Multi-platform commodity search engine creating method, system and equipment and storage medium

By creating display classification tables and building index data structure tables, the problem of indiscriminate multi-platform product classification standards is solved, efficient and unified management of multi-platform product search engines is realized, and search efficiency and user experience are improved.

CN118916562APending Publication Date: 2024-11-08BEIJING AIFUKE NETWORK TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410696063.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-05-31
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The product coding specifications and classification rules of multiple e-commerce platforms are different, resulting in the classification standards being not unique when displayed in a single search engine, resulting in too long search requests, large resource consumption, long response time, and affecting the user experience.

Method used

Create a display classification table, bind the product categories of multiple e-commerce platforms with display category ids, build an index data structure table, configure search service instances, and set search sorting, category prediction and query analysis functions to realize Multi-platform product search engine.

Benefits of technology

It simplifies the management rules of the search platform, improves search efficiency, reduces server resource consumption, and improves user experience and the accuracy of search results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118916562A_ABST
    Figure CN118916562A_ABST
Patent Text Reader

Abstract

The invention provides a multi-platform commodity search engine creating method, system and device and a storage medium, and belongs to the technical field of computers, and the method comprises the steps of creating a display classification table, carrying out many-to-many relation binding on platform commodity category ids of commodity categories of multiple e-commerce platforms and display category ids, and re-maintaining a set of classification system, so that the commodity search engine is created. Therefore, commodity search of multiple e-commerce platforms can be integrated into one search engine, the management rule of the search platform is simplified, the index data structure table is created according to the data information of each commodity in the bound display classification table on each e-commerce platform so as to create a search service instance, a multi-task and multi-service processing mode is realized, and the search efficiency is improved. Changed commodity information can be found in time and pushed to multiple servers for parallel processing, and the number of the servers can be dynamically controlled according to the data size to achieve cost reduction and benefit improvement.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of computer technology, and in particular relates to a method, system, device and storage medium for creating a multi-platform commodity search engine. Background Art

[0002] There are many e-commerce platforms nowadays, the most famous ones are Taobao, JD.com, NetEase Yanxuan, etc. These e-commerce platforms will provide users with product search functions, allowing users to quickly find the products they need to buy. However, shopping malls like Taobao, JD.com, and NetEase Yanxuan sell products from separate shopping malls that they manage themselves. Their products have unified product coding specifications, and the product sku_id (stock unit identifier) ​​is unique; the product classification coding method is unique, and a unified product classification scheme is used for management. When converting to search engines, a unified classification processing scheme and a unified product coding specification can also be used for processing.

[0003] However, for e-commerce platforms that have both self-managed self-operated products and products connected to the malls of multiple e-commerce platforms, the products on the entire platform will have multiple different product coding specifications and multiple different product classification rules. Putting multiple products with different specifications in the same search engine and displaying them to users at the same time is much more complicated than a platform with a single management rule.

[0004] For example, if you store products from multiple shopping malls in the same search engine instance at the same time, the isolation of each independent product can be achieved by using sku_id and the product's respective shopping mall ID. However, when searching for products by category, the classification standard may not be unique. For example, if a user wants to search for products in the "food" category, he needs to first find all possible food-related categories in multiple shopping malls, and then search for products through these categories. This will cause the problem of too long parameters. A food category may correspond to more than a dozen categories in each shopping mall. Each category in a different shopping mall must be distinguished by the ID of the shopping mall to which it belongs. The search conditions to be spliced ​​in a search request may be very long. After submitting it to the search engine, it will report an error of "6009" that the query clause is too long. At the same time, searching for related categories of "food" before querying will also consume a lot of resources, resulting in a long response time for querying products, affecting customer experience. Summary of the invention

[0005] Based on the above status of existing commodity searches, the present invention provides a method, system, device and storage medium for creating a multi-platform commodity search engine, so as to overcome at least one technical problem existing in the prior art.

[0006] To achieve the above object, the present invention provides a method for creating a multi-platform commodity search engine, comprising:

[0007] Based on the commodity categories of multiple e-commerce platforms connected to the basic search engine, a display category table is created; wherein the display category table includes a display category ID and a category name corresponding to the display category ID;

[0008] Based on the category names of the product categories, the platform product category IDs of the product categories of multiple e-commerce platforms are bound to the display category IDs in a many-to-many relationship to obtain a bound display category table;

[0009] Based on the data information of each product name in the binding display classification table on each e-commerce platform, an index data structure table is constructed for the basic search engine;

[0010] Creating a search service instance according to the index data structure table, and configuring the search service instance to the basic search engine to obtain a primary search engine;

[0011] The primary search engine is provided with a search ranking function, a category prediction function and an optional current query analysis function to obtain a multi-platform commodity search engine.

[0012] In order to solve the above problems, the present invention also provides a multi-platform commodity search engine creation system, the system comprising:

[0013] A display category table creation module, used to create a display category table based on the commodity categories of multiple e-commerce platforms connected to the basic search engine; wherein the display category table includes a display category ID and a category name corresponding to the display category ID;

[0014] A relationship binding module, used for binding the platform commodity category IDs of commodity categories of multiple e-commerce platforms with the display category IDs in a many-to-many relationship based on the category names of the commodity categories, to obtain a binding display category table;

[0015] An index data creation module, used to construct an index data structure table for the basic search engine based on the data information of each product name in the binding display classification table on each e-commerce platform;

[0016] An instance configuration module, used to create a search service instance according to the index data structure table, and configure the search service instance to the basic search engine to obtain a primary search engine;

[0017] The function creation module is used to set a search ranking function, a category prediction function and a current query analysis optional function for the primary search engine to obtain a multi-platform commodity search engine.

[0018] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0019] at least one processor; and,

[0020] a memory communicatively connected to the at least one processor; wherein,

[0021] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps in the multi-platform product search engine creation method as described above.

[0022] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. When the at least one instruction is executed by a processor in an electronic device, the above-mentioned multi-platform product search engine creation method is implemented.

[0023] The multi-platform commodity search engine creation method, system, device and storage medium provided by the present invention can integrate commodity searches of multiple e-commerce platforms into one search engine by creating a display classification table, binding the platform commodity category IDs of commodity categories of multiple e-commerce platforms with the display category IDs in a many-to-many relationship, and re-maintaining a set of classification systems, thereby simplifying the management rules of the search platform, and creating an index data structure table based on the data information of each commodity in the bound display classification table on each e-commerce platform to create a search service instance, thereby realizing a multi-task, multi-service processing method, being able to find changed commodity information in a timely manner, and pushing it to multiple servers for parallel processing, and being able to dynamically control the number of servers according to the amount of data to achieve cost reduction and efficiency improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 A schematic diagram of a process for creating a multi-platform product search engine according to an embodiment of the present invention;

[0026] Figure 2 A schematic diagram of modules of a multi-platform commodity search engine creation system provided by an embodiment of the present invention;

[0027] Figure 3 A schematic diagram of the internal structure of an electronic device for implementing a method for creating a multi-platform product search engine provided by an embodiment of the present invention;

[0028] Figure 4 A schematic diagram of the configuration of a category prediction function provided by an embodiment of the present invention.

[0029] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0030] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0031] Based on the above-mentioned problems existing in the prior art, the present invention mainly provides a multi-platform commodity search engine creation method, system, device and storage medium, whose main purpose is to solve the problems in the prior art that the generated summary is prone to repetition, disordered word order, etc. due to the lack of ability to obtain graph structure information in the automatic summary generation model.

[0032] Figure 1 A flowchart of a method for creating a multi-platform product search engine provided by an embodiment of the present invention is provided. The method may be executed by a system, which may be implemented by software and / or hardware.

[0033] Figure 1 This paper gives an overall description of the method for creating a multi-platform product search engine. Figure 1 As shown, in this embodiment, the multi-platform product search engine creation method includes steps S110 to S150.

[0034] Step S110: Create a display category table based on the commodity categories of multiple e-commerce platforms connected to the basic search engine; wherein the display category table includes a display category ID and a category name corresponding to the display category ID.

[0035] Specifically, in the present invention, multiple e-commerce platforms refer to at least two e-commerce platforms, each of which has its own commodity category and category code (commodity category id), and the commodity categories of all e-commerce platforms connected to the basic search engine are counted, so as to create a display classification table containing all the display categories connected to the basic search engine. The display classification table at least includes a display category id, which is generally a string of numbers or letters plus numbers (not particularly limited to this), and the category name corresponding thereto. Among them, the docking method of the basic search engine with multiple e-commerce platforms includes direct docking and may also include indirect docking. For example, the basic search engine directly docks with the shopping malls of multiple e-commerce platforms through a data interface (direct docking), or a certain e-commerce platform uses a certain search engine, and other e-commerce platforms establish a data interface with the e-commerce platform. Therefore, other e-commerce platforms indirectly dock with the basic search engine through the platform. The following is an example of indirect docking.

[0036] Taking the Aifuke e-commerce platform as an example, the products sold in its platform mall include self-operated products, and are connected to JD.com and NetEase. The mall products of the Aifuke e-commerce platform use 32-bit UUID (universally unique identifier) ​​as the product's sku_id (stock unit identifier) ​​as shown in Table 1.

[0037]

[0038] Table 1

[0039] Commodity classification adopts two-level classification management, and the classification code adopts 6-digit combination processing, as shown in Table 2.

[0040]

[0041] Table 2

[0042] The operation personnel control the addition and modification of goods through the operation management background. Except for inventory, each change of Aifuke's mall goods will modify the update_time field of the corresponding product table. The system will establish multiple xxl-job scheduled tasks, which will be triggered every 10 minutes. By scanning the product data of update_time between the last execution time and the current time, it will be pushed to the search engine and the index will be updated for users to query products.

[0043] The products in NetEase Mall use 1-32 digits as the product sku_id, as shown in Table 3.

[0044]

[0045] Table 3

[0046] Commodity classification adopts two-level classification management, and the classification code adopts 1-7 digit combination processing, as shown in Table 4.

[0047]

[0048] Table 4

[0049] Changes to NetEase products are triggered by NetEase Yanxuan calling the NetEase message receiving interface of the Aifuke e-commerce platform. NetEase will call this interface to push product change information. After receiving the information, the interface will also use RabbitMQ to distribute the product change information to multiple consumers for processing. After the update is completed, NetEase will also be notified that the message has been consumed. Finally, the update_time of the NetEase product table will be updated, which is used as an identifier to update the search engine index used by the Aifuke e-commerce platform.

[0050] To store the mall products of the above three e-commerce platforms into the same search engine instance at the same time, the isolation of each independent product can be achieved by using sku_id and the mall identifier of each product. However, when searching for products by category, the classification standard will not be unique. For example, if a user wants to search for products classified as "food", he needs to first find all the categories that may be related to food in the malls of the three platforms, and then search for products through these categories. This will cause the problem of too long parameters. A food category may correspond to more than a dozen categories of each mall. Each category of different malls must be distinguished by the identifier of the mall to which it belongs. The search conditions to be spliced ​​in a search request may be very long. After submitting it to the search engine, it will report the error "6009" that the query clause is too long. At the same time, searching for related categories of "food" before querying will also consume a lot of resources, resulting in a long response time for querying products, affecting customer experience.

[0051] In order to solve the above problems, the present invention can create a display classification table, which can be set in the platform's mall as a new "display classification" function, and re-maintain a classification system. The different classifications of the malls of the three platforms are bound to this "display classification". In this way, when a user searches for "food", he can directly find all the products that have been bound to the "display classification-food".

[0052] As an optional embodiment of the present invention, the display classification table is a three-level display classification table, and the method for creating the three-level display classification table includes:

[0053] Create a basic classification code with a preset number of digits, and divide the basic classification code into primary classification digits, secondary classification digits, and tertiary classification digits from front to back;

[0054] Set the numbers of the basic classification code to all 0 as the parent classification id of the first-level classification, combine the numbers of the first-level classification digits according to the first preset number combination method, and set the numbers of the second-level classification digits and the third-level classification digits to 0 as the first-level classification id;

[0055] The first-level classification id is used as the parent classification id of the corresponding second-level classification, the number on the first-level classification position of the basic classification code is set as the parent classification id of the second-level classification, the numbers on the second-level classification positions are combined according to the second preset number combination method, and the numbers on the third-level classification positions are set to 0 as the second-level classification id;

[0056] The secondary classification id is used as the parent classification id of the corresponding third-level classification, the numbers on the primary classification digits and the numbers on the secondary classification digits of the basic classification code are set as the parent classification id of the third-level classification, and the numbers on the third-level classification digits are combined in a third preset digital combination as the third-level classification id;

[0057] Establish a correspondence between the first-level classification id, the second-level classification id and the third-level classification id and the name of the product category to obtain a three-level display classification table.

[0058] Specifically, the display classification table in the present invention is preferably a three-level display classification table, and of course it can also be other multi-level display classification tables, such as a four-level display classification table, a five-level display classification table, a six-level display classification table, etc., and the present invention does not make special restrictions on this. Among them, the first level is a large category, the second level is a subcategory under the first level, and the third level is a subcategory under the second level. For example, clothing is the first level, women's clothing is the second level, and skirts are the third level. The classification ids corresponding to clothing, women's clothing, and skirts are the first-level classification id, the second-level classification id, and the third-level classification id, respectively. Among them, the preset number of digits of the basic classification code can be set as needed, and there is no special restriction on this. The first-level classification bit, the second-level classification bit, and the third-level classification bit can also be set as needed. The first preset digital combination method, the second preset digital combination method, and the third preset digital combination method can also be set as needed. For example, if the first-level classification bit is two digits, then the digital combination can start from "01" to "99". Take the first-level classification bit, the second-level classification bit, and the third-level classification bit with 6 as the number as an example.

[0059] First-level classification: The parent category id (parent_id) of the first-level classification is "000000", the classification level (category_level) is 1, and the first-level classification id (category_id) is a 6-digit number. The combination of the first 2 digits starts from "01" to "99", and the last 4 digits are 0. 99 first-level categories can be recorded.

[0060] Secondary category: The parent category id (parent_id) of the secondary category is the id (category_id) of the corresponding primary category. The category level (category_level) is 2. The id (category_id) of the secondary category is a 6-digit number. The first 2 digits are the first 2 digits of the parent category id, the middle 2 digits are a digital combination starting from "01" to "99", and the last 2 digits are 0. Each primary category can have 99 secondary categories, and each secondary category can only have one parent primary category.

[0061] Three-level classification: The parent category id (parent_id) of the three-level classification is the id (category_id) of the corresponding second-level classification, the classification level (category_level) is 3, and the id (category_id) of the three-level classification is a 6-digit number, the first 4 digits are the first 4 digits of the parent category id, and the last 2 digits are a combination of numbers starting from "01" to "99". Each second-level classification can have 99 three-level classifications, and each three-level classification can only have one parent second-level classification. Table 5 creates a table for basic classification coding. Table 6 is an example of a three-level display classification table.

[0062]

[0063] Table 5

[0064]

[0065] Table 6

[0066] As an optional embodiment of the present invention, the basic search engine is an OpenSearch search engine.

[0067] Specifically, Alibaba's OpenSearch is used as the basic search engine. OpenSearch has built-in rich customized algorithm models, and combines the search characteristics of different industries to launch industry recall and sorting algorithms to ensure better search results. Based on its own business characteristics and data, it can customize the corresponding algorithm model, application structure, data processing, query analysis, sorting and other configurations to meet personalized search needs and improve the click-through rate of search results.

[0068] Step S120: Based on the category names of the product categories, the platform product category IDs of the product categories of multiple e-commerce platforms are bound to the display category IDs in a many-to-many relationship to obtain a bound display category table.

[0069] Specifically, based on the category name of the product category, the platform product category ids of the product categories of multiple e-commerce platforms are bound to the display category ids in a many-to-many relationship to obtain a bound display classification table. Take the binding of the three malls of the above three e-commerce platforms to the display classification table as an example. The product categories of the three malls each establish a binding relationship with the display classification, and the relationship between the product classification of each mall and the display classification is many-to-many. One display classification can correspond to multiple mall product categories, and one mall product classification can also correspond to multiple display categories. Table 7 is a binding relationship table between the product classification of Aifuke Mall and the display classification table; Table 8 is a binding relationship table between the product classification of JD Mall and the display classification table; Table 9 is a binding relationship table between the product classification of NetEase Mall and the display classification table. Among them,

[0070]

[0071] Table 7

[0072]

[0073] Table 8

[0074]

[0075] Table 9

[0076] Step S130: construct an index data structure table for a basic search engine based on the data information of each product name in the binding display classification table on each e-commerce platform.

[0077] Specifically, based on the data information of each product name in the binding display classification table on each e-commerce platform, a corresponding index is created for each product, and the application structure table is obtained and the data information is filled into the application structure table, thereby constructing an index data structure table. An example of the application structure table is shown in Table 10.

[0078]

[0079]

[0080]

[0081] Table 10

[0082] Step S140: Create a search service instance according to the index data structure table, and configure the search service instance to the basic search engine to obtain a primary search engine.

[0083] Specifically, a search service instance is created according to the index data structure table, wherein an instance is a set of data configurations of a user, which may include a data source structure, an index structure and other attribute configurations. An instance is a search service.

[0084] As an optional embodiment of the present invention, after creating a search service instance according to the index data structure table and configuring the search service instance to the basic search engine to obtain a primary search engine, it also includes: creating a scheduled index information update task and a scheduled removal of delisted products task in the primary search engine.

[0085] Specifically, the primary search engine writes data. In order to flexibly handle the updates of the primary search engine by products from multiple platform malls, multiple xxl-job (scheduled) tasks are established to push according to different malls and different task responsibilities, and to be efficient and parallel. The first step is to add / update index information. Taking the products of the Aifuke Mall as an example, first establish the last update time of the index with the key "afk_delta_sku_lasttime" in redis; establish a scheduled task in xxl-job to trigger the incremental update of the Aifuke Mall product index every 10 minutes, and select the product information with update_time within the range of "afk_delta_sku_lasttime" and the task trigger time to be pushed to OpenSearch. If the execution is successful, "afk_delta_sku_lasttime" is assigned to the new task trigger time as the next task. The start time of the next execution. If the execution fails, the value of "afk_delta_sku_lasttime" is retained and waits for the next re-execution. Each product information to be updated is pushed to the index service as an independent RabbitMQ message for processing. If the business is heavy and there is too much backlog of product index update information, you can add an index service at any time to solve the problem; the processing mode for removing delisted products is the same as adding / updating index information, except that the key of redis that records the last update time of the index is changed to "afk_delta_remove_sku_lasttime", and another xxl-job is created for processing. The processing of JD.com and NetEase product indexes is the same as that of Aifuke products. Adding / updating and deleting belong to different tasks, so that each task does not interfere with each other, and the overall situation will not be affected by the problem of a certain task. The index service that processes messages can also be added and subtracted dynamically in real time to control costs while processing data as quickly as possible.

[0086] Step S150: Setting a search ranking function, a category prediction function and a current query analysis optional function for the primary search engine to obtain a multi-platform product search engine.

[0087] Specifically, after the primary search engine is established, users will have basic product search capabilities and can retrieve product information through keywords, product categories, brands, and affiliated shopping malls. But this is only the basis. Users can retrieve products that meet the search criteria, but it is not certain whether they are the products the user wants to classify. For example, if a user wants to buy a mobile phone, he will use the keyword "mobile phone" to search. The search will include products such as mobile phone cases and mobile phone charging cables, which also contain the keyword "mobile phone", but they are not the products the user wants. In order to improve the quality of user search results, it is also necessary to equip the primary search engine index with relevance configuration to control the order of returned results to improve the accuracy of search results. At the same time, configure intervention words and other settings to give customers a better search experience.

[0088] As an optional embodiment of the present invention, the search sorting function includes: a basic sorting function and a default business sorting function; wherein the basic sorting function is used when the query conditions contain keywords; and the default business sorting function is used when there are no conditions for the query.

[0089] Specifically, the basic sorting category_base_pro is set according to the product classification. It can also be set by setting a business sorting, which is used together with the business sorting category_biz_pro, and then the category prediction intervention dictionary is set to take effect. In actual use, it is used only when the query condition contains keywords and is configured on first_rank_name. Among them, the setting method of the basic sorting function is shown in Table 11, and the setting method of the business sorting function is shown in Table 12.

[0090] Scoring Features Search Fields Weight category_score() category_id_int 5

[0091] Table 11

[0092] Sorting Configuration category_score(category_id_int)

[0093] Table 12

[0094] The default business sorting hot_category_pro is set according to the product background classification. It is used when there is no conditional query. It can be customized according to the different needs of the business operator. This is just a simple setting. In the future, you can configure multiple such sorting rules according to your needs, and then the front-end parameters will control the influence of sorting, which is configured on second_rank_name. The setting method of the default business sorting function is shown in Table 13.

[0095]

[0096]

[0097] Table 13

[0098] As an optional embodiment of the present invention, a method for setting a category prediction function for a primary search engine includes:

[0099] Create a category prediction model in the basic sorting function;

[0100] The training methods of the category prediction model include:

[0101] Taking the pre-collected query field data as input and the pre-collected display category ID corresponding to the query field data as output, training the preset basic model to obtain a training category prediction model;

[0102] When the training category prediction model reaches the preset training conditions, the category prediction model is obtained.

[0103] Specifically, the category prediction function is simply as follows: when a user enters a query, the relevance between the category to which the item belongs and the query is calculated. The higher the relevance between the category and the query, the higher the ranking score of the item will be, and the higher the item will be ranked.

[0104] Model training requires a data source. Before training a model, you must first bind an application. Application data, application query data, and application behavior data are all data required by the category prediction model. If you have not uploaded click behavior data, or do not use click behavior data for training, or the click behavior data does not meet the training conditions, you can train the model without using click behavior data. At this time, the category prediction model needs to select data from three fields in the application, namely: category id (display category id), item title, and category name. Category id and item title are required, and category name is optional. After the model training is completed, the prediction results of some models will be output for effect evaluation. The category name will be used on the effect evaluation page to evaluate whether the relevance between the query and the category meets expectations. Therefore, it is recommended to provide the category name field during training. If click behavior data has been uploaded, in addition to selecting the above field content, you can also select the option of associating behavior data when training the model. As long as the behavior data meets the training conditions, this part of data will be used to train the model. Create a category prediction model under the application. Apply the category prediction model: First, you need to apply the model in the query analysis, and then make the model effective in the basic sorting and business sorting. Create a query analysis, configure the category prediction and select the model created in 1. Make the category prediction model effective in the query: SDK calls the query interface and enters the raw_query parameter. Figure 4 shown.

[0105] As an optional embodiment of the present invention, the optional functions of query analysis currently include a synonym expansion function, a stop word omission function, a spelling correction function and a word weight analysis function.

[0106] Specifically, the intention judgment of the query keywords in the search directly determines whether the search results can meet the needs. The query semantic understanding (Query Planner) in OpenSearch (search engine) is a function used to understand the query search intention; through a series of intelligent analysis of the query, the query is rewritten and then retrieved and sorted in the engine. Currently, the optional functions of query analysis include synonym expansion, stop word omission, spelling correction, word weight analysis, category prediction, and entity recognition in e-commerce scenarios. Among them, the stop word function: filter meaningless words in the query according to the system's built-in stop word dictionary (generally words with high frequency of use but do not affect the query results, such as punctuation, modal particles, etc.). This can be done using the default function and finally configured on the query analyzer; spelling correction function: the query entered by the user is not always correct. Incorrect input may cause the query result to be inconsistent with expectations or no result, so the user's input needs to be spell checked. The spelling check function provided in the query analysis of OpenSearch corrects errors in the query terms and gives the correct query terms. And according to the credibility of the error correction, decide whether to use the corrected word for the current query. This can be done by using the default function, and finally configured on the query analyzer; word weight function: This function mainly analyzes the importance of each word in the query in the text and quantifies it into weights. Words with lower weights may not participate in the recall. This can avoid limiting the recall according to the query words entered by the user when the query words entered by the user contain some low-weight words, resulting in too few hit results. First establish a word weight dictionary; synonym function: In actual search scenarios, expressions containing synonyms often appear. For example, it is hoped that when users search for Apple phones, content containing iPhones can also be retrieved and presented. In real life, there are often many expressions with the same semantics, and it is difficult for users to reflect them all in one query when searching, so identifying and providing synonym retrieval can obviously achieve a higher recall rate. The synonym function mainly expands the query words synonymously and expands the recall of documents that are synonymous with the query words. This also requires first establishing a synonym dictionary, and finally configuring it to take effect on the query analyzer; Entity recognition function: Entity recognition, the full name of which is Named Entity Recognition (NER), refers to the recognition of semantic entities with specific meanings in query terms. Query analysis rewrites query terms based on the results of the recognition and the weight of the entity type, so that the recalled documents meet the intent of the query. Query rewriting mainly rewrites the query based on the importance of the entity. When recalling, the entity terms with high importance are retained, and the low-importance terms are not affected in the recall, but only in the algorithm sorting. Just keep the default here.

[0107] The function of query analysis intervention can also be achieved by setting up an intervention dictionary. First, create an intervention dictionary. The user enters the query analysis intervention dictionary page through the console ---> function extension ---> dictionary management. After entering the page, click "Create" in the upper right corner of the page. After selecting the dictionary type, name the dictionary. After the intervention dictionary is created, the dictionary will appear in the dictionary list on the page. Add and manage the intervention entries in the intervention dictionary. After the dictionary is created, click on the dictionary name in the list or click "Manage" corresponding to the dictionary to enter the details page of the intervention dictionary. The user can add and manage the intervention entries on the details page. The user can adjust the relevance of the categories in the recall results of a certain Query intervention, correct the categories with incorrect predicted relevance by the model, or supplement the categories with no predicted relevance by the model. Use the intervention dictionary. After creating and populating the intervention dictionary for category prediction, it can be used in any category prediction model. In an actual query, if there are calculation results of the relevance of the same category under a Query in both the intervention dictionary and the category prediction model, the results in the intervention dictionary will be merged. For example: when searching for "A Smile Is Beautiful", the category prediction model calculates that the relevance of this Query to category 20 is 0 (irrelevant), and the relevance to category 10 is 2 (relevant). The relevance of this Query to category 20 in the intervention dictionary is intervened to 2 (relevant), and the relevance to category 30 is 1 (slightly relevant). Then the final result is: the relevance of category 20 is 2, the relevance of category 30 is 1, and the relevance of category 10 is 2. The classification id in the intervention dictionary uses the id field in the index_category_map table (as shown in Table 14). The reason is that the classification id in the configuration of Alibaba's category prediction, basic sorting, and business sorting must be of int type. Currently, the classification ids in actual business are all of string type, and there are 3 types of mall product classifications: iFengKe, JD.com, and NetEase. Therefore, the index_category_map table is used to normalize the classification id. The logic is that when writing to the Open Search index, it is judged whether the classification id and the supplier of the current product exist in the index_category_map table. If not, a new record is inserted. category_id_int must be added to the list of attribute fields. In this way, the classification mapping is done by mall and is only used in sorting, which can adjust the sorting rules of the same classification products in different malls and provide a strategic tool for operation.

[0108]

[0109] Table 14

[0110] Such as Figure 2As shown, the present invention provides a multi-platform commodity search engine creation system 200, which can be installed in an electronic device. According to the functions implemented, the multi-platform commodity search engine creation system 200 may include: a display classification table creation module 210, a relationship binding module 220, an index data creation module 230, an instance configuration module 240 and a function creation module 250. The unit of the present invention can also be called a module, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0111] In this embodiment, the functions of each module / unit are as follows:

[0112] A display category table creation module 210 is used to create a display category table based on the commodity categories of multiple e-commerce platforms connected to the basic search engine; wherein the display category table includes a display category ID and a category name corresponding to the display category ID;

[0113] The relationship binding module 220 is used to perform a many-to-many relationship binding between the platform product category IDs and the display category IDs of the product categories of multiple e-commerce platforms based on the category names of the product categories, and obtain a binding display category table;

[0114] An index data creation module 230 is used to construct an index data structure table for a basic search engine based on the data information of each product name in the binding display classification table on each e-commerce platform;

[0115] An instance configuration module 240 is used to create a search service instance according to the index data structure table, and configure the search service instance to a basic search engine to obtain a primary search engine;

[0116] The function creation module 250 is used to set a search ranking function, a category prediction function and a current query analysis optional function for the primary search engine to obtain a multi-platform product search engine.

[0117] The multi-platform product search engine creation system 200 of the present invention, by creating a display classification table, binds the platform product category IDs of the product categories of multiple e-commerce platforms with the display category IDs in a many-to-many relationship, and re-maintains a set of classification systems, so as to integrate the product searches of multiple e-commerce platforms into one search engine, simplify the management rules of the search platform, and create an index data structure table based on the data information of each product in the bound display classification table on each e-commerce platform to create a search service instance, thereby realizing a multi-task, multi-service processing method, being able to find the changed product information in time, and push it to multiple servers for parallel processing, and being able to dynamically control the number of servers according to the amount of data to achieve cost reduction and efficiency improvement.

[0118] like Figure 3As shown, the present invention provides an electronic device 3 for a method of creating a multi-platform commodity search engine.

[0119] The electronic device 3 may include a processor 30, a memory 31 and a bus, and may also include a computer program stored in the memory 31 and executable on the processor 30, such as a multi-platform commodity search engine creation program 32. The memory 31 may also include both an internal storage unit of the multi-platform commodity search engine creation system and an external storage device. The memory 31 may be used not only to store the code installed in the application software and various data, such as the multi-platform commodity search engine creation program, but also to temporarily store data that has been output or is to be output.

[0120] Among them, the memory 31 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. The memory 31 may be an internal storage unit of the electronic device 3 in some embodiments, such as a mobile hard disk of the electronic device 3. The memory 31 may also be an external storage device of the electronic device 3 in other embodiments, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 3. Further, the memory 31 may also include both an internal storage unit of the electronic device 3 and an external storage device. The memory 31 may not only be used to store application software and various types of data installed in the electronic device 3, such as a multi-platform commodity search engine creation method code, etc., but may also be used to temporarily store data that has been output or is to be output.

[0121] The processor 30 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The processor 30 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules (such as a multi-platform product search engine creation program, etc.) stored in the memory 31, and calls data stored in the memory 31 to execute various functions of the electronic device 3 and process data.

[0122] The bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 31 and at least one processor 30, etc.

[0123] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3 The structure shown does not constitute a limitation on the electronic device 3, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0124] For example, although not shown, the electronic device 3 may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 30 through a power management system, so that the power management system can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging systems, power failure detection circuits, power converters or inverters, and power status indicators. The electronic device 3 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0125] Furthermore, the electronic device 3 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 3 and other electronic devices.

[0126] Optionally, the electronic device 3 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 3 and to display a visual user interface.

[0127] It should be understood that the embodiments are for illustration purposes only and the scope of the invention application is not limited by this structure.

[0128] The multi-platform product search engine creation program 32 stored in the memory 31 of the electronic device 3 is a combination of multiple instructions. When running in the processor 30, it can achieve:

[0129] Based on the commodity categories of multiple e-commerce platforms connected to the basic search engine, a display category table is created; wherein the display category table includes a display category ID and a category name corresponding to the display category ID;

[0130] Based on the category names of the product categories, the platform product category IDs of the product categories of multiple e-commerce platforms are bound to the display category IDs in a many-to-many relationship to obtain a bound display category table;

[0131] Based on the data information of each product name in the binding display classification table on each e-commerce platform, an index data structure table is constructed for the basic search engine;

[0132] Creating a search service instance according to the index data structure table, and configuring the search service instance to the basic search engine to obtain a primary search engine;

[0133] The primary search engine is provided with a search ranking function, a category prediction function and an optional current query analysis function to obtain a multi-platform commodity search engine.

[0134] Specifically, the specific implementation method of the processor 30 for the above instructions can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0135] Furthermore, if the module / unit integrated in the electronic device 3 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable medium may include: any entity or system capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0136] An embodiment of the present invention further provides a computer-readable storage medium, which may be non-volatile or volatile, and stores a computer program, which, when executed by a processor, implements:

[0137] Based on the commodity categories of multiple e-commerce platforms connected to the basic search engine, a display category table is created; wherein the display category table includes a display category ID and a category name corresponding to the display category ID;

[0138] Based on the category names of the product categories, the platform product category IDs of the product categories of multiple e-commerce platforms are bound to the display category IDs in a many-to-many relationship to obtain a bound display category table;

[0139] Based on the data information of each product name in the binding display classification table on each e-commerce platform, an index data structure table is constructed for the basic search engine;

[0140] Creating a search service instance according to the index data structure table, and configuring the search service instance to the basic search engine to obtain a primary search engine;

[0141] The primary search engine is provided with a search ranking function, a category prediction function and an optional current query analysis function to obtain a multi-platform commodity search engine.

[0142] Specifically, the specific implementation method when the computer program is executed by the processor can refer to the description of the relevant steps in the method for creating a multi-platform product search engine in the embodiment, and will not be repeated here.

[0143] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0144] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0145] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0146] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0147] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0148] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or systems stated in a system claim can also be implemented by one unit or system through software or hardware. Secondary words are used to indicate names, but not to indicate any particular order.

[0149] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for creating a multi-platform commodity search engine, characterized in that: The steps include: Based on the commodity categories of multiple e-commerce platforms connected to the basic search engine, a display category table is created; wherein the display category table includes a display category ID and a category name corresponding to the display category ID; Based on the category names of the product categories, the platform product category IDs of the product categories of multiple e-commerce platforms are bound to the display category IDs in a many-to-many relationship to obtain a bound display category table; Based on the data information of each product name in the binding display classification table on each e-commerce platform, an index data structure table is constructed for the basic search engine; Creating a search service instance according to the index data structure table, and configuring the search service instance to the basic search engine to obtain a primary search engine; The primary search engine is provided with a search ranking function, a category prediction function and an optional current query analysis function to obtain a multi-platform commodity search engine.

2. The method for creating a multi-platform product search engine according to claim 1, characterized in that: The display classification table is a three-level display classification table, and the method for creating the three-level display classification table includes: Create a basic classification code with a preset number of digits, and divide the basic classification code into a primary classification digit, a secondary classification digit, and a tertiary classification digit from front to back; Set the numbers of the basic classification code to all 0 as the parent classification id of the first-level classification, combine the numbers of the first-level classification digits according to the first preset number combination method, and set the numbers of the second-level classification digits and the third-level classification digits to 0 as the first-level classification id; The first-level classification id is used as the parent classification id of the corresponding second-level classification, the number on the first-level classification position of the basic classification code is set as the parent classification id of the second-level classification, the numbers on the second-level classification positions are combined according to the second preset number combination method, and the numbers on the third-level classification positions are set to 0 as the second-level classification id; The secondary classification id is used as the parent classification id of the corresponding third-level classification, the numbers on the primary classification digits and the numbers on the secondary classification digits of the basic classification code are set as the parent classification id of the third-level classification, and the numbers on the third-level classification digits are combined in a third preset digital combination as the third-level classification id; A correspondence is established between the first-level classification id, the second-level classification id and the third-level classification id and the name of the commodity category to obtain a three-level display classification table.

3. The method for creating a multi-platform product search engine according to claim 1, characterized in that: The basic search engine is the OpenSearch search engine.

4. The method for creating a multi-platform product search engine according to claim 1, characterized in that: After creating a search service instance according to the index data structure table and configuring the search service instance to the basic search engine to obtain a primary search engine, the method further includes: A task of regularly updating index information and a task of regularly clearing off-shelf commodities are created in the primary search engine.

5. The method for creating a multi-platform product search engine according to claim 1, characterized in that: The search ranking function includes: Basic sorting function and default business sorting function; among them, The basic sorting function is used when the query conditions include keywords; the default service sorting function is used when there are no conditions for the query.

6. The method for creating a multi-platform product search engine according to claim 5, characterized in that: The method for setting a category prediction function for the primary search engine comprises: Creating a category prediction model in the basic sorting function; wherein, The training method of the category prediction model includes: Taking the pre-collected query field data as input and the pre-collected display category ID corresponding to the query field data as output, training the preset basic model to obtain a training category prediction model; When the training category prediction model reaches the preset training condition, a category prediction model is obtained.

7. The method for creating a multi-platform product search engine according to claim 1, characterized in that: The optional functions of the current query analysis include a synonym expansion function, a stop word omission function, a spelling correction function and a word weight analysis function.

8. A multi-platform commodity search engine creation system, characterized in that: The system comprises: A display category table creation module, used to create a display category table based on the commodity categories of multiple e-commerce platforms connected to the basic search engine; wherein the display category table includes a display category ID and a category name corresponding to the display category ID; A relationship binding module, used for binding the platform commodity category IDs of commodity categories of multiple e-commerce platforms with the display category IDs in a many-to-many relationship based on the category names of the commodity categories, to obtain a binding display category table; An index data creation module, used to construct an index data structure table for the basic search engine based on the data information of each product name in the binding display classification table on each e-commerce platform; An instance configuration module, used to create a search service instance according to the index data structure table, and configure the search service instance to the basic search engine to obtain a primary search engine; The function creation module is used to set a search ranking function, a category prediction function and a current query analysis optional function for the primary search engine to obtain a multi-platform commodity search engine.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps in the multi-platform product search engine creation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing at least one instruction, characterized in that: When the at least one instruction is executed by a processor in an electronic device, the multi-platform commodity search engine creation method as described in any one of claims 1 to 7 is implemented.