Disclosure of Invention
In view of the above, the present invention provides a method and apparatus for reproducing a commodity list page, a storage medium, and an electronic device, in which URL, user characteristics, and commodity characteristics can be obtained through association identifiers, and the commodity list page causing user complaints is accurately reproduced according to the obtained URL, user characteristics, and commodity characteristics.
In order to achieve the above object, the embodiment of the present invention provides the following technical solutions:
the first aspect of the invention discloses a method for reproducing a commodity list page, which comprises the following steps:
Acquiring complaint pictures, wherein the complaint pictures are pictures acquired when a user complains about a commodity list page displayed by application software;
processing the complaint pictures to obtain commodity information sets, wherein the commodity information sets comprise commodity information of each commodity in the complaint pictures;
Determining a target buried point record from a plurality of pre-generated buried point records based on the commodity information set, wherein the target buried point record is a record generated in the process of displaying the commodity list page to the user;
Acquiring an associated identifier in the target buried point record, and acquiring user characteristics, commodity characteristics and uniform resource locators associated with the associated identifier in a preset historical snapshot database;
And based on the user characteristics, the commodity characteristics and the uniform resource locator, reconstructing a commodity list page when the complaint picture is generated.
The method, optionally, of processing the complaint image to obtain a commodity information set, includes:
Identifying each commodity in the complaint picture using a picture identification algorithm;
determining commodity information of each commodity in a preset commodity system;
And ordering the commodity information of each commodity according to the ordering position of each commodity in the complaint picture, so as to generate a commodity information set.
In the above method, optionally, the determining, based on the commodity information set, a target buried point record from a plurality of pre-generated buried point records includes:
determining a commodity set recorded by each buried point;
comparing the commodity information set with each commodity set;
determining a commodity set containing the commodity information set as a target commodity set;
and determining the buried point record of the target commodity set as a target buried point record.
The method, optionally, includes a process of pre-generating the target buried point record, including:
receiving a commodity browsing request sent by a user;
Acquiring user information, commodity information and a uniform resource locator based on the commodity browsing request, wherein the uniform resource locator comprises an association identifier;
Acquiring user characteristics corresponding to the user information and commodity characteristics corresponding to the commodity information;
associating the user characteristic and the merchandise characteristic with the association identifier;
Determining a commodity list page corresponding to the commodity browsing request based on the user characteristics and the commodity characteristics;
Displaying a commodity list page corresponding to the commodity browsing request to the user, monitoring whether a preset event embedded point is triggered in the process of displaying the commodity list page to the user in real time, and generating the target embedded point record based on the associated identifier when the event embedded point is triggered.
In the above method, optionally, the determining, based on the user feature and the commodity feature, a commodity list page corresponding to the commodity browsing request includes:
inputting the user features and the commodity features into a pre-trained commodity prediction model;
Triggering the commodity prediction model to process the user characteristics and the commodity characteristics to obtain a commodity prediction list output by the commodity prediction model;
And generating a commodity list page corresponding to the commodity browsing request based on the commodity prediction list.
The second aspect of the present invention discloses a reproduction apparatus of a commodity list page, comprising:
the first acquisition unit is used for acquiring complaint pictures, wherein the complaint pictures are pictures acquired when a user complains about a commodity list page displayed by the application software;
the processing unit is used for processing the complaint pictures to obtain commodity information sets, wherein the commodity information sets comprise commodity information of each commodity in the complaint pictures;
A first determining unit, configured to determine, based on the commodity information set, a target buried point record from a plurality of buried point records generated in advance, where the target buried point record is a record generated in a process of displaying the commodity list page to the user;
The second acquisition unit is used for acquiring the association identifier in the target buried point record and acquiring the user characteristics, commodity characteristics and uniform resource locators associated with the association identifier in a preset historical snapshot database;
and the reproduction unit is used for reproducing the commodity list page when the complaint picture is generated based on the user characteristics, the commodity characteristics and the uniform resource locator.
The above apparatus, optionally, the processing unit includes:
an identification subunit for identifying each commodity in the complaint picture using a picture identification algorithm;
the first determining subunit is used for determining commodity information of each commodity in a preset commodity system;
The first generation subunit is used for sequencing the commodity information of each commodity according to the sequencing position of each commodity in the complaint picture to generate a commodity information set.
The above apparatus, optionally, the first determining unit includes:
the second determining subunit is used for determining the commodity set recorded by each buried point;
a comparison subunit, configured to compare the commodity information set with each commodity set;
a third determining subunit configured to determine a commodity set including the commodity information set as a target commodity set;
and the fourth determination subunit is used for determining the buried point record to which the target commodity set belongs as a target buried point record.
The above device, optionally, further comprises:
the receiving unit is used for receiving a commodity browsing request sent by a user;
the third acquisition unit is used for acquiring user information, commodity information and a uniform resource locator based on the commodity browsing request, wherein the uniform resource locator comprises an association identifier;
a fourth obtaining unit, configured to obtain a user feature corresponding to the user information and a commodity feature corresponding to the commodity information;
An association unit for associating the user characteristic and the merchandise characteristic with the association identifier;
a second determining unit configured to determine a commodity list page corresponding to the commodity browsing request based on the user feature and the commodity feature;
and the generation unit is used for displaying the commodity list page corresponding to the commodity browsing request to the user, monitoring whether a preset event embedded point is triggered in the process of displaying the commodity list page to the user in real time, and generating the target embedded point record based on the association identifier when the event embedded point is triggered.
The above apparatus, optionally, the second determining unit includes:
An input subunit, configured to input the user feature and the commodity feature into a pre-trained commodity prediction model;
The triggering subunit is used for triggering the commodity prediction model to process the user characteristics and the commodity characteristics to obtain a commodity prediction list output by the commodity prediction model;
And the second generation subunit is used for generating a commodity list page corresponding to the commodity browsing request based on the commodity prediction list.
A third aspect of the present invention discloses a storage medium, wherein the storage medium includes stored instructions, and wherein the instructions, when executed, control a device in which the storage medium is located to perform a method for reproducing a merchandise list page as described above.
A fourth aspect of the present invention discloses an electronic device, characterized by comprising a memory, and one or more instructions, wherein the one or more instructions are stored in the memory and configured to perform a method of reproduction of a merchandise list page as described above by one or more processors.
Compared with the prior art, the invention has the following advantages:
The invention provides a method and a device for reproducing a commodity list page, a storage medium and electronic equipment, wherein the method comprises the steps of obtaining complaint pictures acquired when a user displays a complaint problem of the commodity list page by application software, and obtaining a commodity information set after processing the complaint pictures; the method comprises the steps of determining a target buried point record in a plurality of buried point records based on a commodity information set, acquiring an associated identifier in the target buried point record, acquiring user characteristics, commodity characteristics and a uniform resource locator associated with the associated identifier, and reproducing a commodity list page when a complaint picture is generated based on the user characteristics, the commodity characteristics and the uniform resource locator. The invention accurately reproduces the commodity list page when the complaint picture is generated by using the user characteristics, commodity characteristics and the uniform resource locator associated with the associated identifier, and provides reliable data for analyzing the problem for the developer.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the present disclosure, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.
In the field of electronic commerce, searching and recommending are two important shopping guides, and the user traffic brought by the two product forms is generally more than 50%. If good experience is provided for the user, and the user order purchase is promoted, the conversion rate (UCVR) and the User Value (UV) of the user access can be improved, and finally the electronic commerce platform is helped to improve the transaction commodity value (GMV), otherwise, if the result of searching and recommending the returned commodity page can influence the user experience, the conversion rate can be reduced, user complaints can be incurred, and even legal liabilities can be caused.
Common user experience problems are exemplified by repeatedly recommending recently purchased commodities to a user, or by having low matching degree between the searched commodities and search words, wasting the positions of the displayed commodities in an App page, and prolonging the decision time of the user.
Therefore, when complaints related to searching and recommending user experience are received, research and development teams need to locate and analyze the root cause as soon as possible, properly solve the problem, shorten the influence time and range of bad user experience, and simultaneously, optimize the algorithm to avoid repeated occurrence. These are one of the key efforts of research and development teams to improve the user experience.
There are three common ways for a user to complain:
a) The user collects complaint pictures of results of searching or recommending unsatisfactory commodity pages, and the complaint pictures are uploaded to customer service through a complaint entrance of electronic commerce software to be fed back directly.
B) And if the users are exactly the users with large influence, larger public opinion fluctuation can be caused, and the reputation of the electronic commerce platform is influenced.
C) In addition, some good users may transfer complaint pictures to staff working on the e-commerce platform or friends and relatives, and the flexible feedback problem is caused.
From the aspect of user experience types, the complaints related to recommendation mainly comprise the recommendation of the purchased goods, the stack recommendation of similar goods, the sensitive goods and the non-compliance of images and texts. The complaints of the search are mainly concentrated, the search accuracy is high, and the searched commodities are irrelevant to the search word.
The current mainstream electronic commerce platform searches and recommends recalled commodities to achieve thousands of people and thousands of faces. The method mainly comprises the steps of taking user portraits, commodity portraits, search words (only suitable for searching), region information, time information and user behavior contexts as input features of a deep neural network, learning commodity recall sets with highest conversion rate through deep learning, and sequencing according to a certain algorithm. Therefore, analyzing the user complaints requires at least a unique identifier (hereinafter referred to as a user ID) of the user, and in addition, the uniform resource locator (Uniform Resource Locator, URL) of the user's access search and recommendation service request also contains much information to help the research team reproduce the request results and analyze the user complaints.
Complaints are now received that search and recommend the user experience, typically through a user ID provided by the household, into the background log system, querying the history request URL containing the user ID. The following is an illustration of an exemplary search request URL, "key=iphone" being a search term entered by the user, "userid= linda" being a user ID, and "area_ids= 1,72,2839,0" being user geographic location information relating to merchandise inventory and distribution areas.
The following is an illustration of the content contained in an exemplary search request URL ：key=iphone&scene=0&log_id=09582426debe49198bf3453e56966420&pvid=09582426debe49198bf3453e56966420&uuid=linda&qp_exclude=24&multi_suppliers=yes&shop_col=yes&area_ids=1,72,2839,0&page=1&pagesize=20.
The whole process of searching, requesting URL, model prediction and returning commodity list to user is shown in figure 1, and figure 1 is a flow chart of a method for showing the page for searching commodity to user. The user searches or opens the commodity recommendation page through the search page, the App sends a URL request to the back-end service, the URL request comprises a search keyword and a user_id, the back-end service obtains user characteristics and commodity characteristics according to the information as inputs of model prediction, and a commodity recall set with the highest conversion rate, namely a commodity list, is output.
In summary, "user features", "merchandise features" are inputs of "model predictions", and "merchandise list" is an output of "model predictions". The user complains about the result of "commodity list", so that the positioning problem is analyzed, and most importantly, the research and development team takes two inputs of "user characteristics" and "commodity characteristics". In the search scenario, "commodity list" and "search word" are generally visible through the user's screen shot (key=iphone), but the user information "user ID" (user_id= linda) is not available, if the problem is that most user IDs exist, the problem can be reproduced and located, otherwise, only the user ID can be found, and the cause can be reproduced and found.
The user experience complaint analysis of the recommended scenario is more complex than searching, as shown in fig. 2, fig. 2 is a flowchart showing a page of recommended goods to the user, the recommended scenario has no search word and has a recommended bit identifier (rec_id). Through the identifier, the recommender system finds a pool of products that can be recommended. In addition, the user has no rec_id in the screen capturing picture, the recommended position information can only be approximately guessed through the picture, and another challenge is that the recommended results of different users are quite different without search words, so that the prior technical means have limitations for reproduction and problem analysis of recommended complaints.
The existing user complaint reproduction and problem positioning methods have the following obvious defects for the three complaint modes:
The ' direct feedback ' type complaints through the complaint entrance ' can be obtained by ' user ID ', the URL accessed by the user at the time can be found from the system log, only a part of the user complaints can be reproduced, and analysis basis is provided. The reason is that the user line taken by the user ID is only the current snapshot and not the historical snapshot at the time of the user complaint. The user behavior and commodity information change very fast, and the algorithm model recalls and arranges the commodities for the user by sensing the changes sharply. Therefore, only if the user behavior and commodity characteristics are not changed or the result of predicting the commodity list is not changed, the research and development team can reproduce the problem, otherwise, the problem cannot be reproduced even if the user ID is taken.
For two complaints, namely 'sending complaint pictures to a social platform' and 'transferring complaint pictures to relatives and friends working on an e-commerce platform', because personal privacy is involved, the user ID of the user is generally inconvenient to directly inquire, and even in the complaints of 'sending complaint pictures to the social platform', inquiring the user ID can arouse the contradiction emotion of the user. In this case, if the scene is a search scene, some analysis clues can be found through the search words contained in the screen capturing picture by the user to try to reproduce the problem, and if the scene is recommended, the problem reproduction difficulty is high.
In a word, in the existing method for reproducing and analyzing the complaint problems of user experience, under the condition that only the complaint picture and the user ID are taken, and no snapshot of the historical behavior of the user and commodity information is provided, the problem is not deterministic, and the result of analysis may not be accurate because of the uncertainty of input.
The invention aims to provide a commodity list page reproduction method, which is characterized in that the user characteristics, commodity characteristics and uniform resource locators associated with the associated identifiers are determined through the associated identifiers in the embedded point records, and the commodity list page causing user complaints can be accurately reproduced through the user characteristics, commodity characteristics and uniform resource locators, so that accurate data reference is provided for research personnel.
The invention is operational with numerous general purpose or special purpose computing device environments or configurations. Such as a personal computer, a server computer, a hand-held or portable device, a tablet device, a multiprocessor apparatus, a distributed computing environment including any of the above, and the like.
The embodiment of the invention provides a method for reproducing a commodity list page, an execution main body of the method is a back-end server, and referring to fig. 3, a method flow chart of the method for reproducing the commodity list page provided for implementation of the invention is specifically described as follows:
s101, acquiring complaint pictures.
The back-end server obtains complaint pictures in various modes, namely three modes are directly obtained through a complaint entrance of the electronic commerce platform, the complaint pictures sent by the user are obtained in various social platforms, and the complaint pictures are transferred to staff of the electronic commerce platform by the user.
The method for collecting the complaint pictures is various, for example, a user can shoot a commodity page with the complaint opinions by using a mobile phone with a shooting function, so that the complaint pictures can be obtained, and the commodity page with the complaint opinions displayed on the intelligent equipment can be directly shot through the self-contained screenshot function in the intelligent equipment, so that the complaint pictures can be obtained.
The complaint pictures are pictures acquired when a user has complaint problems on a commodity list page displayed by application software, the application software is software capable of carrying out data interaction with an electronic commerce platform, and preferably, the complaint pictures comprise a plurality of commodities which are predicted by the electronic commerce platform according to commodity browsing requests sent by the user.
S102, processing the complaint pictures to obtain a commodity information set.
In the method provided by the embodiment of the invention, the commodity information set comprises commodity information of each commodity in the complaint picture.
Referring to fig. 4, a flowchart of a method for processing complaint pictures to obtain a commodity information set according to an embodiment of the present invention is specifically described below:
S201, identifying each commodity in the complaint picture by using a picture identification algorithm.
The image recognition algorithm is a preset algorithm, and the image recognition algorithm is used for processing the complaint images, so that each commodity in the complaint images can be recognized.
S202, determining commodity information of each commodity in a preset commodity system.
The commodity system includes, but is not limited to, a commodity retrieval system and a commodity pool, wherein the commodity retrieval system and the commodity pool contain information of a plurality of commodities.
When the commodity information of each commodity is determined in a preset commodity system, the commodity information of each commodity can be determined in a commodity retrieval system according to the scene type of the complaint picture, when the scene type of the complaint picture is the searched commodity type, the commodity information of each commodity can be determined in a commodity pool, and when the scene type of the complaint picture is the recommended commodity type, the commodity information of each commodity can be determined in the commodity pool.
The commodity information includes, but is not limited to, information such as commodity ID, commodity name, commodity category, and the like.
S203, sorting the commodity information of each commodity according to the sorting position of each commodity in the complaint picture, and generating a commodity information set.
The commodities are distributed and displayed in the complaint pictures according to a preset arrangement mode, for example, the commodities can be arranged and displayed in a sequence mode.
According to the ordering position of each commodity in the complaint picture, the commodity information of each commodity is arranged, so that the ordering of each commodity information in the generated commodity information set is the same as the ordering position of each commodity in the complaint picture, the consistency and the relevance between the extracted data and the data provided by the user are ensured, and the accuracy is improved for the subsequent reproduction of the commodity list page corresponding to the complaint picture.
S103, determining a target buried point record from a plurality of pre-generated buried point records based on the commodity information set.
In the method provided by the embodiment of the invention, the target embedded point record is a record generated in the process of displaying the commodity list page to the user.
Referring to fig. 5, a flowchart of a method for determining a target buried point record based on a commodity information set according to an embodiment of the present invention is specifically described below:
S301, determining a commodity set recorded by each buried point.
The buried point records are stored in ELASTIC SEARCH search service, ELASTIC SEARCH search service user stores buried point records generated for each user.
The commodity set of the embedded point record comprises a plurality of commodity IDs and an associated identifier, wherein the associated identifier is used for searching data applied when a commodity list page is generated for a user, and further, each commodity ID in the commodity set is presented in a sequence mode, namely, each commodity ID in the commodity set forms a commodity ID sequence.
S302, comparing the commodity information set with each commodity set.
And comparing the commodity information set with each commodity set so as to determine the commodity set containing the commodity information set from the commodity sets.
Preferably, the commodity information set includes commodity information arranged in a sequence, that is, each commodity information in the commodity information set forms a commodity information sequence, and it should be noted that a sequence segment corresponding to the commodity information sequence in the commodity information set exists in the commodity ID sequence in the commodity set including the commodity information set.
S303, determining the commodity set containing the commodity information set as a target commodity set.
S304, determining the buried point record of the target commodity set as a target buried point record.
The target buried point record is determined through the commodity information set obtained after the complaint picture is processed, so that records generated in the process of displaying the commodity list page to the user can be found in the buried point records, and the complaint picture is a picture collected according to the commodity list page, and therefore the target buried point record is associated with the complaint picture.
According to the invention, the target buried point record corresponding to the complaint picture can be queried by using ELASTIC SEARCH search service, so that the commodity list page displayed for the user can be reproduced conveniently through the searched record.
S104, acquiring the associated identifier in the target buried point record, and acquiring the user characteristic, the commodity characteristic and the uniform resource locator associated with the associated identifier in a preset historical snapshot database.
The target embedded point record comprises an associated identifier, a user characteristic, a commodity characteristic and a uniform resource locator corresponding to the associated identifier are determined in a historical characteristic snapshot database, the user characteristic and the commodity characteristic are characteristic data applied when a commodity list page displayed to a user is generated, the uniform resource locator is characteristic data applied when the commodity list page displayed to the user is generated, and further, the uniform resource locator also comprises page information of a page type of the commodity list page displayed to the user.
And S105, reproducing the commodity list page when the complaint picture is generated based on the user characteristics, the commodity characteristics and the uniform resource locator.
When the commodity list page when the complaint picture is reproduced, commodity prediction lists of a plurality of commodities can be obtained based on the user characteristics and the commodity characteristics, and the commodity list page is generated according to the page information contained in the uniform resource locator and the commodity prediction list, wherein the generated commodity list page is the reproduced commodity list page.
The page types are divided into a search type and a recommendation type, wherein a commodity list page corresponding to the search type is a commodity search page obtained after a user inputs a search keyword, and a commodity list page corresponding to the recommendation type is a commodity recommendation page actively recommended to the user by an electronic commerce platform, and can be generated without the search keyword.
The method comprises the steps of obtaining complaint pictures, processing the complaint pictures to obtain commodity information sets, determining target embedded point records in a plurality of embedded point records based on the commodity information sets, wherein the target embedded point records are records generated in the process of displaying the commodity list pages to users, obtaining associated identifiers in the target embedded point records, obtaining user features, commodity features and uniform resource locators associated with the associated identifiers, and reconstructing the commodity list pages when the complaint pictures are generated based on the user features, the commodity features and the uniform resource locators. According to the invention, the user characteristics, the commodity characteristics and the uniform resource locators applied to the commodity list page displayed to the user can be determined according to the associated identifiers in the target embedded point records by acquiring the target embedded point records corresponding to the complaint pictures, so that the commodity list page displayed to the user can be accurately reproduced according to the user characteristics, the commodity characteristics and the uniform resource locators when the complaint pictures are generated, the reproduced commodity list page can provide reliable data basis for research personnel, so that the research personnel can accurately optimize the recommendation algorithm and the search method, and provide higher-quality service for the user.
Referring to fig. 6, a flowchart of a method for pre-generating a target buried point record according to an embodiment of the present invention is provided, where the target buried point record is a buried point record, and therefore, the process of generating the buried point record may also be referred to as a flow shown in fig. 6, and the specific description of fig. 6 is as follows:
s401, receiving a commodity browsing request sent by a user.
The commodity browsing request is divided into two types, wherein one type is a commodity searching request and the other type is a commodity recommending request, the commodity searching request comprises but is not limited to search keywords input by a user, and the commodity recommending request is a request generated when the user clicks a recommending page and does not comprise the search keywords.
Preferably, the commodity browsing request includes an initial uniform resource locator, and the initial uniform resource locator does not include an association identifier.
And S402, acquiring user information, commodity information and a uniform resource locator based on the commodity browsing request, wherein the uniform resource locator comprises an association identifier.
It should be noted that, the uniform resource locator may be obtained by inserting the association identifier into the initial uniform resource locator, where the association identifier is a unique identifier.
The user information comprises but is not limited to user identity information, names and other contents, when the commodity browsing request is a commodity searching request, the commodity information is a searching keyword, and when the commodity browsing request is a commodity recommending request, the commodity information is a commodity recommending identifier.
S403, acquiring user characteristics corresponding to the user information and commodity characteristics corresponding to the commodity information.
Searching a preset user characteristic database for user characteristics corresponding to the user information, wherein the user characteristics comprise, but are not limited to, behavior information of commodities recently browsed by a user, information of recently purchased commodities and the like.
When the commodity information is a commodity recommendation identifier, the commodity characteristics corresponding to the commodity recommendation identifier are extracted from a commodity recommendation pool, and the commodity characteristics include but are not limited to commodity category, commodity production place and other information.
S404, associating the user characteristics and the commodity characteristics with the association identifier and storing the association identifier into a history snapshot database.
The history snapshot database may be comprised of a kafka messaging system, associating the user features and the merchandise features with the associated identifiers, and saving the user features and merchandise features associated with the associated identifiers to the history snapshot database.
By associating the user features and the merchandise features with the associated identifiers, the user features and the merchandise features can be conveniently found out when complaints occur later.
It should be noted that, the historical feature snapshot database is used for storing the user behavior historical snapshot and the commodity information historical snapshot, where the user features stored in the historical feature snapshot database can be understood as the user behavior historical snapshot, and the commodity features stored in the historical feature snapshot database can be understood as the commodity information historical snapshot.
S405, determining a commodity list page corresponding to the commodity browsing request based on the user characteristics and the commodity characteristics.
Referring to fig. 7, a flowchart of a method for generating a commodity list page corresponding to a commodity browsing request according to an embodiment of the present invention is specifically described below:
s501, inputting the user characteristics and the commodity characteristics into a pre-trained commodity prediction model.
The commodity prediction model is used for predicting each commodity displayed to the user, wherein the commodity prediction model can be constructed by using a deep neural network.
S502, triggering the commodity prediction model to process the user characteristics and the commodity characteristics to obtain a commodity prediction list output by the commodity prediction model.
After the user features and the commodity features are input into the commodity prediction model, the commodity prediction model processes the user features and the commodity features and outputs a prediction result, wherein the prediction result comprises a commodity prediction list, and each commodity predicted is contained in the commodity prediction list.
S503, generating a commodity list page corresponding to the commodity browsing request based on the commodity prediction list.
According to the embodiment of the invention, the commodity prediction model is obtained by processing the user characteristics and the commodity characteristics by using the commodity prediction model, and the commodity list page is generated according to the commodity prediction model.
And S406, displaying the commodity list page corresponding to the commodity browsing request to a user, monitoring whether a preset event embedded point is triggered in the process of displaying the commodity list page to the user in real time, and generating the target embedded point record based on the association identifier when the trigger event embedded point is monitored.
The method includes the steps that when the intelligent terminal displaying the commodity list page is monitored, the triggering event embedding point is determined, and when the intelligent terminal displaying the commodity list page is monitored, the intelligent terminal displaying the commodity list page is used for capturing complaint pictures by default, wherein the complaint pictures are acquired by the screenshot function of the intelligent terminal when the complaint problem of the user on the commodity list page exists. For example, when the commodity list page is displayed to the user, the event embedding point is triggered immediately, and the event embedding point is triggered when the commodity list page is displayed to the user, so that the situation that the user uses other picture acquisition equipment to acquire complaint pictures is covered, and the event embedding point can be triggered when the commodity list page is displayed to the user.
And after the trigger event buries, generating a target buries record associated with the commodity list page displayed to the user according to the association identifier.
In the method provided by the embodiment of the invention, the uniform resource locator containing the association identifier is obtained, the association identifier is used for associating the user characteristic with the commodity characteristic and then storing the same, the commodity list page is generated by using the user characteristic and the commodity characteristic, and when the event buried point is triggered in the process of displaying the commodity list page to the user, the buried point record can be generated, so that when the complaint occurs later, the uniform resource locator, the user characteristic and the commodity characteristic can be determined based on the buried point record, the commodity list page displayed to the user is reproduced based on the uniform resource locator, the user characteristic and the commodity characteristic, and data required for solving the complaint of the user is provided for the research personnel.
Referring to fig. 8, an exemplary diagram of a scenario provided by an embodiment of the present invention is illustrated, in which a scenario in which a user searches for a commodity is illustrated, and the specific description is as follows:
"searchkey =iphone & user_id= linda" in the figure can be understood as a programming expression form of a commodity browsing request sent by a user, "ser_id= linda" can be understood as user information, and "key=iphone" can be understood as commodity information, wherein URL in the figure is a uniform resource locator containing an associated identifier, a commodity retrieval system obtains user characteristics according to the user information, processes commodity information to obtain commodity characteristics, inputs the user characteristics and the commodity characteristics into a commodity prediction model, processes the user characteristics and the commodity characteristics by the commodity prediction model, outputs a prediction result, wherein the prediction result comprises a commodity list, generates a commodity list page according to the commodity list in the prediction result, and feeds the commodity list page back to the user. In the process of displaying the commodity list to the user, triggering an event buried point when the user captures a screen, and associating the uniform resource locator, the user characteristic, the commodity characteristic and the commodity list by using the association identifier to generate a buried point record corresponding to a commodity list page displayed to the user. After receiving the screenshot picture provided by the user, carrying out recognition processing on the complaint picture to obtain a commodity list (the commodity list at the moment is equivalent to a commodity information set), determining a corresponding buried point record according to the commodity list, determining user characteristics, commodity characteristics and uniform resource locators associated with the associated identifiers based on the associated identifiers in the buried point record, and reconstructing a commodity list page of the screenshot picture in the process of acquisition according to the user characteristics, commodity characteristics and uniform resource locators.
In the method provided by the embodiment of the invention, after the URL is obtained, the URL can be stored in a real-time log stream, a unique identification page_id (equivalent to an association identifier) is added to a commodity list page displayed to a user and stored in the real-time log stream, the user uses a screen capture triggering event buried point to generate a buried point record and stores the buried point record in the real-time log stream, wherein the buried point record is obtained by associating the URL with information related to the commodity list page by using the association identifier, and the record is a buried point record of a user screen capture associated with user characteristics, commodity list and the URL.
After the research and development team obtains the screenshot picture (corresponding to the complaint picture) of the user complaint, the commodity ID list (corresponding to the commodity information set) can be obtained through manual or image recognition technology according to the commodity retrieval system, so that the information can be correlated to the record (target buried point record) of the stored screenshot related information, and the information is enough to help the research and development team to quickly and accurately reproduce and provide clues for analyzing the problems.
Preferably, the scheme provided by the invention can be divided into three parts, namely, storing commodity page related information, burying points by an App trigger event, automatically associating the commodity page related information by a background, identifying commodity IDs in a screen capturing picture and searching related characteristic information.
When a user uses an e-commerce App to search for goods or opens a pushing page to browse goods (corresponding to receiving a goods browsing request sent by the user), the background system automatically stores relevant information of a goods page, and specifically comprises the following steps:
The URL requested by App to the search page or the recommended page is stored in the kafka publish/subscribe messaging system, and a page_id field (which is used to hold an association identifier) is added to the URL as a unique identifier.
The user characteristics are obtained from the user information storage system, the commodity characteristics are obtained from the commodity retrieval system or the commodity pool, the user characteristics and the commodity characteristics are respectively stored in the kafka message system, and the page_id field is used as a unique identifier for association when the user characteristics and the commodity characteristics are stored.
The commodity prediction model processes commodity characteristics and user characteristics, outputs a commodity list, feeds back to a user, generates a commodity list page based on the commodity list, uses a page_id field as a unique identifier of the commodity list page, and stores the unique identifier in the kafka message system.
When a user intercepts a commodity list page by using a screen capturing function, triggering an event buried point, and automatically associating related information of the commodity page by a background:
and the user uses a screen capturing function to capture the picture of the commodity list page, and the captured picture is stored in the local storage of the intelligent device.
The user screen capture event, the unique identifier page_id of the merchandise list page are all stored to the kafka messaging system.
The page_id of the screen capturing event is associated with the URL, the user characteristic and the commodity characteristic stored in the last part to form a buried point record, and the buried point record is stored in ELASTIC SEARCH retrieval service for inquiring when a user complains.
When the research and development team receives the screenshot picture (corresponding to the complaint picture) sent by the user, the record (corresponding to the target embedded point record) in the retrieval service can be searched ELASTIC SEARCH by using the commodity ID (corresponding to the commodity information set), the URL is obtained to reproduce the problem, the parameters of the user characteristic and the commodity characteristic are used to analyze the problem, and the obtained URL, the user characteristic and the commodity characteristic are used to reproduce the commodity list page corresponding to the screen resolution picture. The method comprises the following specific steps:
Manually analyzing commodity images in the screen capturing pictures, and combining a commodity retrieval system to obtain commodity IDs, or automatically obtaining the commodity IDs through an image recognition technology;
And searching ELASTIC SEARCH the retrieval service by using the commodity ID list to obtain a corresponding commodity page related record (corresponding to the target buried point record).
After the event buried point is triggered, the URL, the user characteristic and the commodity characteristic are associated into a buried point record which is stored in ELASTIC SEARCH retrieval service for inquiring when the user complains. After the complaint picture is received, the complaint picture is processed to obtain commodity ID, corresponding buried point records are queried in ELASTIC SEARCH search service to obtain corresponding URL, user characteristics and commodity characteristics, the URL is used for reproducing the problem, the model input parameters of the user characteristics and the commodity characteristics are used for analyzing the problem, and a commodity list page of a user when collecting the complaint picture can be reproduced according to the URL, the user characteristics and the commodity characteristics.
The reproduction scheme of the commodity list page provided by the embodiment of the invention can be suitable for various user complaint scenes, can comprehensively cover the complaint paths of users, can obtain the user characteristics, the URL and the commodity characteristics when the user complaints are caused no matter what paths the users complain through, and can rapidly and accurately reproduce the commodity list page when the user complaints are caused through the obtained user characteristics, the URL and the commodity characteristics, thereby providing the data basis required by the development team during repair and improving the efficiency of repairing and solving the complaints.
Corresponding to the method shown in fig. 1, the embodiment of the invention also provides a reproduction device of the commodity list page, the device can be applied to a back-end server, the structure schematic diagram of the device is shown in fig. 9, and the specific description is as follows:
The first obtaining unit 601 is configured to obtain a complaint picture, where the complaint picture is a picture collected when a user complains about a commodity list page displayed by application software;
the processing unit 602 is configured to process the complaint image to obtain a commodity information set, where the commodity information set includes commodity information of each commodity in the complaint image;
A first determining unit 603 configured to determine, based on the commodity information set, a target buried point record, which is a record generated in a process of showing the commodity list page to the user, from among a plurality of buried point records generated in advance;
A second obtaining unit 604, configured to obtain an association identifier in the target embedded point record, and obtain, in a preset historical snapshot database, a user feature, a commodity feature, and a uniform resource locator associated with the association identifier;
And a reproducing unit 605, configured to reproduce a commodity list page when the complaint image is generated, based on the user feature, the commodity feature, and the uniform resource locator.
The device provided by the embodiment of the invention comprises the steps of obtaining complaint pictures, processing the complaint pictures to obtain commodity information sets, determining target embedded point records in a plurality of embedded point records based on the commodity information sets, wherein the target embedded point records are records generated in the process of displaying the commodity list pages to users, obtaining associated identifiers in the target embedded point records, obtaining user characteristics, commodity characteristics and uniform resource locators associated with the associated identifiers, and reconstructing the commodity list pages when the complaint pictures are generated based on the user characteristics, the commodity characteristics and the uniform resource locators. According to the invention, the user characteristics, the commodity characteristics and the uniform resource locators applied to the commodity list page displayed to the user can be determined according to the associated identifiers in the target embedded point records by acquiring the target embedded point records corresponding to the complaint pictures, so that the commodity list page displayed to the user can be accurately reproduced according to the user characteristics, the commodity characteristics and the uniform resource locators when the complaint pictures are generated, the reproduced commodity list page can provide reliable data basis for research personnel, so that the research personnel can accurately optimize the recommendation algorithm and the search method, and provide higher-quality service for the user.
In the apparatus provided by the embodiment of the present invention, the processing unit 602 may be configured to:
an identification subunit for identifying each commodity in the complaint picture using a picture identification algorithm;
the first determining subunit is used for determining commodity information of each commodity in a preset commodity system;
The first generation subunit is used for sequencing the commodity information of each commodity according to the sequencing position of each commodity in the complaint picture to generate a commodity information set.
The apparatus provided by the embodiment of the present invention, the first determining unit 603 may be configured to:
the second determining subunit is used for determining the commodity set recorded by each buried point;
a comparison subunit, configured to compare the commodity information set with each commodity set;
a third determining subunit configured to determine a commodity set including the commodity information set as a target commodity set;
and the fourth determination subunit is used for determining the buried point record to which the target commodity set belongs as a target buried point record.
The device provided by the embodiment of the invention can be further configured to:
the receiving unit is used for receiving a commodity browsing request sent by a user;
the third acquisition unit is used for acquiring user information, commodity information and a uniform resource locator based on the commodity browsing request, wherein the uniform resource locator comprises an association identifier;
a fourth obtaining unit, configured to obtain a user feature corresponding to the user information and a commodity feature corresponding to the commodity information;
An association unit for associating the user characteristic and the merchandise characteristic with the association identifier;
a second determining unit configured to determine a commodity list page corresponding to the commodity browsing request based on the user feature and the commodity feature;
and the generation unit is used for displaying the commodity list page corresponding to the commodity browsing request to the user, monitoring whether a preset event embedded point is triggered in the process of displaying the commodity list page to the user in real time, and generating the target embedded point record based on the association identifier when the event embedded point is triggered.
The device provided by the embodiment of the present invention, the second determining unit may be configured to:
An input subunit, configured to input the user feature and the commodity feature into a pre-trained commodity prediction model;
The triggering subunit is used for triggering the commodity prediction model to process the user characteristics and the commodity characteristics to obtain a commodity prediction list output by the commodity prediction model;
And the second generation subunit is used for generating a commodity list page corresponding to the commodity browsing request based on the commodity prediction list.
The embodiment of the invention also provides a storage medium, which comprises stored instructions, wherein the instructions are used for controlling equipment where the storage medium is located to execute the reproduction method of the commodity list page.
The embodiment of the present invention further provides an electronic device, whose structural schematic diagram is shown in fig. 10, specifically including a memory 701, and one or more instructions 702, where the one or more instructions 702 are stored in the memory 701, and configured to be executed by the one or more processors 703 to perform the method for reproducing the commodity list page.
The specific implementation process and derivative manner of the above embodiments are all within the protection scope of the present invention.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for a system or system embodiment, since it is substantially similar to a method embodiment, the description is relatively simple, with reference to the description of the method embodiment being made in part. The systems and system embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of functionality in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.