Method and device for generating design document in low-code development environment

Abstract

The invention relates to the technical field of seal collection document planning, in particular to a method and device for generating a design document in a low-code development environment, and the method comprises the steps: obtaining the approval rule data of a task region collected by an approval node through a low-code service, and converting the approval rule data into a two-dimensional table; scanning a rendering environment in real time, generating a real-time transaction form, and performing spatial registration on the real-time transaction form and the two-dimensional table to generate a document template; based on the passable area label of the two-dimensional table and the document template, a seal starting set decision state space is constructed; dividing an enabled seal set of a client into seals of a plurality of mechanisms, and performing adaptive dynamic planning in a decision state space of the enabled seal set to generate an enabled seal set sequence; and generating a rendering design document of the client according to the started seal set sequence. According to the method, information searching, screening and problem positioning are carried out by consuming less resources, the page debugging efficiency is high, real-time self-adaptive adjustment in a large-scale environment can be efficiently supported, and the response speed and safety of the client are improved.

Description

Technical Field

[0001] The present application relates to the technical field of seal collection document planning, and in particular to a method and device for generating design documents in a low-code development environment. Background Art

[0002] In low-code development environments, automated design document generation is crucial for improving development efficiency and lowering technical barriers to entry. Traditional design document generation methods often rely on manual editing or automated tools based on fixed templates, making them difficult to adapt to the dynamic business rules and user needs of low-code platforms. Existing technologies are particularly incapable of responding to environmental changes in real time, leading to discrepancies between generated documents and actual business logic.

[0003] Currently, low-code platforms typically use rule engines or workflow engines to handle approval processes, but these methods have the following limitations: Traditional methods hard-code or store approval rules in static configuration files, which makes it difficult to adapt to dynamically changing business needs, especially in scenarios of multi-task area collaboration. Conflicts and priority issues between rules cannot be effectively resolved; the generation of design documents often lags behind the adjustment of business logic, and lacks the ability to map dynamic permissions and seal sequences in real time, resulting in reduced document readability and practicality.

[0004] Therefore, there is an urgent need for a method that can dynamically integrate approval rules, real-time transaction forms and permission management in a low-code environment, and generate highly available design documents through intelligent algorithms to fill the gaps in existing technologies. Summary of the Invention

[0005] The present application provides a method and device for generating design documents in a low-code development environment, which is used to solve the problems of poor code-document coupling capability for dynamic permissions and low computational efficiency of spatial seal set search in the prior art.

[0006] In a first aspect, the present application provides a method for generating design documents in a low-code development environment, comprising: Use low-code services to obtain approval rule data for the task area collected by the approval node, and convert the approval rule data into a two-dimensional table; The page component carried by the client scans the rendering environment in real time to generate a real-time transaction form, and spatially aligns the real-time transaction form with the two-dimensional table to generate a document template containing dynamic permissions; Constructing a seal activation set decision state space based on the passable area annotations of the two-dimensional table and the document template; A deep Q-network model is used to divide the client's enabled seal set into seals of multiple institutions. Adaptive dynamic programming is performed in the enabled seal set decision state space in combination with a greedy strategy to generate an enabled seal set sequence. A rendering design document of the client is generated according to the activation seal set sequence.

[0007] Optionally, the deep Q network model is used to divide the client's enabled seal set into seals of multiple institutions, and a greedy strategy is used to perform adaptive dynamic programming in the enabled seal set decision state space to generate an enabled seal set sequence, including: A deep Q-network model is used to process the address sequence of the enabled seal set to generate a set of segmentation point location addresses; Based on the segmentation point location address set, segmenting the activated seal set from a starting point to an end point into a set of seals comprising multiple institutions; For each institution's seal, extract the seal's starting point address, construct a local address system with the starting point address as the origin, discretize the seal under the local address system, and generate a seal set point sequence; Defining authority level features according to the seal set point sequence, extracting corresponding area data from the enabled seal set decision state space, and obtaining a local seal set decision state space as the seal set decision state space associated with the seal; Using a greedy strategy, adaptive dynamic programming is performed in the seal set decision state space associated with the seal of each institution to generate a planned seal address sequence; According to the arrangement order of the split points, the corresponding institution index is assigned to the planned seal address sequence of all institutions to form an enabled seal set sequence.

[0008] Optionally, the method of utilizing a greedy strategy to perform adaptive dynamic programming in the seal set decision state space associated with the seal of each institution to generate a planned seal address sequence includes: Using a greedy strategy, traverse adjacent states in the local stamp set decision state space, calculate the action value function value, and select the state corresponding to the maximum value of the action value function to form a state transition chain consisting of an action selection identifier sequence; Generate a state transfer stamp set based on an identifier sequence selected from actions in the state transfer chain; Converting the state transfer stamp set into the original address space stamp set through the mapping relationship between the local address system and the enabled address system; Based on the original address space stamp set, a planned stamp address sequence is generated.

[0009] Optionally, the greedy strategy is used to traverse adjacent states in the local stamp set decision state space, calculate the action value function value, and select the state corresponding to the maximum value of the action value function to form a state transition chain consisting of an action selection identifier sequence, including: Converting the enabled address of the current state in the local seal set decision state space into a local relative address, and generating a mapping table between state identifiers and local relative addresses; Extracting eight neighborhood address points according to the local relative address of the current state in the mapping table and screening the adjacent states that fall within the local seal set decision state space to generate an adjacent state identifier set; Calculating the absolute deviation value and the inverse of the nearest authority Euclidean distance of each element of each state in the adjacent state identifier set, and performing weighted summation to generate an action value function value; Comparing the action-value function values, and selecting a state identifier corresponding to a maximum value of the action-value function to generate a selected state identifier for inter-state transition; The selected state identifier is set as a new current state, and the screening, calculation, and selection operations are repeatedly performed until the end point state is reached, and a state transfer chain consisting of a sequence of action selection identifiers is output.

[0010] Optionally, constructing a seal activation set decision state space based on the passable area annotations in the two-dimensional table and the document template includes: Calculating the approval rule change rates of adjacent grids in the two-dimensional table to generate an approval rule change rate grid graph; Screening candidate grids whose approval rule change rates are lower than a preset change rate threshold from the approval rule change rate grid map, performing connected area detection and area threshold filtering on adjacent candidate grids, and generating a passable area annotation set; Performing spatial clustering identification on the transaction cluster center addresses of the document template to obtain a transaction cluster set, and performing calculation based on the two-dimensional geometric center address of each cluster in the transaction cluster set to generate a transaction cluster center address set; Based on the transaction cluster center address set, the inverse of the Euclidean distance from each grid to the nearest transaction cluster center is calculated and risk analysis is performed to generate a risk level identifier set; The grid addresses of the approval rule change rate grid graph are associated with the corresponding pass states in the passable area annotation set and the corresponding risk levels in the risk level identifier set to obtain the seal activation set decision state space.

[0011] Optionally, associating the grid addresses of the approval rule change rate grid graph with the corresponding pass states in the passable area annotation set and the corresponding risk levels in the risk level identifier set to obtain the seal activation set decision state space includes: Creating a state attribute structure set according to the passable area label set and the risk level identifier in the risk level identifier set; Establish a mapping relationship between the grid address of the two-dimensional table and the state attribute structure set, and construct the decision state space for enabling the seal set.

[0012] Optionally, based on the transaction cluster center address set, the inverse of the Euclidean distance from each grid to the nearest transaction cluster center is calculated and risk analysis is performed to generate a risk level identifier set, including: Calculating the minimum Euclidean distance from each grid in the two-dimensional table to the nearest transaction cluster center address set and taking the reciprocal to generate a risk degree coefficient set; The risk degree coefficient set is divided according to the preset interval to generate a risk level identifier set.

[0013] Furthermore, the generating of the rendering design document of the client according to the activation seal set sequence further includes: Construct at least one low-code data segment, wherein each of the low-code data segments includes basic attributes, and the basic attributes include at least a primary key value, an attribute name, and an attribute content; Obtaining a business document uploaded by a user, parsing the business document into encryption components, and rendering the encryption components into an online editable document, wherein each editing unit of the business document has a unique identification in the encryption component; Selecting an editing unit to be edited as a selected editing unit, and selecting at least one low-code data segment corresponding to the selected editing unit as a selected low-code data segment; The selected low-code data segment and the selected editing unit are bound according to the unique identification identifier of the selected editing unit and the primary key value of the selected low-code data segment, and the attribute content of the selected low-code data segment is inserted into the corresponding encrypted component of the selected editing unit to form an updated component; the non-updated component and the updated component are rendered to form a personalized document.

[0014] In a second aspect, the present application provides a device for generating design documents in a low-code development environment, comprising: An acquisition module is used to use a low-code service to acquire approval rule data of a task area collected by an approval node, and convert the approval rule data into a two-dimensional table; A matching module is used to scan the rendering environment in real time through a page component carried by the client, generate a real-time transaction form, and spatially align the real-time transaction form with the two-dimensional table to generate a document template containing dynamic permissions; A construction module, configured to construct a seal activation set decision state space based on the passable area annotations in the two-dimensional table and the document template; A generation module is used to divide the client's enabled seal set into seals of multiple institutions using a deep Q-network model, and perform adaptive dynamic programming in the enabled seal set decision state space in combination with a greedy strategy to generate an enabled seal set sequence; The document design module generates a rendering design document for the client according to the activation seal set sequence.

[0015] In the present application, a method for generating design documents in a low-code development environment is provided, the method comprising: using a low-code service to obtain approval rule data of a task area collected by an approval node, and converting the approval rule data into a two-dimensional table; using a page component carried by a client to scan the rendering environment in real time, generate a real-time transaction form, and spatially align the real-time transaction form with the two-dimensional table to generate a document template containing dynamic permissions; based on the passable area annotation of the two-dimensional table and the document template, construct an enabled seal set decision state space; using a deep Q network model, divide the client's enabled seal set into seals of multiple institutions, and combine the greedy strategy to perform adaptive dynamic planning in the enabled seal set decision state space to generate an enabled seal set sequence; based on the enabled seal set sequence, generate a rendered design document for the client.

[0016] This application constructs a high-precision static environment model through the two-dimensional grid conversion of approval node approval rule data, providing a reliable spatial benchmark for the planning of activated seal sets; with the help of spatial alignment of page component real-time transactions and static data tables, it realizes the fusion perception of dynamic and static environments, and forms a dynamic code document coupling layer containing sudden permissions; the activated seal set decision state space constructed based on the passable area annotation and real-time code document coupling layer unifies the environmental constraints and dynamic variables; finally, the deep Q network's branch seal set cutting mechanism is adopted, combined with the greedy strategy to perform adaptive dynamic planning in the state space, and simultaneously achieve the triple goals of adaptability to complex terrain, real-time avoidance of dynamic obstacles and optimal energy of activated seal sets, thereby improving the client's autonomous decision-making efficiency and rendering security in unknown dynamic environments.

[0017] Furthermore, a deep Q-network model is first used to process the enabled seal set address sequence to generate a key segmentation point position address set, and the enabled seal set is divided into seal sets of multiple continuous institutions accordingly; for each seal, a local address system is constructed with its starting point as the origin, the seal is discretized to generate a seal set point sequence, and then the authority level characteristics are defined based on the seal set point sequence, and the corresponding area data is extracted from the enabled seal set decision state space to form a local seal set decision state space; then, a greedy strategy is used to perform adaptive dynamic planning in the local state space, and the action value function value is calculated by traversing adjacent states and selecting the state corresponding to the maximum value to generate a state transition chain consisting of an action selection identifier sequence, based on which the state transition seal set is derived, and then converted into the original address space seal set through the mapping relationship between the local and enabled address systems, and finally a planned seal address sequence is generated; the planned seals of all institutions are assigned institution indexes and reorganized in the order of segmentation points to form a complete enabled seal set sequence. The computational dimension is greatly reduced through the seal set segmentation driven by the deep Q network and the state space construction under the local address system. Combined with the greedy strategy, the state transition seal set that maximizes the action value is quickly selected in the local space to achieve millisecond-level dynamic response capability. At the same time, the feature constraints based on the permission level ensure the smoothness and robustness of the seal set, and the local to enabled address mapping and institutional index reorganization mechanism ensure that each segmented seal set is seamlessly connected to enable the optimal seal set, thereby simultaneously improving planning efficiency, security and adaptability in a complex dynamic environment.

[0018] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 A flowchart of a method for generating design documents in a low-code development environment provided in an embodiment of the present application; Figure 2 A structural diagram of an apparatus for generating design documents in a low-code development environment provided in an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0022] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 11, 12, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0023] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0024] Figure 1 A flowchart of a method for generating design documents in a low-code development environment provided in an embodiment of the present application is shown as follows: Figure 1 As shown, the method includes: S11. Use low-code services to obtain the approval rule data of the task area collected by the approval node, and convert the approval rule data into a two-dimensional table.

[0025] S12. Scan the rendering environment in real time through the page component carried by the client to generate a real-time transaction form, and spatially align the real-time transaction form with the two-dimensional table to generate a document template containing dynamic permissions.

[0026] S13. Based on the passable area annotation and document template of the two-dimensional table, a decision state space for enabling the seal set is constructed.

[0027] S14. Using the deep Q network model, the client's enabled seal set is divided into seals of multiple institutions. Combined with the greedy strategy, adaptive dynamic programming is performed in the enabled seal set decision state space to generate an enabled seal set sequence.

[0028] The Deep Q-Network model is a reinforcement learning algorithm that combines deep neural networks and Q-learning to learn optimal action strategies. The enabled stamp set refers to the complete rendering route of the client from the starting point to the end point. Stamps refer to the subunits into which the enabled stamp set is divided, with each segment representing a decision-making mechanism, including short-range stamp set segments. A greedy strategy is a method for selecting the optimal action in a decision, including selecting the action with the maximum Q value in each state to locally optimize the stamp set. Adaptive dynamic programming is an algorithm that adjusts the planning process based on environmental changes. The enabled stamp set sequence is an ordered list of generated stamp set points.

[0029] In an embodiment of the present application, a deep Q-network model is first used to divide the enabled seal set into multiple continuous seals according to a distance threshold; secondly, for each seal, a local decision domain is delineated in the decision state space of the enabled seal set with the starting point as the reference; then a greedy strategy is used to select the state node with the highest action value in the local decision domain; finally, the optimized seal sets of each mechanism are generated through state node backtracking, and connected in series to form an enabled seal set sequence.

[0030] S15, generating a rendering design document of the client according to the activation seal set sequence; Among them, low-code components that meet business needs are designed, and the low-code components and basic functional components are integrated based on the business to form a business form interface; the data of the basic attributes corresponding to the low-code components on the business form interface are collected to obtain a low-code data segment.

[0031] Fill in the attribute content in the basic attributes of the low-code component of the business form interface, and collect the attribute content to obtain the low-code data segment. The data content of the low-code data segment is stored in the database table.

[0032] The same low-code data segment includes at least one basic attribute, and the low-code data segment includes an attribute combination method that defines the combination method of the basic attributes.

[0033] The unique identification mark is composed of a timestamp of importing the business document and a position of the editing unit in the business document.

[0034] Identify the text and the default configuration of the text in the business document, and convert the text into an encryption component according to the default configuration; identify the table, sub-cell and the default configuration of the table in the business document, and convert the table and the sub-cell into an encryption component according to the default configuration of the table.

[0035] A binding relationship string is generated based on the unique identification identifier and the primary key value, wherein the binding relationship string includes at least the unique identification identifier and the primary key value, and the attribute content of the corresponding selected low-code data segment is retrieved based on the binding relationship string, and the attribute content is dynamically inserted into the corresponding encryption component to form an update component.

[0036] By executing S11 to S15, the embodiment of the present application realizes efficient and secure rendering of the client in a dynamic environment by integrating approval node data and real-time page component scanning, can automatically adapt to changes in permissions, optimize and enable seal collection planning, reduce document generation risks, and improve task execution efficiency and reliability.

[0037] Optionally, the deep Q network model is used to divide the client's enabled seal set into seals of multiple institutions, and a greedy strategy is used to perform adaptive dynamic programming in the enabled seal set decision state space to generate an enabled seal set sequence, including: A deep Q-network model is used to process the address sequence of the enabled seal set to generate a set of segmentation point location addresses; Based on the segmentation point location address set, segmenting the activated seal set from a starting point to an end point into a set of seals comprising multiple institutions; For each institution's seal, extract the seal's starting point address, construct a local address system with the starting point address as the origin, discretize the seal under the local address system, and generate a seal set point sequence; Defining authority level features according to the seal set point sequence, extracting corresponding area data from the enabled seal set decision state space, and obtaining a local seal set decision state space as the seal set decision state space associated with the seal; Using a greedy strategy, adaptive dynamic programming is performed in the seal set decision state space associated with the seal of each institution to generate a planned seal address sequence; According to the arrangement order of the split points, the corresponding institution index is assigned to the planned seal address sequence of all institutions to form an enabled seal set sequence.

[0038] In this example, the low-code platform needs to automatically generate design documents for a multi-level approval system. This system involves multiple approval nodes (such as submission, preliminary review, re-review, and final review), each with different permission rules and seal call logic. Traditional methods struggle to dynamically optimize the seal call sequence, resulting in inefficient document generation. This example uses a deep Q-network (DQN) coupled with a greedy strategy for adaptive planning to generate the optimal seal sequence.

[0039] Optionally, the method of utilizing a greedy strategy to perform adaptive dynamic programming in the seal set decision state space associated with the seal of each institution to generate a planned seal address sequence includes: Using a greedy strategy, traverse adjacent states in the local stamp set decision state space, calculate the action value function value, and select the state corresponding to the maximum value of the action value function to form a state transition chain consisting of an action selection identifier sequence; Generate a state transfer stamp set based on an identifier sequence selected from actions in the state transfer chain; Converting the state transfer stamp set into the original address space stamp set through the mapping relationship between the local address system and the enabled address system; Based on the original address space stamp set, a planned stamp address sequence is generated.

[0040] Optionally, the greedy strategy is used to traverse adjacent states in the local stamp set decision state space, calculate the action value function value, and select the state corresponding to the maximum value of the action value function to form a state transition chain consisting of an action selection identifier sequence, including: Converting the enabled address of the current state in the local seal set decision state space into a local relative address, and generating a mapping table between state identifiers and local relative addresses; Extracting eight neighborhood address points according to the local relative address of the current state in the mapping table and screening the adjacent states that fall within the local seal set decision state space to generate an adjacent state identifier set; Calculating the absolute deviation value and the inverse of the nearest authority Euclidean distance of each element of each state in the adjacent state identifier set, and performing weighted summation to generate an action value function value; Comparing the action-value function values, and selecting a state identifier corresponding to a maximum value of the action-value function to generate a selected state identifier for inter-state transition; The selected state identifier is set as a new current state, and the screening, calculation, and selection operations are repeatedly performed until the end point state is reached, and a state transfer chain consisting of a sequence of action selection identifiers is output.

[0041] Optionally, constructing a seal activation set decision state space based on the passable area annotations in the two-dimensional table and the document template includes: Calculating the approval rule change rates of adjacent grids in the two-dimensional table to generate an approval rule change rate grid graph; Screening candidate grids whose approval rule change rates are lower than a preset change rate threshold from the approval rule change rate grid map, performing connected area detection and area threshold filtering on adjacent candidate grids, and generating a passable area annotation set; Performing spatial clustering identification on the transaction cluster center addresses of the document template to obtain a transaction cluster set, and performing calculation based on the two-dimensional geometric center address of each cluster in the transaction cluster set to generate a transaction cluster center address set; Based on the transaction cluster center address set, the inverse of the Euclidean distance from each grid to the nearest transaction cluster center is calculated and risk analysis is performed to generate a risk level identifier set; The grid addresses of the approval rule change rate grid graph are associated with the corresponding pass states in the passable area annotation set and the corresponding risk levels in the risk level identifier set to obtain the seal activation set decision state space.

[0042] Optionally, associating the grid addresses of the approval rule change rate grid graph with the corresponding pass states in the passable area annotation set and the corresponding risk levels in the risk level identifier set to obtain the seal activation set decision state space includes: Creating a state attribute structure set according to the passable area label set and the risk level identifier in the risk level identifier set; Establish a mapping relationship between the grid address of the two-dimensional table and the state attribute structure set, and construct the decision state space for enabling the seal set.

[0043] Optionally, based on the transaction cluster center address set, the inverse of the Euclidean distance from each grid to the nearest transaction cluster center is calculated and risk analysis is performed to generate a risk level identifier set, including: Calculating the minimum Euclidean distance from each grid in the two-dimensional table to the nearest transaction cluster center address set and taking the reciprocal to generate a risk degree coefficient set; The risk degree coefficient set is divided according to the preset interval to generate a risk level identifier set.

[0044] Furthermore, the generating of the rendering design document of the client according to the activation seal set sequence further includes: Construct at least one low-code data segment, wherein each of the low-code data segments includes basic attributes, and the basic attributes include at least a primary key value, an attribute name, and an attribute content; Obtaining a business document uploaded by a user, parsing the business document into encryption components, and rendering the encryption components into an online editable document, wherein each editing unit of the business document has a unique identification in the encryption component; Selecting an editing unit to be edited as a selected editing unit, and selecting at least one low-code data segment corresponding to the selected editing unit as a selected low-code data segment; The selected low-code data segment and the selected editing unit are bound according to the unique identification identifier of the selected editing unit and the primary key value of the selected low-code data segment, and the attribute content of the selected low-code data segment is inserted into the corresponding encrypted component of the selected editing unit to form an updated component; the non-updated component and the updated component are rendered to form a personalized document.

[0045] Among them, in this embodiment, the scenario is that a development zone in a certain city needs to obtain approval from three departments: land, environmental protection, and fire protection.

[0046] During specific operations, the low-code platform obtains the rules of each department: the Land Bureau requires "non-agricultural land", the Environmental Protection Bureau requires "noise <60 decibels", and the Fire Department requires "channels ≥2".

[0047] Map the rules to a virtual coordinate grid: for example, coordinate (1,1) represents industrial land (which complies with the Land Bureau's rules), and coordinate (1,2) represents agricultural land (which violates the Land Bureau's rules). This forms a regular grid map covering the entire area. Engineers scan the code and fill in the form on site; Scan the construction site with your mobile phone to generate a real-time form: including location coordinates (1,1), industrial land attributes, noise level 55 decibels, 3 channels and other data.

[0048] The system matches the form coordinates (1,1) with the regular grid: it automatically unlocks the approval fields for the Land, Environmental Protection, and Fire Departments, and generates a dynamic permission document template. If the coordinates (1,2) (agricultural land) are matched, only the Land Bureau field is unlocked.

[0049] When constructing the decision state space, the rate of change of adjacent areas of the regular grid is calculated. For example, industrial land areas tend to be regular and stable (with a low rate of change), while the boundary between agriculture and industry has a high rate of change. Areas with a rate of change below a threshold are screened and marked as "traversable" (e.g., areas with concentrated industrial land).

[0050] Identify clusters of approval fields in a document template. For example, the Land Bureau fields are clustered in the upper left area, while the Fire Department fields are in the lower right area. Calculate the coordinates of each cluster center.

[0051] Using the gathering center as a benchmark, calculate the inverse distance from the grid to the nearest center. The closer the distance, the higher the risk (for example, areas with concentrated fire stalls require special review), and the risk levels are categorized as high, medium, and low.

[0052] Bind the "passable state" and "risk level" of each grid to form a decision state space. For example, coordinate (1,1) is marked as "passable + low risk".

[0053] When generating a stamp set sequence, the deep Q network divides the stamp set into 20 locations (e.g., 20 coordinate points) by department. For example, the first 5 points belong to the Land Bureau, the middle 10 points to the Environmental Protection Bureau, and the last 5 points to the Fire Department.

[0054] Local path planning: A local coordinate system is established with the first coordinate point (1,1) of the Land Bureau as the origin, and the five points of the bureau are converted to relative positions (e.g. point A: (0,0), point B: (0,1)).

[0055] In the local decision space, the greedy strategy is used to select the optimal path: Starting from point A, detect the adjacent points in 8 directions around it.

[0056] Calculate the value of each action (for example, moving due north shortens the distance to the target and has the lowest risk, so it has the highest value); Select the point with the highest value to move to, gradually forming a path chain (such as A→B→C); Convert the local path coordinates back to the original grid coordinates; Generate sequence: Merge paths in the order of Land Bureau → Environmental Protection Bureau → Fire Bureau, and output a complete stamp sequence; When users upload project plans, they need to insert approval data.

[0057] The system parses the planning map into editable units (such as the "Land Use Type" and "Fire Egress" columns), and each unit is assigned an independent ID.

[0058] Create a low-code data segment: for example, the data segment primary key is "LAND-001", the attribute name is "land use nature", and the attribute value is "industrial land".

[0059] Bind data: Select the "Land Use Type" column (ID: AREA-01), associate the data segment "LAND-001", and insert the "Industrial Land" value into the column.

[0060] Rendering result: "Industrial Land" is automatically displayed in the "Land Use Type" column of the planning map, and departmental electronic seals are added to other columns according to the seal sequence to generate the final approval document; After the construction site workers scan the code, the system automatically generates a document containing approval columns for the three departments and adds an electronic seal according to the optimal path. There is no need to manually configure the code throughout the process.

[0061] Figure 2 A schematic diagram of a structure of a device for generating design documents in a low-code development environment provided in an embodiment of the present application is shown as follows: Figure 2 As shown, the device includes: Acquisition module 21 is used to use low-code services to obtain the approval rule data of the task area collected by the approval node, and convert the approval rule data into a two-dimensional table.

[0062] The matching module 22 is used to scan the rendering environment in real time through the page component carried by the client, generate a real-time transaction form, and spatially align the real-time transaction form with the two-dimensional table to generate a document template containing dynamic permissions.

[0063] The construction module 23 is used to construct the decision state space of the enabled seal set based on the passable area annotation and the document template in the two-dimensional table.

[0064] A generation module 24 is configured to use a deep Q-network model to divide the client's enabled seal set into seals of multiple institutions, and to perform adaptive dynamic programming in the enabled seal set decision state space in combination with a greedy strategy to generate an enabled seal set sequence; The document design module 25 generates a rendering design document for the client according to the activation seal set sequence.

[0065] Figure 2 The device for generating design documents in a low-code development environment can execute Figure 1 The implementation principle and technical effects of the method for generating design documents in a low-code development environment described in the embodiment shown will not be repeated here. The specific manner in which each module and unit performs operations in the device for generating design documents in a low-code development environment in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0066] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0067] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

[0069] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some seals, devices or units, which can be electrical, mechanical or other forms.

[0070] In addition, the functional units in the various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units. The above is only an implementation method of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the description and drawings of this application, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

[0071] The above detailed description of the specific embodiments of the invention is intended only as an example, and the present application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions of the invention are also within the scope of the present application. Therefore, equivalent changes, modifications, and improvements made without departing from the spirit and scope of the present application should be included within the scope of the present application.

Claims

1. A method for generating design documents in a low-code development environment, characterized in that: include: Use low-code services to obtain approval rule data for the task area collected by the approval node, and convert the approval rule data into a two-dimensional table; The page component carried by the client scans the rendering environment in real time to generate a real-time transaction form, and spatially aligns the real-time transaction form with the two-dimensional table to generate a document template containing dynamic permissions; Constructing a seal activation set decision state space based on the passable area annotations of the two-dimensional table and the document template; A deep Q-network model is used to divide the client's enabled seal set into seals of multiple institutions. Adaptive dynamic programming is performed in the enabled seal set decision state space in combination with a greedy strategy to generate an enabled seal set sequence. A rendering design document of the client is generated according to the activation seal set sequence.

2. The method according to claim 1, characterized in that The deep Q network model is used to divide the client's enabled seal set into seals of multiple institutions, and a greedy strategy is combined to perform adaptive dynamic programming in the enabled seal set decision state space to generate an enabled seal set sequence, including: A deep Q-network model is used to process the address sequence of the enabled seal set to generate a set of segmentation point location addresses; Based on the segmentation point location address set, segmenting the activated seal set from a starting point to an end point into a set of seals comprising multiple institutions; For each institution's seal, extract the seal's starting point address, construct a local address system with the starting point address as the origin, discretize the seal under the local address system, and generate a seal set point sequence; Defining authority level features according to the seal set point sequence, extracting corresponding area data from the enabled seal set decision state space, and obtaining a local seal set decision state space as the seal set decision state space associated with the seal; Using a greedy strategy, adaptive dynamic programming is performed in the seal set decision state space associated with the seal of each institution to generate a planned seal address sequence; According to the arrangement order of the split points, the corresponding institution index is assigned to the planned seal address sequence of all institutions to form an enabled seal set sequence.

3. The method according to claim 2, characterized in that The method utilizes a greedy strategy to perform adaptive dynamic programming in the seal set decision state space associated with the seal of each institution to generate a planned seal address sequence, including: Using a greedy strategy, traverse adjacent states in the local stamp set decision state space, calculate the action value function value, and select the state corresponding to the maximum value of the action value function to form a state transition chain consisting of an action selection identifier sequence; Generate a state transfer stamp set based on an identifier sequence selected from actions in the state transfer chain; Converting the state transfer stamp set into the original address space stamp set through the mapping relationship between the local address system and the enabled address system; Based on the original address space stamp set, a planned stamp address sequence is generated.

4. The method according to claim 3, characterized in that The greedy strategy is used to traverse adjacent states in the local stamp set decision state space, calculate the action value function value, and select the state corresponding to the maximum value of the action value function to form a state transfer chain consisting of an action selection identifier sequence, including: Converting the enabled address of the current state in the local seal set decision state space into a local relative address, and generating a mapping table between state identifiers and local relative addresses; Extracting eight neighborhood address points according to the local relative address of the current state in the mapping table and screening the adjacent states that fall within the local seal set decision state space to generate an adjacent state identifier set; Calculating the absolute deviation value and the inverse of the nearest authority Euclidean distance of each element of each state in the adjacent state identifier set, and performing weighted summation to generate an action value function value; Comparing the action-value function values, and selecting a state identifier corresponding to a maximum value of the action-value function to generate a selected state identifier for inter-state transition; The selected state identifier is set as a new current state, and the screening, calculation, and selection operations are repeatedly performed until the end point state is reached, and a state transfer chain consisting of a sequence of action selection identifiers is output.

5. The method according to claim 1, characterized in that The step of constructing a seal activation set decision state space based on the passable area annotation of the two-dimensional table and the document template includes: Calculating the approval rule change rates of adjacent grids in the two-dimensional table to generate an approval rule change rate grid graph; Screening candidate grids whose approval rule change rates are lower than a preset change rate threshold from the approval rule change rate grid map, performing connected area detection and area threshold filtering on adjacent candidate grids, and generating a passable area annotation set; Performing spatial clustering identification on the transaction cluster center addresses of the document template to obtain a transaction cluster set, and performing calculation based on the two-dimensional geometric center address of each cluster in the transaction cluster set to generate a transaction cluster center address set; Based on the transaction cluster center address set, the inverse of the Euclidean distance from each grid to the nearest transaction cluster center is calculated and risk analysis is performed to generate a risk level identifier set; The grid addresses of the approval rule change rate grid graph are associated with the corresponding pass states in the passable area annotation set and the corresponding risk levels in the risk level identifier set to obtain the seal activation set decision state space.

6. The method according to claim 5, characterized in that The step of associating the grid address of the approval rule change rate grid graph with the corresponding pass status in the passable area annotation set and the corresponding risk level in the risk level identifier set to obtain the seal activation set decision state space includes: Creating a state attribute structure set according to the passable area label set and the risk level identifier in the risk level identifier set; Establish a mapping relationship between the grid address of the two-dimensional table and the state attribute structure set, and construct the decision state space for enabling the seal set.

7. The method according to claim 5, characterized in that The method of calculating the inverse of the Euclidean distance from each grid to the nearest transaction cluster center based on the transaction cluster center address set and performing risk analysis to generate a risk level identifier set includes: Calculating the minimum Euclidean distance from each grid in the two-dimensional table to the nearest transaction cluster center address set and taking the reciprocal to generate a risk degree coefficient set; The risk degree coefficient set is divided according to the preset interval to generate a risk level identifier set.

8. The method according to claim 5, characterized in that The step of generating the rendering design document of the client according to the activation seal set sequence further includes: Construct at least one low-code data segment, wherein each of the low-code data segments includes basic attributes, and the basic attributes include at least a primary key value, an attribute name, and an attribute content; Obtaining a business document uploaded by a user, parsing the business document into encryption components, and rendering the encryption components into an online editable document, wherein each editing unit of the business document has a unique identification in the encryption component; Selecting an editing unit to be edited as a selected editing unit, and selecting at least one low-code data segment corresponding to the selected editing unit as a selected low-code data segment; The selected low-code data segment and the selected editing unit are bound according to the unique identification identifier of the selected editing unit and the primary key value of the selected low-code data segment, and the attribute content of the selected low-code data segment is inserted into the corresponding encrypted component of the selected editing unit to form an updated component; the non-updated component and the updated component are rendered to form a personalized document.

9. A device for generating design documents in a low-code development environment, used to execute a method for generating design documents in a low-code development environment as described in any one of claims 1 to 8, characterized in that: include: An acquisition module is used to use a low-code service to acquire approval rule data of a task area collected by an approval node, and convert the approval rule data into a two-dimensional table; A matching module is used to scan the rendering environment in real time through a page component carried by the client, generate a real-time transaction form, and spatially align the real-time transaction form with the two-dimensional table to generate a document template containing dynamic permissions; A construction module, configured to construct a seal activation set decision state space based on the passable area annotations in the two-dimensional table and the document template; A generation module is used to divide the client's enabled seal set into seals of multiple institutions using a deep Q-network model, and perform adaptive dynamic programming in the enabled seal set decision state space in combination with a greedy strategy to generate an enabled seal set sequence; The document design module generates a rendering design document for the client according to the activation seal set sequence.

Images