Distribution method, device, electronic equipment and storage medium of gray scale test program

Abstract

This invention provides a method, apparatus, electronic device, and storage medium for distributing grayscale testing programs, relating to the field of video technology. In this method, in response to a test join request sent by a target requesting end, target identification information is determined for a target tool from the target requesting end's identification information. The target tool is a tag generation tool capable of outputting tags including predefined tags corresponding to the function under test. The predefined tags are usage characteristic tags defined in preset conditions that the requesting end must meet to use the function under test. Based on the target identification information, specified types of data related to the target requesting end are obtained, and the target tool generates usage characteristic tags for the target requesting end according to the specified types of data. If the usage characteristic tags of the target requesting end meet the preset conditions corresponding to the function under test, the experimental group configuration corresponding to the function under test is distributed to the target requesting end. This ensures the accuracy of the test results.

Description

Technical Field

[0001] This invention relates to the field of video technology, and in particular to a method, apparatus, electronic device, and storage medium for distributing a grayscale testing program. Background Technology

[0002] Currently, applications (APPs) are widely used. To enrich the functionality of applications, technical personnel often develop new features. Before a new feature is launched, it usually needs to undergo A / B testing. In A / B testing, experimental group configurations and control group configurations are set up for the feature to be tested. Some requests need to be diverted to the experimental group, that is, the requesting parties of some requests are returned with the experimental group configuration to instruct these requesting parties to enable the feature to be tested.

[0003] In related technologies, for a test join request, the system directly returns the experimental group configuration corresponding to the function under test to the requester if the last digit of the user identifier on the requesting end is a pre-specified fixed value. This approach has the potential to enable the function under test on requesting ends that are unsuitable for it, leading to lower accuracy of the test results. Summary of the Invention

[0004] The purpose of this invention is to provide a method, apparatus, electronic device, and storage medium for distributing grayscale testing programs, so as to improve the accuracy of test results.

[0005] In a first aspect of this invention, a method for distributing a grayscale testing program is provided, the method comprising: In response to a test join request sent by the target requesting end, target identification information is determined for the target tool from the identification information of the target requesting end; the target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test, and the predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test; Based on the target identification information, specified type data related to the target request client is obtained, and the target tool generates usage characteristic tags for the target request client according to the specified type data; the specified type data includes the type of data required by the target tool when generating tags; If the usage characteristic tag of the target request terminal meets the preset conditions corresponding to the function to be tested, then the experimental group configuration corresponding to the function to be tested is distributed to the target request terminal.

[0006] In a second aspect of the invention, a distribution device for a grayscale testing program is also provided, the device comprising: The first acquisition module is used to respond to a test join request sent by the target requesting end and determine target identification information for the target tool from the identification information of the target requesting end; the target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test, and the predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test; The generation module is used to obtain specified type data related to the target request client based on the target identification information, and generate usage characteristic tags for the target request client according to the specified type data through the target tool; the specified type data includes the type of data required by the target tool when generating tags; The first distribution module is used to distribute the experimental group configuration corresponding to the function under test to the target requesting end if the usage characteristic tag of the target requesting end meets the preset conditions corresponding to the function under test.

[0007] In another aspect of the present invention, an electronic device is also provided, characterized in that it includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. Memory, used to store computer programs; When a processor executes a program stored in memory, it implements any of the above-described grayscale test program distribution methods.

[0008] In another aspect of the present invention, a computer-readable storage medium is also provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the distribution method of any of the grayscale test programs described above.

[0009] In another aspect of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the distribution method of any of the grayscale test programs described above.

[0010] This invention provides a method, apparatus, electronic device, and storage medium for distributing a grayscale testing program. In response to a test join request sent by a target requesting end, target identification information is determined for a target tool from the target requesting end's identification information. The target tool is a tag generation tool capable of outputting tags including predefined tags corresponding to the function under test. The predefined tags are usage characteristic tags defined in preset conditions that the requesting end must meet to use the function under test. Based on the target identification information, specified types of data related to the target requesting end are obtained, and the target tool generates usage characteristic tags for the target requesting end according to the specified types of data. The specified types of data include the types of data required by the target tool when generating tags. If the usage characteristic tags of the target requesting end meet the preset conditions corresponding to the function under test, then the experimental group configuration corresponding to the function under test is distributed to the target requesting end. Thus, only when the preset conditions corresponding to the function under test are met, i.e., when the target requesting end is suitable for using the function under test, is the experimental group configuration corresponding to the function under test returned to the target requesting end. This avoids enabling the function under test for requesting ends that are not suitable for using the function under test, thereby ensuring the accuracy of the test results. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0012] Figure 1 This is a flowchart of the steps of a grayscale testing program distribution method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a request processing framework provided in an embodiment of the present invention; Figure 3 This is a structural diagram of a grayscale testing program distribution device provided in an embodiment of the present invention; Figure 4 This is a structural diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention.

[0014] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0015] The distribution method of the grayscale testing program provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0016] Figure 1 This is a flowchart illustrating the steps of a grayscale testing program distribution method provided in an embodiment of the present invention, as follows: Figure 1 As shown, the method may include: Step 101: In response to the test join request sent by the target requesting end, determine the target identification information for the target tool from the identification information of the target requesting end; the target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test, and the predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test.

[0017] Step 102: Based on the target identification information, obtain specified type data related to the target request end, and generate usage characteristic tags for the target request end using the target tool according to the specified type data; the specified type data includes the type of data required by the target tool when generating tags.

[0018] Step 103: If the usage characteristic tag of the target request terminal meets the preset conditions corresponding to the function to be tested, then the experimental group configuration corresponding to the function to be tested is distributed to the target request terminal.

[0019] The execution entity of the gray-scale testing program distribution method provided in this embodiment of the invention can be a server, specifically an intelligent distribution system within the server. In one application scenario, the function to be tested in this embodiment can be a function of a video application that needs to be tested. For example, the function to be tested can be a new function developed for the video application, or an updated function of the video application. The video application can be various end-user products, such as a mobile video application, a computer video application, or a television video application; this embodiment of the invention does not limit this. In this application scenario, the video application of the video platform is the gray-scale testing program. Of course, the gray-scale testing program can also be an application with the function to be tested on other platforms; for example, it can be a music player software on an audio platform, a shopping software on a shopping platform, etc.; this embodiment of the invention does not limit this.

[0020] The target requesting end is the requesting end that sends the test join request. The requesting end can be a user terminal. The test join request indicates that the target requesting end agrees to participate in the test. The test join request can be a request triggered when using a gray-scale testing program. For example, in a video application scenario, the test join request can be a request triggered within the video application (hereinafter referred to as a video-related request). The specific request type of the video-related request can be related to the function under test. For example, a video-related request can be a video playback request or a video pause request. If the function under test is used when the video is playing, then the video-related request can be a video playback request. If the function under test is used when the video is paused, for example, if the function under test includes recommending anime-related content when the video is paused, then the video-related request can be a video pause request. The identification information of the target requesting end can include identifiers used to search for various types of information related to the requesting end. The relevant data of the target requesting end can include specified types of data related to the target requesting end. These specified types of data can be the types of data required when the target tool generates tags. For example, the specified type of data may include information that characterizes the content viewing preferences of the requesting user, information that characterizes the hardware and network capabilities of the viewing device used by the requesting user, information that characterizes the viewing behavior of the requesting user, and so on.

[0021] In this embodiment of the invention, multiple tag generation tools can be pre-integrated. Each tag generation tool generates a usage characteristic tag, and its output tags include at least one usage characteristic tag. A usage characteristic tag can be used to measure the capabilities of a user's device, content preference characteristics during use, or operational behavior characteristics during use, etc. For example, taking a video application as an example, the multiple tag generation tools may include a fluctuation analysis tool (Tool_VV_Fluctuation), an interruption tolerance tool (Tool_Interrupt), a content preference tool (Tool_Content_Pref), and a device capability tool (Tool_Device_Cap). The fluctuation analysis tool is used to output a stable viewing tag ([Stable_Viewer]) or a roaming browsing tag ([Surfing_Viewer]) based on the completion rate variance. The completion rate variance characterizes viewing stability; the larger the completion rate variance, the higher the viewing stability. Accordingly, the output tags of the fluctuation analysis tool include stable viewing tags and roaming browsing tags. If the requesting client displays the "Stable Viewing" tag, it indicates that the user on that client does not frequently switch while watching videos, demonstrating high viewing stability and exhibiting the characteristic of stable viewing. If the requesting client displays the "Roaming Browsing" tag, it indicates that the user on that client frequently switches while watching videos, demonstrating low viewing stability and exhibiting the characteristic of frequent switching.

[0022] The interruption tolerance tool determines the duration of a single playback session based on pause / exit timestamps in a user's playback history. It then outputs a low-patience label ([Low_Patience]) and a high-patience label ([High_Engagement]) based on this duration. The duration of a single playback session characterizes viewing patience; a longer session indicates higher patience. Correspondingly, the interruption tolerance tool can output low-patience and high-patience labels. If a user has a low-patience label, it indicates that they watch videos for short sessions, have low patience, and high interruption tolerance, exhibiting the characteristic of low viewing patience. Conversely, if a user has a high-patience label, it indicates that they watch videos for long sessions, have high patience, and low interruption tolerance, exhibiting the characteristic of high viewing patience.

[0023] The content preference tool determines the viewing percentage of specific content (e.g., anime) based on the requester's historical viewing history. A higher viewing percentage indicates a greater user preference for that specific content. It outputs specific content preference tags based on the viewing percentage, such as the "Anime Core" tag ([Anime_Core]). Accordingly, the content preference tool can output tags including the "Anime Core" tag. If the requester has the "Anime Core" tag, it indicates that the user on that requester prefers anime content and exhibits this viewing behavior characteristic. The device capability tool outputs a high device capability tag ([High_End_Env]) assuming the viewing device's hardware and network capabilities meet preset requirements. The device capability tool can output tags including the "High Device Capability" tag. If the requester has the "High Device Capability" tag, it indicates that the requester's device has strong capabilities and exhibits the characteristic of a high-performance device.

[0024] Furthermore, the predefined tags corresponding to the function under test are set by technicians based on the function under test. Specifically, usage characteristic tags required by the requesting end suitable for using the function under test and / or usage characteristic tags that the requesting end suitable for using the function under test cannot have can be used as predefined tags. For a predefined tag, a condition that needs to be met is set for the function under test. The condition to which the predefined tag belongs is a preset condition that needs to be met for that predefined tag. For example, suppose the function under test is an anime bullet screen enhancement function, which is used to display bullet screens related to anime content in a more prominent way, for example, in bold. Then, if the anime core tag is available, it is suitable to use the function under test. Accordingly, the anime core tag can be used as a predefined tag for the anime bullet screen enhancement function, and a preset condition that needs to be met can be set for this predefined tag: containing [Anime_Core]. Suppose that for a function under test, it is suitable to use the function if the requesting end does not have a certain performance tag (represented by the 'a' tag). Then, the 'a' tag can be used as a predefined tag for the function under test, and a preset condition can be set for this predefined tag: it must not contain the 'a' tag. Conversely, suppose that for a function under test, it is only suitable to use the function if the requesting end does not have a certain performance tag (represented by the 'a' tag) but has a certain performance tag (represented by the 'b' tag). Then, the 'a' tag and the 'b' tag can be used as predefined tags for the function under test, and preset conditions can be set: it must not contain the 'a' tag and must contain the 'b' tag.

[0025] We can first use a tag generation tool that includes predefined tags corresponding to the function under test in its output tags as the target tool. Upon receiving a video-related request, we obtain the identifier type information required by the target tool (i.e., target identifier information), and based on the target identifier information, we obtain the specified type of data related to the target requesting end. We then obtain the usage characteristic tags generated by the target tool based on the specified type of data, thus obtaining the usage characteristic tags of the target requesting end. Furthermore, for any function under test, if the usage characteristic tags of the target requesting end meet the preset conditions corresponding to that function, it indicates that the target requesting end is suitable for using that function. Therefore, we can return the experimental group configuration corresponding to the function under test to the target requesting end, i.e., decide whether the target client should use the experimental group configuration test branch. The experimental group configuration is used to instruct the target requesting end to enable the function to be tested. The experimental group configuration includes some information required to enable the function to be tested. For example, the experimental group configuration for the anime bullet screen enhancement function may include bullet screen text display attributes (e.g., bolding), the experimental group configuration for the high bitrate decoding algorithm function may include the high bitrate decoding algorithm, the experimental group configuration for the content recommendation function may include the features required for the recommended content, the experimental group configuration for the user interface interaction mode function may include the display attribute information of the user interface, and so on.

[0026] In summary, the grayscale testing program distribution method provided in this embodiment of the invention determines target identification information for the target tool from the identification information of the target requesting end in response to a test join request sent by the target requesting end. The target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test. The predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test. Based on the target identification information, specified type data related to the target requesting end is obtained, and the target tool generates usage characteristic tags for the target requesting end according to the specified type data. The specified type data includes the type of data required by the target tool when generating tags. If the usage characteristic tags of the target requesting end meet the preset conditions corresponding to the function under test, the experimental group configuration corresponding to the function under test is distributed to the target requesting end. In this way, the experimental group configuration corresponding to the function under test is returned to the target requesting end only when the preset conditions corresponding to the function under test are met, that is, when the target requesting end is suitable to use the function under test. This can avoid enabling the function under test for requesting ends that are not suitable to use the function under test, thereby ensuring the accuracy of the test results.

[0027] Optionally, in embodiments of the present invention, after step 102, the following steps may also be included: Step S21: If the usage characteristic tag of the target request terminal does not meet the preset conditions corresponding to the function under test, then the control group configuration corresponding to the function under test is distributed to the target request terminal; wherein, the control group configuration is used to instruct the target request terminal not to enable the function under test.

[0028] For any feature to be tested, if the usage characteristic tags of the target request client do not meet the predefined tag conditions corresponding to that feature, it means that the target request client is not suitable for using the feature under test. Therefore, the target request client can be added to the control group, and the control group configuration corresponding to the feature under test can be returned to the target request client, thereby enabling real-time routing of the request client to the appropriate strategy group (experimental group or control group). The control group configuration is used to instruct the target request client not to enable the feature under test, avoiding the application of the feature under test to the target request client's users. The control group configuration can include default configurations, such as the default display attributes of the bullet screen text, the default decoding algorithm, the default recommendation strategy, etc.

[0029] In practical applications, we can collect user feedback information (referred to as first feedback information) from the requesting end that received the configuration of the experimental group (within a test period) and user feedback information (referred to as second feedback information) from the requesting end that received the configuration of the control group (within a test period). For example, we can collect the video completion rate. If the first feedback information is better than the second feedback information, then the function under test is determined to have passed the test.

[0030] In this embodiment of the invention, if the preset conditions corresponding to the function under test are not met, that is, if the target requesting end is not suitable for using the function under test, the control group configuration corresponding to the function under test is returned to the target requesting end, so as to ensure that the function under test is not enabled for the requesting end that is not suitable for using the function under test, thereby ensuring the accuracy of the test results.

[0031] Optionally, in this embodiment of the invention, before the step of determining the target identification information for the target tool from the identification information of the target requesting end, the following steps may also be included: Step S31: In response to the function under test setting operation, obtain the preset conditions set for the function under test.

[0032] Step S32: Determine the usage characteristic tags defined in the preset conditions as the predefined tags corresponding to the set test function.

[0033] Step S33: For any of the predefined tags, determine the target tool from the preset tag generation tool library that can output tags including the predefined tags.

[0034] The "setting up a function under test" operation is an operation on the server side that defines the function under test. For example, this operation can be an operation that stores the function identifier of the function under test on the server side. When the server detects this operation, it can receive the required conditions (i.e., preset conditions) input for the newly added function under test, thus obtaining the preset conditions set for the function under test. For example, suppose the preset conditions set for the function under test (function X) include: not containing the 'a' tag and containing the 'b' tag. Then, the usage characteristic tags defined in the preset conditions can be extracted to obtain the predefined tags corresponding to the function under test: the 'a' tag and the 'b' tag.

[0035] In this embodiment of the invention, a tag generation tool library can be pre-built, including various tag generation tools. For example, the tag generation tools in the tool library can be ModelContext Protocol (MCP) tools. These tools are lightweight and follow a unified interface standard (input is identification information in the context, output is a tag). Internally, the tag generation tools can be based on rules, statistical models, or lightweight small models (e.g., Logistic Regression) to generate tags, ensuring millisecond-level response. The tag generation tool library can be mounted to an AI Agent. After the server-side gateway listens for video-related requests sent by the target requester, the gateway intercepts the video-related requests and forwards them to the AI ​​Agent, triggering the AI ​​Agent to execute the distribution method of the aforementioned gray-scale testing program to distribute the requests. In this embodiment, the distribution decision-making power is transferred from static rules to the AI ​​Agent. This AI Agent can be deployed at the gateway layer of the server. Of course, tag generation tools can also be integrated into other components and used through interface interaction; this embodiment of the invention does not limit this approach.

[0036] After building the tag generation tool library, a tool registry can be configured during the agent's initialization phase. The tool registry defines the mapping between tag generation tools and output tags, as well as the mapping between tag generation tools and identifier types. Configuring the tool registry fulfills the context data requirements (i.e., the required identifier types) for declaring tag generation tools and their invocation.

[0037] Furthermore, for any tag in the predefined tags, the tag generation tool whose corresponding output tags include that tag can be found from the correspondence between tag generation tools and output tags, and used as the target tool. Assuming function Y is also set as the function to be tested, and the preset conditions for function Y include: containing tag c and containing tag d, then the usage characteristic tags defined in the preset conditions can be extracted to obtain the predefined tags corresponding to the set function to be tested: tag c and tag d. For tag a, assuming that the output tags corresponding to tool A include tag a; for tag b, assuming that the output tags corresponding to tool B include tag b; for tag c, assuming that the output tags corresponding to tool C include tag c; for tag d, assuming that the output tags corresponding to tool D include tag d, then tools A, B, C, and D are determined as target tools. That is, in this embodiment of the invention, there can be multiple functions to be tested (e.g., functions X and Y in the above example), and correspondingly, there are multiple target tools corresponding to the functions to be tested (e.g., tools A, B, C, and D in the above example). In this embodiment of the invention, tool identifiers for each target tool can be defined in a tool identifier table. Correspondingly, the label generation tool represented by each tool identifier defined in the tool identifier table can be used as the target tool. After a function under test is tested, the target tool determined for that function can be modified to a non-target tool. For example, the tool identifier of the target tool determined for that function can be deleted from the tool identifier table.

[0038] In this embodiment of the invention, in response to a function-to-be-tested setting operation, preset conditions set for the function to be tested are obtained. The usage characteristic tags defined in the preset conditions are determined as predefined tags corresponding to the function to be tested. For any predefined tag, a tool in a preset tag generation tool library that can output tags including the predefined tags is determined as the target tool. This ensures that the target tools cover the tag generation tools required to generate each predefined tag, thereby ensuring that the tags subsequently generated by the target tool for the target request end are comparable to the predefined tags defined in the preset conditions.

[0039] Optionally, in this embodiment of the invention, the step of returning the experimental group configuration corresponding to the function under test to the target requesting end if the usage feature tag of the target requesting end meets the preset conditions corresponding to the function under test may specifically include: Step 1031: For any of the preset conditions corresponding to the function to be tested, if the preset condition represents a predefined label required by the requesting end, then if the usage characteristic label of the target requesting end includes the predefined label, it is determined that the usage characteristic label of the target requesting end satisfies the preset condition.

[0040] Step 1032: If the preset condition represents a predefined tag that the requesting end does not need to have, then if the predefined tag is not included in the usage characteristic tags of the target requesting end, it is determined that the usage characteristic tags of the target requesting end meet the preset condition.

[0041] Step 1033: If the characteristic tags used on the target request end meet all the preset conditions corresponding to the function under test, return the experimental group configuration corresponding to the function under test to the target request end.

[0042] In this embodiment of the invention, the usage characteristic tags of the target request end can be compared with the preset conditions corresponding to the function under test using regular expression matching or logical judgment to determine whether the preset conditions are met. Here, a preset condition corresponding to the function under test is a condition that must be met to define a predefined tag corresponding to the function under test.

[0043] When there are multiple functions to be tested, step 103 described above can be performed for any function to be tested. For example, corresponding to the above example, the predefined tags for function X include: tag a and tag b, and the preset conditions for function X include: not containing tag a and containing tag b. The predefined tags for function Y include: tag c and tag d, and the preset conditions for function Y include: containing tag c and containing tag d. The usage characteristic tags of the target request end include the outputs of tools A, B, C, and D respectively. Assuming function X corresponds to experimental group 1 and function Y corresponds to experimental group 2, for function X, by comparing the usage characteristic tags of the target request end with the preset conditions corresponding to function X, if the usage characteristic tags of the target request end satisfy all the preset conditions corresponding to function X, then the configuration for experimental group 1 is returned to the target request end. For function Y, by comparing the usage characteristic tags of the target request end with the preset conditions corresponding to function Y, if the usage characteristic tags of the target request end satisfy all the preset conditions corresponding to function Y, then the configuration for experimental group 2 is returned to the target request end.

[0044] In this embodiment of the invention, for any predefined tag belonging to the function under test, it is determined whether the usage characteristic tag of the target requesting end meets the preset conditions. If the usage characteristic tag of the target requesting end meets all the preset conditions corresponding to the function under test, the experimental group configuration corresponding to the function under test is returned to the target requesting end. Thus, only when the usage characteristic tag of the target requesting end meets all the predefined tag belonging conditions corresponding to the function under test is the request routed to the experimental group of the service under test, and the experimental group configuration corresponding to the function under test is returned to the target requesting end. This ensures the compatibility between the target requesting end and the experimental group configuration corresponding to the function under test, determines suitable users to enter the corresponding experimental group, and achieves accurate testing.

[0045] Furthermore, if the preset conditions meet the first format, predefined tags representing the requirements of the requesting party are determined. If the preset conditions meet the second format, predefined tags representing the requirements of the requesting party are determined. The first format can be: containing…, and the second format can be: not containing….

[0046] Accordingly, if the preset conditions represent predefined tags required by the requesting end, it indicates that the requesting end with the predefined tags is suitable for using the function under test. If the preset conditions represent predefined tags not required by the requesting end, it indicates that the requesting end with the predefined tags is not suitable for using the function under test. For example, taking function X as an example, the preset conditions corresponding to function X include: not containing the 'a' tag and containing the 'b' tag. For not containing the 'a' tag, this preset condition represents a predefined tag not required by the requesting end. Therefore, if the usage characteristic tags of the target requesting end do not include the 'a' tag, it can be determined that the usage characteristic tags of the target requesting end satisfy the preset condition: not containing the 'a' tag. For containing the 'b' tag, this preset condition represents a predefined tag required by the requesting end. Therefore, if the usage characteristic tags of the target requesting end include the 'b' tag, it can be determined that the usage characteristic tags of the target requesting end satisfy the preset condition: containing the 'b' tag. Finally, if the usage characteristic tags of the target requesting end do not contain the 'a' tag and contain the 'b' tag, it is determined that the usage characteristic tags of the target requesting end satisfy all the preset conditions corresponding to function X.

[0047] In this embodiment of the invention, if the preset condition represents a predefined tag required by the requesting end, then if the usage characteristic tags of the target requesting end include the predefined tag, the preset condition is determined to be satisfied. If the preset condition represents a predefined tag not required by the requesting end, then if the usage characteristic tags of the target requesting end do not include the predefined tag, the preset condition is determined to be satisfied. Thus, by determining whether the usage characteristic tags of the target requesting end include the predefined tag, it can be determined whether the usage characteristic tags of the target requesting end satisfy the preset condition. The implementation difficulty of the determination logic is low, thereby ensuring determination efficiency.

[0048] Optionally, the step of determining the target identifier information for the target tool from the identifier information of the target requesting end in the embodiments of the present invention may specifically include: Step 1011: Identify the identification information in the identification information of the target request terminal that matches the target type as the target identification information; the target type is the identification type corresponding to the target tool.

[0049] Specifically, for any target tool, the corresponding identifier type can be found from the preset mapping between tag generation tools and identifier types, and used as the target type. The mapping between tag generation tools and identifier types can be defined in the aforementioned tool call table. The identifier information of the target requesting end can be carried in the video-related request. For example, the identifier information of the target requesting end may include the user identifier (UID) information, the device identifier (DeviceID) information, and the video identifier (VideoID) information, etc. The intelligent agent can parse the user context carried in the video-related request to obtain the identifier information of the target requesting end. For example, assuming the identifier type corresponding to the target tool is DeviceID, the device identifier information carried in the video-related request can be used as the target identifier information of that target tool.

[0050] The steps described above, including obtaining specified type data related to the target request client based on the target identification information and generating usage feature tags for the target request client using the target tool according to the specified type data, may specifically include: Step 1021: Invoke the target tool and input the target identification information into the target tool so that the target tool can obtain the relevant data from the current background data based on the target identification information and generate the usage characteristic tag based on the relevant data.

[0051] Step 1022: Based on the return result of the target tool's call, determine the usage characteristic tags of the target requesting end.

[0052] In this embodiment of the invention, the logic for obtaining relevant data from the target requesting end based on target identification information can be implemented by the target tool. In this way, the server only needs to input the target identification information into the target tool to obtain the tags returned by the target tool, thereby reducing the processing burden on the server. Of course, in one implementation, the server can also obtain relevant data from the target requesting end based on the target identification information. Inputting the relevant data into the target tool is not limited in this embodiment of the invention.

[0053] The target identifier information input into the target tool can serve as a retrieval basis. The target tool can use the target identifier information to retrieve specified types of data related to the target request from the current background data. The specific types of data are determined by the actual processing logic of the target tool, and this invention does not impose any limitations on this. For example, the processing logic of the fluctuation analysis tool includes: reading the user's video viewing (VV) logs from the current background data within the past N minutes based on the UID, and calculating the user's completion rate variance based on the VV logs. If the user's completion rate variance is lower than a first preset threshold T1, a stable viewing label is output; otherwise, a roaming browsing label is output. N can be set according to requirements; for example, N can be 5-30 minutes to cover short-term user viewing behavior and adapt to long-video viewing scenarios. The specific value of the first preset threshold can include a threshold preset for short videos (e.g., videos less than 5 minutes in length) and a threshold preset for long videos (e.g., videos longer than 30 minutes in length). If the VV logs within the last N minutes indicate that a user is watching short videos, a pre-set threshold for short videos is selected as T1. If the VV logs within the last N minutes indicate that a user is watching long videos, a pre-set threshold for long videos is selected as T1. The pre-set threshold for short videos is higher than the pre-set threshold for long videos; for example, the pre-set threshold for short videos is 0.15, and the pre-set threshold for long videos is 0.8. This provides a higher tolerance for user switching behavior in short video scenarios and is more suitable for short video viewing. Correspondingly, the identifier type for the fluctuation analysis tool can be UID, and the specified information for the fluctuation analysis tool can include: video viewing logs within the last N minutes.

[0054] Assume the interruption tolerance tool's processing logic includes: retrieving pause / exit timestamps from the user's historical playback data based on the UID; and determining the duration of a single playback session based on these pause / exit timestamps. If the single playback duration is less than a first preset duration, a low patience label is output; otherwise, a high patience label is output. The first preset duration can be set as needed; for example, it can be 1 minute, but this embodiment of the invention does not impose any limitations on this. Correspondingly, the identifier type for the interruption tolerance tool can be a UID, and the specified category information for the interruption tolerance tool can include: pause / exit timestamps from the user's historical playback.

[0055] The content preference tool's processing logic includes: retrieving the user's historical viewing records from the current backend data based on VideoID and UID; and searching for the video metadata indicated by the VideoID. Based on the historical viewing records, it determines the viewing percentage of specific content (e.g., anime content) and checks whether the video tags in the video metadata indicate specific content. If the video tag indicates specific content and the viewing percentage is greater than a preset percentage, it outputs the anime core tag; otherwise, it outputs nothing, i.e., no tag is output. The preset percentage can be set according to requirements; for example, it can be 50%, but this embodiment of the invention does not limit this.

[0056] Assume the processing logic of the device capability tool includes: reading the device model information and network information of the requesting end from the current background data based on DeviceID; determining the hardware performance score, decoding capability score, remaining storage space score, and system version score of the viewing device based on the device model information; and weighting and summing these scores to obtain the device score. The network information can be the network round-trip time (RTT). If the device score is greater than a preset score and the network information is less than a preset duration, a high device capability tag is output; otherwise, the output is empty, i.e., no tag is output. The preset score and preset duration can be set according to requirements. For example, the preset score can be 90 and the preset duration can be 50ms; this embodiment of the invention does not impose limitations on these settings. In this embodiment, the complex user analysis task is broken down into several independent, lightweight logical units (i.e., tag generation tools), each tool only responsible for judging one dimension, thus facilitating subsequent expansion and reuse.

[0057] Accordingly, the return result of the fluctuation analysis tool can be a roaming browsing tag or a stable viewing tag; the return result of the interruption tolerance tool can be a high patience tag or a low patience tag; the return result of the content preference tool can be an anime core tag or empty; and the return result of the device capability tool can be a high device capability tag or empty. If the content preference tool or device capability tool fails to return a tag within the agreed time, the agent can determine that the content preference tool or device capability tool has no matching tag, i.e., the output tag is empty. Assuming the target tools include the fluctuation analysis tool, interruption tolerance tool, content preference tool, and device capability tool, the agent can invoke these target tools upon receiving a request to obtain their return results for decision-making. Specifically, the usage characteristic tags of the target request end can be obtained based on the return results of these four tools. These usage characteristic tags are dynamically generated based on specified types of data obtained in real-time from the current backend data, and do not rely on offline big data profiles updated T+1, thus representing a real-time user profile set. As the backend data changes, the usage characteristic tags of the same target request end will change dynamically.

[0058] In this embodiment of the invention, a target tool is invoked, and target identification information is input into the target tool. The target tool then retrieves relevant data from the current background data based on the target identification information and generates usage characteristic tags based on the relevant data. The usage characteristic tags of the target requesting end can be obtained based on the return result of the target tool invocation. In this way, the usage characteristic tags of the target requesting end can be obtained through tool invocation, which is relatively easy and thus ensures request processing efficiency.

[0059] Since the current background data can reflect the user's latest status, changes in user behavior, such as changes in user viewing preferences, switching networks, or skipping intros, will be reflected in the background data in real time. Therefore, in this embodiment of the invention, obtaining relevant data based on the current background data can ensure the real-time nature of the relevant data and ensure that the generated tags can reflect the user's latest status, avoiding data lag and inaccurate tags.

[0060] Furthermore, this invention allows for precise testing targeting specific demographics, resulting in more accurate experimental data, reduced waste of testing resources, and enhanced testing efficiency. Taking the anime comment enhancement feature as an example, related technologies might force this feature on users who dislike anime content. Since users themselves dislike anime content, their feedback on this feature is not representative, diluting the experimental results. In this invention, based on current backend data, the usage characteristic tags of the target request end are determined, allowing for timely identification of the user's current intent or preferences. The anime comment enhancement feature is only enabled for the request end when it contains the core anime tag, achieving precise testing targeting the core anime audience, resulting in more accurate experimental data, reduced waste of testing resources, and enhanced testing efficiency.

[0061] Optionally, the tag generation toolkit in this embodiment of the invention includes at least one or more of a first generation tool, a second generation tool, a third generation tool, and a fourth generation tool; wherein, the first generation tool is used to generate tags characterizing the viewing stability of the requesting user, the second generation tool is used to generate tags characterizing the viewing patience of the requesting user, the third generation tool is used to generate tags characterizing the viewing content preferences of the requesting user, and the fourth generation tool is used to generate tags characterizing the device capabilities of the requesting device. The first generation tool can be the aforementioned fluctuation analysis tool, the second generation tool can be the aforementioned interruption tolerance tool, the third generation tool can be the aforementioned content preference tool, and the fourth generation tool can be a device capability tool. This allows for the acquisition of tags covering different dimensions using the tag generation toolkit, thereby ensuring the comprehensiveness of tag comparison.

[0062] Optionally, the steps of calling the target tool described above in this embodiment of the invention may specifically include: step 1021a, in the case of multiple target tools, calling each target tool in parallel; different target tools have different output tags.

[0063] The steps described above for determining the usage characteristic tag of the target requesting end based on the call return result of the target tool may specifically include: step 1022a, taking the union of the call return results of multiple target tools as the usage characteristic tag of the target requesting end.

[0064] Specifically, multiple target tools can be invoked simultaneously using asynchronous concurrency (Async / Await), achieving parallel scheduling of each target tool. Sequential invocation of target tools results in significant latency. In this embodiment, the target tools can be scheduled in parallel, allowing multiple target tools to run concurrently. This reduces the overall execution time to approximately the maximum execution time of a single tool, enabling the generated tags to capture the user's real-time status promptly and supporting instantaneous scenario testing. Furthermore, it can shorten the traffic splitting time, reduce the testing cycle, improve request response speed, and avoid request processing lag issues.

[0065] Furthermore, since the target tools identified for different services under test may overlap, upon receiving the call return results from all target tools, the union of the call return results from multiple target tools can be calculated to achieve deduplication, i.e., aggregating the call return results to obtain the usage characteristic tags of the target request end. For example, assuming the UID is User_A, and the deduplicated call return results include Stable_Viewer, Anime_Core, and High_End_Env, then the usage characteristic tags of the target request end can be represented as: {User_A_Tags: ["Stable_Viewer", "Anime_Core", "High_End_Env"]}.

[0066] Furthermore, suppose there exists a preset condition for a function to be tested: containing [High_End_Env] and containing [Stable_Viewer], corresponding to the high bitrate configuration in experimental group A. There exists another preset condition for a function to be tested: containing [Anime_Core], corresponding to the anime bullet screen enhancement configuration in experimental group B. There is also a control group configuration. If the preset conditions for the two functions to be tested are not met, the control group configuration is returned. It is understood that the number of functions to be tested and their corresponding preset conditions can be set as needed. For example, a short video fast-scrolling function can also be included, and the preset conditions for the short video fast-scrolling function may include: containing [Low_Patience] and containing [Surfing_Viewer]. After obtaining the usage characteristic tags of the target request end, the usage characteristic tags of the target request end can be matched in real time with the preset conditions corresponding to each function to be tested. If both [High_End_Env] and [Stable_Viewer] are met, the high bitrate configuration is returned via the interface; if [Anime_Core] is met, the anime bullet screen enhancement configuration is returned via the interface. If neither is met, the control group configuration is returned.

[0067] In this embodiment of the invention, by deduplicating the return results of multiple target tools, duplicate tags can be avoided from appearing in the usage feature tags of the target requesting end, thereby reducing data redundancy.

[0068] It should be noted that, in another implementation, the target tools can be deduplicated first. In this way, the return results of multiple target tools can be directly used as the usage characteristic tags of the target request end. This embodiment of the invention does not limit this.

[0069] Related technologies assume uniform traffic. Upon receiving a video-related request, a hash algorithm or random number is used to map the user's identifier to a bucket, with each bucket corresponding to a single digit within a preset range (e.g., 0-9). A fixed value is pre-specified based on a fixed traffic ratio. For example, if the experimental group's traffic ratio is fixed at 20%, then a fixed value of 0-1 is pre-specified. Thus, throughout the testing period, requests with digits ending in 0-1 are by default added to the experimental group for the function under test, while requests with digits ending in other than 0-1 are by default added to the control group. This approach statically distributes traffic according to static rules, making traffic allocation independent of user characteristics. The traffic entering the experimental group is fixed, resulting in a rigid and coarse distribution mechanism that fails to consider the user's latest status and preferences. This can lead to the activation of the function under test for unsuitable users, resulting in low accuracy of the test results. For example, if an anime comment enhancement feature designed for anime enthusiasts is applied to users who dislike anime, their feedback on the feature will be unrepresentative. Similarly, applying high-bitrate decoding algorithms to users with weaker devices (e.g., low-end models or weak network environments) will also result in unrepresentative feedback. Therefore, this leads to inaccurate and unreliable test results, resulting in poor performance and negatively impacting the user's viewing experience, such as causing playback stuttering.

[0070] Figure 2 This is a schematic diagram of a request processing framework provided in an embodiment of the present invention, as shown below. Figure 2As shown, the video-related requests sent by the target requester are forwarded to the agent by the gateway in the server. The agent can parse the user context carried in the video-related requests through the context resolver to obtain the target requester's identification information and determine the target identification information. Next, the multi-tool parallel scheduler calls each target tool in parallel, inputting its respective target identification information and obtaining the call return results of multiple target tools. When a target tool generates usage feature tags, it can obtain the required specified type of data from the current background data. The dynamic tag aggregator deduplicates the call return results of multiple target tools to obtain the usage feature tags of the target requester. Finally, the traffic splitting decision engine matches the preset conditions corresponding to the function under test stored in the preset condition configuration center with the usage feature tags of the target requester. If the usage feature tags of the target requester meet the preset conditions corresponding to the function under test, the experimental group configuration corresponding to the function under test is obtained and returned; otherwise, the control group configuration is obtained and returned.

[0071] In this embodiment of the invention, there is no need to pre-set the traffic ratio of each experimental group. The traffic allocation is dynamically determined by the matching result of the usage characteristic tags generated for the requesting end in real time and the preset conditions corresponding to the function to be tested. The traffic splitting ratio between the experimental group and the control group changes dynamically with the change of the usage characteristic tags of the requesting end, thus realizing personalized traffic splitting.

[0072] Figure 3 This is a structural diagram of a grayscale testing program distribution device provided in an embodiment of this application, with reference to... Figure 3 The device may include: The first acquisition module 301 is used to respond to a test join request sent by the target requesting end and determine target identification information for the target tool from the identification information of the target requesting end; the target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test, and the predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test; The generation module 302 is used to obtain specified type data related to the target request terminal based on the target identification information, and generate usage characteristic tags for the target request terminal according to the specified type data through the target tool; the specified type data includes the type of data required by the target tool when generating tags; The first distribution module 303 is used to distribute the experimental group configuration corresponding to the function under test to the target requesting end if the usage characteristic tag of the target requesting end meets the preset conditions corresponding to the function under test.

[0073] This invention provides a distribution device for a grayscale testing program. In response to a test join request sent by a target requesting end, target identification information is determined for a target tool from the target requesting end's identification information. The target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test. The predefined tags are usage characteristic tags defined in preset conditions that the requesting end needs to meet to use the function under test. Based on the target identification information, specified types of data related to the target requesting end are obtained, and the target tool generates usage characteristic tags for the target requesting end according to the specified types of data. The specified types of data include the types of data required by the target tool when generating tags. If the usage characteristic tags of the target requesting end meet the preset conditions corresponding to the function under test, then the experimental group configuration corresponding to the function under test is distributed to the target requesting end. Thus, only when the preset conditions corresponding to the function under test are met, i.e., when the target requesting end is suitable for using the function under test, is the experimental group configuration corresponding to the function under test returned to the target requesting end. This avoids enabling the function under test for requesting ends that are not suitable for using the function under test, thereby ensuring the accuracy of the test results.

[0074] Optionally, the first distribution module 303 is specifically used for: For any of the preset conditions corresponding to the function to be tested, if the preset condition represents a predefined label that the requesting end needs to have, then if the usage characteristic label of the target requesting end includes the predefined label, it is determined that the usage characteristic label of the target requesting end satisfies the preset condition. If the preset condition represents a predefined tag that the requesting end does not need to have, then if the predefined tag is not included in the usage characteristic tags of the target requesting end, it is determined that the usage characteristic tags of the target requesting end satisfy the preset condition. If the characteristic tags used by the target requesting end satisfy all the preset conditions corresponding to the function to be tested, the experimental group configuration corresponding to the function to be tested is returned to the target requesting end.

[0075] Optionally, the first acquisition module 301 is specifically used to: determine the identification information that matches the target type from the identification information of the target request terminal as the target identification information; the target type is the identification type corresponding to the target tool; The generation module 302 is specifically used for: The target tool is invoked, and the target identification information is input into the target tool so that the target tool can obtain the specified type of data from the current background data based on the target identification information, and generate the usage characteristic tag based on the specified type of data; Based on the return result of the target tool's call, the usage characteristic tags of the target requesting end are determined.

[0076] Optionally, the generation module 302 is further configured to: When there are multiple target tools, each target tool is invoked in parallel; different target tools can output different labels. The union of the return results of multiple calls to the target tools is used as the usage characteristic tag of the target requesting end.

[0077] Optionally, the device further includes: The second distribution module is used to distribute the control group configuration corresponding to the function under test to the target requesting end if the usage characteristic tag of the target requesting end does not meet the preset conditions corresponding to the function under test. The control group configuration is used to instruct the target requesting end not to enable the function to be tested.

[0078] Optionally, the device further includes: The second acquisition module is used to acquire preset conditions set for the set function before the first acquisition module 301 acquires the target identification information required by the target tool from the identification information of the target request terminal, in response to the function setting operation. The first determining module is used to determine the usage characteristic tags defined in the preset conditions as the predefined tags corresponding to the set test functions; The second determining module is used to determine, for any of the predefined tags, a tool in a preset tag generation tool library that can output tags including the predefined tags as the target tool.

[0079] Optionally, the tag generation tool library includes at least one or more of the following: a first generation tool, a second generation tool, a third generation tool, and a fourth generation tool; The first generation tool is used to generate tags that characterize the viewing stability of users on the requesting end, the second generation tool is used to generate tags that characterize the viewing patience of users on the requesting end, the third generation tool is used to generate tags that characterize the viewing preferences of users on the requesting end, and the fourth generation tool is used to generate tags that characterize the capabilities of the device on the requesting end.

[0080] This invention also provides an electronic device, such as... Figure 4 As shown, it includes a processor 901, a communication interface 902, a memory 903, and a communication bus 904, wherein the processor 901, the communication interface 902, and the memory 903 communicate with each other through the communication bus 904. Memory 903 is used to store computer programs; When processor 901 executes a program stored in memory 903, it performs the following steps: In response to a video-related request sent by the target requesting end, the target identification information required by the target tool is obtained from the identification information of the target requesting end; the target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test, and the predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test; Based on the target identification information, relevant data of the target requesting end is obtained, and the target tool generates usage characteristic tags for the target requesting end according to the relevant data. If the usage feature tags of the target request terminal meet the preset conditions corresponding to the function to be tested, then the experimental group configuration corresponding to the function to be tested is returned to the target request terminal.

[0081] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0082] The communication interface is used for communication between the aforementioned terminal and other devices.

[0083] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0084] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0085] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to perform the grayscale test program distribution method described in any of the above embodiments.

[0086] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the grayscale test program distribution method described in any of the above embodiments.

[0087] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0088] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply 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. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0089] The various embodiments in this specification are described in a related 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 system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0090] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for distributing a grayscale testing program, characterized in that, The method includes: In response to a test join request sent by the target requesting end, target identification information is determined for the target tool from the identification information of the target requesting end; the target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test, and the predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test; Based on the target identification information, specified type data related to the target request client is obtained, and the target tool generates usage characteristic tags for the target request client according to the specified type data; the specified type data includes the type of data required by the target tool when generating tags; If the usage characteristic tag of the target request terminal meets the preset conditions corresponding to the function to be tested, then the experimental group configuration corresponding to the function to be tested is distributed to the target request terminal.

2. The method according to claim 1, characterized in that, If the usage characteristic tag of the target requesting end meets the preset conditions corresponding to the function under test, then the experimental group configuration corresponding to the function under test is distributed to the target requesting end, including: For any of the preset conditions corresponding to the function to be tested, if the preset condition represents a predefined label that the requesting end needs to have, then if the usage characteristic label of the target requesting end includes the predefined label, it is determined that the usage characteristic label of the target requesting end satisfies the preset condition. If the preset condition represents a predefined tag that the requesting end does not need to have, then if the predefined tag is not included in the usage characteristic tags of the target requesting end, it is determined that the usage characteristic tags of the target requesting end satisfy the preset condition. If the characteristic tags used by the target requesting end satisfy all the preset conditions corresponding to the function to be tested, the experimental group configuration corresponding to the function to be tested is returned to the target requesting end.

3. The method according to claim 1, characterized in that, Determining target identification information for the target tool from the identification information of the target requesting end includes: identifying the identification information in the identification information of the target requesting end that matches the target type as the target identification information; the target type is the identification type corresponding to the target tool; The step of obtaining specified type data related to the target requester based on the target identification information, and generating usage characteristic tags for the target requester according to the specified type data using the target tool, includes: The target tool is invoked, and the target identification information is input into the target tool so that the target tool can obtain the specified type of data from the current background data based on the target identification information, and generate the usage characteristic tag based on the specified type of data; Based on the return result of the target tool's call, the usage characteristic tags of the target requesting end are determined.

4. The method according to claim 3, characterized in that, The invocation of the target tool includes: When there are multiple target tools, each target tool is invoked in parallel; different target tools can output different labels; The step of determining the usage characteristic tag of the target requesting end based on the call return result of the target tool includes: taking the union of the call return results of multiple target tools as the usage characteristic tag of the target requesting end.

5. The method according to claim 1, characterized in that, The method further includes: If the usage feature tag of the target request terminal does not meet the preset conditions corresponding to the function to be tested, then the control group configuration corresponding to the function to be tested is distributed to the target request terminal. The control group configuration is used to instruct the target requesting end not to enable the function to be tested.

6. The method according to claim 1, characterized in that, The tag generation tool library includes at least one or more of the following: a first generation tool, a second generation tool, a third generation tool, and a fourth generation tool; The first generation tool is used to generate tags that characterize the viewing stability of users on the requesting end, the second generation tool is used to generate tags that characterize the viewing patience of users on the requesting end, the third generation tool is used to generate tags that characterize the viewing preferences of users on the requesting end, and the fourth generation tool is used to generate tags that characterize the capabilities of the device on the requesting end.

7. A distribution device for a grayscale testing program, characterized in that, The device includes: The first acquisition module is used to respond to a test join request sent by the target requesting end and determine target identification information for the target tool from the identification information of the target requesting end; the target tool is a tag generation tool that can output tags including predefined tags corresponding to the function under test, and the predefined tags are usage characteristic tags defined in the preset conditions that the requesting end needs to meet to use the function under test; The generation module is used to obtain specified type data related to the target request client based on the target identification information, and generate usage characteristic tags for the target request client according to the specified type data through the target tool; the specified type data includes the type of data required by the target tool when generating tags; The first distribution module is used to distribute the experimental group configuration corresponding to the function under test to the target requesting end if the usage characteristic tag of the target requesting end meets the preset conditions corresponding to the function under test.

8. The apparatus according to claim 7, characterized in that, The first distribution module is specifically used for: For any of the preset conditions corresponding to the function to be tested, if the preset condition represents a predefined label that the requesting end needs to have, then if the usage characteristic label of the target requesting end includes the predefined label, it is determined that the usage characteristic label of the target requesting end satisfies the preset condition. If the preset condition represents a predefined tag that the requesting end does not need to have, then if the predefined tag is not included in the usage characteristic tags of the target requesting end, it is determined that the usage characteristic tags of the target requesting end satisfy the preset condition. If the characteristic tags used by the target requesting end satisfy all the preset conditions corresponding to the function to be tested, the experimental group configuration corresponding to the function to be tested is returned to the target requesting end.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When a processor executes a program stored in memory, it implements the grayscale test program distribution method according to any one of claims 1-6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the distribution method of the grayscale test program as described in any one of claims 1-6.

Images