Cross-platform cad drawing intelligent collaborative management method and system and medium

Abstract

The application discloses a cross-platform CAD drawing intelligent collaborative management method and system and a medium, relates to the related technical field of data management, and comprises the following steps: uploading original CAD files and management tasks; developing a cross-platform intermediate engine and establishing information interaction of the cross-platform intermediate engine with each target platform; scanning the original files and the management tasks, driving a coder to perform platform classification coding based on CAD element characteristics, and driving a decoder to perform optimal expression collapse based on platform migration operators; issuing N decoding operators to perform platform graph reconstruction and storage; performing graph matching calling and uniform template integration, and visualizing as real-time project graphs. The application solves the technical problems of information loss, semantic attenuation, low collaborative efficiency and difficulty in realizing real-time synchronization and visual management of design elements in a heterogeneous environment during cross-platform data exchange in the prior art, and achieves the technical effects of improving the fidelity, intelligent interoperability and real-time collaborative management of cross-platform CAD design data.

Description

Technical Field

[0001] This invention relates to the field of data management technology, specifically to a cross-platform intelligent collaborative management method, system, and medium for CAD drawings. Background Technology

[0002] CAD is being widely applied in engineering fields such as architecture, machinery, electronics, and infrastructure. Design activities are becoming increasingly complex, distributed, and collaborative. Modern engineering projects often involve collaboration among multiple professional teams, different functional units, and even across regions and organizations. Each team may choose different CAD software platforms and versions based on historical background, professional preferences, project requirements, or cost considerations. Platform heterogeneity leads to data silos in design data in terms of format, data structure, element expression, and functional logic. Currently, cross-platform CAD collaborative management mainly uses standard neutral formats (such as DXF, STEP, IFC) for data exchange, or requires all collaborators to use a specific platform or forces data migration. However, the conversion process is often accompanied by problems such as loss of geometric information, decay of attribute data, destruction of hierarchical relationships, and platform-specific intelligent objects becoming simple geometric shapes. This seriously restricts the reusability and intelligent processing capabilities of data, while also bringing high software licensing costs and training expenses. Furthermore, the differences in platform functions may affect design efficiency and creative expression, making it difficult to implement in practice. In addition, traditional CAD file management focuses on version control and centralized storage, lacking a deep understanding and dynamic maintenance of design intent, element-level changes, and cross-platform semantic consistency. It is unable to extract, match, and integrate key information in real time and accurately from design results generated by heterogeneous platforms.

[0003] Therefore, current technologies suffer from technical problems such as information loss, semantic decay, low collaboration efficiency, and difficulty in achieving real-time synchronization and visual management of design elements in heterogeneous environments during cross-platform data exchange. Summary of the Invention

[0004] This application provides a cross-platform CAD drawing intelligent collaborative management method, system, and medium, which solves the technical problems of information loss, semantic decay, low collaborative efficiency, and difficulty in achieving real-time synchronization and visual management of design elements in heterogeneous environments during cross-platform data exchange. It achieves the technical effect of improving the fidelity, intelligent interoperability, and real-time collaborative management of cross-platform CAD design data.

[0005] This application provides a cross-platform intelligent collaborative management method for CAD drawings. The method includes: uploading original CAD files and management tasks; developing a cross-platform intermediate engine and establishing information interaction between the cross-platform intermediate engine and each target platform; the cross-platform intermediate engine scans the original CAD files and management tasks, drives the encoder to perform platform inclusion encoding based on CAD element features, drives the decoder to perform optimal expression collapse based on platform migration operators, and generates N decoding operators, wherein each target platform corresponds to one decoding operator; the N decoding operators are distributed to each target platform to perform platform diagram reconstruction and storage; and, along with the real-time project management process, the diagram matching and unified template integration based on each target platform are performed to visualize the real-time project diagram.

[0006] In one possible implementation, a cross-platform intermediate engine is developed, comprising: building an encoder based on the extraction of CAD element features and platform attribution encoding as the underlying logic; constructing a platform migration operator for each target platform by superimposing the state of the platform representation of CAD elements; building a decoder array by collapsing the optimal expression based on the platform migration operator, wherein each target platform corresponds to one decoder; introducing a multi-head attention mechanism, wherein the multi-head attention is defined based on geometry, topology, process, and constraints; and constructing a cross-platform intermediate engine based on the encoder, the multi-head attention mechanism, and the decoder array.

[0007] In one possible implementation, the driver encoder performs platform inclusion encoding based on CAD element features, including: scanning the original CAD file and extracting CAD element features; scanning the management task and classifying the CAD element features based on each target platform to determine platform CAD element features, wherein there are multiple target platforms with the same CAD element features; and determining the platform encoding sequence by integrating the platform CAD element features based on multi-head attention, wherein the platform encoding sequence corresponds one-to-one with the target platform.

[0008] In one possible implementation, driving the decoder to perform optimal representation collapse based on platform migration operators includes: sending a first platform encoded sequence to the corresponding first decoder, performing parallel collapse processing based on multi-head attention on the first platform migration operator to determine a group of decoding operators; and integrating the group of decoding operators as the first decoding operator.

[0009] In one possible implementation, the cross-platform CAD drawing intelligent collaborative management method further includes: sending the platform encoding sequence to the corresponding decoder to obtain N decoding operators; and sending the N decoding operators to the corresponding target platform according to the data interface interaction between the cross-platform intermediate engine and each target platform to perform platform drawing reconstruction processing and storage.

[0010] In one possible implementation, the cross-platform CAD drawing intelligent collaborative management method further includes: performing intent recognition on the original CAD file to determine intent tags; embedding the intent tags as additional semantic features into the platform encoding sequence; wherein, intent recognition is performed based on a preset rule set: if the feature sequence contains a preset combination, it is marked as a standard connector intent; if the surface continuity and surface change rate meet preset conditions, it is marked as a fluid dynamics intent.

[0011] In one possible implementation, as the real-time project management process progresses, graph matching calls based on each target platform are executed, including: the user client uploading project management data, wherein the project management data includes graph element requirements based on the project stage; based on the project management data, graph matching based on each target platform is executed to determine the data to be retrieved by the platform; and the platform-retrieved data is sent to the corresponding target platform to determine the platform graph set.

[0012] In one possible implementation, the unified template integration includes: determining a diagram display template based on the project management data; scanning the platform diagram set, performing CAD element feature extraction and unified structure transformation based on the diagram display template to determine the project management diagram; and integrating the platform diagram set and the project management diagram for visualization on a cross-platform display interface.

[0013] This application also provides a cross-platform CAD drawing intelligent collaborative management system, the system comprising: a data upload module for uploading original CAD files and management tasks; an information interaction establishment module for developing a cross-platform intermediate engine and establishing information interaction between the cross-platform intermediate engine and each target platform; a decoding operator generation module for the cross-platform intermediate engine to scan the original CAD files and management tasks, drive the encoder to perform platform inclusion encoding based on CAD element features, drive the decoder to perform optimal expression collapse based on platform migration operators, and generate N decoding operators, wherein each target platform corresponds to one decoding operator; a graph reconstruction and storage module for distributing the N decoding operators to each target platform to perform platform graph reconstruction and storage; and a graph matching and integration module for performing graph matching calls and unified template integration based on each target platform as the real-time project management process progresses, and visualizing it as a real-time project graph.

[0014] This application also provides a computer-readable storage medium, including: a computer program stored thereon, which, when executed by a processor, implements a cross-platform intelligent collaborative management method for CAD drawings.

[0015] This application proposes a cross-platform CAD drawing intelligent collaborative management method, system, and medium. The method involves uploading original CAD files and management tasks; developing a cross-platform intermediate engine and establishing information interaction between it and various target platforms; scanning the original files and management tasks; driving the encoder to perform platform incorporation encoding based on CAD element features; driving the decoder to perform optimal expression collapse based on platform migration operators; distributing N decoding operators to perform platform drawing reconstruction and storage; and performing drawing matching and unified template integration for visualization as a real-time project drawing. This solves the technical problems of information loss, semantic decay, low collaborative efficiency, and difficulty in achieving real-time synchronization and visualization management of design elements in heterogeneous environments during cross-platform data exchange. It achieves the technical effects of improving the fidelity, intelligent interoperability, and real-time collaborative management of cross-platform CAD design data. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a schematic diagram of the cross-platform CAD drawing intelligent collaborative management method provided in the embodiments of this application.

[0018] Figure 2 This is a schematic diagram of the cross-platform CAD drawing intelligent collaborative management system provided in an embodiment of this application.

[0019] Figure labeling: Data upload module 10, information interaction establishment module 20, decoding operator generation module 30, graph reconstruction and storage module 40, graph matching and integration module 50. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description is provided in conjunction with the accompanying drawings and preferred embodiments, based on the specific implementation methods, structure, features, and effects of the present invention.

[0021] This application provides a cross-platform intelligent collaborative management method for CAD drawings, such as... Figure 1 As shown, the method includes: Step S100: Upload the original CAD file and manage the task.

[0022] Preferably, the original design documents from a specific CAD software platform (such as AutoCAD, SolidWorks, CATIA, Revit, etc.) that have not been processed by the cross-platform intermediate engine are uploaded to the system. These documents contain complete design data in their native format, specifically including geometric data, such as geometric elements that constitute the graphics, such as points, lines, surfaces, and volumes, and their precise coordinates and dimensions; topological data, such as structural relationships between geometric elements, such as connections, adjacencies, and inclusions; attribute data, i.e., non-geometric information attached to geometric elements or primitives, such as materials, colors, layers, line types, part numbers, and manufacturer information; constraint and relationship data, such as parametric constraints defining the relative positions and dimensions between geometric elements, assembly relationships, etc.; and view and annotation data, such as view layouts, dimensioning, text annotations, and symbols in engineering drawings. Simultaneously, directive, configurable, or requirement information related to project management and collaborative design processes will be uploaded to the system to guide and constrain data processing, transformation, and integration behaviors. This includes specifying the CAD software platform for which the original CAD files need to be converted or adapted for this collaboration, specifying the conversion rules or preference settings when processing specific types of elements, defining the data extraction scope in cross-platform processing, clarifying the collaborative stage information of the current task, and describing the visualization requirements of the final real-time project diagram.

[0023] Step S200 involves developing a cross-platform middleware engine and establishing information interaction between the cross-platform middleware engine and each target platform.

[0024] Step S200 further includes: building an encoder based on the extraction of CAD element features and platform attribution encoding as the underlying logic; constructing a platform migration operator for each target platform by superimposing the state of the platform representation of CAD elements; building a decoder array by collapsing the optimal expression based on the platform migration operator, wherein each target platform corresponds to one decoder; introducing a multi-head attention mechanism, wherein the multi-head attention is defined based on geometry, topology, process and constraints; and constructing a cross-platform intermediate engine based on the encoder, the multi-head attention mechanism and the decoder array.

[0025] Preferably, from the original CAD file, the most basic constituent unit's various element features are parsed and extracted. These include not only geometric features such as shape, size, and position, but also topological features such as connections with other elements, attribute features such as material, color, and number, and semantic features such as design intent, such as load-bearing components and fluid channels. Platform attribution coding converts the extracted general element features into neutral, platform-independent coding vectors or sequences according to predefined rules. Specifically, an encoder is built to identify the features or feature combinations of the element and determine which target platform it is assigned to for priority processing. For example, a complex parametric surface feature may be assigned with high weight to the CATIA and NX platforms, while its simplified boundary representation features may be assigned to Auto CAD platform; For each target platform, the corresponding platform migration operator is predefined by superimposing the state of the CAD element's platform representation. Here, the platform representation refers to the set of all possible and compliant data representations of the same CAD element in a specific target platform. For example, in Revit, it may be a "family" instance containing material, fire rating, and manufacturer information; in AutoCAD, it may be a block reference or polyline with attributes. State superposition is an analogy from the concept of quantum computing, indicating that before the transformation occurs, the system does not only pre-set a fixed transformation path, but also maintains a set of multiple possible and probabilistic transformation options for the element in the target platform. The platform migration operator is used to encapsulate all possible options and their transformation rules.

[0026] Preferably, each target platform is equipped with a dedicated decoder to perform the final conversion from "general encoding" to "platform-specific data". The optimal representation collapse refers to the decoder receiving the general encoding from the encoder, calling the corresponding platform's migration operator, and evaluating and selecting the most specific and optimal representation from the various possible representations held by the migration operator in the current context, combined with the current specific management task and context. This "collapses" the data into a definite decoding operator that can be directly reconstructed by the target platform. The decoder array refers to a set of multiple decoders in parallel, with each decoder responsible for generating a data output that can be recognized by a specific target platform. A multi-head attention mechanism is introduced during the encoding and decoding process to improve the accuracy of feature analysis and transformation decisions. This mechanism allows for the simultaneous weighing of information importance from multiple different "representation subspaces" or "interest angles" when processing data. Based on the definitions of geometry, topology, process, and constraints, at least four attention heads are established in parallel: a geometry head, a topology head, a process head, and a constraint head. These heads focus on geometric attributes such as element shape, size, and precision; connections, assembly, and hierarchical relationships between elements; manufacturing and construction-related characteristics; and parametric relationships, design rules, and engineering constraints. The multiple attention heads work together to determine key information during feature extraction and to determine the representation that best meets all dimensional requirements during decoding "collapse," thereby making more comprehensive and balanced decisions.

[0027] Preferably, the encoder responsible for feature extraction and general encoding, the multi-head attention mechanism responsible for multi-dimensional intelligent analysis, and the decoder array responsible for generating platform-specific instruction sets are integrated according to specific data flow and control logic to form a cross-platform intermediate engine. This engine receives raw CAD files and manages tasks, processes them through an internal pipeline, and outputs instructions to drive data reconstruction on each target platform. Next, through standardized data interface APIs and a unified interaction protocol layer, information interaction between the cross-platform middleware engine and various target platforms is established. This enables a stable, bidirectional data and command communication channel between the cross-platform middleware engine and various heterogeneous CAD software platforms. The interactive information includes two flows: downlink commands and data from the cross-platform middleware engine to the platform, including precise command sequences or structured data packets generated by the decoder that are platform-recognizable for reconstructing CAD graphics, as well as management commands such as graphics call requests, version synchronization commands, and attribute update notifications; and uplink feedback and data from the platform to the engine. When the target platform connects, it registers metadata such as the types of graphics elements it supports, data formats, API interface versions, and functional limitations with the cross-platform middleware engine. It also provides feedback on the processing status, such as whether the commands were successfully executed, the progress of graphics reconstruction, and error or warning messages. Furthermore, after the user modifies the graphics issued by the cross-platform middleware engine locally on the target platform, the element-level change summary is synchronized back to the engine to update the central status.

[0028] Preferably, the information interaction between the cross-platform middleware engine and each target platform runs through the entire collaborative management lifecycle. Specifically, in the initialization and registration phase, the target platform starts its adapter, registers its capabilities with the cross-platform middleware engine, and establishes a connection; in the data distribution and reconstruction phase, the cross-platform middleware engine calls the platform API through the adapter, sends the "decoding operator" to the target platform, drives it to automatically create or update graphic objects in the local session, and stores them as files or database entries in the platform's native format; in the dynamic invocation and matching phase, when a real-time project view needs to be generated, the cross-platform middleware engine initiates a query to the target platform through the adapter, matches and extracts specific graphic element data stored locally based on feature descriptions; in the change synchronization and closed-loop phase, the target platform reports the user's local modifications to the cross-platform middleware engine incrementally through the adapter and updates the central code accordingly, which may trigger a new round of distribution to synchronize to other target platforms, thereby ensuring accurate, reliable, and automated two-way drive and feedback from the general instructions of the intelligent engine to the native operations of specific CAD platforms.

[0029] In step S300, the cross-platform intermediate engine scans the original CAD file and management tasks, drives the encoder to perform platform inbound encoding based on CAD element features, drives the decoder to perform optimal expression collapse based on platform migration operators, and generates N decoding operators, wherein each target platform corresponds to one decoding operator.

[0030] Preferably, after the cross-platform intermediate engine starts, it scans the input original CAD files and management tasks, performs structured parsing and information fusion. Specifically, it calls the parsing library of the corresponding CAD format to perform deep reading of the files, including deconstructing their internal data structure, identifying all graphic entities, layers, attribute sets, constraint relationships, views, etc., and converting them into an intermediate representation that can be processed within the cross-platform intermediate engine. At the same time, it parses the task instructions submitted by the user or system, extracts key parameters such as the target platform list, conversion preferences, data range, and project stage, and uses them as control variables and contextual basis for the encoding and decoding process. Then, the parsed data is sent to the encoder to perform platform-based encoding based on CAD element features, that is, to perform feature analysis and evaluation on CAD elements, determine geometric complexity, parameterization degree, semantic labels, etc., and determine the compatibility priority of the element with each target platform based on the feature analysis results. An element may be "classified" to multiple target platforms at the same time or mainly classified to one target platform. In turn, a platform encoding sequence is generated for each target platform, that is, a compressed, semantically rich mathematical vector or structured description, representing the core features and importance summary of the element for the target platform.

[0031] Preferably, the platform encoding sequences generated by the encoder for each platform are sent to the dedicated decoder of the corresponding platform to perform optimal expression collapse based on the platform migration operator. Specifically, the decoder activates multiple candidate expression forms that meet the conditions from the migration operator according to the input encoding. Combining the current management task context and based on the multi-head attention mechanism, it evaluates and weighs the candidate expression forms from multiple dimensions such as geometric accuracy, topological integrity, process requirements, and constraint relationships. Through rule scoring decision, it selects a comprehensive optimal specific expression form from multiple candidate states. Finally, the selected optimal expression form is converted into precise and sequential operation instructions or data objects that the target platform can directly understand and execute, as decoding operators, such as a macro command, a series of API call parameters, or a specific data exchange format fragment. Finally, the cross-platform intermediate engine outputs N decoding operators, where N is a positive integer representing the total number of target platforms.

[0032] Furthermore, step S300 also includes scanning the original CAD file and extracting CAD element features; scanning the management task and classifying the CAD element features based on each target platform to determine the platform CAD element features, wherein multiple target platforms with the same CAD element features are included; and determining the platform encoding sequence by integrating the platform CAD element features based on multi-head attention, wherein the platform encoding sequence corresponds one-to-one with the target platform.

[0033] Preferably, the encoder receives the original CAD file and delves into its internal data structure, decomposing it into the smallest independently identifiable design unit, i.e., CAD element. For each decomposed CAD element, it extracts its multi-dimensional attributes, such as geometric features like shape, size, position coordinates, and boundary representation data; topological features like connection relationships with adjacent elements, hierarchy and mating relationships in the assembly; attribute features like layers, colors, line types, materials, names, and user-defined attributes; and constraint parameter features like driving dimensions and geometric constraints. Simultaneously, it receives management task instructions to specify the target platform list for this collaboration. Based on the platform knowledge base with thresholds and the current task requirements, the encoder examines and classifies the extracted CAD element features, determining the most relevant CAD element features for each target platform. This means that the same CAD element may generate multiple different feature views, and there may be multiple target platform inclusions with the same CAD element features, meaning that a CAD element can be simultaneously classified into multiple target platforms for processing.

[0034] Preferably, a multi-head attention mechanism is used to integrate the platform CAD element features of each element. Specifically, multiple parallel attention heads are set up for each target platform. Each head independently analyzes the feature view of the platform and evaluates the importance weight of different feature dimensions. For example, when generating codes for Revit, the semantic head may assign a very high weight to the "door family type", while the "geometric head" pays more attention to its bounding box size. The weighted and merged scoring results are then fused to generate a unified, fixed-length, and dense platform coding sequence for each target platform. This sequence is used to encapsulate the operations that should be understood and reconstructed for the CAD element of the target platform. One target platform corresponds to one coding sequence. The coding sequences of all elements are combined in structural order to form a complete input data stream that drives the platform decoder.

[0035] Furthermore, step S300 also includes sending the first platform encoded sequence to the corresponding first decoder, performing parallel collapse processing based on multi-head attention on the first platform migration operator to determine the decoding operator group; and integrating the decoding operator group as the first decoding operator.

[0036] Preferably, the cross-platform intermediate engine routes the first platform encoding sequence to a first decoder specifically configured for it, and performs parallel collapse processing based on multi-head attention on the first platform transfer operator. The first platform transfer operator defines all legal graphical element representations for the platform and their mapping relationships with general features. Specifically, the first decoder uses the received platform encoding sequence as a query to retrieve and activate relevant information in the first platform transfer operator and performs collapse processing. This means simultaneously activating multiple candidate representations in the first transfer operator that match the input encoding, and enabling its multi-head attention mechanism to perform parallel evaluation of all candidate representations, including evaluating the fidelity of each candidate representation to the original design's size and shape, and evaluating the topology and adjacency. The system assesses the correctness of element connections, whether semantics / process meets management requirements, and evaluates the efficiency and lightweight nature of constraint performance modeling on the target platform. Each attention head then outputs a score for each candidate representation. The decoder integrates multiple scores through weighted summation and selects the highest-scoring candidate representation based on its overall score, determining it as the optimal representation collapse. Based on this candidate representation, it generates a set of ordered, low-level, platform-specific operation instructions, determines the decoding operator group, and integrates these ordered instructions into the final first decoding operator. This avoids the single, pre-defined conversion path in traditional conversions, dynamically selecting the optimal path through real-time evaluation, thereby achieving intelligence, adaptability, and high fidelity in the conversion process.

[0037] Furthermore, step S300 also includes performing intent recognition on the original CAD file to determine intent tags; embedding the intent tags as additional semantic features into the platform encoding sequence; wherein, intent recognition is performed based on a preset rule set: if the feature sequence contains a preset combination, it is marked as a standard connector intent; if the surface continuity and surface change rate meet preset conditions, it is marked as a fluid dynamics intent.

[0038] Preferably, intent tag information is introduced and integrated during the encoding process to enhance the semantic fidelity of cross-platform conversion. Specifically, after the encoder extracts basic CAD element features, the intent recognition unit is activated to analyze the design data, infer the specific engineering purpose, function, or standard represented by each CAD element feature or combination, and generate semantically categorized intent tags. For example, a specific set of holes and threads may be labeled as "standard connector intent," and a smoothly transitioning complex surface may be labeled as "fluid dynamics intent." The intent recognition unit has built-in preset rules, which are logical judgment conditions defined by domain experts or learned from historical data. If the feature sequence contains a preset combination, i.e., a cylinder is detected, the intent tag is determined. If a component has a chamfered end, a threaded side conforming to a certain standard, and its dimensional parameters match an entry in the standard library, it is considered a threaded cylinder and marked as a standard connector intent. Analyzing surface properties, if the surface exhibits high continuity and its curvature changes smoothly, conforming to typical drag reduction and flow guidance patterns in aerodynamics or fluid dynamics, it is considered a smooth surface and marked as a fluid dynamics intent. The generated intent label is then transformed into structured additional semantic features and fused with basic features such as geometry and topology extracted from the same element. This is embedded into the platform encoding sequence. The encoder's multi-head attention mechanism uses the intent label as enhanced semantic input, ultimately generating the platform encoding sequence and improving its semantic level.

[0039] Step S400: Distribute the N decoding operators to each target platform and perform platform graph reconstruction and storage.

[0040] Step S400 further includes: for the platform encoding sequence, sending it to the corresponding decoder to obtain N decoding operators; according to the data interface interaction between the cross-platform intermediate engine and each target platform, sending the N decoding operators to the corresponding target platform, and performing platform graph reconstruction processing and storage.

[0041] Preferably, after the encoder generates its own "platform encoding sequence" for each target platform, the routing mechanism within the cross-platform intermediate engine sends the platform encoding sequence to the corresponding decoder. Each decoder, upon receiving its platform encoding sequence, independently executes its internal optimal expression collapse logic, ultimately outputting a complete instruction set that can be directly understood and executed by its corresponding platform. This determines N decoding operators, the number of which is strictly equal to the number of target platforms N. Each decoding operator is unique in content and format and specifically designed for its target platform. Then, the cross-platform intermediate engine initiates communication with each target platform through a data interface, including sending the decoding operator A to the running CAD software A (such as Revit) through platform A's data interface, and sending the decoded operator... The code operator B is sent to the CAD software B (such as AutoCAD) to perform a reconstruction process. This means that the CAD software automatically remodels the model without human intervention, such as "creating a new part", "drawing a sketch outline", "performing an extrusion feature", "adding fillets", "applying material properties", and "inserting into a specific position in the assembly". The final result is a complete graphic model that is highly consistent with the original design intent, and the platform drawing is determined. Finally, it is automatically stored in the specified project path or database. For example, it is saved as a .rvt file in Revit and as .sldasm and .sldprt files in SOLIDWORKS. At the same time, the storage path, version information and other metadata are fed back to the cross-platform intermediate engine so that the engine can perform unified management and tracking.

[0042] In step S500, as the real-time project management progresses, graph matching and unified template integration based on various target platforms are performed to visualize the real-time project graph.

[0043] Step S500 further includes: the user client uploading project management data, wherein the project management data includes graph element requirements based on project phases; performing graph matching based on each target platform according to the project management data to determine the platform-retrieved data; and sending the platform-retrieved data to the corresponding target platform to determine the platform graph set.

[0044] Preferably, as the real-time project management progresses, i.e., the project enters different stages, such as scheme design, construction drawing design, multi-disciplinary coordination meetings, quantity calculation, and construction handover, the user uploads project management data, including the graphic element requirements based on the project stage, i.e. the content that needs to be viewed and analyzed at different stages. For example, in the multi-disciplinary coordination meeting stage, the wall outline of the architectural profession, the beam and column positioning of the structural profession, and the main pipeline routing of the mechanical and electrical profession are extracted and overlaid to check for collisions; in the quantity calculation stage, the volume and material information of all walls and the model and quantity of all doors and windows are extracted. After receiving project management data, the cross-platform middleware engine parses the graph element requirements and performs graph matching based on each target platform. Specifically, it uses the platform coding sequence library or feature index library generated and maintained during the coding phase to perform graph matching, transforming the graph element requirements into a set of feature query conditions. It then performs parallel matching searches on the stored records of all relevant target platforms in the platform coding sequence library or feature index library to quickly locate all feature elements that meet the conditions, determine the platform data to be retrieved, clarify the list of element IDs to be retrieved from each target platform to meet the current requirements, and specify the specific feature data corresponding to each element. Then, the cross-platform middleware engine sends the platform-retrieved data to the corresponding target platform through a data interface. The platform's locally stored graph file accurately extracts the requested specific data based on the element ID list and returns it to the cross-platform middleware engine, integrating and determining the platform graph set.

[0045] Furthermore, step S500 also includes: determining a diagram display template based on the project management data; scanning the platform diagram set, performing CAD element feature extraction and unified structure conversion based on the diagram display template to determine the project management diagram; and integrating the platform diagram set and the project management diagram for visualization on a cross-platform display interface.

[0046] Preferably, the system parses the project management data uploaded by the user and matches a diagram display template from a pre-set template library. The diagram display template defines a set of view generation rules, including: layout structure (the arrangement of various disciplines and systems on the drawing, such as floor plans, elevations, 3D isometrics, and system diagrams); visual style (the display color, line type, line width, fill mode, and transparency of different categories of elements, for example, using solid blue lines for water supply pipes, dashed yellow lines for drainage pipes, and semi-transparent gray fill for structural components); information annotation (the attribute labels to be displayed, such as equipment numbers, pipe diameters, component dimensions, annotation styles, and positions); and filtering and simplification rules (determining which view details to display, simplify, or hide based on the view scale and purpose, for example, not displaying light fixtures and sockets on a site plan, and only displaying the connection principle rather than precise coordinates on a system diagram). Then, the platform diagram set is scanned and parsed a second time to extract the core features that meet the template requirements. Strictly following the rules of the diagram display template, it is converted into a neutral, unified data structure for display purposes only, and the project management diagram is determined, retaining only the geometry, style, and key attribute information required for the view. Finally, the project management diagram is used as the underlying foundation and logically associated with the complete reference data in the platform diagram set to generate a real-time project diagram, which is then visualized in a cross-platform display interface. The main view seen by the user is the standardized project management diagram. When the user clicks on or queries an element in the project management diagram, the detailed information of that element in the original platform diagram set can be located and displayed through the association relationship. At the same time, it is ensured that any device (PC, tablet, mobile phone) and operating system can access the view through a standard browser.

[0047] In the above text, refer to Figure 1 This paper describes in detail a cross-platform intelligent collaborative management method for CAD drawings according to embodiments of the present invention. Next, we will refer to... Figure 2 This invention describes a cross-platform intelligent collaborative management system for CAD drawings according to embodiments of the present invention.

[0048] The cross-platform CAD drawing intelligent collaborative management system according to embodiments of the present invention addresses the technical problems in existing technologies, such as information loss, semantic decay, low collaborative efficiency, and difficulty in achieving real-time synchronization and visual management of design elements in heterogeneous environments during cross-platform data exchange. It achieves the technical effects of improving the fidelity, intelligent interoperability, and real-time collaborative management of cross-platform CAD design data. Figure 2 As shown, the cross-platform CAD drawing intelligent collaborative management system includes: a data upload module 10, an information interaction establishment module 20, a decoding operator generation module 30, a drawing reconstruction and storage module 40, and a drawing matching and integration module 50.

[0049] The data upload module 10 is used to upload the original CAD files and management tasks; the information interaction establishment module 20 is used to develop a cross-platform intermediate engine and establish information interaction between the cross-platform intermediate engine and each target platform; the decoding operator generation module 30 is used by the cross-platform intermediate engine to scan the original CAD files and management tasks, drive the encoder to perform platform ingress encoding based on CAD element features, drive the decoder to perform optimal expression collapse based on platform migration operators, and generate N decoding operators, where each target platform corresponds to one decoding operator; the graph reconstruction and storage module 40 is used to distribute the N decoding operators to each target platform and perform platform graph reconstruction and storage; the graph matching and integration module 50 is used to perform graph matching calls and unified template integration based on each target platform as the real-time project management process progresses, and visualize it as a real-time project graph.

[0050] The specific configuration of the information interaction establishment module 20 will be described in detail below. The information interaction establishment module 20 further includes: building an encoder based on the extraction of CAD element features and platform attribution encoding as the underlying logic; constructing a platform migration operator for each target platform by superimposing the state of the platform representation of CAD elements; building a decoder array by collapsing the optimal expression based on the platform migration operator, wherein each target platform corresponds to one decoder; introducing a multi-head attention mechanism, wherein the multi-head attention is defined based on geometry, topology, process, and constraints; and constructing a cross-platform intermediate engine based on the encoder, the multi-head attention mechanism, and the decoder array.

[0051] The specific configuration of the decoding operator generation module 30 will be described in detail below. The decoding operator generation module 30 further includes: scanning the original CAD file and extracting CAD element features; scanning the management task and classifying the CAD element features based on each target platform to determine platform CAD element features, wherein multiple target platforms with the same CAD element features are included; and determining the platform encoding sequence by integrating the platform CAD element features based on multi-head attention, wherein the platform encoding sequence corresponds one-to-one with the target platform.

[0052] The specific configuration of the decoding operator generation module 30 will be described in detail below. The decoding operator generation module 30 further includes: sending the first platform encoded sequence to the corresponding first decoder; performing parallel collapse processing based on multi-head attention on the first platform migration operator to determine the decoding operator group; and integrating the decoding operator group as the first decoding operator.

[0053] The specific configuration of the graph reconstruction storage module 40 will be described in detail below. The graph reconstruction storage module 40 further includes: for the platform encoding sequence, sending it to the corresponding decoder to obtain N decoding operators; according to the data interface interaction between the cross-platform intermediate engine and each target platform, sending the N decoding operators to the corresponding target platform to perform platform graph reconstruction processing and storage.

[0054] The specific configuration of the decoding operator generation module 30 will be described in detail below. The decoding operator generation module 30 further includes: performing intent recognition on the original CAD file to determine intent tags; embedding the intent tags as additional semantic features into the platform encoding sequence; wherein, intent recognition is performed based on a preset rule set: if the feature sequence contains a preset combination, it is marked as a standard connector intent; if the surface continuity and surface change rate meet preset conditions, it is marked as a fluid dynamics intent.

[0055] The specific configuration of the graph matching and integration module 50 will be described in detail below. The graph matching and integration module 50 further includes: uploading project management data from the user terminal, wherein the project management data contains graph element requirements based on project stages; performing graph matching based on each target platform according to the project management data to determine the platform-retrieved data; and distributing the platform-retrieved data to the corresponding target platform to determine the platform graph set.

[0056] The specific configuration of the graph matching and integration module 50 will be described in detail below. The graph matching and integration module 50 further includes: determining a graph display template based on the project management data; scanning the platform graph set, performing CAD element feature extraction and unified structure conversion based on the graph display template to determine the project management graph; and integrating the platform graph set and the project management graph for visualization on a cross-platform display interface.

[0057] The cross-platform CAD drawing intelligent collaborative management system provided in this embodiment of the invention can execute the cross-platform CAD drawing intelligent collaborative management method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0058] Based on the foregoing embodiments, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, enables the implementation of the cross-platform CAD drawing intelligent collaborative management method as described in any of the preceding embodiments.

[0059] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations 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 scope of the present invention.

Claims

1. A cross-platform intelligent collaborative management method for CAD drawings, characterized in that: The method includes: Upload original CAD files and manage tasks; By developing a cross-platform middleware engine and establishing information exchange between the cross-platform middleware engine and various target platforms; The cross-platform intermediate engine scans the original CAD file and management tasks, drives the encoder to perform platform inbound encoding based on CAD element features, drives the decoder to perform optimal expression collapse based on platform migration operators, and generates N decoding operators, where each target platform corresponds to one decoding operator. The N decoding operators are distributed to each target platform to perform platform graph reconstruction and storage. As the real-time project management progresses, graph matching and unified template integration are performed across various target platforms to visualize the project as a real-time project graph.

2. The cross-platform CAD drawing intelligent collaborative management method as described in claim 1, characterized in that, Develop a cross-platform middleware engine, including: An encoder is built based on the underlying logic of CAD element feature extraction and platform attribution encoding; For each target platform, a platform migration operator is constructed by superimposing the state representation of CAD elements on the platform. A decoder array is constructed using optimal expression collapse based on the platform migration operator, wherein each target platform corresponds to one decoder; A multi-head attention mechanism is introduced, wherein the multi-head attention is defined based on geometry, topology, process, and constraints; Based on the encoder, multi-head attention mechanism, and decoder array, a cross-platform intermediate engine is constructed.

3. The cross-platform CAD drawing intelligent collaborative management method as described in claim 2, characterized in that, The driver encoder performs platform inbound encoding based on CAD element features, including: Scan the original CAD file and extract CAD element features; The management task is scanned, and the CAD element features are classified based on each target platform to determine the platform CAD element features. Among them, multiple target platforms with the same CAD element features are included. By integrating the CAD element features of the platform based on multi-head attention, the platform encoding sequence is determined, wherein the platform encoding sequence corresponds one-to-one with the target platform.

4. The cross-platform CAD drawing intelligent collaborative management method as described in claim 3, characterized in that, The driver decoder performs optimal representation collapse based on platform migration operators, including: The first platform encoded sequence is sent to the corresponding first decoder, and parallel collapse processing based on multi-head attention is performed on the first platform transfer operator to determine the decoding operator group; The aforementioned decoding operator group is integrated as the first decoding operator.

5. The cross-platform CAD drawing intelligent collaborative management method as described in claim 4, characterized in that, For the platform's encoded sequence, each sequence is sent to the corresponding decoder to obtain N decoding operators; Based on the data interface interaction between the cross-platform intermediate engine and each target platform, the N decoding operators are sent to the corresponding target platforms to perform platform graph reconstruction processing and storage.

6. The cross-platform CAD drawing intelligent collaborative management method as described in claim 3, characterized in that, The original CAD file is subjected to intent recognition to determine intent tags; The intent label is embedded as an additional semantic feature into the platform encoding sequence; Among them, intent recognition is performed based on a preset rule set: if the feature sequence contains a preset combination, it is marked as a standard connector intent; If the surface continuity and the surface rate of change meet the preset conditions, it is marked as a fluid dynamics intention.

7. The cross-platform CAD drawing intelligent collaborative management method as described in claim 1, characterized in that, As the real-time project management process progresses, graph matching calls based on various target platforms are executed, including: The user uploads project management data, which includes graphical element requirements based on project phases; Based on the project management data, perform graph matching based on each target platform to determine the data to be retrieved by the platform; The platform retrieves data and sends it to the corresponding target platform to determine the platform graph set.

8. The cross-platform CAD drawing intelligent collaborative management method as described in claim 7, characterized in that, Perform unified template integration, including: Based on the project management data, determine the diagram display template; Scan the platform diagram set, perform CAD element feature extraction and unified structure conversion based on the diagram display template to determine the project management diagram; The platform atlas and the project management diagram are integrated and visualized in a cross-platform display interface.

9. A cross-platform intelligent collaborative management system for CAD drawings, characterized in that: The system is used to implement the cross-platform CAD drawing intelligent collaborative management method according to any one of claims 1 to 8, and the system includes: The data upload module is used to upload original CAD files and manage tasks; The information interaction establishment module is used to develop a cross-platform middleware engine and establish information interaction between the cross-platform middleware engine and various target platforms. The decoding operator generation module is used by the cross-platform intermediate engine to scan the original CAD file and management tasks, drive the encoder to perform platform inbound encoding based on CAD element features, drive the decoder to perform optimal expression collapse based on platform migration operators, and generate N decoding operators, where each target platform corresponds to one decoding operator. The graph reconstruction and storage module is used to distribute the N decoded operators to each target platform and perform platform graph reconstruction and storage. The graph matching and integration module is used to perform graph matching calls and unified template integration based on various target platforms as the real-time project management process progresses, and to visualize the real-time project graph.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the cross-platform intelligent collaborative management method for CAD drawings as described in any one of claims 1-8.

Images