A method, system, and device for intelligent connection of task flow processing nodes.
By using deep learning models and intelligent matching algorithms, efficient and automatic connection of task flow processing nodes is achieved, solving the problems of low efficiency and high error rate under traditional manual operation, and providing an efficient and accurate node link establishment and verification mechanism.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- LU ZE TECH CO LTD
- Filing Date
- 2025-09-23
- Publication Date
- 2026-06-30
AI Technical Summary
In traditional task flow processing, node connection relies on manual operation, lacks adaptive adjustment, resulting in low efficiency, high matching error rate, and imperfect verification mechanism, making it difficult to meet the needs of complex link construction.
A deep learning-based node matching model is used to extract and process node feature vectors. Intelligent matching and connection are performed by combining similarity calculation and operation menu/drag and drop methods. Link verification is performed by combining depth-first search and a verification rule base.
It improves the matching accuracy of processing nodes, reduces the human error rate, shortens connection time, improves connection efficiency, and provides intuitive user interaction and error reporting to adapt to diverse marketing needs.
Smart Images
Figure CN121256385B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data analysis, and in particular to a method, system and device for intelligent connection of task flow processing nodes. Background Technology
[0002] In current task flow processing, multiple processing nodes are typically involved (for example, in bulk marketing campaigns, the marketing process often involves multiple nodes, such as audience acquisition, computation, and bulk SMS sending). Traditional node connection methods mainly rely on manual operation, requiring manual searching of corresponding preceding and following nodes and setting up associations. This approach has several technical bottlenecks:
[0003] The navigation method lacks adaptive adjustment, relying entirely on manual connection and fixed verification methods. It cannot dynamically optimize connection strategies based on node type and data characteristics, resulting in low efficiency in complex link construction scenarios. Node matching is prone to errors due to the lack of unified feature extraction and quantification standards. When the number of nodes exceeds 20, the manual matching error rate is as high as 35%. The verification mechanism lacks systematic rule base support and mainly relies on manual experience judgment, making it difficult to fully detect data flow errors, type mismatches, and other problems in the link, resulting in a high task execution failure rate. The node connection operation method does not conform to user operating habits and lacks convenient interaction methods, increasing the operational burden on personnel. Summary of the Invention
[0004] The purpose of this application is to provide a method, system, and device for intelligent connection of task flow processing nodes, which can realize intelligent matching and efficient connection of processing nodes.
[0005] To achieve the above objectives, this application provides the following solution:
[0006] Firstly, this application provides a method for intelligent connection of task flow processing nodes, including:
[0007] Obtain basic information about each processing node in the task to be processed;
[0008] Based on the basic information of each processing node, a node matching model based on deep learning is used to extract the feature vectors of each processing node.
[0009] Calculate the similarity between the feature vectors of each pair of processing nodes, and match the processing nodes according to the similarity and a preset threshold to obtain the matching result;
[0010] Based on the matching results, the processing nodes in the task flow are connected by using the operation menu and / or drag and drop to establish node links and display them.
[0011] In one embodiment, the basic information includes a unique node identifier, node type, functional description text, data input format description, data output format description, and a list of association rules.
[0012] In one embodiment, before extracting the feature vectors of each processing node, the method further includes: performing deduplication, standardization, and dimensional unification processing on the basic information of each processing node in sequence.
[0013] In one implementation, the deep learning-based node matching model is obtained by pre-tuning and training the BERT model.
[0014] In one implementation, the similarity between the feature vectors of two processing nodes is calculated using the following formula: Where sim represents the similarity between the feature vectors of the two processing nodes, V1 and V2 are the feature vectors of the two processing nodes respectively, n is the dimension of the feature vector, and V 1i For the i-th component of V1, V 2i Let i be the i-th component of V2.
[0015] In one embodiment, based on the matching result, processing nodes in the task flow are connected using an operation menu and / or drag-and-drop method, including: obtaining a target node selected by the user, and receiving the user's operation menu instruction and / or drag-and-drop instruction; the operation menu instruction is for the user to select the node to be connected from the operation menu that pops up when right-clicking on the target node; the drag-and-drop instruction is for the user to drag the target node to the display area and release it at the node to be connected; determining the matchable nodes of the target node based on the matching result, and sorting and displaying the matchable nodes in the operation menu according to the similarity from high to low, and establishing a connection link between the target node and the node to be connected according to the operation menu instruction; based on the matching result, highlighting processing nodes with a similarity greater than a set threshold to the target node in the display area, the highlighting intensity being proportional to the similarity, and establishing a connection link between the target node and the node to be connected according to the drag-and-drop instruction.
[0016] In one embodiment, the method further includes: traversing the node links using a depth-first search algorithm, recording node connection relationships and data flow directions to obtain link information; matching the link information with rules in a pre-established verification rule base to determine the matching degree of the node links; the verification rule base includes node type matching rules, data flow rules, and node dependency relationship rules; if the matching degree is less than a set matching threshold, it is determined that the node links have rule violations, and an error report is generated.
[0017] In one embodiment, the method further includes: using a relational database to store the basic information and feature vectors of each processing node, and using a graph database to store the node links.
[0018] Secondly, this application provides an intelligent connection system for task flow processing nodes, comprising:
[0019] The node information acquisition module is used to acquire basic information about each processing node in the task to be processed.
[0020] The automatic node matching module is used to extract the feature vectors of each processing node based on the basic information of each processing node, using a deep learning-based node matching model, and to calculate the similarity between the feature vectors of each pair of processing nodes. The processing nodes are then matched according to the similarity and a preset threshold to obtain the matching result.
[0021] The connection operation optimization module is used to connect the processing nodes in the task to be processed flow according to the matching results by using operation menus and / or drag and drop, so as to establish node links and display them.
[0022] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described intelligent connection method for task flow processing nodes.
[0023] According to the specific embodiments provided in this application, this application has the following technical effects:
[0024] This application provides a method, system, and device for intelligent connection of task flow processing nodes. It extracts feature vectors of each processing node through a node matching model based on deep learning and combines similarity calculation to achieve automatic matching of processing nodes, thereby improving the matching accuracy of processing nodes and significantly reducing the error rate of manual matching. Furthermore, it uses operation menus and / or drag-and-drop methods to connect processing nodes, shortening the connection time and improving the connection efficiency. Attached Figure Description
[0025] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0026] Figure 1 This is an application environment diagram of a task flow processing node intelligent connection method according to an embodiment of this application.
[0027] Figure 2 This is a flowchart illustrating a method for intelligent connection of task flow processing nodes according to an embodiment of this application.
[0028] Figure 3 This is a schematic diagram of the functional modules of an intelligent connection system for task flow processing nodes provided in an embodiment of this application.
[0029] Figure 4 This is a schematic diagram of the node information acquisition module in one embodiment of this application.
[0030] Figure 5 This is a schematic diagram of the node automatic matching module in one embodiment of this application.
[0031] Figure 6 This is a schematic diagram of the connection operation optimization module in one embodiment of this application.
[0032] Figure 7 This is a schematic diagram of a node link rationality verification module in one embodiment of this application.
[0033] Figure 8 This is a schematic diagram of a data storage module in one embodiment of this application.
[0034] Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation
[0035] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0036] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0037] The intelligent connection method for task flow processing nodes provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, terminal 101 communicates with server 102 via a network. A data storage system can store the data that server 102 needs to process. The data storage system can be set up independently, integrated into server 102, or placed in the cloud or on another server. Terminal 101 can send basic information about each processing node in the task flow to server 102. After receiving the basic information of each processing node, server 102 automatically establishes node links. Server 102 can provide feedback on the established node links to terminal 101. Furthermore, in some embodiments, the intelligent connection method for task flow processing nodes can also be implemented independently by server 102 or terminal 101.
[0038] The terminal 101 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The server 102 can be implemented using a standalone server or a server cluster composed of multiple servers, or it can be a cloud server.
[0039] In one exemplary embodiment, such as Figure 2 As shown, a method for intelligent connection of task flow processing nodes is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 102 as an example, the explanation includes the following steps 201 to 206.
[0040] Step 201: Obtain basic information about each processing node in the task to be processed flow.
[0041] Specifically, basic information of each processing node is collected in real time through a visual canvas interface, with a collection frequency of 1Hz. This basic information includes a unique node identifier, node type, functional description text, data input format description, data output format description, and a list of association rules.
[0042] In an exemplary embodiment, step 201 further performs deduplication, standardization, and dimensional unification processing on the basic information of each processing node in sequence.
[0043] The deduplication process involves using a hash algorithm to verify the basic information of the nodes being processed and removing duplicate node information. The hash function is: Where 'a' represents the basic information of the processing node, 'J' represents the dimension of the basic information, and 'w' represents the dimension of the processing node. j For feature weights, a jLet be the j-th dimension in the basic information, and m be the length of the hash table.
[0044] The standardization process is as follows: unify the format of similar data from different processing nodes, perform word segmentation and stop word removal on the functional description text, and use the TF-IDF algorithm to convert the text into a vector. The formula is: TF-IDF(t,d)=TF(t,d)×IDF(t); Where TF-IDF(t,d) is the weight of word t in document d, TF(t,d) is the term frequency of word t in document d, IDF(t) is the inverse document frequency, and n t,d Let ∑ be the number of times word t appears in document d. k n k,d Let be the sum of the occurrences of all words in document d, |D| be the total number of documents, and |d∈D:t∈d| be the number of documents containing word t.
[0045] The process of unifying dimensions is as follows: The numerical node attributes are normalized using the following formula: Where x is the original value, min(x) is the minimum value of the attribute, and max(x) is the maximum value of the attribute. This is the normalized value.
[0046] Step 202: Based on the basic information of each processing node, a deep learning-based node matching model is used to extract the feature vectors of each processing node. The deep learning-based node matching model is obtained by pre-tuning and training the BERT model.
[0047] The functional description text of the processing node is matched with the processed input node matching model to obtain a feature vector with a dimension of 768.
[0048] Step 203: Calculate the similarity between the feature vectors of every two processing nodes, and match the processing nodes according to the similarity and a preset threshold to obtain the matching result. Specifically, when the similarity is greater than the preset similarity threshold (e.g., 0.7), the two processing nodes are considered to be matched.
[0049] In a specific application example, the cosine similarity algorithm is used to calculate the similarity between the feature vectors of two processing nodes, and the formula is: Where sim represents the similarity between the feature vectors of the two processing nodes, V1 and V2 are the feature vectors of the two processing nodes respectively, n is the dimension of the feature vector, and V 1i For the i-th component of V1, V 2i Let i be the i-th component of V2.
[0050] Step 204: Based on the matching results, connect the processing nodes in the task flow to be processed by using the operation menu and / or drag and drop to establish node links and display them.
[0051] In a specific application example, step 204 includes steps 41 to 43.
[0052] Step 41: Obtain the target node selected by the user, and receive the user's operation menu command and / or drag command. The operation menu command is for the user to select the node to be connected from the operation menu that pops up when right-clicking on the target node. The drag command is for the user to drag the target node to the display area and release it at the node to be connected.
[0053] Step 42: Determine the matchable nodes of the target node based on the matching results, and sort and display the matchable nodes in the operation menu according to the similarity from high to low. Establish the connection link between the target node and the node to be connected according to the operation menu instructions.
[0054] Specifically, when a user selects multiple processing nodes of the same type, right-clicks to bring up the operation menu, selects "Insert Post-Node," and displays a list of matching post-nodes based on the matching results, sorted from high to low similarity. After the user selects a node to be connected from the post-node list, the connection relationship between the two nodes is automatically established. The connection relationship data format is: R = (S, N) p N s ,T); where S is the unique identifier for the connection, N p N is the ID of the preceding node. s T is the ID of the subsequent node, and T is the timestamp of the connection establishment.
[0055] Step 43: Based on the matching result, in the display area, the processing nodes whose similarity to the target node is greater than a set threshold are highlighted. The highlighting intensity is proportional to the similarity. And according to the drag command, a connection link is established between the target node and the node to be connected.
[0056] Specifically, when a user clicks on a target node and drags it onto the visualization canvas, the similarity between the target node and other processed nodes is calculated in real time. Processed nodes with a similarity greater than a preset similarity threshold are highlighted. The highlight intensity is proportional to the similarity, and the formula is: I = 255 × cos(V1, V2); where I is the highlight intensity value, which is 0-255. The user drags the target node to the node to be connected and releases the mouse to complete the connection.
[0057] Step 205: Perform a rationality check on the node link.
[0058] In a specific application example, step 205 includes steps 51 to 53.
[0059] Step 51: Use a depth-first search algorithm to traverse the node links, record the node connection relationships and data flow directions, and obtain the link information. The traversal path is recorded as Path = (N1, N2, ..., N...). K ); where K is the total number of processing nodes, N k For the k-th processing node, k = 1 to K, N k With N k+1 There is a connection.
[0060] Step 52: Match the link information with the rules in the pre-established verification rule base to determine the matching degree of the node link. Specifically, a fuzzy matching algorithm is used to calculate the matching degree, and the formula is: Where M(Path,R) is the matching degree of the node link Path, Matched(Path,R) is the number of nodes in the link Path that match rule R, and Total(R) is the total number of rules R.
[0061] The verification rule base includes node type matching rules, data flow rules, and node dependency relationship rules.
[0062] The node type matching rules define the types of preceding nodes that different types of subsequent nodes are allowed to connect to, and are expressed using rule expressions. For example, the type matching rule for SMS bulk sending nodes is: R type (N s )={N p |Type(N p )∈{crowd acquisition node, crowd operation node}}. Where, R type (N s ) represents the type matching rules for bulk SMS sending nodes, N s N is a node for bulk SMS sending. p For the preceding nodes that the SMS bulk sending node is allowed to connect to, Type(N) p ) is N p The type.
[0063] Data flow rules ensure that data flows from the preceding node to the following node, prohibiting reverse or cross-flow. The rule expression is: R flow (N p N s ) = Direction(N p N s ) = Forward. Where R... flow (N p N s ) is Np and N s Data flow rules between them, Direction(N) p N s ) represents N p and N s Direct connections are made between them, while "Forward" indicates a forward connection.
[0064] The node dependency rule states that for nodes with dependencies, a dependency condition must be met before a connection can be established. For example, if a computation node depends on a specific data input node, the rule expression is: Among them, R dep (N s ) is N s Dependency rules, Dependent(N) s ) is N s The dependent data set, Provide(N) p ) is N p The provided data set.
[0065] Step 53: If the matching degree is less than the set matching threshold (e.g., 1), it is determined that there is a rule violation in the node link, and an error report is generated. The user can adjust the node link according to the error report. After the modification is completed, repeat step 205 above until the matching degree is greater than or equal to the set matching threshold.
[0066] The error report includes the error location, error type, error cause, and suggested corrections. Error location is determined using a link node indexing method, error types are categorized by rule type, and suggested corrections are generated based on a rule base.
[0067] Step 206: A relational database is used to store the basic information and feature vectors of each processing node, and a graph database is used to store the node links. The table structure of the relational database includes fields for node ID, type, feature vector, and creation time. The graph database uses nodes as vertices and connections as edges, supporting efficient link query and traversal operations. Furthermore, an XML format is used to store the validation rule base, facilitating rule expansion and maintenance.
[0068] This application also provides an application scenario in which the above-described intelligent connection method for task flow processing nodes is applied. Specifically, the intelligent connection method for task flow processing nodes provided in this embodiment can be applied in marketing scenarios in the financial field. For example, in a bank's bulk marketing scenario, it is necessary to send interest rate adjustment notification SMS messages to mortgage customers.
[0069] The system hardware configuration is as follows: the server uses an Intel Xeon E5-2680v4 processor, 32GB of memory, and a 1TB SSD; the client uses a regular personal computer equipped with a 1920×1080 resolution monitor.
[0070] Software environment: The server operating system is Linux CentOS7, and the database uses MySQL 8.0 and Neo4j 4.0; the client browser is Chrome 90.0, which supports WebGL rendering.
[0071] The implementation process is as follows.
[0072] (1) Collect information from the population acquisition node (extract mortgage customer information), the calculation node (screen mortgage customers), and the batch SMS sending node (send interest rate adjustment notifications). Process the function description text "extract customer information that meets mortgage conditions", "screen mortgage customers with good credit", and "send interest rate adjustment SMS to mortgage customers" and convert them into feature vectors.
[0073] (2) The similarity between the crowd acquisition node and the operation node and the SMS batch sending node was calculated to be 0.85 and 0.81, respectively, both of which are greater than the threshold of 0.7, and a match was determined.
[0074] (3) Users can use the drag-and-drop connection method. Click the SMS batch sending node and drag it to the canvas. The system highlights the crowd acquisition node (highlight intensity I = 255 × 0.85 = 217) and the calculation node (highlight intensity I = 255 × 0.81 = 207). Users can drag the node to the crowd acquisition node and the calculation node and release it to complete the connection.
[0075] (4) Verify the node link. The rule matching degree is 1, and the link is reasonable.
[0076] (5) After the verification is passed, the marketing campaign is allowed to proceed and interest rate adjustment notification SMS messages are sent to the target customers.
[0077] Based on the same inventive concept, this application also provides a task flow processing node intelligent connection system for implementing the above-described task flow processing node intelligent connection method. The solution provided by this system is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more task flow processing node intelligent connection system embodiments provided below can be found in the limitations of the task flow processing node intelligent connection method described above, and will not be repeated here.
[0078] In one exemplary embodiment, such as Figure 3As shown, a task flow processing node intelligent connection system is provided, including: a node information acquisition module 301, a node automatic matching module 302, a connection operation optimization module 303, a node link rationality verification module 304, a data storage module 305, and a user interaction module 306.
[0079] The node information acquisition module 301 is used to acquire basic information of each processing node in the process of the task to be processed.
[0080] like Figure 4 As shown, the node information acquisition module 301 includes an information acquisition unit and a preprocessing unit. The information acquisition unit is used to acquire basic information of the processing nodes in real time through a visual canvas interface. The preprocessing unit is used to perform deduplication, standardization, and dimension unification processing on the basic information of each processing node in sequence, and then output it to the node feature extraction module 302 and the data storage module 305.
[0081] The node automatic matching module 302 is used to extract the feature vectors of each processing node based on the basic information of each processing node, using a node matching model based on deep learning, and calculate the similarity between the feature vectors of each pair of processing nodes. The processing nodes are then matched according to the similarity and a preset threshold to obtain the matching result.
[0082] like Figure 5 As shown, the automatic node matching module 302 includes a model building unit, a feature extraction unit, and a similarity calculation unit. The model building unit constructs a deep learning-based node matching model, using the BERT model as its basic architecture, and fine-tunes the training to adapt to the node matching task. The feature extraction unit extracts feature vectors from the nodes using the trained BERT model. The similarity calculation unit calculates the cosine similarity between feature vectors, performs a threshold judgment, and outputs the matching results to the connection operation optimization module 303 and the data storage module 305.
[0083] The connection operation optimization module 303 is used to connect the processing nodes in the task to be processed flow according to the matching result by using the operation menu and / or drag and drop, so as to establish node links and display them.
[0084] like Figure 6As shown, the connection operation optimization module 303 includes a right-click insertion unit and a drag-and-drop connection unit. The right-click insertion unit allows the user to select multiple target nodes, right-click to bring up an operation menu, and automatically establish a connection after selecting a subsequent node. The drag-and-drop connection unit calculates the similarity between the node and other preceding nodes in real time after the user clicks on a subsequent node and drags it to the visualization canvas. Preceding nodes with similarity greater than a set threshold are highlighted. The user drags the node to the target preceding node and releases the mouse to complete the connection. The node link is then output to the node link validity verification module 304 and the data storage module 305.
[0085] The node link rationality verification module 304 is used to verify the rationality of node links.
[0086] like Figure 7 As shown, the node link rationality verification module 304 includes a rule base construction unit, a link traversal unit, a rule matching unit, and an error handling unit. The rule base construction unit is used to construct a verification rule base. The link traversal unit uses a depth-first search algorithm to traverse the node links, recording node connection relationships and data flow directions to obtain link information. The rule matching unit matches the link information with rules in the pre-established verification rule base to determine the matching degree of the node links. The error handling unit generates an error report when the matching degree is less than a set matching threshold and outputs it to the user interaction module 306 and the data storage module 305.
[0087] The data storage module 305 is used to store the basic information and feature vectors of each processing node using a relational database and to store the node links using a graph database.
[0088] like Figure 8 As shown, the data storage module 305 includes a node information storage unit, a rule information storage unit, and a link data storage unit. These three units support other modules through a data interaction interface. The node information storage unit stores the unique identifier of each node, the node type, and the feature vector. The rule information storage unit stores the verification rules, and the link data storage unit stores the node links.
[0089] The user interaction module 306 is used for user interaction. Specifically, the user interaction module is responsible for rendering and displaying the visual canvas, supporting node scaling and panning operations, and using WebGL technology for efficient rendering; it provides a rich set of user operation menus, including node creation, deletion, editing, and connection functions; and it displays verification error reports in a visual manner, using a combination of charts and text to highlight the error location and cause.
[0090] This application improves the matching accuracy of processed nodes and significantly reduces the error rate of manual matching by using a node matching mechanism based on BERT model feature extraction and cosine similarity calculation. Optimized right-click insertion and drag-and-drop connection operations, combined with similarity highlighting, shorten the connection time and improve efficiency. A multi-dimensional verification rule base is constructed, employing depth-first search traversal and rule matching algorithms to achieve comprehensive verification of node links, resulting in high error detection coverage. User-friendly interaction methods and intuitive error report display reduce user difficulty and enhance the system's user experience. Modular design and standardized data storage format support the flexible addition of new node types and matching rules, adapting to the diverse marketing needs of different enterprises.
[0091] In summary, this application effectively solves the problems of low connection efficiency and high error rate of traditional batch marketing nodes through node matching algorithm, optimized connection operation method and comprehensive verification mechanism.
[0092] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 9 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores basic information about each processing node in the task flow. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for intelligent connection of task flow processing nodes.
[0093] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0094] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.
[0095] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.
[0096] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0097] In this application, all actions to acquire signals, information, or data are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with the authorization granted by the owner of the relevant device.
[0098] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).
[0099] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.
[0100] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0101] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A method for intelligent connection of task flow processing nodes, characterized in that, The method includes: Obtain basic information about each processing node in the task to be processed; Based on the basic information of each processing node, a node matching model based on deep learning is used to extract the feature vectors of each processing node. Calculate the similarity between the feature vectors of each pair of processing nodes, and match the processing nodes according to the similarity and a preset threshold to obtain the matching result; Based on the matching results, the processing nodes in the task flow to be processed are connected by using the operation menu and / or drag and drop to establish node links and display them; Specifically, based on the matching results, processing nodes in the task flow are connected using an operation menu and / or drag-and-drop method, including: The system acquires the target node selected by the user and receives the user's operation menu command and / or drag command; the operation menu command is for the user to select the node to be connected from the operation menu that pops up when right-clicking on the target node; the drag command is for the user to drag the target node to the display area and release it at the node to be connected. Based on the matching results, the matchable nodes of the target node are determined, and the matchable nodes are sorted and displayed in the operation menu in descending order of similarity. The connection link between the target node and the node to be connected is established according to the operation menu instructions. Based on the matching results, processing nodes with a similarity greater than a set threshold to the target node are highlighted in the display area. The highlighting intensity is proportional to the similarity. A connection link between the target node and the node to be connected is established according to the drag command.
2. The intelligent connection method for task flow processing nodes according to claim 1, characterized in that, The basic information includes a unique node identifier, node type, functional description text, data input format description, data output format description, and a list of association rules.
3. The intelligent connection method for task flow processing nodes according to claim 1, characterized in that, Before extracting the feature vectors of each processing node, the method further includes: performing deduplication, standardization, and dimensional unification on the basic information of each processing node in sequence.
4. The intelligent connection method for task flow processing nodes according to claim 1, characterized in that, The deep learning-based node matching model is obtained by pre-tuning and training the BERT model.
5. The intelligent connection method for task flow processing nodes according to claim 1, characterized in that, The similarity between the feature vectors of two processing nodes is calculated using the following formula: ； in, The similarity between the feature vectors of two processing nodes. and These are the feature vectors of the two processing nodes. n The dimension of the feature vector. for The One portion, for The Each component.
6. The intelligent connection method for task flow processing nodes according to claim 1, characterized in that, The method further includes: The node links are traversed using a depth-first search algorithm to record the node connection relationships and data flow directions, thereby obtaining the link information. The link information is matched with the rules in a pre-established verification rule base to determine the matching degree of the node link; the verification rule base includes node type matching rules, data flow rules, and node dependency relationship rules. If the matching degree is less than the set matching threshold, it is determined that there is a rule violation in the node link, and an error report is generated.
7. The intelligent connection method for task flow processing nodes according to claim 1, characterized in that, The method further includes: A relational database is used to store the basic information and feature vectors of each processing node, and a graph database is used to store the node links.
8. A task flow processing node intelligent connection system, characterized in that, The system is applied to the intelligent connection method for task flow processing nodes according to any one of claims 1-7, and the system comprises: The node information acquisition module is used to acquire basic information about each processing node in the process of the task to be processed. The automatic node matching module is used to extract the feature vectors of each processing node based on the basic information of each processing node, using a deep learning-based node matching model, and to calculate the similarity between the feature vectors of each pair of processing nodes. The processing nodes are then matched according to the similarity and a preset threshold to obtain the matching result. The connection operation optimization module is used to connect the processing nodes in the task to be processed flow according to the matching results by using the operation menu and / or drag and drop, so as to establish node links and display them.
9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the intelligent connection method for task flow processing nodes according to any one of claims 1-7.