Information processing method and device, readable storage medium and program product

Abstract

The embodiment of the invention provides an information processing method and device, a readable storage medium and a program product, and the method comprises the steps: obtaining a first rendering code through a Web container, and obtaining a first resolution; wherein the first rendering code is a code used for rendering the first Web page; the first resolution is used for displaying the resolution of the first Web page; sending the first rendering code and the first resolution; associated information related to the exception in the first Web page is obtained, the associated information is determined based on the first rendering code, and the associated information comprises at least one of exception information, or an exception repair code or a processing instruction based on the exception information; wherein the exception repair code comprises a code for repairing the exception in the first Web page; wherein the exception in the first Web page is determined by the first model according to the visual analysis of the second Web page; the second Web page is obtained after page rendering is conducted on the basis of the first rendering code.

Description

Technical Field

[0001] This disclosure relates to the field of Internet technology, and in particular to an information processing method, device, readable storage medium, and program product. Background Technology

[0002] Web containers can be used to load web page content on terminal devices. For example, applications using a hybrid development model can use web containers to load web pages.

[0003] The stability and integrity of a web page display are closely related to user experience. However, in some application scenarios, web pages may malfunction, causing them to fail to display accurately or even provide the page's functionality to the user. Summary of the Invention

[0004] This disclosure provides an information processing method, apparatus, readable storage medium, and program product.

[0005] In a first aspect, embodiments of this disclosure provide an information processing method, including:

[0006] The first rendering code and the first resolution are obtained through the Web container; wherein, the first rendering code is code used to render the first Web page; and the first resolution is the resolution used to display the first Web page.

[0007] Send the first rendering code and the first resolution;

[0008] Obtain association information related to the anomaly in the first Web page. The association information is determined based on the first rendering code. The association information includes at least one of the following: anomaly information, anomaly repair code, or processing instructions based on the anomaly information. The anomaly repair code includes code for repairing the anomaly in the first Web page. The anomaly in the first Web page is determined by the first model based on visual analysis of the second Web page. The second Web page is obtained after page rendering based on the first rendering code.

[0009] Secondly, embodiments of this disclosure provide an information processing method, including:

[0010] Receive first rendering information; wherein the first rendering information is information for rendering a first Web page, including first rendering code and a first resolution; the first rendering code is obtained via a Web container, and the first resolution is the resolution used to display the first Web page;

[0011] Based on the first resolution, the cloud rendering service renders the page according to the first rendering code to obtain the second web page;

[0012] A first image of the second web page is obtained, and the first image is sent to a first model for visual analysis. The first model determines whether there is an anomaly in the first web page based on the visual analysis.

[0013] Thirdly, embodiments of this disclosure provide an information processing apparatus, including:

[0014] The first acquisition unit is configured to acquire first rendering code and first resolution via a Web container; wherein the first rendering code is code used to render a first Web page; and the first resolution is the resolution used to display the first Web page.

[0015] The first sending unit is configured to send the first rendering code and the first resolution;

[0016] The second acquisition unit is used to acquire association information related to the anomaly in the first Web page. The association information is determined based on the first rendering code and includes at least one of the following: anomaly information, anomaly repair code, or processing instructions based on the anomaly information. The anomaly repair code includes code for repairing the anomaly in the first Web page. The anomaly in the first Web page is determined by the first model based on visual analysis of the second Web page. The second Web page is obtained after page rendering based on the first rendering code.

[0017] Fourthly, embodiments of this disclosure provide an information processing apparatus, including:

[0018] A first receiving unit is configured to receive first rendering information; wherein the first rendering information is information for rendering a first Web page, including first rendering code and a first resolution; the first rendering code is obtained via a Web container, and the first resolution is the resolution used to display the first Web page;

[0019] A rendering unit is used to render a page based on the first resolution using a cloud rendering service according to the first rendering code, to obtain a second Web page;

[0020] The parsing unit is used to acquire a first image of the second web page, send the first image to a first model for visual analysis, and the first model determines whether there is an anomaly in the first web page based on the visual analysis.

[0021] Fifthly, embodiments of this disclosure provide an electronic device, including: a processor and a memory;

[0022] The memory stores computer-executed instructions;

[0023] The processor executes computer execution instructions stored in the memory, causing the at least one processor to perform the methods described in the first aspect, the second aspect, and various possible designs of the first and second aspects.

[0024] In a sixth aspect, embodiments of this disclosure provide a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the methods described in the first aspect, the second aspect, and various possible designs of the first and second aspects.

[0025] In a seventh aspect, embodiments of this disclosure provide a computer program product, including a computer program that, when executed by a processor, implements the methods described in the first aspect, the second aspect, and various possible designs of the first and second aspects.

[0026] The information processing method, apparatus, readable storage medium, and program product provided in this embodiment can accurately identify web page anomalies that are difficult to determine based on code rules or image comparison schemes by utilizing the visual effects of a reproduced web page rendered in a web container and calling a model to detect the visual performance of these visual effects. Furthermore, this solution can achieve precise correlation analysis between anomalies and the display environment, solving the problem of reproduction and localization. It can achieve web page anomaly detection over a wide range at a lower cost, improving the efficiency of web page detection. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 A flowchart illustrating the information processing method provided in this embodiment of the disclosure;

[0029] Figure 2 A flowchart illustrating the information processing method provided in this embodiment of the disclosure;

[0030] Figure 3 Another schematic flowchart of the information processing method provided in this disclosure embodiment;

[0031] Figure 4A structural block diagram of the information processing apparatus provided in the embodiments of this disclosure;

[0032] Figure 5 A structural block diagram of the information processing apparatus provided in the embodiments of this disclosure;

[0033] Figure 6 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0035] Terminal devices, including mobile devices, can load web pages as needed, such as within a web container (e.g., a WebView). In some cases, web pages may malfunction, even rendering their functionality unusable. For example, a mismatch between the page's layout and / or style and the resolution corresponding to the rendered page can cause rendering errors, such as misaligned layouts, missing styles, or misplaced elements, potentially affecting user comprehension of the page content or hindering user operations. Furthermore, page functionality may fail due to page content version updates, API call identification issues, or process blockages. Additionally, attacks targeting the web container or malicious script injection can cause page loading failures.

[0036] Some related technologies can perform rule-based rendering anomaly detection by parsing the Document Object Model (DOM) tree structure and verifying for missing or redundant nodes. For example, DOM path analysis can be used to determine if there are missing elements. However, such methods can only cover obvious code anomalies (such as missing buttons and broken links), and cannot effectively identify pure style anomalies (such as font size or color deviations) or minor layout differences, and they are also costly to maintain.

[0037] Some related technologies compare rendered page screenshots with standard screenshots to identify web page anomalies, but this only covers style and layout errors. Furthermore, this approach is sensitive to environmental factors, such as resolution and font size; minor style or layout differences can easily generate false positives. It also relies heavily on standard web page images, resulting in high maintenance costs.

[0038] Therefore, there is a need for a low-cost web page anomaly identification solution that can cover a wide range of anomalies.

[0039] The solution disclosed herein sends the rendering code of a web page to a processing end via a web container. The processing end then renders the page, obtains an image of the rendered result (e.g., a screenshot), and calls a first model for visual analysis to parse web page anomalies. The first model can also provide repair strategies for these anomalies. Leveraging the image recognition capabilities of the first model, page anomalies can be accurately located, enabling the detection of a wide range of web page anomalies at a relatively low cost.

[0040] Please refer to Figure 1 , Figure 1 Flowchart of the information processing method provided in this disclosure Figure 1 ,like Figure 1 As shown, the method includes the following steps:

[0041] S101: Obtain the first rendering code and the first resolution via the Web container; wherein, the first rendering code is the code used to render the first Web page; and the first resolution is the resolution used to display the first Web page.

[0042] The execution subject of the information processing method provided in this embodiment can be an electronic device. In one example, the execution subject can be an application running on the electronic device. The application can use a web container to load web pages. Web pages can be, for example, payment pages.

[0043] In response to a user's command to open a first web page within the application's interface, the web container can retrieve the page resources corresponding to the first web page based on the command and the corresponding Uniform Resource Locator (URL). These page resources include one or more of the following: HyperText Markup Language (HTML) files, Cascading Style Sheets (CSS) files, logic or interactive code (e.g., written in JavaScript), and media files such as images and fonts.

[0044] The aforementioned HTML file includes the page skeleton of the first web page. The CSS file includes the stylesheets of the first web page. Logic code or interaction code may include, but is not limited to: page initialization logic, data processing and transformation logic, business processes, state management code, user interaction and event handling code such as UI component behaviors, communication and data interaction code such as application programming interface calls, and dynamic content and structure operation code such as dynamic content generation and insertion, and dynamic style control. After downloading the page resources of the first web page, the web container can parse and calculate these page resources to obtain the first rendering code. The first rendering code is used to render the first web page.

[0045] The first rendering code includes, but is not limited to: structural state snapshots, visual style snapshots, layout geometry snapshots, and rendering environment snapshots.

[0046] The structural state snapshot includes, but is not limited to, at least one of the following: all nodes of the current DOM tree and their hierarchical relationships; content dynamically added, deleted, or modified by the logic code; and the current value, selection state, and other interactive states of form elements.

[0047] A visual style snapshot includes at least one of the following: the final calculated style value applied to each element; the absolute unit conversion result of style properties (such as relative units em (relative to the font size of the parent element), rem (relative to the font size of the root element), vw (viewport width percentage unit), etc. converted to pixels); and a set of visually relevant CSS properties.

[0048] A layout geometry snapshot includes at least one of the following: the precise pixel position of the element in the viewport; bounding box size, padding, border, and margin; stacking context order and layer information.

[0049] A snapshot of the rendering environment includes at least one of the following: viewport size, device pixel ratio, orientation; active media query breakpoint; font loading status, image loading status.

[0050] The aforementioned execution entity can read the width and height properties of the Web container control to obtain its actual display size on the screen, which serves as the first resolution.

[0051] S102: Send the first rendering code and the first resolution.

[0052] S103: Obtain association information related to the anomaly in the first Web page. The association information is determined based on the first rendering code and includes at least one of the following: anomaly information, anomaly repair code, or processing instructions based on the anomaly information. The anomaly repair code includes code for repairing the anomaly in the first Web page. The anomaly in the first Web page is determined by the first model based on visual analysis of the second Web page. The second Web page is obtained after page rendering based on the first rendering code.

[0053] The aforementioned execution entity or web container can send the first rendering code and the first resolution to the server. After receiving the first rendering code and the first resolution, the server can invoke the cloud rendering service to render the web page according to the first resolution and the first rendering code, thus obtaining the second web page.

[0054] The server can take a screenshot of the second web page to obtain the first image, or save the second web page as the first image by calling an algorithm that stores web pages as images. Then, it calls the first model to perform visual analysis on the first image and determines whether the second web page has any anomalies based on the analysis results.

[0055] In some implementations, the method further includes:

[0056] The second web page is obtained by rendering the web page according to the first resolution based on the first rendering code.

[0057] The first image corresponding to the second web page is sent to the first model, wherein the first image records the display content of the second web page, and the first model performs the following first operation:

[0058] Perform visual analysis on the first image and determine whether there is an anomaly in the first image based on the visual analysis results; in response to determining that there is an anomaly in the first image, generate related information based on the anomaly.

[0059] In these implementations, the aforementioned execution entity can render a web page in the background according to the first rendering code and at the first resolution to obtain a second web page. It can then take a screenshot of the second web page or obtain a first image of the second web page using a preset method. The first image is sent to a first model through a preset interface. The first model can run on the current electronic device or on other electronic devices that have established a communication connection with the electronic device. The first model can perform visual analysis on the first image to detect any anomalies. If an anomaly is found, relevant related information can be generated based on the anomaly.

[0060] Here, the aforementioned first model can be any model capable of feature extraction and semantic understanding of images. In some implementations, a first model capable of semantic understanding and visual processing of web pages can be obtained by fine-tuning a model with visual processing capabilities.

[0061] The first model can visually analyze the first image to obtain anomalies in the second web page, and then output related information about the anomalies. This related information may include, but is not limited to, anomaly information, anomaly repair code, or processing instructions based on the anomaly information. Anomalies in the second web page can be considered as anomalies corresponding to the first web page displayed on the electronic device.

[0062] Anomalies on the first web page include one or more of the following:

[0063] Layout error, missing elements, missing styles, abnormal content, or logical or interactive errors.

[0064] The aforementioned layout anomalies refer to unreasonable spatial arrangement, relative positions, and size ratios of page elements, resulting in visual structural chaos or impaired functional usability. Examples include element misalignment and overlap, incorrect size calculations, and abnormal alignment and spacing.

[0065] Among these, element misalignment and overlap include, but are not limited to: key interactive elements (such as buttons and input boxes) being obscured by other content; abnormal overlap between columns in a multi-column layout; and floating elements being removed from the normal document flow, disrupting the overall layout structure.

[0066] Size calculation errors include, but are not limited to: element width or height abnormally expanding or shrinking beyond container boundaries; responsive layouts not adjusting size as expected in a specific viewport; percentage dimensions and viewport units not calculated according to mathematical expectations.

[0067] Alignment and spacing anomalies include, but are not limited to: elements in the same row or group not aligned as designed (left alignment, center alignment, etc.); inconsistent or missing vertical or horizontal spacing between elements; and deviations in text baseline alignment, resulting in uneven text.

[0068] The aforementioned missing elements refer to HTML elements that should exist on the page and be visible to the user failing to render and display correctly. These can include: complete absence, partial absence, conditional absence, and temporary absence. Complete absence includes: an entire functional block being blank, and key interactive elements such as navigation menus and operation buttons being completely invisible. Partial absence includes: some child elements of a composite element being missing (e.g., a form only displaying labels without input fields), and list items being incomplete. Conditional absence includes: elements that only appear on a specific device or after a specific interaction not being displayed, and lazy-loaded elements not being loaded even after the trigger condition is met. Temporary absence includes: elements that appear briefly during page loading and then disappear.

[0069] The aforementioned style deficiencies refer to page elements failing to correctly apply preset CSS style rules, resulting in a significant deviation from design specifications in terms of visual appearance. This can include, but is not limited to: styles completely failing, partial style properties failing, and style priority conflicts. Complete style failures include: elements displaying with the browser's default styles (no custom styles), and the entire stylesheet not loading or failing to load. Partial style property failures include: key visual properties such as color, font, and background not taking effect; decorative styles such as borders, shadows, and rounded corners being missing; and animations or transitions not being triggered. Style priority conflicts include: important styles being unexpectedly overridden by lower-priority rules, and conflicts between inline styles and external stylesheets.

[0070] Content anomalies refer to abnormal states where the text, images, multimedia, and other content within page elements are incorrect, incomplete, formatted incorrectly, or inconsistent with the context. Content anomalies include, but are not limited to, text content anomalies, media content anomalies, and data binding anomalies. Text content anomalies may include, for example, garbled characters or incorrect character encoding, missing key information, improper text truncation leading to incomplete information, and placeholder text not being replaced with actual data. Media content anomalies may include, for example, image loading failure (displaying a split image or a default placeholder image), image distortion (aspect ratio distortion), video or audio not playing or playback control malfunction, and icon fonts not rendering correctly (displayed as squares or garbled characters). Data binding anomalies may include, for example, dynamic data not being correctly bound to the template, incorrect data format (e.g., displaying a date as a timestamp), and multilingual content not being displayed according to the user's language preference.

[0071] Logical or interactive anomalies refer to errors in page interaction logic, state management, or business processes that cause functionality to fail as expected, although the visual appearance may be completely normal. Logical or interactive anomalies include, but are not limited to: interaction logic failure, state management anomalies, business process interruptions, and performance logic anomalies. For example, interaction logic failures may include: unresponsive or incorrect click events, incorrect form validation logic (allowing what should be blocked, and blocking what should be allowed), malfunctioning keyboard navigation or shortcuts, and drag-and-drop operations not working properly. State management anomalies may include: incorrect page state initialization or reset, state synchronization errors (e.g., inconsistent states across multiple tabs), and user session management anomalies (automatic logout, permission errors). Business process interruptions may include: multi-step processes failing to proceed to the next step or prematurely ending, failures in critical business operations such as payment and submission without clear error messages, and looping or incorrect page redirects. Performance logic anomalies may include: infinite loops causing page freezes, memory leaks causing page responsiveness to gradually slow down, and unoptimized rendering leading to frequent redraws and reflows.

[0072] Error correction code can include visual correction code such as layout error correction code, missing element correction code, missing style correction code, and content error correction code, as well as logic correction code or interaction correction code.

[0073] Error handling instructions can include, for example, reloading web page resources and rendering the page, using repair code to repair the first web page, indicating that the page is abnormal and needs to be repaired, etc.

[0074] In this embodiment, by utilizing the visual effects of a reproduced web page rendered in a web container and calling a first model to detect the visual performance of these effects, web page anomalies that are difficult to identify based on code rules or image comparison schemes can be accurately identified, thus solving the problem of reproducing and locating web page anomalies. This achieves coverage of a wide range of web anomaly detection at a relatively low cost, improving the efficiency and accuracy of web page anomaly detection.

[0075] Please refer to Figure 2 , Figure 2 Flowchart of the information processing method provided in this disclosure Figure 2 ,like Figure 2 As shown, the method includes, in addition to, with Figure 1 In addition to step S201, which is the same as step S101 in the illustrated embodiment, the following steps are also included:

[0076] S202: Desensitize the user information in the first rendering code, and send the desensitized first rendering code and the first resolution.

[0077] The information processing method provided in this embodiment can be executed by an application running on an electronic device. This application can use a web container to load web pages. Web pages can be, for example, payment pages.

[0078] In one example, once the client has captured the rendering state of the first web page and obtained the first rendering code, a de-identification process can be triggered. For instance, this code can be loaded into a secure, isolated temporary processing environment. Within this temporary environment, a multi-layered strategy can be employed to scan the rendering code to identify all possible user personal information. This can be done through pattern matching based on preset rules and regular expressions, and / or contextual recognition based on semantic tags and DOM attributes. Pattern matching based on preset rules and regular expressions can include regular expressions and contextual rules set separately for various types of sensitive information, such as regular expressions for identity information, financial accounts, amounts, contact information, addresses, and personal information.

[0079] The above regular expressions can be used to perform a global scan of the first rendered code text, matching all data fragments that conform to these regular expressions.

[0080] Contextual recognition based on semantic tags and DOM attributes can, for example, parse the DOM structure of the rendered code to identify HTML elements and attributes with explicit semantics. In one example, it can... <input> Values ​​with the `type` attribute of an element that are "password", "phone", or "email", or their corresponding input values ​​associated with `<tag>`, are marked as sensitive. In one example, text content containing keywords such as "card number", "username", or "address" in the element's identifier or class attribute is identified as sensitive.

[0081] In some examples, sensitive user information can be removed. However, element tags and key style attributes can be retained to ensure the layout remains unchanged.

[0082] In some implementations, this step further includes: identifying the area where the user information is located from the first rendering code, replacing the user information with replacement content, and obtaining the desensitized first rendering code.

[0083] For example, a sensitive information filtering service can be integrated into the aforementioned execution entity. This sensitive information filtering service could be a local SDK or a trusted edge service. For instance, this service could use a pre-trained named entity recognition model to identify sensitive user information in the first rendering code and pinpoint the regions containing that user information within the first rendering code. The service could then output information identifying the regions containing the user information (e.g., starting index, length, and type).

[0084] In this example, the replacement content can be randomly generated placeholders without any real meaning. That is, the identified real sensitive data is replaced with placeholders that conform to the original data format but lack authenticity. For example, a user's real card number is replaced with a test card number of the same length or a general placeholder, such as "1111111111111111". Another example is replacing the real name "Zhang San" with the general placeholder "*Name*" or "TestUser".

[0085] In these implementations, by identifying the areas where user information is located and replacing sensitive user information with alternative content, the page's layout structure and style calculation basis can be preserved to the greatest extent. This avoids CSS layout errors caused by drastic changes in text length and ensures the accuracy of subsequent cloud rendering.

[0086] The aforementioned executing entity can upload the anonymized first rendering code and first resolution to the server via encrypted channels such as HTTPS.

[0087] S203: Obtain association information related to the exception in the first Web page. The association information is determined based on the first rendering code. The association information includes at least one of the following: exception information, exception repair code, or processing instructions based on exception information; wherein, the exception repair code includes code for repairing the exception in the first Web page.

[0088] The anomalies in the first web page are determined by the first model based on visual analysis of the second web page; the second web page is obtained after page rendering based on the first rendering code.

[0089] For the specific implementation of step S203 above, please refer to Figure 1 Description of relevant parts of the illustrated embodiment.

[0090] S204: Correct the first web page according to the exception repair code; wherein, the associated information includes the exception repair code.

[0091] Visual fixes include code to correct layout errors, missing elements, missing styles, and content errors. It can also include logic or interaction fixes. These fixes can be used to correct the primary web page.

[0092] In some implementations, the repair code includes visual repair code, and / or logical repair code or interactive repair code; step S204 above includes:

[0093] The visual fix code is injected into the first web page via a web container;

[0094] The logic fix code or interactive fix code is applied to the first web page via the web container.

[0095] For example, the layout error correction code and style missing correction code mentioned above are usually CSS code. These two code snippets can be encapsulated in a new... <style>元素中，并为该元素添加唯一标识符。元素缺失修复代码与内容异常修复代码通常为JavaScript代码，用于动态创建DOM节点或更新元素内容。可以将其封装在一个立即执行函数表达式或一个待触发的函数中。

[0096] 对于CSS代码，Web容器中的浏览器渲染引擎将封装好的<style>元素插入到当前第一渲染代码<head>的末尾，利用CSS层叠规则确保其生效。Web容器的浏览器渲染引擎随即重新计算样式并布局，即时修正布局错乱或样式缺失问题。

[0097] 对于修复元素缺失或内容异常的JS代码，Web容器的浏览器渲染引擎可以在如当前宏任务末尾或下一个动画帧执行该代码，动态创建缺失的元素或更正错误的文本、图像内容。

[0098] 对于逻辑修复代码或交互修复代码，可以将逻辑修复代码或交互修复代码注入到第一Web页面。具体地，可以在严格的安全策略下或动态创建<script>标签的方式执行修复代码。此代码可直接操作当前页面的DOM、样式或应用状态。例如，代码可能重新计算并设置某个元素的尺寸，或重新发起失败的数据请求。

[0099] 本实施例中，描述了对第一渲染代码进行用户信息脱敏，发送脱敏后的第一渲染代码和第一分辨率，以及在获取到的关联信息中包括异常修复代码时，根据异常修复代码对第一Web页面进行修正，从而在绝对保障用户隐私的前提下，实现了对Web页面异常的监控与实时修复，将页面异常与具体的显示环境绑定，使虚拟问题可被物理复现，无需用户参与的情况下可以实现Web页面异常的快速修复。

[0100] 在一些实施方式中，该方法还包括：

[0101] 获取与第一Web页面关联的历史跳转记录；历史跳转记录包括多个页面的打开时间和 / 或顺序；

[0102] 发送第一渲染代码和第一分辨率，包括发送历史跳转记录、第一渲染代码及第一分辨率；以及，

[0103] 处理指令包括与第一功能的异常对应的第一处理指令；第一功能为第一Web页面支持的功能；第一处理指令由第一功能的异常及第一功能对应的跳转路径确定，跳转路径包括从历史跳转记录中解析出的多个页面的打开顺序。

[0104] 上述执行主体可以采集与第一Web页面或用户标识关联的历史跳转记录。每一历史跳转记录是一个有序序列，记录了第一Web页面的第一功能（例如支付功能）中，用户依次打开过的多个页面的信息，至少包括每个页面的打开时间戳与打开顺序。历史跳转记录的来源包括：前端路由日志，页面可见性API接口，通过埋点记录的页面跳转数据等。上述执行主体可以将历史跳转记录、脱敏后的第一渲染代码以及第一分辨率，发送至服务端，由服务端将调用第一模型对历史跳转记录进行解析。

[0105] 利用第一模型的归纳能力，过滤掉非主流的分支页面（例如"查看协议”、"找回密码”等页面），抽象出第一功能的跳转路径。该跳转路径以页面序列的形式定义，例如对于一个支付功能，其跳转路径可能为：[收银台页面A，渠道选择页面B，密码输入页面C， 支付成功页面D]。

[0106] 可以对失败的历史跳转记录与上述跳转路径进行比对，通过序列对齐算法，定位失败路径在何处偏离了跳转路径。例如，失败路径为[A，B，X]，其中"X”是一个错误页或未知页面，上述第一模型可以判定异常发生在从页面B到页面C的跳转环节。

[0107] 在这些实施方式中，通过引入历史跳转记录，将异常监控的维度从单个页面的静态状态扩展到了跨多个页面的动态业务流程。从而实现了对不可控第三方页面功能链路的非侵入式检测，以及业务功能异常的精准定位。

[0108] 请参考图3，图3为本公开提供的信息处理方法的流程示意图。如图3所示，该方法包括如下步骤：

[0109] S301：接收第一渲染信息；其中，第一渲染信息为第一Web页面的渲染信息，包括第一渲染代码及第一分辨率；第一渲染代码由Web容器在加载第一Web页面时获取，第一分辨率为用于显示第一Web页面的分辨率。

[0110] 本实施例中，信息处理方法的执行主体可以是应用程序服务端。可以理解的是，本实施例中，从客户端采集第一Web页面对应的第一渲染信息可以按照一定的采集策略进行采集，例如在不同地理区域按照不同的采集比率进行采集，又例如根据按照一定的采集比率从全量服务对象采集等，可以减少服务端的负载。

[0111] 上述第一渲染信息可以是应用程序客户端发送的，或者网页客户端发送的。上述第一渲染信息可以是Web容器对第一Web页面的页面资源进行解析和计算之后确定的。具体地，第一渲染代码和第一分辨率的相关内容可以参考图1所示实施例相关部分的描述，此处不赘述。

[0112] S302：基于第一分辨率利用云端渲染服务对第一渲染代码进行渲染，得到第二Web页面。

[0113] 上述执行主体在接收到上述第一渲染信息后，可以根据第一渲染代码和第一分辨率使用云端渲染服务进行页面渲染，得到第二Web页面。可以理解的是，由于云端渲染服务使用与Web容器相同的第一渲染代码与第一分辨率，因此，云端渲染服务可以再现Web容器渲染第一Web页面的视觉效果。

[0114] S303：截取第二Web页面的第一图像，将第一图像发送至第一模型进行视觉分析，获取第一模型基于视觉分析确定的第一Web页面的渲染异常。

[0115] 上述服务端可以对已渲染的第二Web页面进行屏幕捕捉，得到一张图像，记为第一图像。截图范围可以覆盖第二Web页面的可视视口，并可选择进行全页面滚动截图以捕获长页面内容。图像格式可为PNG或JPEG，并需保证足够的清晰度以供后续分析。

[0116] 可以将第一图像与相关的分析上下文（可选的包括页面URL、分辨率信息、业务类型等）输入至第一模型。该第一模型为一个语言模型与视觉模型的融合体，具备多模态理解能力。第一模型基于其训练所得的网页视觉知识、设计规范和常见缺陷模式，对第一图像进行分析。可以识别像素级的视觉异常（如内容撕裂），理解语义级的异常（如"登录按钮被底部导航栏遮挡导致无法点击”、"价格数字因字体回退而显示为乱码”）。第一模型输出的分析结果即为对第一Web页面渲染异常的确定性描述，包括异常类型、位置坐标和推测的根因。

[0117] 本实施例中，通过利用再现Web页面在Web容器中渲染后的视觉效果，并对该视觉效果调用模型进行视觉表现的检测，能够精准识别基于代码规则或图像对比方案难以确定的Web页面异常，此外，该方案可以实现异常与显示环境的精准关联分析，解决复现定位的问题。可以在较低的成本下实现较大类型的Web页面异常检测，提高了Web页面检测的效率。

[0118] 在一些实施例中，该方法还包括：调用第一模型生成与第一Web页面中的异常相关的关联信息，关联信息包括如下至少之一：异常信息、或异常修复代码、或基于异常信息的处理指令。

[0119] 异常信息、异常修复代码、基于异常信息的处理指令，可以参考图1所示实施例相关部分的描述。

[0120] 在一些实施方式中，异常修复代码包括：

[0121] 视觉修复代码，和 / 或，逻辑修复代码或交互修复代码；

[0122] 基于异常信息的处理指令，包括以下至少之一：

[0123] 重新加载第一Web页面；

[0124] 或者，调整与第一Web页面关联的流程信息。

[0125] 在这些实施方式中，流程调整例如建议前端交互流程的更改（如下发指令引导用户执行某些操作）等。

[0126] 例如第一模型确定第一Web页面的异常原因复杂、无法通过注入修复代码即时解决，或检测到页面核心资源加载失败时，可以生成重新加载指令。该指令下发至客户端。客户端接收到指令后，可以通过调用Web容器提供的页面刷新API接口，或引导用户执行特定交互以触发刷新，从而重新发起对第一Web页面资源的请求与渲染流程，以尝试恢复页面至正常状态。

[0127] 又例如，当第一模型基于对历史跳转记录的分析，确定异常源于相关功能的流程本身（例如跳转路径错误、流程节点配置失效）时，可以生成调整流程信息指令。具体地，可以将失败跳转记录与由成功跳转记录解析的跳转路径进行比对，定位偏离节点，并据此决策流程调整内容。流程调整内容包括：替换后续流程目标，例如将后续应跳转至故障第三方页面的URL，替换为功能等效的备用页面URL；更新流程参数，例如，修改嵌入在跳转链接中的会话标识、业务编码或渠道参数等。

[0128] 在这些实施方式中，通过资源重置与逻辑重配两个层面，为不同根因的异常提供了差异化的恢复机制，共同保障了相关功能的最终可用性。

[0129] 在一些实施方式中，该方法还包括以下一项或多项：

[0130] 将视觉修复代码发送到Web容器；其中，Web容器将视觉修复代码注入到第一Web页面；

[0131] 将逻辑修复代码或交互修复代码发送到Web容器；其中，Web容器将逻辑修复代码或交互修复代码应用到第一Web页面。

[0132] 上述执行主体可以将视觉修复代码发送到Web容器，上述Web容器将视觉修复代码注入到第一Web页面。

[0133] 上述执行主体可以将逻辑修复代码或交互修复代码发送到Web容器，上述Web容器将逻辑或交互修复代应用到第一Web页面，以对第一Web页面进行动态更新。

[0134] 详情可以参考图1或图2所示实施例的相关部分。

[0135] 在这些实施方式中，通过将生成的修复代码定向发送至客户端Web容器并自动执行，建立了从异常感知、分析决策到修复执行的完整自动化闭环，将依赖人工干预的运维流程转变为瞬间响应的智能修复系统，极大提升了运维效率与问题解决速度。通过下发视觉修复代码与逻辑修复代码或交互修复代码，可以针对不同类型的异常根源，提供CSS样式注入或逻辑或交互脚本执行等适配的修复手段，实现了修复策略的精准化与较宽的修复范围覆盖。

[0136] 在一些实施例中，第一Web页面包括与第一功能关联的多个第二页面；该方法还包括：

[0137] 获取与第一功能关联的历史跳转记录，历史跳转记录包括第一功能成功完成的第一跳转记录，以及第一功能未成功完成的第二跳转记录；第一跳转记录包括多个第二页面间的历史跳转记录，第二跳转记录包括多个第二页面间的历史跳转记录；

[0138] 基于第一跳转记录，解析第一功能所需的页面跳转路径；

[0139] 基于页面跳转路径，确定第二跳转记录对应的异常以及异常位置；

[0140] 发送异常的描述信息及异常位置至第一功能的提供方。

[0141] 下面以第一功能为支付功能为例，上述支付功能例如可以由不同支付方式支持。上述执行主体以预设时间长度（例如7天）为时间窗口，基于上一时间窗口内各支付方式下支付成功订单的Web页面跳转记录，归纳出每种支付方式对应的页面跳转路径（同时过滤掉找回密码、查看协议、图形验证码等分支页面），最终结果将以JSON格式输出。

[0142] 例如根据支付成功的跳转记录可以解析出如下的多个跳转记录：跳转记录一：a->b->c，跳转记录二：a->a1->b->c，跳转记录三：a->b->b1->c，跳转记录四：a->a1->d->c，跳转记录五：a->d->d1->c，跳转记录六：a->d->c。根据上述跳转记录可以计算出以下两个跳转路径：跳转路径一：a->b->c，跳转路径二：a->d->c。

[0143] 当支付失败时，将支付失败路径，与跳转路径一和跳转路径二进行对比，可以计算出失败的位置，失败前的页面转化率。

[0144] 在这些实施方式中，上述执行主体可以将上述历史跳转记录以及提示信息发送至第一模型。上述提示信息包括从历史跳转记录中提取第一功能成功完成记录和第一功能未成功完成记录。根据第一功能成功完成记录提取共同的页面，根据这些共同的页面确定跳转路径。然后根据跳转路径解析跳转失败记录对应的异常和异常位置。从而实现了对不可控的页面功能流程的非侵入式精准监控，对功能流程异常进行节点级精准定位与根因关联，提升了问题诊断的精度和效率。此外，基于历史成功记录动态解析出的跳转路径，能够自动跟随功能相关渠道或页面的更新而变化，克服了基于静态配置的监控规则易过时、误报率高的问题，确保了异常分析的时效性与准确性。另外，通过明确异常关联页面为功能提供方提供了异常解决与定位的信息，将跨团队沟通与问题解决从模糊描述推进至基于具体页面和步骤的精准协作，缩短了异常排查与解决的周期。

[0145] 对应于上文图1所示实施例的信息处理方法，图4为本公开实施例提供的信息处理装置的结构框图。为了便于说明，仅示出了与本公开实施例相关的部分。参照图4，信息处理装置40包括：第一获取单元401、第一发送单元402和第二获取单元403。

[0146] 其中，第一获取单元401，用于经由Web容器获取第一渲染代码，并获取第一分辨率；其中，第一渲染代码为用于对第一Web页面进行渲染的代码；第一分辨率为用于显示第一Web页面的分辨率；

[0147] 第一发送单元402，用于发送第一渲染代码和第一分辨率；

[0148] 第二获取单元403，用于获取与第一Web页面中的异常相关的关联信息，关联信息基于第一渲染代码确定，关联信息包括如下至少之一：异常信息、或异常修复代码、或基于异常信息的处理指令；其中，异常修复代码包括对第一Web页面中的异常进行修复的代码；其中，第一Web页面中的异常，由第一模型根据对第二Web页面进行视觉分析确定；第二Web页面基于第一渲染代码进行页面渲染后得到。

[0149] 本公开实施例提供的信息处理装置，通过利用再现Web页面在Web容器中渲染后的视觉效果，并对该视觉效果调用第一模型进行视觉表现的检测，能够精准识别基于代码规则或图像对比方案难以确定的Web页面异常，可以解决Web页面异常复现定位的问题。实现了在较低成本的前提下，实现较大范围的Web异常检测覆盖。提高了Web页面异常检测的效率。

[0150] 在本公开的一个实施例中，关联信息包括异常修复代码；装置40还包括修复单元（图中未示出），修复单元用于：

[0151] 根据异常修复代码对第一Web页面进行修正。

[0152] 在本公开的一个实施例中，第一Web页面中的异常包括以下一项或多项：

[0153] 布局异常、元素缺失、样式缺失、内容异常，或者逻辑或交互异常。

[0154] 在本公开的一个实施例中，第一发送单元402进一步用于：

[0155] 对第一渲染代码进行用户信息脱敏，发送脱敏后的第一渲染代码。

[0156] 在本公开的一个实施例中，第一发送单元402进一步用于：

[0157] 从第一渲染代码中识别出用户信息所在区域，使用替换内容对用户信息进行替换，得到脱敏后的第一渲染代码。

[0158] 在本公开的一个实施例中，第一获取单元401进一步用于：获取与第一Web页面关联的历史跳转记录；历史跳转记录包括多个页面的打开时间和 / 或顺序；

[0159] 第一发送单元402进一步用于：发送历史跳转记录、第一渲染代码及第一分辨率；以及，

[0160] 处理指令包括与第一功能的异常对应的第一处理指令；第一功能为第一Web页面支持的功能；第一处理指令由第一功能的异常及第一功能对应的跳转路径确定，跳转路径包括从历史跳转记录中解析出的多个页面的打开顺序。

[0161] 在本公开的一个实施例中，装置40还包括第二发送单元（图中未示出），第二发送单元用于：

[0162] 根据第一渲染代码按照第一分辨率进行Web页面渲染，得到第二Web页面；

[0163] 将第二Web页面对应的第一图像发送至第一模型，其中，第一图像记录有第二Web页面的显示内容，第一模型执行如下第一操作：

[0164] 对第一图像进行视觉分析，根据视觉分析结果确定第一图像是否有异常；响应于确定第一图像有异常，根据异常生成关联信息。

[0165] 在本公开的一个实施例中，修复代码包括视觉修复代码，和 / 或，逻辑修复代码或交互修复代码；其中，修复单元进一步用于：

[0166] 经由Web容器将视觉修复代码注入到第一Web页面；

[0167] 经由Web容器将逻辑修复代码或交互修复代码应用到第一Web页面。

[0168] 对应于上文图3所示实施例的信息处理方法，图5为本公开实施例提供的信息处理装置的结构框图。为了便于说明，仅示出了与本公开实施例相关的部分。参照图5，信息处理装置50包括：第一接收单元501、渲染单元502和解析单元503。

[0169] 其中，第一接收单元501，用于接收第一渲染信息；其中，第一渲染信息为用于对第一Web页面进行渲染的信息，包括第一渲染代码及第一分辨率；第一渲染代码经由Web容器获取，第一分辨率为用于显示第一Web页面的分辨率；

[0170] 渲染单元502，用于基于第一分辨率利用云端渲染服务根据第一渲染代码渲染页面，得到第二Web页面；

[0171] 解析单元503，用于获取第二Web页面的第一图像，将第一图像发送至第一模型进行视觉分析，由第一模型基于视觉分析确定第一Web页面是否存在异常。

[0172] 本实施例中，通过利用再现Web页面在Web容器中渲染后的视觉效果，并对该视觉效果调用模型进行视觉表现的检测，能够精准识别基于代码规则或图像对比方案难以确定的Web页面异常，此外，该方案可以实现异常与显示环境的精准关联分析，解决复现定位的问题。可以在较低的成本下实现较大范围的Web页面异常检测，提高了Web页面检测的效率。

[0173] 在本公开的一个实施例中，装置50还包括生成单元（图中未示出），生成单元用于：

[0174] 调用第一模型生成与第一Web页面中的异常相关的关联信息，关联信息包括如下至少之一：异常信息、或异常修复代码、或基于异常信息的处理指令。

[0175] 在本公开的一个实施例中，异常修复代码包括：

[0176] 视觉修复代码，和 / 或，逻辑修复代码或交互修复代码；

[0177] 基于异常信息的处理指令，包括以下至少之一：

[0178] 重新加载第一Web页面；

[0179] 或者，调整与第一Web页面关联的流程信息。

[0180] 在本公开的一个实施例中，装置50还包括第二发送单元（图中未示出），第二发送单元用于：

[0181] 将视觉修复代码发送到Web容器；其中，Web容器将视觉修复代码注入到第一Web页面；

[0182] 将逻辑修复代码或交互修复代码发送到Web容器；其中，Web容器将逻辑修复代码或交互修复代码应用到第一Web页面。

[0183] 在本公开的一个实施例中，第一Web页面包括与第一功能关联的多个第二页面；装置50还包括功能异常检测单元（图中未示出），功能异常检测单元用于：

[0184] 获取与第一功能关联的历史跳转记录，历史跳转记录包括第一功能成功完成的第一跳转记录，以及第一功能未成功完成的第二跳转记录；第一跳转记录包括多个第二页面间的打开时间或顺序，第二跳转记录包括多个第二页面间的打开时间或顺序；

[0185] 基于第一跳转记录，解析第一功能所需的跳转路径；跳转路径包括多个关联页面的打开顺序；

[0186] 基于跳转路径，确定第二跳转记录对应的异常以及异常关联页面；

[0187] 发送异常的描述信息及异常关联页面的信息至第一功能的提供方。

[0188] 为了实现上述实施例，本公开实施例还提供了一种电子设备。

[0189] 参考图6，其示出了适于用来实现本公开实施例的电子设备600的结构示意图，该电子设备600可以为终端设备或服务器。其中，终端设备可以包括但不限于诸如移动电话、笔记本电脑、数字广播接收器、个人数字助理（Personal Digital Assistant，简称PDA）、平板电脑（ Tablet Computer）、便携式多媒体播放器（Portable Media Player，简称PMP）、车载终端（例如车载导航终端）等等的移动终端以及诸如数字TV、台式计算机等等的固定终端。图6示出的电子设备仅仅是一个示例，不应对本公开实施例的功能和使用范围带来任何限制。

[0190] 如图6所示，电子设备600可以包括处理装置（例如中央处理器、图形处理器等）601，其可以根据存储在只读存储器（Read Only Memory ，简称ROM）602中的程序或者从存储装置608加载到随机访问存储器（Random Access Memory ，简称RAM）603中的程序而执行各种适当的动作和处理。在RAM 603中，还存储有电子设备600操作所需的各种程序和数据。处理装置601、ROM 602以及RAM 603通过总线604彼此相连。输入 / 输出（I / O）接口605也连接至总线604。

[0191] 通常，以下装置可以连接至I / O接口605：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置606；包括例如液晶显示器（Liquid CrystalDisplay ，简称LCD）、扬声器、振动器等的输出装置607；包括例如磁带、硬盘等的存储装置608；以及通信装置609。通信装置609可以允许电子设备600与其他设备进行无线或有线通信以交换数据。虽然图6示出了具有各种装置的电子设备600，但是应理解的是，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。

[0192] 特别地，根据本公开的实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的实施例包括一种计算机程序产品，其包括承载在计算机可读存储介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的实施例中，该计算机程序可以通过通信装置609从网络上被下载和安装，或者从存储装置608被安装，或者从ROM 602被安装。在该计算机程序被处理装置601执行时，执行本公开实施例的方法中限定的上述功能。

[0193] 需要说明的是，本公开上述的计算机可读存储介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器（RAM）、只读存储器（ROM）、可擦式可编程只读存储器（EPROM或闪存）、光纤、便携式紧凑磁盘只读存储器（CD-ROM）、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读存储介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读存储介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、RF（射频）等等，或者上述的任意合适的组合。

[0194] 上述计算机可读存储介质可以是上述电子设备中所包含的；也可以是单独存在，而未装配入该电子设备中。

[0195] 上述计算机可读存储介质承载有一个或者多个程序，当上述一个或者多个程序被该电子设备执行时，使得该电子设备执行上述实施例所示的方法。

[0196] 可以以一种或多种程序设计语言或其组合来编写用于执行本公开的操作的计算机程序代码，上述程序设计语言包括面向对象的程序设计语言—诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如"C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网（LocalArea Network ，简称LAN）或广域网（Wide Area Network ，简称WAN）—连接到用户计算机，或者，可以连接到外部计算机（例如利用因特网服务提供商来通过因特网连接）。

[0197] 附图中的流程图和框图，图示了按照本公开各种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和 / 或流程图中的每个方框、以及框图和 / 或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。

[0198] 描述于本公开实施例中所涉及到的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。其中，单元的名称在某种情况下并不构成对该单元本身的限定，例如，第一获取单元还可以被描述为"获取至少两个网际协议地址的单元”。

[0199] 本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列（FPGA）、专用集成电路（ASIC）、专用标准产品（ASSP）、片上系统（SOC）、复杂可编程逻辑设备（CPLD）等等。

[0200] 本公开实施例提供的电子设备、计算机可读存储介质、计算机程序产品，通过利用再现Web页面在Web容器中渲染后的视觉效果，并对该视觉效果调用第一模型进行视觉表现的检测，能够精准识别基于代码规则或图像对比方案难以确定的Web页面异常，可以解决Web页面异常复现定位的问题。实现了在较低成本的前提下，实现较大范围的Web异常检测覆盖。提高了Web页面异常检测的效率。

[0201] 第一方面，根据本公开的一个或多个实施例，提供了一种信息处理方法，包括：经由Web容器获取第一渲染代码，并获取第一分辨率；其中，第一渲染代码为用于对第一Web页面进行渲染的代码；第一分辨率为用于显示第一Web页面的分辨率；

[0202] 发送第一渲染代码和第一分辨率；

[0203] 获取与第一Web页面中的异常相关的关联信息，关联信息基于第一渲染代码确定，关联信息包括如下至少之一：异常信息、或异常修复代码、或基于异常信息的处理指令；其中，异常修复代码包括对第一Web页面中的异常进行修复的代码；

[0204] 其中，第一Web页面中的异常，由第一模型根据对第二Web页面进行视觉分析确定；第二Web页面基于第一渲染代码进行页面渲染后得到。

[0205] 根据本公开的一个或多个实施例，关联信息包括异常修复代码；方法还包括：

[0206] 根据异常修复代码对第一Web页面进行修正。

[0207] 根据本公开的一个或多个实施例，第一Web页面中的异常包括以下一项或多项：

[0208] 布局异常、元素缺失、样式缺失、内容异常或者逻辑或交互异常。

[0209] 根据本公开的一个或多个实施例，发送第一渲染代码，包括：

[0210] 对第一渲染代码进行用户信息脱敏，发送脱敏后的第一渲染代码。

[0211] 根据本公开的一个或多个实施例，对第一渲染代码进行用户信息脱敏，包括：

[0212] 从第一渲染代码中识别出用户信息所在区域，使用替换内容对用户信息进行替换，得到脱敏后的第一渲染代码。

[0213] 根据本公开的一个或多个实施例，方法还包括：

[0214] 获取与第一Web页面关联的历史跳转记录；历史跳转记录包括多个页面的打开时间和 / 或顺序；

[0215] 发送第一渲染代码和第一分辨率，包括发送历史跳转记录、第一渲染代码及第一分辨率；以及，

[0216] 处理指令包括与第一功能的异常对应的第一处理指令；第一功能为第一Web页面支持的功能；第一处理指令由第一功能的异常及第一功能对应的跳转路径确定，跳转路径包括从历史跳转记录中解析出的多个页面的打开顺序。

[0217] 根据本公开的一个或多个实施例，方法还包括：

[0218] 根据第一渲染代码按照第一分辨率进行Web页面渲染，得到第二Web页面；

[0219] 将第二Web页面对应的第一图像发送至第一模型，其中，第一图像记录有第二Web页面的显示内容，第一模型执行如下第一操作：

[0220] 对第一图像进行视觉分析，根据视觉分析结果确定第一图像是否有异常；响应于确定第一图像有异常，根据异常生成关联信息。

[0221] 根据本公开的一个或多个实施例，修复代码包括视觉修复代码，和 / 或，逻辑修复代码或交互修复代码；其中，根据异常修复代码对显示的第一Web页面进行修正，包括：

[0222] 经由Web容器将视觉修复代码注入到第一Web页面；

[0223] 经由Web容器将逻辑修复代码或交互修复代码应用到第一Web页面。

[0224] 第二方面，根据本公开的一个或多个实施例，提供了一种信息处理方法，包括：接收第一渲染信息；其中，第一渲染信息为用于对第一Web页面进行渲染的信息，包括第一渲染代码及第一分辨率；第一渲染代码经由Web容器获取，第一分辨率为用于显示第一Web页面的分辨率；

[0225] 基于第一分辨率利用云端渲染服务根据第一渲染代码渲染页面，得到第二Web页面；

[0226] 获取第二Web页面的第一图像，将第一图像发送至第一模型进行视觉分析，由第一模型基于视觉分析确定第一Web页面是否存在异常。

[0227] 根据本公开的一个或多个实施例，方法还包括：

[0228] 调用第一模型生成与第一Web页面中的异常相关的关联信息，关联信息包括如下至少之一：异常信息、或异常修复代码、或基于异常信息的处理指令。

[0229] 根据本公开的一个或多个实施例，异常修复代码包括：

[0230] 视觉修复代码，和 / 或，逻辑修复代码或交互修复代码；

[0231] 基于异常信息的处理指令，包括以下至少之一：

[0232] 重新加载第一Web页面；

[0233] 或者，调整与第一Web页面关联的流程信息。

[0234] 根据本公开的一个或多个实施例，方法还包括以下一项或多项：

[0235] 将视觉修复代码发送到Web容器；其中，Web容器将视觉修复代码注入到第一Web页面；

[0236] 将逻辑修复代码或交互修复代码发送到Web容器；其中，Web容器将逻辑修复代码或交互修复代码应用到第一Web页面。

[0237] 根据本公开的一个或多个实施例，第一Web页面包括与第一功能关联的多个第二页面；方法还包括：

[0238] 获取与第一功能关联的历史跳转记录，历史跳转记录包括第一功能成功完成的第一跳转记录，以及第一功能未成功完成的第二跳转记录；第一跳转记录包括多个第二页面间的打开时间或顺序，第二跳转记录包括多个第二页面间的打开时间或顺序；

[0239] 基于第一跳转记录，解析第一功能所需的跳转路径；跳转路径包括多个关联页面的打开顺序；

[0240] 基于跳转路径，确定第二跳转记录对应的异常以及异常关联页面；

[0241] 发送异常的描述信息及异常关联页面的信息至第一功能的提供方。

[0242] 第三方面，根据本公开的一个或多个实施例，提供了一种信息处理装置，包括：第一获取单元，用于经由Web容器获取第一渲染代码，并获取第一分辨率；其中，第一渲染代码为用于对第一Web页面进行渲染的代码；第一分辨率为用于显示第一Web页面的分辨率；

[0243] 第一发送单元，用于发送第一渲染代码和第一分辨率；

[0244] 第二获取单元，用于获取与第一Web页面中的异常相关的关联信息，关联信息基于第一渲染代码确定，关联信息包括如下至少之一：异常信息、或异常修复代码、或基于异常信息的处理指令；其中，异常修复代码包括对第一Web页面中的异常进行修复的代码；其中，第一Web页面中的异常，由第一模型根据对第二Web页面进行视觉分析确定；第二Web页面基于第一渲染代码进行页面渲染后得到。

[0245] 根据本公开的一个或多个实施例，关联信息包括异常修复代码；装置还包括修复单元，修复单元用于：

[0246] 根据异常修复代码对第一Web页面进行修正。

[0247] 根据本公开的一个或多个实施例，第一Web页面中的异常包括以下一项或多项：

[0248] 布局异常、元素缺失、样式缺失、内容异常或者逻辑或交互异常。

[0249] 根据本公开的一个或多个实施例，第一发送单元进一步用于：

[0250] 对第一渲染代码进行用户信息脱敏，发送脱敏后的第一渲染代码。

[0251] 根据本公开的一个或多个实施例，第一发送单元进一步用于：

[0252] 从第一渲染代码中识别出用户信息所在区域，使用替换内容对用户信息进行替换，得到脱敏后的第一渲染代码。

[0253] 根据本公开的一个或多个实施例，第一获取单元进一步用于：获取与第一Web页面关联的历史跳转记录；历史跳转记录包括多个页面的打开时间和 / 或顺序；

[0254] 第一发送单元进一步用于：发送历史跳转记录、第一渲染代码及第一分辨率；以及，

[0255] 处理指令包括与第一功能的异常对应的第一处理指令；第一功能为第一Web页面支持的功能；第一处理指令由第一功能的异常及第一功能对应的跳转路径确定，跳转路径包括从历史跳转记录中解析出的多个页面的打开顺序。

[0256] 根据本公开的一个或多个实施例，装置还包括第二发送单元，第二发送单元用于：

[0257] 根据第一渲染代码按照第一分辨率进行Web页面渲染，得到第二Web页面；

[0258] 将第二Web页面对应的第一图像发送至第一模型，其中，第一图像记录有第二Web页面的显示内容，第一模型执行如下第一操作：

[0259] 对第一图像进行视觉分析，根据视觉分析结果确定第一图像是否有异常；响应于确定第一图像有异常，根据异常生成关联信息。

[0260] 根据本公开的一个或多个实施例，修复代码包括视觉修复代码，和 / 或，逻辑修复代码或交互修复代码；其中，修复单元进一步用于：

[0261] 经由Web容器将视觉修复代码注入到第一Web页面；

[0262] 经由Web容器将逻辑修复代码或交互修复代码应用到第一Web页面。

[0263] 第四方面，根据本公开的一个或多个实施例，提供了一种信息处理装置，包括：

[0264] 第一接收单元，用于接收第一渲染信息；其中，第一渲染信息为用于对第一Web页面进行渲染的信息，包括第一渲染代码及第一分辨率；第一渲染代码经由Web容器获取，第一分辨率为用于显示第一Web页面的分辨率；

[0265] 渲染单元，用于基于第一分辨率利用云端渲染服务根据第一渲染代码渲染页面，得到第二Web页面；

[0266] 解析单元，用于获取第二Web页面的第一图像，将第一图像发送至第一模型进行视觉分析，由第一模型基于视觉分析确定第一Web页面是否存在异常。

[0267] 根据本公开的一个或多个实施例，装置还包括生成单元，生成单元用于：

[0268] 调用第一模型生成与第一Web页面中的异常相关的关联信息，关联信息包括如下至少之一：异常信息、或异常修复代码、或基于异常信息的处理指令。

[0269] 根据本公开的一个或多个实施例，异常修复代码包括：

[0270] 视觉修复代码，和 / 或，逻辑修复代码或交互修复代码；

[0271] 基于异常信息的处理指令，包括以下至少之一：

[0272] 重新加载第一Web页面；

[0273] 或者，调整与第一Web页面关联的流程信息。

[0274] 根据本公开的一个或多个实施例，装置50还包括第二发送单元，第二发送单元用于：

[0275] 将视觉修复代码发送到Web容器；其中，Web容器将视觉修复代码注入到第一Web页面；

[0276] 将逻辑修复代码或交互修复代码发送到Web容器；其中，Web容器将逻辑修复代码或交互修复代码应用到第一Web页面。

[0277] 根据本公开的一个或多个实施例，第一Web页面包括与第一功能关联的多个第二页面；装置还包括功能异常检测单元，功能异常检测单元用于：

[0278] 获取与第一功能关联的历史跳转记录，历史跳转记录包括第一功能成功完成的第一跳转记录，以及第一功能未成功完成的第二跳转记录；第一跳转记录包括多个第二页面间的打开时间或顺序，第二跳转记录包括多个第二页面间的打开时间或顺序；

[0279] 基于第一跳转记录，解析第一功能所需的跳转路径；跳转路径包括多个关联页面的打开顺序；

[0280] 基于跳转路径，确定第二跳转记录对应的异常以及异常关联页面；

[0281] 发送异常的描述信息及异常关联页面的信息至第一功能的提供方。

[0282] 第五方面，根据本公开的一个或多个实施例，提供了一种电子设备，包括：至少一个处理器和存储器；

[0283] 存储器存储计算机执行指令；

[0284] 至少一个处理器执行存储器存储的计算机执行指令，使得至少一个处理器执行如上第一方面、第二方面以及第一方面和第二方面各种可能的设计的方法。

[0285] 第六方面，根据本公开的一个或多个实施例，提供了一种计算机可读存储介质，计算机可读存储介质中存储有计算机执行指令，当处理器执行计算机执行指令时，实现如上第一方面、第二方面以及第一方面和第二方面各种可能的设计的方法。

[0286] 第五方面，根据本公开的一个或多个实施例，提供了一种计算机程序产品，包括计算机程序，计算机程序被处理器执行时实现上第一方面、第二方面以及第一方面和第二方面各种可能的设计的方法。

[0287] 综上可知，通过利用再现Web页面在Web容器中渲染后的视觉效果，并对该视觉效果调用第一模型进行视觉表现的检测，能够精准识别基于代码规则或图像对比方案难以确定的Web页面异常，可以解决Web页面异常复现定位的问题。实现了在较低成本的前提下，实现较大范围的Web异常检测覆盖。提高了Web页面异常检测的效率。

[0288] 以上描述仅为本公开的较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开中所涉及的公开范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述公开构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开中公开的（但不限于）具有类似功能的技术特征进行互相替换而形成的技术方案。

[0289] 此外，虽然采用特定次序描绘了各操作，但是这不应当理解为要求这些操作以所示出的特定次序或以顺序次序执行来执行。在一定环境下，多任务和并行处理可能是有利的。同样地，虽然在上面论述中包含了若干具体实现细节，但是这些不应当被解释为对本公开的范围的限制。在单独的实施例的上下文中描述的某些特征还可以组合地实现在单个实施例中。相反地，在单个实施例的上下文中描述的各种特征也可以单独地或以任何合适的子组合的方式实现在多个实施例中。

[0290] 尽管已经采用特定于结构特征和 / 或方法逻辑动作的语言描述了本主题，但是应当理解所附权利要求书中所限定的主题未必局限于上面描述的特定特征或动作。相反，上面所描述的特定特征和动作仅仅是实现权利要求书的示例形式。< / style>

Claims

1. An information processing method, characterized in that, include: The first rendering code and the first resolution are obtained through the Web container; wherein, the first rendering code is code used to render the first Web page; and the first resolution is the resolution used to display the first Web page. Send the first rendering code and the first resolution; Obtain association information related to the anomaly in the first Web page. The association information is determined based on the first rendering code. The association information includes at least one of the following: anomaly information, anomaly repair code, or processing instructions based on the anomaly information. The anomaly repair code includes code for repairing the anomaly in the first Web page. The anomalies in the first web page are determined by the first model based on visual analysis of the second web page; the second web page is obtained after page rendering based on the first rendering code.

2. The method according to claim 1, characterized in that, The associated information includes exception repair code; the method further includes: The first web page is corrected according to the aforementioned anomaly repair code.

3. The method according to claim 1, characterized in that, The exception in the first web page includes one or more of the following: Layout error, missing elements, missing styles, abnormal content, or logical or interactive errors.

4. The method according to claim 1, characterized in that, Sending the first rendering code includes: User information is desensitized in the first rendering code, and the desensitized first rendering code is sent.

5. The method according to claim 4, characterized in that, User information anonymization is performed on the first rendering code, including: The user information area is identified from the first rendering code, and the user information is replaced with replacement content to obtain the de-identified first rendering code.

6. The method according to claim 1, characterized in that, The method further includes: Obtain the historical redirection records associated with the first web page; the historical redirection records include the opening time and / or order of multiple pages; Sending the first rendering code and the first resolution includes sending the historical jump record, the first rendering code, and the first resolution; and... The processing instructions include a first processing instruction corresponding to an exception of the first function; the first function is a function supported by the first web page; the first processing instruction is determined by the exception of the first function and the jump path corresponding to the first function, and the jump path includes the opening order of multiple pages parsed from the historical jump records.

7. The method according to claims 1-6, characterized in that, The method further includes: The second web page is obtained by rendering the web page according to the first resolution based on the first rendering code. The first image corresponding to the second web page is sent to the first model, wherein the first image records the display content of the second web page, and the first model performs the following first operation: The first image is subjected to visual analysis, and the presence of any anomalies in the first image is determined based on the visual analysis results; in response to the determination that the first image is abnormal, the associated information is generated based on the anomalies.

8. The method according to claim 2, characterized in that, The repair code includes visual repair code, and / or logical repair code or interactive repair code; wherein, the step of correcting the displayed first Web page according to the anomaly repair code includes: The visual repair code is injected into the first web page via the web container; The logic fix code or interactive fix code is applied to the first web page via the web container.

9. An information processing method, characterized in that, include: Receive first rendering information; wherein the first rendering information is information for rendering a first Web page, including first rendering code and a first resolution; the first rendering code is obtained via a Web container, and the first resolution is the resolution used to display the first Web page; Based on the first resolution, the cloud rendering service renders the page according to the first rendering code to obtain the second web page; A first image of the second web page is obtained, and the first image is sent to a first model for visual analysis. The first model determines whether there is an anomaly in the first web page based on the visual analysis.

10. The method according to claim 9, characterized in that, The method further includes: The first model is invoked to generate associated information related to the anomaly in the first web page. The associated information includes at least one of the following: anomaly information, anomaly repair code, or processing instructions based on anomaly information.

11. The method according to claim 10, characterized in that, The exception repair code includes: Visual fix code, and / or, logic fix code or interaction fix code; The processing instructions based on anomaly information include at least one of the following: Reload the first web page; Alternatively, adjust the process information associated with the first web page.

12. The method according to claim 11, characterized in that, The method also includes one or more of the following: The visual repair code is sent to the web container; wherein the web container injects the visual repair code into the first web page; The logic repair code or interactive repair code is sent to the Web container; wherein the Web container applies the logic repair code or interactive repair code to the first Web page.

13. The method according to any one of claims 9-12, characterized in that, The first web page includes multiple second pages associated with the first function; the method further includes: Obtain historical redirection records associated with the first function, including first redirection records where the first function was successfully completed and second redirection records where the first function was not successfully completed; the first redirection records include the opening time or order of multiple second pages, and the second redirection records include the opening time or order of multiple second pages; Based on the first jump record, the jump path required by the first function is parsed; the jump path includes the opening order of multiple related pages; Based on the jump path, determine the exception corresponding to the second jump record and the page associated with the exception; Send the description of the exception and the information of the page associated with the exception to the provider of the first function.

14. An electronic device, characterized in that, include: Processor and memory; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1 to 13.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the method as described in any one of claims 1 to 13.

Images