A multi-platform commodity publishing method and system based on field mapping

Abstract

The present application relates to product data management technical field, specifically to a kind of multi-platform commodity publishing method and system based on field mapping, comprising the following steps: obtaining commodity attribute construction source commodity directed acyclic graph, target rule is generated target constraint topological graph and template, two graphs are matched to generate primary mapping graph, complete mapping graph is generated by template splicing text, encapsulation generates multi-platform commodity publishing data set.In the present application, by constructing source commodity directed acyclic graph, constraint topological graph and subgraph template are generated according to target field rule, and then graph topological distance matching, node screening and text dynamic splicing are used to realize the automatic accurate adaptation of complex hierarchical commodity attribute and multi-platform differentiated publishing rule, which plays a role in getting rid of artificial tedious intervention and greatly reducing cross-platform commodity circulation error rate, solves the pain point of non-uniform field format in multi-channel sales, and improves multi-platform commodity batch listing publishing efficiency and data structure consistency.

Description

Technical Field

[0001] This invention relates to the field of product data management technology, and in particular to a multi-platform product publishing method and system based on field mapping. Background Technology

[0002] Product data management technology primarily involves a system for the centralized integration and control of various data and related processes throughout the product lifecycle. Its core content covers the centralized storage and daily maintenance of product structure information, configuration specifications, image standards, and related attribute statuses. Its main purpose is to establish a unified data standard and specification system, ensuring the consistency of information flow across various business processes within an enterprise. It is widely used in industries such as e-commerce, software development, and retail that require handling large-scale product information. Among these, the multi-platform product publishing method based on field mapping refers to an operational process for synchronizing and listing product information across different online trading platforms. Traditional multi-platform product publishing methods involve merchants manually entering basic information, descriptions, and specifications of the same product according to the requirements of each online platform, or using basic batch import tools for format conversion before uploading. This is typically used to address the business need for establishing and displaying product information on different platforms during multi-channel sales. Existing technologies usually use application programming interface (API) connections, web data scraping scripts, or basic spreadsheet conversion tools to achieve the management purpose of distributing source product data to various target online platforms.

[0003] In the traditional multi-platform product publishing process, although basic batch import tools or spreadsheet conversions are used to complete the establishment of product profiles and distribution of basic data across channels, there are still technical problems when dealing with the synchronization of large-scale product information with complex hierarchical structures and huge differences in rules between multiple platforms. These problems include low efficiency in data format conversion, high mapping error rate caused by manual input, and difficulty in dealing with complex attribute splicing logic. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a multi-platform product publishing method based on field mapping.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a multi-platform product publishing method based on field mapping, comprising the following steps: S1: Obtain product attribute entry data, and perform graph structure construction based on the attribute name text, attribute content text, and parent-level attribute name text in the product attribute entry data to generate a directed acyclic graph of the source product; S2: Obtain the target field rule requirement data, perform a topology transformation process based on the target field name text, required identifier, and hierarchical code in the target field rule requirement data, generate a target constraint topology diagram, and generate a target subgraph template based on the sub-attribute name text and text concatenation order number in the target field rule requirement data, and store the target subgraph template in the target constraint topology diagram; S3: Based on the source product directed acyclic graph and the target constraint topology graph, perform graph topology distance matching. When a match is found, write the attribute content text into the target constraint topology graph to generate a primary mapping target topology graph. S4: Based on the source product directed acyclic graph and the target subgraph template in the primary mapping target topology graph, perform target source node filtering and extract the attribute content text, perform a splicing and combination process on the attribute content text according to the text splicing order number to obtain the spliced ​​total string, write the spliced ​​total string into the primary mapping target topology graph to generate a complete mapping target topology graph; S5: Based on the required identifier and the target field name text, attribute content text and concatenated string in the complete mapping target topology map, establish key-value correspondence and perform network protocol encapsulation to generate a multi-platform product release data set.

[0006] As a further embodiment of the present invention, the source product directed acyclic graph includes a set of entity nodes, a set of directed edges, and a source graph topology layer; the target constraint topology includes target placeholder nodes, constraint status flags, and dependency hierarchy chains; the target subgraph template includes a local aggregation framework, a sorting mapping sequence, and sub-element slots; the primary mapping target topology includes single-point assignment nodes, basic associated edges, and a first-round mapping area; the complete mapping target topology includes composite reconstruction nodes, merged data blocks, and a global filling network; and the multi-platform product publishing data set includes a protocol message header, a key-value payload, and a communication verification code.

[0007] As a further aspect of the present invention, step S1 specifically comprises: S111: Obtain product attribute entry data from multiple platforms, parse the product attribute entry data line by line to extract the attribute name text and attribute content text within the product attribute entry data, establish corresponding combination nodes for attribute name text and attribute content text and set them as source nodes, calculate the total number of source nodes and summarize the source node distribution frequency, and generate source node distribution frequency data. S112: Based on the source node distribution frequency data, extract the upper-level attribute name text from the product attribute entry data, find the parent node corresponding to the upper-level attribute name text in the source node, establish a connection line between the parent node and the source node and set it as the first directed edge, calculate the connection span and connection depth parameters of the first directed edge, and generate the hierarchical association directed edge span. S113: Based on the stride of the hierarchical associated directed edge, extract the source node and the first directed edge element parameters, merge the source node and the first directed edge parameters to construct a basic topology, count the number of levels and node depth in the topology, arrange the source node and the first directed edge combination order according to the number of levels, execute the graph structure construction, and establish a directed acyclic graph of the source product.

[0008] As a further aspect of the present invention, step S2 specifically comprises: S211: Obtain the target field rule requirement data in the same time dimension, parse the target field rule requirement data and extract the target field name text, required identifier and hierarchical code in the target field rule requirement data, convert the target field name text into target nodes, summarize the various parameters of the target nodes and calculate the number of target nodes, and generate the target node count value; S212: Based on the number of target nodes, establish a second directed edge between target nodes according to the inheritance relationship before and after hierarchical encoding, extract the sub-attribute name text and text concatenation order number in the target field rule requirement data, construct a basic subgraph structure based on the sub-attribute name text and combine it with the text concatenation order number to reorganize the structural elements and generate a target subgraph template. S213: Extract the target subgraph template, target node, and second directed edge; splice the target subgraph template, target node, and second directed edge to construct the overall network structure; store the overall network structure in the network medium; calculate the network medium storage space occupancy ratio and perform topology transformation to obtain the target constraint topology graph.

[0009] As a further aspect of the present invention, step S3 specifically comprises: S311: Based on the source product directed acyclic graph and the target constraint topology graph, select the highest-level source node in the source product directed acyclic graph as the starting source node, select the highest-level target node in the target constraint topology graph as the starting target node, compare the attribute name text in the starting source node with the target field name text in the starting target node character by character, and generate the node character comparison matching degree. S312: Based on the matching degree of the node characters, when it is determined that the comparison is consistent, the attribute content text in the starting source node is written into the starting target node, the source node is traversed one level down along the first directed edge, the number of the first topological edges between the source node and the starting source node is calculated and set as the first topological distance value, and the first topological distance evaluation value is obtained. S313: For the first topological distance evaluation value, obtain the second topological distance value between the target node and the starting target node, compare the first topological distance value with the second topological distance value, and when the two are equal and the parent node has completed the text writing, write the attribute content text into the target node to complete the graph topological matching and generate the primary mapping target topological graph.

[0010] As a further aspect of the present invention, step S4 specifically comprises: S411: Collect composite field concatenation processing parameters for product publishing on multiple platforms, extract the target subgraph template within the primary mapping target topology graph, count the total number of nodes and node connection pointers within the target subgraph template, sort out the relationship network of the total number of nodes and node connection pointers, calculate the distribution ratio of the total number of nodes, and generate the node distribution ratio of the target subgraph template. S412: Based on the distribution ratio of the target subgraph template nodes, according to the total number of nodes and the node connection direction, perform target source node filtering and extract the target source node set in the directed acyclic graph of the source product, obtain the attribute content text in the target source node set and count the byte length of the attribute content text, and obtain the byte length of the source node text. S413: Based on the length of the source node text bytes, perform an ascending sort operation on the attribute content text in the target source node set according to the text concatenation order number, perform a concatenation and combination process of concatenating the first and last ends of the sorted attribute content text to obtain the concatenated total string, write the concatenated total string into the target node of the primary mapping target topology graph, and establish a complete mapping target topology graph.

[0011] As a further aspect of the present invention, step S5 specifically comprises: S511: Obtain cross-platform data transmission process elements, extract target nodes in the complete mapping target topology map, extract target field name text and attribute content text, as well as concatenated total string and required identifier in the target node, establish key-value correspondence between target field name text, attribute content text and concatenated total string, and generate field key-value correspondence ratio value. S512: For the key-value corresponding ratio of the field, check the status of missing data items in the key-value correspondence relationship in combination with the required identifier, summarize the number of missing data items and calculate the ratio of each missing item, assemble the checked and corrected key-value correspondence relationship according to the communication protocol format, perform network protocol encapsulation operation to construct the basic underlying data transmission structure, and obtain the network protocol transmission data packet volume. S513: Based on the data packet volume transmitted according to the network protocol, obtain the basic parameter indicators of the data packet to be sent, open the network communication interface to allocate communication resources, send the data packet to be sent to the target receiving end through the network communication interface, verify the data receiving status of the target receiving end, summarize the data of various product release information records, and establish a multi-platform product release data set.

[0012] A multi-platform product publishing system based on field mapping includes: The source product graph construction module obtains product attribute entry data, and performs graph structure construction based on the attribute name text, attribute content text, and parent-level attribute name text in the product attribute entry data to generate a source product directed acyclic graph; The rule topology transformation module obtains the target field rule requirement data, performs a topology transformation process based on the target field name text, required identifier, and hierarchical code in the target field rule requirement data, generates a target constraint topology diagram, and generates a target subgraph template based on the sub-attribute name text and text concatenation order number in the target field rule requirement data, and stores the target subgraph template in the target constraint topology diagram; The topology distance matching module performs graph topology distance matching based on the source product directed acyclic graph and the target constraint topology graph. When a match is found, the attribute content text is written into the target constraint topology graph to generate a primary mapping target topology graph. The subgraph isomorphic splicing module performs target source node filtering and extracts the attribute content text based on the target subgraph template in the source product directed acyclic graph and the primary mapping target topology graph. It then performs a splicing and combination process on the attribute content text according to the text splicing order number to obtain the spliced ​​total string. Finally, it writes the spliced ​​total string into the primary mapping target topology graph to generate a complete mapping target topology graph. The product data publishing module establishes a key-value correspondence and performs network protocol encapsulation based on the required identifier and the target field name text, attribute content text, and concatenated string in the complete mapping target topology map, generating a multi-platform product publishing data set.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, a directed acyclic graph of source products is constructed, and a constraint topology graph and subgraph template are generated based on the target field rules. Then, by using graph topology distance matching, node filtering, and dynamic text splicing techniques, the invention achieves automated and accurate adaptation of complex hierarchical product attributes and differentiated publishing rules across multiple platforms. This eliminates tedious manual intervention, significantly reduces the error rate of cross-platform product circulation, effectively solves the pain point of inconsistent field formats when selling through multiple channels, and significantly improves the publishing efficiency and data structure consistency of batch synchronous product listing across multiple platforms. Attached Figure Description

[0014] Figure 1 This is a flowchart of the main steps of the present invention; Figure 2 This is a flowchart of steps S1 of the present invention; Figure 3 This is a flowchart of steps S2 of the present invention; Figure 4 This is a flowchart of steps S3 of the present invention; Figure 5 This is a flowchart of step S4 of the present invention; Figure 6 This is a flowchart of step S5 of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0016] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0017] Please see Figure 1 A multi-platform product publishing method based on field mapping includes the following steps: S1: Obtain product attribute entry data, and perform graph structure construction based on the attribute name text, attribute content text, and parent attribute name text in the product attribute entry data to generate a directed acyclic graph of the source product; S2: Obtain the target field rule requirement data, perform a topology transformation process based on the target field name text, required identifier, and hierarchical code in the target field rule requirement data, generate the target constraint topology diagram, and generate the target subgraph template based on the sub-attribute name text and text concatenation order number in the target field rule requirement data, and store the target subgraph template in the target constraint topology diagram; S3: Based on the source product directed acyclic graph and the target constraint topology graph, perform graph topology distance matching. When a match is found, write the attribute content text into the target constraint topology graph to generate a primary mapping target topology graph. S4: Based on the target subgraph templates in the source product directed acyclic graph and the primary mapping target topology graph, perform target source node filtering and extract attribute content text. Perform a splicing and combination process on the attribute content text according to the text splicing order number to obtain the spliced ​​total string. Write the spliced ​​total string into the primary mapping target topology graph to generate the complete mapping target topology graph. S5: Based on the required identifier and the target field name text, attribute content text, and concatenated string in the complete mapping target topology diagram, establish key-value correspondence and perform network protocol encapsulation to generate a multi-platform product publishing data set.

[0018] The source product directed acyclic graph includes a set of entity nodes, a set of directed edges, and a source graph topology layer; the target constraint topology includes target placeholder nodes, constraint status flags, and dependency hierarchy chains; the target subgraph template includes a local aggregation framework, a sorting mapping sequence, and sub-element slots; the primary mapping target topology includes single-point assignment nodes, basic associated edges, and the first-round mapping area; the complete mapping target topology includes composite reconstruction nodes, merged data blocks, and a global filling network; and the multi-platform product publishing data set includes protocol message headers, key-value payloads, and communication verification codes.

[0019] Please see Figure 2 Step S1 is as follows: S111: Obtain product attribute entry data from multiple platforms, parse the product attribute entry data line by line to extract the attribute name text and attribute content text within the product attribute entry data, establish corresponding combination nodes for attribute name text and attribute content text and set them as source nodes, calculate the total number of source nodes and summarize the source node distribution frequency, and generate source node distribution frequency data. A distributed data crawling engine retrieved a total of 15,000 product listing attribute entries from three external e-commerce platform interfaces. This massive amount of basic data is the core driving force behind the multi-platform product listing flow. The data format is JSON string. The JSON string is parsed line by line, separating key-value pairs and extracting the attribute name text on the left (e.g., size) and the attribute content text on the right (e.g., XL). The extracted attribute name text and corresponding attribute content text are bound and combined, and then instantiated into independent data structure nodes in the memory of the product data management system, which are then set as source nodes. During this process, a node frequency threshold of 5 occurrences is set to filter out redundant nodes that appear less than 5 times. The instantiated source node list is traversed one by one, and an incrementing variable operator is used to perform an accumulation count operation to calculate the total number of source nodes. After a complete traversal and calculation, the final number of valid source nodes is 12,450. Subsequently, the attribute name text within each source node is extracted, and a frequency variable is initialized in the hash map table. The attribute name text with the same name is used as the hash key for lookup; for each match, the corresponding frequency variable value is incremented by 1, thus summarizing the distribution frequency of each source node. For example, summarizing the occurrences of source nodes with the attribute name "color" results in a frequency variable that stops at 4500 times, indicating a distribution frequency of 4500 times for that source node. All key-value pairs are then sorted in descending order according to the value of the frequency variable corresponding to each hash key, and the hash structure is serialized and stored to generate the source node distribution frequency data. Based on the auto-incrementing and descending sorting operations of the hash table, extremely low-frequency long-tail attributes are effectively eliminated, improving the processing efficiency of subsequent topology-level extraction. This lays the structural foundation for master data cleaning and deduplication in product data management, ensuring the consistency and purity of upstream data sources when publishing products across multiple platforms.

[0020] S112: Based on the frequency distribution data of the source node, extract the text of the upper-level attribute name in the product attribute entry data, find the parent node corresponding to the upper-level attribute name text in the source node, establish a connection line between the parent node and the source node and set it as the first directed edge, calculate the connection span and connection depth parameters of the first directed edge, and generate the hierarchical association directed edge span. Read the frequency distribution data of the 12,450 source nodes generated above. For high-frequency source nodes with a frequency value greater than 100, extract the attribute name text of the next higher level from the product attribute entry data. For example, extract the next higher level category, men's clothing, from the subcategory "shirts". Traverse the memory address space of the source node and perform string comparison operations to find a node with the exact same attribute name text as the next higher level. Mark the located node as the parent node. Establish a unidirectional data link between the parent node and the source node by allocating a memory address pointer to the parent node, and set this link as the first directed edge. Extract the level index values ​​of the nodes connected to both ends of the first directed edge. Subtract the level index value of the parent node from the level index value of the current source node to obtain the level span difference, and set this difference as the connection span of the first directed edge. For example, if the level index value of the source node is 4 and the level index value of the parent node is 2, perform a subtraction operation between 4 and 2 to calculate the connection span as 2. The logical operation is 4-2=2. Similarly, the total number of unidirectional data links from the top-level root node to the current source node in the entire link is extracted and set as the connection depth parameter. For example, if the total number of unidirectional links from the highest level to the current node is 3, then the depth parameter is 3. A string concatenation function is called to merge the calculated connection span and connection depth parameter in an underscore-separated format. For example, 2 and 3 are concatenated as 2 underscore 3, generating the directed edge span of the hierarchical association. The specific span value is calculated using the subtraction difference of the hierarchical index, quantifying the hierarchical distance between attributes. This allows for precise measurement of the vertical association of the tree structure. This provides a quantitative coordinate space for dimensionality reduction or dimensionality upgrade alignment at the product data management level, addressing the mapping challenges arising from differences in the category tree hierarchy across multiple platforms when publishing products.

[0021] S113: Based on the stride of the directed edge in the hierarchical association, extract the element parameters of the source node and the first directed edge, merge the parameters of the source node and the first directed edge to construct the basic topology, count the number of levels and node depth in the topology, arrange the combination order of the source node and the first directed edge according to the number of levels, execute the graph structure construction, and establish the directed acyclic graph of the source product. Retrieve the aforementioned hierarchical directed edge span data containing the underscore "3". Extract the source node memory address sequence associated with this data and the connection pointer parameter of the first directed edge. In the working memory area, encapsulate the source node memory address sequence and the connection pointer parameter of the first directed edge into the same topology data structure object, merging isolated nodes and edges to construct a basic topology. Perform a breadth-first search traversal operation on this basic topology, counting the maximum number of unidirectional links required to reach any leaf node from the root node within the entire network, and set this as the number of levels. For example, if a maximum of 5 links are required, the number of levels is set to 5. Simultaneously, record the specific level value of each topology data structure object, extract this value, and set it as the node depth index. Obtain the parameter value of the set level number of 5, and based on the node depth index of each object, combine the source node and the first directed edge in ascending order of value from 1 to 5 and store them in the construction execution queue. The graph structure is constructed according to the execution queue. Node entities and connecting edges are sequentially established in the graph storage database. The generation of closed loops is constrained by pointers to unidirectional links, thus establishing a directed acyclic graph (DAG) for the source products. This process employs a strict ascending queuing mechanism determined by the maximum number of links, preventing reverse referencing from the bottom-level node creation order and ensuring the topological stability of the generated DAG. This data is then used as the foundational material graph knowledge base for the product data management center. This allows the graph network matching process for multi-platform product releases to readily access and utilize a stable, acyclic, standardized source-end data dictionary, significantly enhancing data reuse.

[0022] Please see Figure 3 Step S2 is as follows: S211: Obtain the target field rule requirement data in the same time dimension, parse the target field rule requirement data and extract the target field name text, required identifier and hierarchical code in the target field rule requirement data, convert the target field name text into target nodes, summarize the various parameters of the target nodes and calculate the number of target nodes, and generate the target node count value; The timestamp collection offset parameter is set to 0 to obtain the product target field rule requirement data from three external e-commerce platforms under the same time dimension. This requirement data uses Extensible Markup Language (XML) format. The Document Object Model (DOM) parser is called to read this rule requirement data, identifying all angle bracket pairs and extracting the target field name text carried within the tags, such as the main image link text. Simultaneously, the required field status flags associated with these flags are extracted and used as required field identifiers (e.g., the number 1 represents mandatory field), and the hierarchical code representing the tree structure level is extracted (e.g., a code value of 102). A new structure storage space is allocated in the data processing cache, and the extracted target field name text is stored in this structure, transforming it into a target node. All instantiated target nodes in the current cache are traversed, summing the string byte space occupied by each target node and the Boolean values ​​of the required field status flags. An accumulator variable is initialized in memory with an initial value of 0. Each time a target node is accessed during traversal, the current value of the accumulator variable is summed by adding a constant 1. For example, the collector reads 850 rule data fragments that meet the criteria. After complete traversal and summation, the accumulator counter variable finally stops at 850, and this final value generates the target node count value of 850. This operation logic ensures the accurate extraction of the target node size by locking the same time dimension and performing object-by-object summation, preventing data size errors caused by dynamic updates of platform rules. It establishes a real-time and accurate target-targeting mechanism for rule parsing of multi-platform product releases, significantly enhancing the automatic detection and dynamic perception capabilities of the entire product data management system in response to changes in external protocols.

[0023] S212: Based on the number of target nodes, establish a second directed edge between target nodes according to the inheritance relationship before and after hierarchical coding, extract the sub-attribute name text and text concatenation order number in the target field rule requirements data, construct the basic subgraph structure based on the sub-attribute name text and combine the text concatenation order number to reorganize the structural elements and generate the target subgraph template. For the 850 target node counts obtained through addition, the hierarchical codes attached to the attribute fields of each target node are extracted. Based on the numerical value of the hierarchical codes, a subtraction difference operation is performed on potentially related node pairs. When the difference in the hierarchical codes of two target nodes is strictly equal to the constant 1, a hierarchical inheritance relationship is determined between them. A directed pointer containing memory address information is declared in the processing cache to connect the two target nodes, and this pointer is set as the second directed edge. The process continues to read the sub-attribute name text attached to specific target nodes within the target field rule requirements data, such as the product subtitle text, and extracts the integer text concatenation sequence number attached to this text segment, such as the sequence number 2. A basic subgraph structure with basic connections is established in the memory working area using the aforementioned sub-attribute name text as structural vertices. For this basic subgraph structure, the text concatenation sequence number corresponding to each node is extracted, and the array sorting interface is called to sort them in ascending order from smallest to largest. For example, if the node sequence numbers are 3, 1, and 2, a comparison operation and address swapping operation are performed on them, reorganizing them into an ascending order structure of 1, 2, and 3. Based on the reorganized basic subgraph structure, the relative memory position information of each vertex is fixed, and the target subgraph template file is generated and exported. This operation logic uses the precise condition of the hierarchical encoding difference equal to 1 to realize the automatic deduction of the relationship between nodes, and combines the ascending numerical sorting to ensure the absolute order of the subgraph splicing logic. Thus, in the product data management platform, the scattered constraints of various business platforms are successfully integrated into a structured skeleton that can be reused across businesses. This is a key pre-emptive measure to solve the problem of disordered order of composite long-tail fields in the process of multi-platform product release.

[0024] S213: Extract the target subgraph template, target node, and second directed edge; splice the target subgraph template, target node, and second directed edge to construct the overall network structure; store the overall network structure in the network medium; calculate the storage space occupancy ratio of the network medium and perform topology transformation to obtain the target constraint topology graph. The exported target subgraph template file is loaded into the CPU's high-speed cache, and 850 instantiated target nodes and all declared second-directed edge pointer variables are extracted simultaneously. Based on the physical address mapping information stored in the pointer variables in the cache, the head node and corresponding target nodes of the target subgraph template are spliced ​​together at the memory level via the second-directed edge pointers, constructing an overall network structure containing 1500 discrete nodes. A serialization function is called to convert this overall network structure into a continuous binary data stream and store it in a non-volatile network storage medium. The total available capacity of the network medium is obtained by calling the system disk status monitoring interface; assuming the total capacity of the medium is 1000 megabytes. Simultaneously, the actual byte size occupied by the aforementioned binary stream file is extracted, for example, its actual occupied space is 20 megabytes. The actual occupied byte size is divided by the total available capacity of the medium to obtain the occupancy quotient, which is then multiplied by the constant 100. For example, dividing 20 by 1000 yields 0.02, and multiplying 0.02 by 100 gives a usage value of 2, indicating a network media storage space usage rate of 2%. A safe conversion rate threshold of 80% is set. When the calculated 2% is less than 80% (logic condition 2% < 80%), the memory topology converter is activated. This deserializes the binary data stream, reconstructing it into a set of memory graph objects, successfully obtaining the target constraint topology. This operational logic, through quotient division of byte capacity and threshold determination, proactively intercepts out-of-specification graph data, avoiding the risk of program crashes due to memory overflow during deserialization. This significantly enhances the continuous high availability of the underlying architecture for product data management, enabling stable operation of multi-platform product publishing microservice applications based on this complex constraint topology, unaffected by peak traffic surges during promotional periods.

[0025] Please see Figure 4 Step S3 is as follows: S311: Based on the source product directed acyclic graph and the target constraint topology graph, select the highest-level source node in the source product directed acyclic graph as the starting source node, and select the highest-level target node in the target constraint topology graph as the starting target node. Compare the attribute name text in the starting source node with the target field name text in the starting target node character by character to generate the node character comparison matching degree. The process involves retrieving the source product directed acyclic graph (DAG) containing 12450 nodes, traversing and scanning the in-degree edge count of all node objects within the graph, filtering out the highest-level source node whose in-degree edge count is strictly equal to 0, extracting its pointer address, and setting it as the starting source node. Simultaneously, the process retrieves the target constraint topology graph generated by the previous reconstruction, using the same in-degree edge count of 0 as above, extracting the highest-level target node, and setting it as the starting target node. The process reads the attribute name text string stored in the memory space of the starting source node, such as the product origin text. Simultaneously, it reads the target field name text string inside the starting target node, such as the manufacturing province text. A character slicing operation is called to split the two text strings into single-character arrays. A loop is used to compare the characters at the same index position in both arrays to see if they are completely identical, and a matching counter is initialized to accumulate the number of identical characters. The number of identical characters obtained from the matching counter is divided by the larger of the two compared strings to obtain the matching quotient. The matching quotient is then multiplied by a constant 100 to obtain the final percentage value. For example, if the starting source node has 2 characters and the starting target node has 4 characters, and 2 of these characters match, dividing 2 by 4 gives 0.5, and multiplying 0.5 by 100 gives 50, resulting in a 50% matching degree for the generated node characters. By traversing and comparing individual characters and calculating the product of division and multiplication, a specific numerical value for matching is obtained. This directly transforms the abstract semantic correlation into a precise numerical measurement standard, greatly reducing the operational cost of manual verification of mapping dictionaries by business personnel within the product data management system. Furthermore, it establishes the first objective and quantitative machine intelligence review defense line in the automated matching process of multi-platform product releases.

[0026] Table 1. Example of character matching degree

[0027] As shown in Table 1, this table presents the statistical results of matching characters and the final distribution of the quantified matching values ​​under different source and target text comparison scenarios through specific text combination examples.

[0028] S312: Based on the matching degree of the node characters, when it is determined that the comparison is consistent, write the attribute content text in the starting source node into the starting target node, traverse the source node one level down along the first directed edge, calculate the number of the first topological edges between the source node and the starting source node and set it as the first topological distance value to obtain the first topological distance evaluation value. The node matching threshold parameter is set to 80%. The calculated character matching scores for each node are extracted, and a comparison operation is performed between this matching score and the threshold parameter. When the matching score is greater than or equal to 80%, the condition ≥80% is met, indicating a complete match. A write protection release instruction is then initiated, and the attribute content text temporarily stored in the address space of the starting source node, such as the Huawei brand logo, is directly written to the null field in the starting target node via a memory overwrite operation. This ensures that the core material data output by the product data management center can flow to the business application without human intervention. After writing, the pointer to the first directed edge, which carries the hierarchical span parameter, is traversed to the next level down in the tree structure to find the next adjacent source node. During the traversal, a topology step accumulator variable is declared to increment the number of entity connections traversed by the traversal pointer. The total number of the first directed edges traversed between the current source node and the initial starting source node is set as the first topology edge number. The specific integer value of the first topological edge is obtained and directly assigned as the first topological distance value. For example, when traversing downwards from the starting source node, a total of 3 connecting edges are passed before reaching the current stopping source node, the value of the first topological edge is 3. Based on this, the integer variable is directly passed to obtain the first topological distance evaluation value of 3. This process performs size comparison and threshold control, intercepting the write operation of low-matching data. At the same time, combined with the step accumulator, it accurately quantifies the physical traversal span between two points within the network, thus enabling multi-platform product publishing tasks to have a clear hierarchical cursor guidance mechanism when performing vertical category exploration mapping.

[0029] S313: For the first topological distance evaluation value, obtain the second topological distance value between the target node and the starting target node, compare the first topological distance value with the second topological distance value, and when the two are equal and the parent node has completed the text writing, write the attribute content text into the target node to complete the graph topological matching and generate the primary mapping target topological graph. Extract the first topological distance evaluation value parameter, which is 3, obtained earlier. Activate the same breadth-first traversal accumulator module in the target constraint topology graph to count the actual number of connections traversed between the current target node and its corresponding starting target node. Obtain the integer value of this connection count and set it as the second topological distance value, assuming that the number of connections measured after breadth-first traversal is also 3. Call the logical comparison operator to perform an equality check operation on the first and second topological distance values. Simultaneously, retrieve the memory status flag of the current target node's corresponding parent node to check if there is a Boolean true status code representing text writing completion. When it is confirmed that the first topological distance value equals the second, and the parent node's memory status flag shows a true value representing completion, release the write lock resource of the target node. Extract the attribute content text carried inside the current source node, such as the red text, and write it completely into the value data segment of the current target node. This completes the graph topology matching of a single node in two independent graph structures. The program iteratively calls the aforementioned equivalence verification and write instructions to process all nodes in the entire graph until all target nodes that meet the equidistant condition have been filled with their corresponding text data. Finally, a primary mapping target topology graph is generated in the graph database. This process performs dual checks—equivalence verification and parent node write status Boolean value verification—effectively preventing erroneous mounting of cross-level structural data and ensuring the consistency of the mapping structure. This signifies that the independent metadata in product data management has achieved deep integration and decoupling of graph structure with the customized field specifications for multi-platform product publishing, creating a closed-loop mapping.

[0030] Please see Figure 5 Step S4 is as follows: S411: Collect composite field concatenation processing parameters for product publishing on multiple platforms, extract target subgraph templates within the primary mapping target topology graph, count the total number of nodes and node connection pointers within the target subgraph templates, sort out the relationship network of the total number of nodes and node connection pointers, calculate the distribution ratio of the total number of nodes, and generate the node distribution ratio of the target subgraph templates. Establish a secure communication application programming interface channel to collect product posting composite field concatenation processing parameters containing data requirements from at least three external e-commerce platforms. Extract the string delimiter types and total character maximum length limit configuration items covered in these parameters. Enter the persistent storage medium area and extract the primary mapping target topology graph data that has completed the write operation. Separate the target subgraph template file object from it using the unique node identification code. Call the node traversal and statistics function for this template file object, and accumulate the valid nodes contained within it to obtain the total number of nodes. Assume that the total number of nodes obtained after traversal is 15. Then, traverse and read the set of edge vectors connected to each node in the template, sort out the start and end states of directed edges, and mark the node where the start of the edge vector is located as the parent node with output connection pointing. By sorting out and constructing the structural network of the total number of nodes and the node connection pointing relationship, extract the value of the parent node marked with output connection pointing in this network. Assume that this value is 3. Dividing the value 3 by the total number of nodes, parameter 15, yields an initial quotient of 0.2. This 0.2 is then multiplied by the constant 100 to calculate the final ratio of 20%. This calculated 20% is set and generated as the distribution ratio of the target subgraph template nodes. In this step, the system uses digital means to accurately characterize the template features in the multi-platform product publishing component. This process performs division and multiplication operations between the number of child and total levels, extracting distribution ratio parameters that represent the characteristics of local graph blocks. This provides an intuitive digital comparison basis for subsequent structural similarity retrieval, thus providing an objective standard for dynamic retrieval and evaluation of subnetworks under the high cohesion requirements of product data management.

[0031] S412: Based on the distribution ratio of the target subgraph template nodes, perform target source node filtering and extract the target source node set in the directed acyclic graph of the source product according to the total number of nodes and the node connection direction. Obtain the attribute content text in the target source node set and count the byte length of the attribute content text. Obtain the byte length of the source node text. Set the ratio comparison tolerance range parameter to 5%, and read the previously calculated 20% target subgraph template node distribution ratio value. Set a dynamically sliding region search window within the source product directed acyclic graph with 12450 valid nodes. Within each local graph area covered by the sliding window, count the total number of nodes in that area and the number of parent nodes with output connections, performing the same division and multiplication operations to obtain the node distribution ratio of that local graph area. When a calculated local ratio for a region falls within the tolerance range of 15% to 25%, i.e., the ratio value P satisfies 15%-25%, trigger the target source node filtering and extraction command. Copy all source node entities contained within this range and store them in a specific memory list, forming a target source node set containing 20 independent nodes. Iterate through each element in this target source node set, using the access interface to read the internal buffered attribute content text strings, such as "pure cotton material." This sliding window exploration mode further reduces the redundant search range of materials published across multiple platforms. The system calls the operating system's standard character encoding length calculation function to count the total number of bytes occupied by each extracted attribute text in memory. For example, when it is determined that the above four Chinese characters occupy 12 bytes of space under the specified encoding format, this value is recorded. This process is repeated to obtain the source node text byte length for each node in the list. This process performs a ratio comparison with a tolerance range, significantly reducing the data volume retrieved from the target node with high efficiency. This significantly reduces the system resource overhead of subsequent text concatenation, while freeing up processing space for character storage limit control and refined memory fragmentation management at the product data management level.

[0032] S413: Based on the length of the source node text bytes, perform an ascending sort operation on the attribute content text in the target source node set according to the text concatenation order number, perform a concatenation and combination process of concatenating the first and last attributes of the sorted attribute content text to obtain the concatenated total string, write the concatenated total string into the target node of the primary mapping target topology graph, and establish a complete mapping target topology graph. The system reads the target source node set containing 20 nodes and extracts the pre-defined text concatenation order numbers from the attribute data segments of each node. For example, the numbers contained in the three core nodes are 5, 2, and 8. It then calls the system's built-in quicksort algorithm, using the text concatenation order numbers attached to each node as the sorting weight, to perform an ascending sort operation on the attribute content text attached to the target source node set and its corresponding source node text byte length. For example, based on the numerical value, the node with the order number 2 and its data are swapped to the first position in memory, the node with the order number 5 is moved to the second position, and the node with the number 8 is placed at the last position. For each attribute content text after the physical address swapping, the string concatenation component is called to perform a concatenation process according to the order from the first to the last position in memory. For example, combining the initial text "slim fit," the middle text "pure cotton," and the final text "breathable" using an additive merging logic yields a complete string of concatenated text: "slim fit, pure cotton, breathable." This concatenation method essentially aligns with the strict validation rules for complex modifier combinations in multi-platform product publishing. Subsequently, this concatenated string, along with the previously obtained overall byte length, is written into the target node content data segment corresponding to the specific business logic position in the primary mapping target topology graph. Simultaneously, the node update status codes recorded within the graph database are refreshed, establishing a complete mapping target topology graph. This process performs numerical benchmark comparison and array position swapping operations, combined with string merging, ensuring the natural semantic order of the composite text synthesis result and guaranteeing that the data arrangement order has strong consistency and fault tolerance characteristics when the product data management module provides the synthesized graph structure downstream.

[0033] Please see Figure 6 The S5 steps are as follows: S511: Obtain cross-platform data transmission process elements, extract target nodes in the complete mapping target topology map, extract target field name text and attribute content text, as well as the concatenated total string and required identifiers in the target nodes, establish key-value correspondence between target field name text, attribute content text and concatenated total string, and generate field key-value correspondence ratio values. Activate the Secure Sockets Layer (SSL) network protocol link to obtain the encryption certificate key and external platform access authorization token elements during cross-platform data transmission. Using these authorization token elements, verify permissions and extract all non-empty target node data objects within the complete mapping target topology diagram with the previously constructed and updated status codes. Use a text extraction tool to extract the target field name text from records within the target nodes, such as product specification description text. Simultaneously extract the associated single-attribute content text or the concatenated string generated after merging operations, such as "slim fit, pure cotton, breathable" text. Extract the required identifier indicating the current field's mandatory entry status, such as a Boolean true / false identifier. In the mapping table of the system data buffer, set the target field name text as the unique lookup key, and set its associated attribute content text and concatenated string as data storage values, establishing a key-value correspondence between the target field name text, the text, and the concatenated string. Count the total number of successfully established key-value correspondence records in the buffer, assuming a count of 800 records. Simultaneously count the total number of target nodes with required identifiers under the original rule constraints, assuming a count of 1000 nodes. Dividing the total number of key-value pairs by the total number of target nodes, and performing a division operation between 800 and 1000 to obtain a quotient of 0.8, the key-value pair ratio is obtained and generated as 0.8. This process performs a division operation between the total number of records and the target base, extracting an intuitive data integrity ratio. This provides a data coverage reference for the final transmission verification and is a core monitoring indicator for the overall product data management efficiency, ensuring that the initiation of multi-platform product release tasks has sufficient data abundance.

[0034] S512: For the proportion of field key-value correspondence, check the status of missing data items in the key-value correspondence relationship in combination with the required identifier, summarize the number of missing data items and calculate the proportion of each missing item, assemble the checked and corrected key-value correspondence relationship according to the communication protocol format, perform network protocol encapsulation operation to build the basic underlying data transmission structure, and obtain the amount of network protocol transmitted data packets. Extract the proportion parameter corresponding to the field key-value pair 0.8 obtained from the aforementioned division operation, and set the transmission warning threshold parameter to 0.95. Perform a logical size comparison; when the parameter value of 0.8 is less than the threshold parameter of 0.95 (0.8 < 0.95), activate the anomaly scanning service. Iterate through the required identifiers bound to each node in the key-value mapping table, comparing and searching for data missing items where the data storage value is blank and the required identifier is true, to intercept fragmented or incomplete data missed in the product data management process. Allocate an error counter variable in memory; each time such a null value state is captured, it increments automatically to summarize the absolute number of such missing data items. Assume that the summarized number of missing items is 50. Divide this 50 missing items by the total number of 800 key-value records established above, and perform division to calculate the data missing item proportion as 0.0625. The formula is: \frac{50}{800} = 0.0625. The update / insert command is invoked in the mapping table to fill in the 50 missing items by writing default placeholder content, such as "No data available," to complete the data correction. The Extensible Markup Language (XML) communication protocol format requirements specification officially released by the target e-commerce platform is retrieved, and the corrected key-value mapping message is assembled according to the tag nesting and header / tail closure order in the specification. The Transmission Control Protocol (TCP) underlying interface is called to perform network protocol encapsulation operations on the above message, serializing it and segmenting it into a standard byte stream format to construct the basic underlying data transmission structure. The data stream counter is called to accumulate the byte stream segment packets to calculate the total number of packets, and the number of network protocol transmission data packets is obtained as 150. This stage of protocol reassembly operation completely eliminates the format compatibility obstacles before the multi-platform product release task is issued.

[0035] Table 2. Statistics on missing data items and examples of repair implementation.

[0036] As shown in Table 2, this table details the results of the actual calculation and deduction examples of performing missing status judgment on different key-value correspondences by combining required identifiers and using placeholders to perform completion correction.

[0037] S513: Based on the data packet volume transmitted according to the network protocol, obtain the basic parameter indicators of the data packet to be sent, open the network communication interface and allocate communication resources, send the data packet to be sent to the target receiving end through the network communication interface, verify the data receiving status of the target receiving end and summarize the data records of various product release information, and establish a multi-platform product release data set. Based on the 150 network protocol data packet volume parameters accumulated by the aforementioned counter, the basic parameter indicators nested in the header of the data packet to be sent, such as the target Internet Protocol routing address and the specified application access port number, are extracted to guide the standard payload generated by the product data management layer to accurately reach each heterogeneous node. The operating system's underlying network socket binding function is called to open the network communication interface according to the specified port number. Based on the size parameter of the network data packet volume, the system's traffic allocation algorithm is invoked, allocating communication resources according to a linear amplification rule of 2 megabytes per second of transmission bandwidth per data packet. For example, multiplying 150 by 2 yields a total allocated transmission bandwidth of 300 megabytes per second, i.e., 150. The queue of the 150 independent data packets to be sent is sequentially transmitted to the target receiving server of the external platform system through the successfully activated and configured network communication interface, following a first-in, first-out queue order. After the transmission is completed, an asynchronous listening proxy service is started to capture the Hypertext Transfer Protocol status response code returned by the target receiving interface. When the specific value of the status response code read is 200, the data reception status of the target receiving end is confirmed and verified as successful. Subsequently, a log parsing script is invoked to extract information such as successfully sent product categories, sending completion timestamps, and total uploaded packets from the transmission logs, summarizing the data for each product release information record. A new data table with an independent auto-incrementing primary key is created in the local relational database, and batch insert statements are executed on the summarized release information records to establish a fully retained multi-platform product release data set. This process performs a product operation of the total number of data packets and the bandwidth of a single packet, achieving precise planning of communication channel resources. Combined with underlying status response code capture and monitoring, it ensures closed-loop archiving of release archive logs, providing a complete source verification link and full lifecycle data tracking guarantee for the collaborative management system of multi-platform product releases.

[0038] A multi-platform product publishing system based on field mapping includes: The source product graph construction module is used to execute S1: obtain product attribute entry data, and perform graph structure construction based on the attribute name text, attribute content text, and parent level attribute name text in the product attribute entry data to generate a source product directed acyclic graph; The rule topology transformation module is used to execute S2: obtain the target field rule requirement data, perform the topology transformation process according to the target field name text, required identifier, and hierarchical code in the target field rule requirement data, generate the target constraint topology diagram, and generate the target subgraph template according to the sub-attribute name text and text concatenation order number in the target field rule requirement data, and store the target subgraph template in the target constraint topology diagram; The topology distance matching module is used to execute S3: based on the source product directed acyclic graph and the target constraint topology graph, perform graph topology distance matching, and when a match is found, write the attribute content text into the target constraint topology graph to generate a primary mapping target topology graph; The subgraph isomorphic splicing module is used to execute S4: Based on the target subgraph template in the source product directed acyclic graph and the primary mapping target topology graph, it performs target source node filtering and extracts attribute content text, performs splicing and combination process on the attribute content text according to the text splicing order number to obtain the spliced ​​total string, writes the spliced ​​total string into the primary mapping target topology graph, and generates the complete mapping target topology graph; The product data publishing module is used to execute S5: based on the required identifier and the target field name text, attribute content text and concatenated string in the complete mapping target topology map, establish key-value correspondence and perform network protocol encapsulation to generate a multi-platform product publishing data set.

[0039] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A multi-platform product publishing method based on field mapping, characterized in that, Includes the following steps: S1: Obtain product attribute entry data, and perform graph structure construction based on the attribute name text, attribute content text, and parent-level attribute name text in the product attribute entry data to generate a directed acyclic graph of the source product; S2: Obtain the target field rule requirement data, perform a topology transformation process based on the target field name text, required identifier, and hierarchical code in the target field rule requirement data, generate a target constraint topology diagram, and generate a target subgraph template based on the sub-attribute name text and text concatenation order number in the target field rule requirement data, and store the target subgraph template in the target constraint topology diagram; S3: Based on the source product directed acyclic graph and the target constraint topology graph, perform graph topology distance matching. When a match is found, write the attribute content text into the target constraint topology graph to generate a primary mapping target topology graph. S4: Based on the source product directed acyclic graph and the target subgraph template in the primary mapping target topology graph, perform target source node filtering and extract the attribute content text, perform a splicing and combination process on the attribute content text according to the text splicing order number to obtain the spliced ​​total string, write the spliced ​​total string into the primary mapping target topology graph to generate a complete mapping target topology graph; S5: Based on the required identifier and the target field name text, attribute content text and concatenated string in the complete mapping target topology diagram, establish key-value correspondence and perform network protocol encapsulation to generate a multi-platform product release data set; The specific steps in S3 are as follows: S311: Based on the source product directed acyclic graph and the target constraint topology graph, select the highest-level source node in the source product directed acyclic graph as the starting source node, select the highest-level target node in the target constraint topology graph as the starting target node, compare the attribute name text in the starting source node with the target field name text in the starting target node character by character, and generate the node character comparison matching degree. S312: Based on the matching degree of the node characters, when it is determined that the comparison is consistent, the attribute content text in the starting source node is written into the starting target node, the source node is traversed one level down along the first directed edge, the number of the first topological edges between the source node and the starting source node is calculated and set as the first topological distance value, and the first topological distance evaluation value is obtained. S313: For the first topological distance evaluation value, obtain the second topological distance value between the target node and the starting target node, compare the first topological distance value with the second topological distance value, and when the two are equal and the parent node has completed the text writing, write the attribute content text into the target node to complete the graph topological matching and generate the primary mapping target topological graph. The specific steps of S4 are as follows: S411: Collect the composite field splicing processing parameters of product release on multiple platforms, extract the target subgraph template in the primary mapping target topology map, count the total number of nodes and node connection points in the target subgraph template, sort out the relationship network of the total number of nodes and node connection points, calculate the distribution ratio of the total number of nodes, and generate the node distribution ratio of the target subgraph template. S412: Based on the distribution ratio of the target subgraph template nodes, according to the total number of nodes and the node connection direction, perform target source node filtering and extract the target source node set in the directed acyclic graph of the source product, obtain the attribute content text in the target source node set and count the byte length of the attribute content text, and obtain the byte length of the source node text. S413: Based on the length of the source node text bytes, perform an ascending sort operation on the attribute content text in the target source node set according to the text concatenation order number, perform a concatenation and combination process of concatenating the first and last ends of the sorted attribute content text to obtain the concatenated total string, write the concatenated total string into the target node of the primary mapping target topology graph, and establish a complete mapping target topology graph.

2. The multi-platform product publishing method based on field mapping according to claim 1, characterized in that: The source product directed acyclic graph includes a set of entity nodes, a set of directed edges, and a source graph topology layer; the target constraint topology includes target placeholder nodes, constraint status flags, and dependency hierarchy chains; the target subgraph template includes a local aggregation framework, a sorting mapping sequence, and sub-element slots; the primary mapping target topology includes single-point assignment nodes, basic associated edges, and a first-round mapping area; the complete mapping target topology includes composite reconstruction nodes, merged data blocks, and a global filling network; the multi-platform product publishing data set includes a protocol message header, a key-value payload, and a communication verification code.

3. The multi-platform product publishing method based on field mapping according to claim 1, characterized in that, The specific steps of S1 are as follows: S111: Obtain product attribute entry data from multiple platforms, parse the product attribute entry data line by line to extract the attribute name text and attribute content text within the product attribute entry data, establish corresponding combination nodes for attribute name text and attribute content text and set them as source nodes, calculate the total number of source nodes and summarize the source node distribution frequency, and generate source node distribution frequency data. S112: Based on the source node distribution frequency data, extract the upper-level attribute name text from the product attribute entry data, find the parent node corresponding to the upper-level attribute name text in the source node, establish a connection line between the parent node and the source node and set it as the first directed edge, calculate the connection span and connection depth parameters of the first directed edge, and generate the hierarchical association directed edge span. S113: Based on the stride of the hierarchical associated directed edge, extract the source node and the first directed edge element parameters, merge the source node and the first directed edge parameters to construct a basic topology, count the number of levels and node depth in the topology, arrange the source node and the first directed edge combination order according to the number of levels, execute the graph structure construction, and establish a directed acyclic graph of the source product.

4. The multi-platform product publishing method based on field mapping according to claim 1, characterized in that, The specific steps of S2 are as follows: S211: Obtain the target field rule requirement data in the same time dimension, parse the target field rule requirement data and extract the target field name text, required identifier and hierarchical code from the target field rule requirement data, convert the target field name text into target nodes, summarize the various parameters of the target nodes and calculate the number of target nodes, and generate the target node count value. S212: Based on the number of target nodes, establish a second directed edge between target nodes according to the hierarchical encoding inheritance relationship, extract the sub-attribute name text and text concatenation order number in the target field rule requirement data, reorganize the structural elements, and generate a target subgraph template. S213: Extract the target subgraph template, target node, and second directed edge; splice the target subgraph template, target node, and second directed edge to construct the overall network structure; store the overall network structure in the network medium; calculate the network medium storage space occupancy ratio and perform topology transformation to obtain the target constraint topology graph.

5. The multi-platform product publishing method based on field mapping according to claim 1, characterized in that, The specific steps of S5 are as follows: S511: Obtain cross-platform data transmission process elements, extract target nodes in the complete mapping target topology map, extract target field name text and attribute content text, as well as the concatenated total string and required identifier in the target node, establish key-value correspondence between target field name text, attribute content text and concatenated total string, and generate field key-value correspondence ratio value. S512: For the key-value corresponding ratio of the field, check the status of missing data items in the key-value correspondence relationship in combination with the required identifier, summarize the number of missing data items and calculate the ratio of each missing item, assemble the checked and corrected key-value correspondence relationship according to the communication protocol format, perform network protocol encapsulation operation to construct the basic underlying data transmission structure, and obtain the network protocol transmission data packet volume. S513: Based on the data packet volume transmitted according to the network protocol, obtain the basic parameter indicators of the data packet to be sent, open the network communication interface to allocate communication resources, send the data packet to be sent to the target receiving end through the network communication interface, verify the data receiving status of the target receiving end, summarize the data of various product release information records, and establish a multi-platform product release data set.

6. The multi-platform product publishing method based on field mapping according to claim 4, characterized in that, During the process of reorganizing structural elements, a basic subgraph structure is constructed based on the text of the sub-attribute names and then reorganized according to the numerical order of the text concatenation.

7. A multi-platform product publishing system based on field mapping, characterized in that, The system is used to implement the method according to any one of claims 1-6, comprising: The source product graph construction module obtains product attribute entry data, and performs graph structure construction based on the attribute name text, attribute content text, and parent-level attribute name text in the product attribute entry data to generate a source product directed acyclic graph. The rule topology transformation module obtains the target field rule requirement data, performs a topology transformation process based on the target field name text, required identifier, and hierarchical code in the target field rule requirement data, generates a target constraint topology diagram, and generates a target subgraph template based on the sub-attribute name text and text concatenation order number in the target field rule requirement data, and stores the target subgraph template in the target constraint topology diagram; The topology distance matching module performs graph topology distance matching based on the source product directed acyclic graph and the target constraint topology graph. When a match is found, the attribute content text is written into the target constraint topology graph to generate a primary mapping target topology graph. The subgraph isomorphic splicing module performs target source node filtering and extracts the attribute content text based on the target subgraph template in the source product directed acyclic graph and the primary mapping target topology graph. It then performs a splicing and combination process on the attribute content text according to the text splicing order number to obtain the spliced ​​total string. Finally, it writes the spliced ​​total string into the primary mapping target topology graph to generate a complete mapping target topology graph. The product data publishing module establishes a key-value correspondence and performs network protocol encapsulation based on the required identifier and the target field name text, attribute content text, and concatenated string in the complete mapping target topology map, generating a multi-platform product publishing data set.

Images