page functional boundaries; The complete requirements information includes at least one of the following: the page type, the page functional modules, the page interaction flow, and the page functional boundaries.

5. The method according to any one of claims 2 to 4, characterized in that, Generate a target page based on the target data and the target style information, including: The target data, the complete requirement information, and the target style information are used to construct a fusion prompt word; The target page is generated by calling a large language model based on the fused prompt words.

6. The method as described in claim 1, characterized in that, Before generating the target page based on the target data and the target style information, the method further includes: Obtain the first image, which is an image containing the first reference page; Determine the component information of each component in the first reference page; Determine the hierarchical relationship between the components; Page layout information is generated based on the component information of each component and the hierarchical relationship between the components. The page layout information is used to describe the page structure of the page to be generated.

7. The method as described in claim 6, characterized in that, Determine the component information of each component in the first reference page, including: The visual component detection model is invoked to identify the first reference page in the first image, and the component information of each component in the first reference page is obtained. The component information includes component category, component position information, component text, component color and component geometric attributes.

8. The method as described in claim 7, characterized in that, Determining the hierarchical relationship between the components includes: Based on the component location information of each component, the inclusion relationship between the components is determined; Based on the inclusion relationship between the components, the hierarchical relationship between the components is determined.

9. The method as described in claim 8, characterized in that, The first reference page includes a first component and a second component; Based on the inclusion relationships between the components, the hierarchical relationships between the components are determined, including: If the first component contains the second component, the first component is determined as the parent and the second component is determined as the child. If the second component contains the first component, the second component is determined as the parent and the first component is determined as the child. If the first component and the second component partially overlap or do not overlap, the first component and the second component are determined to be at the same level.

10. The method according to any one of claims 6 to 9, characterized in that, Generate a target page based on the target data and the target style information, including: The target data, the page layout information, and the target style information are used to construct a fusion prompt word; The target page is generated by calling a large language model based on the fused prompt words.

11. The method as described in claim 1, characterized in that, The method further includes: Obtain the second image, which is an image containing the second reference page; The target style information is obtained by performing visual analysis on the second reference page; Receive the feature information associated with the target style information input by the user; The mapping relationship between the feature information and the target style information is stored in the style specification library.

12. The method as described in claim 1, characterized in that, Determine the target style information in the style specification library that matches the feature information, including: Identify multiple design style information entries in the style specification library that match the feature information; The target style information with the highest similarity to the feature information is determined from the plurality of design style information.

13. The method as described in claim 1, characterized in that, Determine the target style information in the style specification library that matches the feature information, including: Identify multiple design style information entries in the style specification library that match the feature information; Displays the multiple design style information; Receive a confirmation command input by the user, the confirmation command being used to indicate that the target style information among the plurality of design style information is selected by the user; The target style information selected from the plurality of design style information is determined.

14. The method as described in claim 1, characterized in that, The feature information includes at least one keyword; Determining feature information based on the target data includes: Semantic recognition is performed on the target data to obtain at least one keyword.

15. The method as described in claim 1, characterized in that, The target style information includes at least one of color specifications, font specifications, spacing specifications, rounded corner specifications, shadow specifications, icon specifications, or button specifications, and the font specifications include at least one of font size, font weight, line height, and character spacing.

16. The method as described in claim 1, characterized in that, Generate a target page based on the target data and the target style information, including: The target data and the target style information are used to construct a fusion prompt word; The target page is generated by calling a large language model based on the fused prompt words.

17. A page generation apparatus, characterized in that, The device includes: The acquisition module is used to acquire target data, which describes the page content of the page to be generated; The determination module is used to determine feature information based on the target data, wherein the feature information is used to characterize the thematic features of the target data; A matching module is used to determine target style information in the style specification library that matches the feature information, wherein the target style information is used to describe the page style of the page to be generated; The generation module is used to generate a target page based on the target data and the target style information.

18. An electronic device, characterized in that, The electronic device includes: Memory, used to store computer program products; A processor is configured to execute a computer program product stored in the memory, wherein, when the computer program product is executed, it implements the method described in any one of claims 1 to 16.

19. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which, when executed, implement the method described in any one of claims 1 to 16.

Images