A scientific research consultation management low code intelligent development and scheduling platform

Abstract

The application discloses a scientific research consultation management low code intelligent development and arrangement platform, relates to the technical field of scientific research consultation management, and constructs a business component related to scientific research consultation business through a visual configuration mode; generates the exposure of the business component based on historical consultation data of the scientific research consultation, constructs a component whiteboard based on the exposure, generates a component matching list based on component matching rules according to the consultation demand input by a user, searches in the component whiteboard based on the component matching list, generates an executable process, a metadata engine module receives the executable process and executes, generates a corresponding business application, dynamically renders an application interface, and constructs an external interface for the interaction between the management platform and an external system; the application realizes the automatic precipitation and contour matching reproduction of business experience through the component whiteboard and the copy layer, effectively reduces the construction threshold of the scientific research consultation management business, and simultaneously improves the arrangement efficiency.

Description

Technical Field

[0001] This invention relates to the field of scientific research consulting management technology, specifically to a low-code intelligent development and orchestration platform for scientific research consulting management. Background Technology

[0002] Existing low-code development platforms abstract business objects, page layouts, process nodes, and business rules into configurable metadata through visual configuration, and encapsulate them into reusable business components. Users can select the required business components on the platform through drag-and-drop and connection operations, arrange the calling order and data flow between components, and generate executable applications for specific business scenarios. However, existing technologies have certain shortcomings. The organization of business components relies on manual maintenance, and the positions of components in directories or lists are fixed, failing to reflect the frequency and importance of component usage in actual business. As the number of business components increases, frequently used core components and infrequently used peripheral components are mixed in the same level, continuously increasing the cost for users to find target components and making the positioning of orchestration starting points inefficient. At the same time, the platform's template reuse mechanism only supports the overall calling of saved applications and cannot automatically capture implicit orchestration experience such as the topological relationships, interaction interfaces, and connection paths between components during each orchestration process. When faced with new consulting needs, users still need to rely on personal memory to determine whether there are historical solutions that can be referenced. It is impossible to achieve automatic accumulation, structural matching, and rapid reproduction of orchestration experience at the system level, making it difficult to continuously accumulate and reuse the methodology of scientific research consulting business. Therefore, developing a low-code intelligent development and orchestration platform for scientific research consulting management is of great significance. Summary of the Invention

[0003] The purpose of this invention is to provide a low-code intelligent development and orchestration platform for scientific research consulting and management, so as to solve the problems in the background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a low-code intelligent development and orchestration platform for scientific research consulting management, comprising: Application Development Module: Used to build business components related to scientific research consulting services through visual configuration; Component Whiteboard Module: Based on historical consulting data from scientific research consulting, the exposure of business components is generated, and a component whiteboard is built based on the exposure. Development support module: Used to generate a component matching list based on the user's input consultation needs and component matching rules; Business orchestration module: Based on the component matching list, it searches the component whiteboard to generate an executable process; Metadata engine module: Used to store and parse metadata, receive and execute executable processes, generate corresponding business applications, and dynamically render the application interface; Integrated open modules: Build external interfaces for managing the interaction between the platform and external systems.

[0005] In a preferred embodiment, the application development module includes: The model building unit, through a graphical configuration interface, defines the business objects, field attributes, and object relationships of scientific research consulting services, and generates data model metadata. The page design unit is used to configure the presentation format of data model metadata and generate presentation metadata. The presentation formats include form format, list format, and Kanban format; The process design unit is used to define the workflow of scientific research consulting business through visual connection configuration and generate process definition metadata; The workflow includes approval nodes, task assignment nodes, and conditional branch nodes. The rule configuration unit is used to define business rules and generate business rule metadata through the rule table. The business rules include business triggering conditions, data verification rules, and automatic allocation rules. The data encapsulation unit is used to encapsulate data model metadata, representation metadata, process definition metadata, and business rule metadata into business components.

[0006] In a preferred embodiment, the component whiteboard module includes: The whiteboard building module constructs a multi-layer mapping space, which, from high to low, includes the component layer, interface layer, connection layer, and overwrite layer. The component layer contains multiple pre-sized square seats; The component deployment unit collects the exposure of each business component based on historical scientific research and consulting data. The exposure is used to characterize the usage of the business component in the preset amount of consulting data. Sorting business components based on exposure, generating a business component sorting table; Based on the business component sorting table, business components are deployed to the grid seats in the component layer. The higher the ranking of the business component, the closer it is to the center area of ​​the component layer. The connection management unit deploys corresponding interaction interfaces for business components deployed at the component layer at the interface layer. Based on the logical association of business components, the interaction interfaces between corresponding business components are connected in the connection layer to generate corresponding business interaction channels; The rewrite building unit is used to map the application profile for each business application build.

[0007] In a preferred embodiment, the step of deploying a rewrite subunit in the rewrite layer to map the application profile for each business application build is as follows: A monitoring subunit is built in the overwrite layer. When the monitoring subunit detects that a business interaction channel in the connection layer is called, the overwrite layer is activated and an overwrite subunit is generated. The copy subunit is used to record the call data in this business application construction. The call data includes the corresponding business components, interaction interfaces and business interaction channels. The call data is mapped to outline data and recorded on the copy sub-unit. The outline data includes the grid seat outline, the interaction interface outline, and the business interaction channel outline. When multiple copy sub-units are generated, similarity matching is performed on the multiple copy sub-units to extract common contours and difference contours. The copying subunit with the largest common contour range is taken as the main copying subunit. Based on the main copying subunit, the differential contours are connected layer by layer with the activation channel to generate the copying template.

[0008] In a preferred embodiment, the development assistance module includes: The intent recognition unit is used to receive the user's consultation request input in natural language, extract the business intent in the consultation request based on the neural network, and generate a structured intent expression. The business intent includes the business type, core entity, target product, and constraints. The requirement expansion unit, based on the structured intent expression and the knowledge graph of the scientific research field, obtains the expansion requirements that are implicitly related to the current requirement. Write the extended requirements into a structured intent statement to generate an extended requirements set; The component mapping unit matches the extended requirement set with the preset component matching rules to generate a component matching list.

[0009] In a preferred embodiment, the service orchestration module includes: The outline retrieval unit generates the corresponding grid seat outline in the component whiteboard based on the component matching list, and matches the corresponding grid seat outline with the copy template; If the match is successful, the application profile data of the copy subunit is extracted and mapped back to the component layer, interface layer and connection layer of the component whiteboard module, and the business components in the corresponding grid seat are activated to generate an executable process. If the matching fails, the corresponding business component and the corresponding interaction interface in the component matching list will be activated, and a business interaction channel between the corresponding business components will be built to generate an executable process.

[0010] In a preferred embodiment, the metadata engine module includes: Metadata storage unit, used to store metadata of business components, build update interface for updating and modifying metadata; The process parsing unit is used to receive executable processes, parse the calling order of the executable processes and the data flow of the business interaction channel, and generate an execution instruction sequence; The rendering execution unit parses the metadata in each business component one by one in the sequence of execution instructions, and renders the data model metadata into the corresponding interface based on the table-format metadata. Based on the process definition metadata, the logic for navigation between interfaces is generated, and user operations and changes in interface data are monitored during interface operation to trigger corresponding business rule metadata.

[0011] In a preferred embodiment, the integrated open module includes: The interface management unit is used to build data interaction interfaces between the management platform and external systems. The data interaction interfaces include interface authentication, authorization, and rate limiting functions. The connector unit is used to encapsulate the interface with external systems related to scientific research, generate standardized connectors, and call the executable processes of external systems.

[0012] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention transforms the traditional static list-based management of scientific research consulting business components into a dynamic spatial layout based on exposure by constructing a component whiteboard module. The component whiteboard uses a multi-layer mapping space architecture. In the component layer, the exposure is calculated based on the frequency of use of the business component in historical consulting data. The higher the exposure, the closer the component is to the central area. This allows frequently used core components to naturally gather in the focus area of ​​search and arrangement. Based on this mechanism, the layout of the component whiteboard will continuously evolve with the accumulation of business practices as the enterprise uses the platform over a long period of time. Frequently used components are automatically placed at the front, while less popular components are automatically marginalized. There is no need for manual maintenance of classification and sorting, which effectively reduces the component search cost and improves the arrangement efficiency of business components. 2. This invention, after each scientific research consulting business application is constructed through the business orchestration module, automatically captures the component set, interaction interface, and business interaction channel called in this orchestration by the copy layer, maps them into lightweight contour data, and persists them. It also extracts common and difference contours from multiple copy sub-units to generate copy templates, forming a unique scientific research consulting business experience library for the enterprise. When a new consulting requirement comes in, the development auxiliary module generates a component matching list, and the business orchestration module retrieves the matching application contours in the copy layer. After a successful match, the contour data is directly reverse-mapped to the component layer, interface layer, and connection layer, allowing for one-click reproduction of historical successful solutions. Based on this mechanism, the implicit orchestration experience in scientific research consulting business is transformed into matching and reusable structured contour experience, effectively simplifying the orchestration process and shortening the orchestration cycle from requirement to delivery. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0014] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1, please refer to Figure 1 As shown in this embodiment, a low-code intelligent development and orchestration platform for scientific research consulting management includes: Application Development Module: Used to build business components related to scientific research consulting services through visual configuration; Component Whiteboard Module: Based on historical consulting data from scientific research consulting, the exposure of business components is generated, and a component whiteboard is built based on the exposure. Development support module: Used to generate a component matching list based on the user's input consultation needs and component matching rules; Business orchestration module: Based on the component matching list, it searches the component whiteboard to generate an executable process; Metadata engine module: Used to store and parse metadata, receive and execute executable processes, generate corresponding business applications, and dynamically render the application interface; Integrated open modules: Build external interfaces for managing the interaction between the platform and external systems; Furthermore, existing low-code development platforms abstract business objects, page layouts, process nodes, and business rules into configurable metadata through visual configuration, and encapsulate them into reusable business components. Users can select the required business components on the platform through drag-and-drop and connection operations, arrange the calling order and data flow between components, and generate executable applications for specific business scenarios. However, existing technologies have certain shortcomings. The organization of business components relies on manual maintenance, and the positions of components in directories or lists are fixed, failing to reflect the frequency and importance of component usage in actual business. As the number of business components increases, frequently used core components and infrequently used peripheral components are mixed in the same level, continuously increasing the cost for users to find target components and making the positioning of orchestration starting points inefficient. At the same time, the platform's template reuse mechanism only supports the overall calling of saved applications and cannot automatically capture implicit orchestration experience such as the topological relationships, interaction interfaces, and connection paths between components during each orchestration process. When faced with new consulting needs, users still need to rely on personal memory to determine whether there are historical solutions that can be referenced. It is impossible to achieve automatic accumulation, structural matching, and rapid reproduction of orchestration experience at the system level, making it difficult to continuously accumulate and reuse the methodology of scientific research consulting business. This invention transforms the traditional static list-based management of scientific research consulting business components into a dynamic spatial layout based on exposure by constructing a component whiteboard module. The component whiteboard uses a multi-layered mapping space architecture. In the component layer, the exposure is calculated based on the frequency of use of the business component in historical consulting data. Components with higher exposure are placed closer to the central area, so that frequently used core components are naturally gathered in the focus area of ​​search and arrangement. Based on this mechanism, the layout of the component whiteboard will continue to evolve with the accumulation of business practices as enterprises use the platform over a long period of time. Frequently used components are automatically placed at the front, and less frequently used components are automatically marginalized. There is no need for manual maintenance of classification and sorting, which effectively reduces the cost of component retrieval and improves the arrangement efficiency of business components. After each scientific research consulting application is built through the business orchestration module, the copy layer automatically captures the set of components, interaction interfaces, and business interaction channels called in this orchestration, maps them into lightweight contour data, and persists them. It also extracts common and differential contours from multiple copy sub-units to generate copy templates, forming a unique scientific research consulting business experience library for the enterprise. When new consulting requirements come in, the development assistance module generates a component matching list, and the business orchestration module retrieves matching application contours in the copy layer. After a successful match, the contour data is directly reverse-mapped to the component layer, interface layer, and connection layer, allowing for one-click reproduction of historical successful solutions. Based on this mechanism, the implicit orchestration experience in scientific research consulting is transformed into matching and reusable structured contour experience, effectively simplifying the orchestration process and shortening the orchestration cycle from requirement to delivery.

[0017] In one embodiment, the application development module includes: The model building unit, through a graphical configuration interface, defines the business objects, field attributes, and object relationships of scientific research consulting services, and generates data model metadata. The page design unit is used to configure the presentation format of data model metadata and generate presentation metadata. The presentation formats include form format, list format, and Kanban format; The process design unit is used to define the workflow of scientific research consulting business through visual connection configuration and generate process definition metadata; The workflow includes approval nodes, task assignment nodes, and conditional branch nodes. The rule configuration unit is used to define business rules and generate business rule metadata through the rule table. The business rules include business triggering conditions, data verification rules, and automatic allocation rules. The data encapsulation unit is used to encapsulate data model metadata, representation metadata, process definition metadata, and business rule metadata into business components. Furthermore, the system administrator inputs business definitions through a graphical configuration interface. This interface includes areas for defining business objects, configuring field attributes, setting relationships, selecting presentation formats, configuring process nodes, drawing connections, and editing rule tables. In the business object definition area, the administrator enters the name of the business object. In the field attribute configuration area, field attributes are added to the business object, including field name, field type, field length, and whether they are required. In the relationship setting area, the administrator specifies the relationship type between the current business object and other business objects, including one-to-one, one-to-many, and many-to-many relationships. Based on the above input, an entity relationship diagram is constructed, and the entity relationship diagram is then transformed... For the structured description language, data model metadata is generated. In the presentation selection area, the page design unit provides presentation options for each data model metadata element. These options include form, list, and dashboard formats. The form format corresponds to the editing interface layout for a single data record, the list format corresponds to a row-column display layout for multiple data records, and the dashboard format displays data records in a grouped card layout. The system maps the field attributes in the data model metadata to the corresponding interface layout description parameters based on the selected presentation format, generating the presentation format metadata. The workflow design unit provides configuration templates for three node types in the workflow node configuration area: approval nodes, task assignment nodes, and conditional branch nodes. Approval nodes are configured... The configuration template includes fields for approver identifier, approval timeout limit, and approval comment feedback. The task assignment node configuration template includes fields for executor identifier, task deadline, and task input / output field mapping. The conditional branch node configuration template includes fields for branch condition expressions and fields for the next node identifier pointed to by each conditional branch. The system determines the flow direction based on the directed connecting lines drawn in the connection drawing area and compiles the configuration information of each node and the connection line topology together into process definition metadata. The rule configuration unit provides an editable rule table in the rule table editing area. Each row of the rule table corresponds to a business rule, including columns for rule name, trigger condition, validation logic, and execution action. The `component` column defines the preconditions or data states for activating the rule; the `validation logic` column defines the expression for validating the specified field; and the `execution action` column defines the response operation automatically executed by the system after the rule is triggered. The rule configuration unit converts each rule in the rule table into a structured rule description statement, generating business rule metadata. The data encapsulation unit maintains a component list index, receives multiple types of metadata, assigns a unique component identifier to multiple types of metadata, aggregates the component identifier and the storage reference addresses of the four types of metadata into a metadata data packet of a business component, and registers the component identifier to the component list index, completing the encapsulation of the business component. The encapsulated business component uses the component identifier as its unique index.

[0018] In one embodiment, the component whiteboard module includes: The whiteboard building module constructs a multi-layer mapping space, which, from high to low, includes the component layer, interface layer, connection layer, and overwrite layer. The component layer contains multiple pre-sized square seats; The component deployment unit collects the exposure of each business component based on historical scientific research and consulting data. The exposure is used to characterize the usage of the business component in the preset amount of consulting data. Sorting business components based on exposure, generating a business component sorting table; Based on the business component sorting table, business components are deployed to the grid seats in the component layer. The higher the ranking of the business component, the closer it is to the center area of ​​the component layer. The connection management unit deploys corresponding interaction interfaces for business components deployed at the component layer at the interface layer. Based on the logical association of business components, the interaction interfaces between corresponding business components are connected in the connection layer to generate corresponding business interaction channels; The rewrite building unit is used to map the application outline for each business application build. Furthermore, the whiteboard building module creates a multi-layer mapping space during platform initialization. This multi-layer mapping space, from top to bottom, consists of a component layer, an interface layer, a connection layer, and a override layer. The construction steps are as follows: Four independent logical storage areas are divided in the system memory, corresponding to the component layer, interface layer, connection layer, and override layer, respectively. Each layer is allocated an independent logical coordinate system, using a two-dimensional Cartesian coordinate system. A grid coordinate space is constructed for the component layer, uniformly dividing its two-dimensional coordinate space into multiple squares of a preset grid size. Each square is uniquely identified by its row and column numbers, which together constitute the coordinates of that square within the component layer. The coordinates of the central region of the component layer are determined by calculating the geometric center position based on the total number of rows and columns in the grid coordinate space. A port coordinate space is constructed for the interface layer, corresponding one-to-one with the component layer coordinates. Each coordinate address in the interface layer corresponds to a grid seat with the same coordinate address in the component layer, ensuring that a business component deployed in a grid seat in the component layer obtains its input and output ports at the same coordinate position in the interface layer. A directed connection coordinate space is constructed for the connection layer. The coordinate system of the connection layer shares the same coordinate reference as the interface layer, ensuring that the starting coordinates of any directed connection line in the connection layer correspond to the coordinates of the output port of the upstream business component, and the ending coordinates correspond to the coordinates of the input port of the downstream business component. Coordinates; An independent contour storage space is constructed for the copy layer. The copy layer does not have a fixed coordinate grid and adopts a dynamically allocated storage unit mechanism. Each time a business application is built, a copy sub-unit is created on demand. The copy sub-unit establishes a mapping relationship with the three layers above by recording the coordinates of the component layer grid seats, the coordinates of the interface layer port, and the channel identifier of the connection layer. The component deployment unit reads the historical consultation data of scientific research consultation and counts the number of times each business component is called within a preset consultation data volume as the exposure. The preset consultation data volume is the total number of scientific research consultation projects completed within a fixed time window. The business components are sorted from high to low exposure to generate a business component sorting table, and the business groups are arranged according to the sorting table. Components are assigned to grid seats in the component layer one by one. The assignment rule is that the business component with the highest exposure is placed in the grid seat closest to the geometric center in the central area of ​​the component layer, the business components with the second highest exposure are placed outwards in sequence, and the business component with the lowest exposure is placed in the grid seat in the edge area of ​​the component layer. When multiple business components have the same exposure, the order is determined by the lexicographical order of the component identifiers. The connection management unit creates an interaction interface at the corresponding coordinate position in the interface layer for each grid seat of the deployed business component in the component layer. The interaction interface includes an input port and an output port. The input port is used to receive data transmitted by the upstream business component, and the output port is used to send data to the downstream business component.The connection management unit further reads the logical relationships between business components. These logical relationships are determined by the component dependencies in the process definition metadata and business rule metadata encapsulated within the business components. Based on these logical relationships, directed connections are established in the connection layer from upstream to downstream business components. These directed connections constitute business interaction channels. Each business interaction channel records a channel identifier, source component identifier, target component identifier, data transmission format, and channel activation status. The rewrite construction unit maps the application profile through deployed rewrite subunits.

[0019] In one embodiment, the step of deploying a rewrite subunit in the rewrite layer to map the application profile for each business application build is as follows: A monitoring subunit is built in the overwrite layer. When the monitoring subunit detects that a business interaction channel in the connection layer is called, the overwrite layer is activated and an overwrite subunit is generated. The copy subunit is used to record the call data in this business application construction. The call data includes the corresponding business components, interaction interfaces and business interaction channels. The call data is mapped to outline data and recorded on the copy sub-unit. The outline data includes the grid seat outline, the interaction interface outline, and the business interaction channel outline. When multiple copy sub-units are generated, similarity matching is performed on the multiple copy sub-units to extract common contours and difference contours. The copying subunit with the largest common contour range is taken as the main copying subunit. Based on the main copying subunit, the differential contours are connected layer by layer with the activation channel to generate a copying template. Furthermore, during the initialization of the overwrite layer, a monitoring subunit is created. This monitoring subunit maintains a channel status listening list, which subscribes to channel activation status change events for all business interaction channels in the connection layer. When the status of any business interaction channel in the connection layer changes to "activated," the connection layer pushes a status change notification to the monitoring subunit. This notification includes the activated channel identifier, source component identifier, and target component identifier. Upon receiving the status change notification, the monitoring subunit activates the overwrite layer and creates an overwrite subunit within it. Each overwrite subunit is assigned a unique subunit identifier, which is obtained by... The data is generated by concatenating a timestamp and an incrementing sequence number. The replication subunit continuously records call data from the start to the end of this business application build process. This call data includes the corresponding business components, interaction interfaces, and business interaction channels called during the build process. The recording method is as follows: the grid coordinates of each business component, the port identifier of the interaction interface, and the channel identifier. The port identifier is formed by concatenating the coordinate address of the interface layer where the interaction interface is located with the port direction identifier. At the end of this business application build, the replication subunit compresses and maps the recorded call data into outline data. The mapping rule is to map all called business components... The coordinates of the agent grid are extracted into agent grid outlines, which are stored as a set of coordinates. The port identifiers of all activated interaction interfaces are extracted into interface outlines, which are also stored as a set of port identifiers. The channel identifiers of all called business interaction channels are extracted into business interaction channel outlines, which are stored as a channel connection table. All outline data is written completely into the storage space of the replication sub-unit. When multiple replication sub-units exist in the replication layer, a similarity matching process is initiated, sequentially traversing all replication sub-units and calculating the coordinate overlap of the agent grid outlines for any two replication sub-units. The port overlap of the degree of interaction interface contour and the channel overlap of the business interaction channel contour are calculated. The three overlap values ​​are weighted and summed according to preset weights to obtain the contour similarity between two copy sub-units. All copy sub-units are clustered according to the contour similarity. Copy sub-units with contour similarity exceeding a preset threshold are grouped into the same similarity group. For copy sub-units in the same similarity group, the grid seat coordinates, port identifiers and channel connection relationships common to all copy sub-units are extracted as common contours. The grid seat coordinates, port identifiers and channel connection relationships unique to each copy sub-unit and not common to other copy sub-units are extracted as difference contours.The primary copying subunit is selected based on the largest common contour range. The common contour range is measured by the number of grid seat coordinates contained within it. Using the common contour of the primary copying subunit as the basic framework, inter-layer connection channels are constructed. Components and channel branches in the differential contours are then attached to the common contour framework via these inter-layer connection channels to generate a copy template. This template is stored in a tree structure, with the root node corresponding to the common contour of the primary copying subunit, branch nodes corresponding to each differential contour, and edges corresponding to the business interaction channels connecting the common contour components and the differential contour components in the connection layer.

[0020] In one embodiment, the development assistance module includes: The intent recognition unit is used to receive the user's consultation request input in natural language, extract the business intent in the consultation request based on the neural network, and generate a structured intent expression. The business intent includes the business type, core entity, target product, and constraints. The requirement expansion unit, based on the structured intent expression and the knowledge graph of the scientific research field, obtains the expansion requirements that are implicitly related to the current requirement. Write the extended requirements into a structured intent statement to generate an extended requirements set; The component mapping unit matches the extended requirement set with preset component matching rules to generate a component matching list; Furthermore, the intent recognition unit receives the consultation request text input by the user in natural language and feeds it into a pre-trained language parsing model. This pre-trained language parsing model, based on the Transformer architecture, is fine-tuned on annotated corpus in the scientific research consultation field. The annotated corpus includes business type labels, core entity labels, target deliverable labels, and constraint labels. Based on the pre-trained language parsing model, the consultation request text is sequence-labeled, extracting the business type field, core entity field, target deliverable field, and constraint field. The business type field represents the business category of the consultation request; the core entity field represents the key business objects involved; the target deliverable field represents the expected deliverable; and the constraint field represents the restrictive conditions of the consultation request. The intent recognition unit combines these four extracted fields into a structured intent representation. The requirement expansion unit maintains a knowledge graph within the scientific research field. The knowledge graph uses core concepts of the scientific research field as entity nodes, connected by semantic relationship edges, including hierarchical, dependency, and collaborative relationships. The requirement expansion unit uses the business type field and core entity field from the structured intent representation as the starting point for retrieval within the knowledge graph. The process involves multi-hop subgraph traversal, with a preset hop count of two. Adjacent entity nodes with semantically related edges at the starting point are collected and retrieved as extended requirement items. The requirement extension unit writes each extended requirement item into a structured intent representation, generating an extended requirement set. The component mapping unit maintains a component matching rule base, stored as a mapping table. Each record in the mapping table includes a business type condition field, a core entity condition field, a target product condition field, a constraint condition field, and a corresponding component identifier field. The business type field, core entity field, target product field, and constraint condition field in the extended requirement set are matched with the component identifier field. The condition fields of each record in the rule library are subjected to multi-field joint matching. The multi-field joint matching adopts a strategy of precise matching as the main method and fuzzy matching as a supplement. When the fields in the extended requirement set match all the condition fields of a record in the component matching rule library, the component identifier field of that record is added to the component matching list. When no record can be found that matches completely, the component mapping unit calculates the semantic similarity of the condition fields and selects the component identifier corresponding to the record with the highest semantic similarity to add to the component matching list. The component mapping unit summarizes all the matched component identifiers to generate the component matching list, which contains all the required business component identifiers.

[0021] In one embodiment, the service orchestration module includes: The outline retrieval unit generates the corresponding grid seat outline in the component whiteboard based on the component matching list, and matches the corresponding grid seat outline with the copy template; If the match is successful, the application profile data of the copy subunit is extracted and mapped back to the component layer, interface layer and connection layer of the component whiteboard module, and the business components in the corresponding grid seat are activated to generate an executable process. If the match fails, activate the business component and the corresponding interaction interface corresponding to the component matching list, build the business interaction channel between the corresponding business components, and generate an executable process; Furthermore, the contour retrieval unit receives the component matching list and initiates a query request to the component deployment unit of the component whiteboard module based on the identifier of each business component. The component deployment unit returns the grid seat coordinates corresponding to each business component identifier. The contour retrieval unit summarizes all the returned grid seat coordinates into a grid seat contour table. The grid seat contour table represents the spatial occupancy of this consultation request in the component layer in the form of a coordinate set. The contour retrieval unit matches the grid seat contour table with the copy templates stored in the copy layer one by one. The matching process involves extracting the grid seat coordinate set contained in the common contour of the main copy sub-unit in the copy template, and calculating the grid seat contour table and the common... The coordinate overlap of the contours is the ratio of the number of elements in the intersection to the number of elements in the union of two coordinate sets. A successful match is determined when the coordinate overlap exceeds a preset contour overlap threshold; a failed match is determined when the coordinate overlap does not exceed the preset threshold. If a match is successful, the contour retrieval unit extracts the application contour data stored in the matched copy subunit and maps the application contour data layer by layer back to the three layers above the component whiteboard module. This process involves activating the deployed business component in the corresponding grid seat in the component layer based on the grid seat coordinate set in the grid seat contour. The activation operation marks the business component's running status as enabled and loads it. The metadata data packet corresponding to the business component is sent to the metadata engine module; based on the port identifier set in the interaction interface profile, the input and output ports are activated at the corresponding coordinate positions in the interface layer, and the activation operation marks the communication status of the port as ready; based on the channel connection relationship table in the business interaction channel profile, each business interaction channel is reconstructed in the connection layer, the directed connection line between the output port of the upstream business component and the input port of the downstream business component is marked as activated, and the data transmission format and channel identifier are filled in the channel attributes. After the channel reconstruction is completed, an executable process is generated; if the matching is unsuccessful, the profile retrieval unit abandons the copy layer retrieval path and uses the whiteboard to construct the path. The specific process of building the whiteboard path is as follows: The contour retrieval unit sends all business component identifiers in the component matching list to the component deployment unit to obtain the coordinates of each grid seat. In the component layer, each business component is activated one by one. The connection management unit requests the creation of interaction interfaces for each business component in the interface layer. The connection management unit also requests the construction of directed connection lines between each business component in the connection layer based on the process definition metadata encapsulated in each business component. After generating all business interaction channels, an executable process is output. That is, when the match is successful, the historical channel connection relationship is directly reused. When the match is unsuccessful, the channel connection relationship is newly constructed through the connection management unit. Both paths eventually generate executable processes.

[0022] In one embodiment, the metadata engine module includes: Metadata storage unit, used to store metadata of business components, build update interface for updating and modifying metadata; The process parsing unit is used to receive executable processes, parse the calling order of the executable processes and the data flow of the business interaction channel, and generate an execution instruction sequence; The rendering execution unit parses the metadata in each business component one by one in the sequence of execution instructions, and renders the data model metadata into the corresponding interface based on the table-format metadata. Based on the process definition metadata, the logic for navigation between interfaces is generated, and user operations and changes in interface data are monitored during interface operation to trigger corresponding business rule metadata. Furthermore, the metadata storage unit constructs a metadata repository during platform initialization, employing a key-value pair storage structure. The key is the component identifier of the business component, and the value is the metadata data packet corresponding to that business component. The metadata data packet contains metadata types and corresponding storage reference addresses. Simultaneously, an update interface is constructed. This interface receives the component identifier and the type of metadata to be updated as input parameters. It locates the metadata data packet through the component identifier and replaces the metadata content pointed to by the corresponding storage reference address according to the type of metadata to be updated, thus completing the metadata update and modification. The process parsing unit receives executable processes, organized in a structured description language, including process identifiers, business component call sequences, and a list of business interaction channels. The process parsing unit expands the business component call sequence into linear execution steps and converts each business interaction channel into a data transmission instruction. The instruction delivery includes the source component output port identifier and the target component input port identifier, merging the linear execution steps and data transfer instructions into an execution instruction sequence. The rendering execution unit processes each business component sequentially according to the linear execution steps in the execution instruction sequence. It reads the corresponding metadata data package from the metadata storage unit through the component identifier, parses the data model metadata to obtain the field attribute list, parses the presentation metadata to determine the interface rendering mode, for example, when the presentation metadata is in form form, it renders it as a single data editing interface, reads the next-hop business component identifier and jump trigger condition from the process definition metadata, and generates an interface jump trigger. During the interface operation, it continuously listens for user operation events and interface data change events. When the captured event meets the business trigger condition in the business rule metadata, it executes the corresponding data verification rules and automatic allocation rules in sequence.

[0023] In one embodiment, the integrated open module includes: The interface management unit is used to build data interaction interfaces between the management platform and external systems. The data interaction interfaces include interface authentication, authorization, and rate limiting functions. The connector unit is used to encapsulate the interface with external systems related to scientific research, and to generate standardized connectors for calling executable processes of external systems; Furthermore, the interface management unit constructs an API gateway between the platform and external systems. This API gateway serves as the unified entry point for all external requests, exposing standardized RESTful interfaces. Each exposed interface is configured with authentication, authorization, and rate limiting policies. The authentication policy employs a token-based authentication mechanism; external systems must request an access token from the interface management unit using a key pair before calling the interface. The access token contains expiration attributes and signature information. The authorization policy is implemented based on a role-based access control model; the interface management unit assigns a role identifier to each external system, binding the role identifier to a list of accessible interfaces and operation permissions. The rate limiting policy uses a sliding window algorithm, configuring a maximum request count threshold for each external system within a unit of time. When the number of requests from an external system exceeds the threshold within a unit of time, the request is blocked. The interface management unit returns a rate-limiting response and temporarily suspends processing subsequent requests. The connector unit maintains a connector repository, which contains pre-set standardized connector templates for interfacing with common external systems in the research field. Each standardized connector template encapsulates the interface address, request format, response parsing rules, and exception retry strategy of the corresponding external system. A unique connector identifier is assigned to each standardized connector template. When the executable process generated by the business orchestration module needs to call an external system, the business interaction channel of the executable process carries the connector identifier. Based on the connector identifier, the corresponding standardized connector template is retrieved from the connector repository, and a request is initiated based on the interface address in the standardized connector template. The request result is converted into a unified data structure within the platform according to the response parsing rules and then returned to the calling business component, completing the interaction with the external system.

[0024] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A low-code intelligent development and orchestration platform for scientific research consulting management, characterized in that, include: Application Development Module: Used to build business components related to scientific research consulting services through visual configuration; Component Whiteboard Module: Connects to the application development module, generates the exposure of business components based on historical consulting data of scientific research consulting, and builds the component whiteboard based on the exposure; Development Support Module: Connects to the Component Whiteboard Module and is used to generate a component matching list based on the user's input consultation needs and component matching rules; Business orchestration module: Connects with the development support module, retrieves components from the component whiteboard based on the component matching list, and generates an executable process; Metadata engine module: Connects to the business orchestration module, used to store and parse metadata, receive and execute executable processes, generate corresponding business applications and dynamically render the application interface; Integrated open modules: Connect with the metadata engine module to build external interfaces for managing the interaction between the platform and external systems.

2. The low-code intelligent development and orchestration platform for scientific research consulting management according to claim 1, characterized in that, The application development module includes: The model building unit, through a graphical configuration interface, defines the business objects, field attributes, and object relationships of scientific research consulting services, and generates data model metadata. The page design unit is used to configure the presentation format of data model metadata and generate presentation metadata. The presentation formats include form format, list format, and Kanban format; The process design unit is used to define the workflow of scientific research consulting business through visual connection configuration and generate process definition metadata; The workflow includes approval nodes, task assignment nodes, and conditional branch nodes. The rule configuration unit is used to define business rules and generate business rule metadata through the rule table. The business rules include business triggering conditions, data verification rules, and automatic allocation rules; The data encapsulation unit is used to encapsulate data model metadata, representation metadata, process definition metadata, and business rule metadata into business components.

3. The low-code intelligent development and orchestration platform for scientific research consulting management according to claim 1, characterized in that, The whiteboard module includes: The whiteboard building module constructs a multi-layer mapping space, which, from high to low, includes the component layer, interface layer, connection layer, and overwrite layer. The component layer contains multiple pre-sized square seats; The component deployment unit collects the exposure of each business component based on historical scientific research and consulting data. The exposure is used to characterize the usage of the business component in the preset amount of consulting data. Sorting business components based on exposure, generating a business component sorting table; Based on the business component sorting table, business components are deployed to the grid seats in the component layer. The higher the ranking of the business component, the closer it is to the center area of ​​the component layer. The connection management unit deploys corresponding interaction interfaces for business components deployed at the component layer at the interface layer. Based on the logical association of business components, the interaction interfaces between corresponding business components are connected in the connection layer to generate corresponding business interaction channels; The rewrite building unit is used to map the application profile for each business application build.

4. The low-code intelligent development and orchestration platform for scientific research consulting management according to claim 3, characterized in that, The step of deploying a rewrite subunit in the rewrite layer to map the application profile for each business application construction is as follows: A monitoring subunit is built in the overwrite layer. When the monitoring subunit detects that a business interaction channel in the connection layer is called, the overwrite layer is activated and an overwrite subunit is generated. The copy subunit is used to record the call data in this business application construction. The call data includes the corresponding business components, interaction interfaces and business interaction channels. The call data is mapped to contour data and recorded on the copy sub-unit. The contour data includes the grid seat contour, the interaction interface contour, and the business interaction channel contour. When multiple copy sub-units are generated, similarity matching is performed on the multiple copy sub-units to extract common contours and difference contours. The copying subunit with the largest common contour range is taken as the main copying subunit. Based on the main copying subunit, the differential contours are connected layer by layer with the activation channel to generate the copying template.

5. The low-code intelligent development and orchestration platform for scientific research consulting management according to claim 1, characterized in that, The development support module includes: The intent recognition unit is used to receive the user's consultation request input in natural language, extract the business intent in the consultation request based on the neural network, and generate a structured intent expression. The business intent includes the business type, core entity, target product, and constraints. The requirement expansion unit, based on the structured intent expression and the knowledge graph of the scientific research field, obtains the expansion requirements that are implicitly related to the current requirement. Write the extended requirements into a structured intent statement to generate an extended requirements set; The component mapping unit matches the extended requirement set with the preset component matching rules to generate a component matching list.

6. The low-code intelligent development and orchestration platform for scientific research consulting management according to claim 1, characterized in that, The service orchestration module includes: The outline retrieval unit generates the corresponding grid seat outline in the component whiteboard based on the component matching list, and matches the corresponding grid seat outline with the copy template; If the match is successful, the application profile data of the copy subunit is extracted and mapped back to the component layer, interface layer and connection layer of the component whiteboard module, and the business components in the corresponding grid seat are activated to generate an executable process. If the matching fails, the corresponding business component and the corresponding interaction interface in the component matching list will be activated, and a business interaction channel between the corresponding business components will be built to generate an executable process.

7. The low-code intelligent development and orchestration platform for scientific research consulting management according to claim 1, characterized in that, The metadata engine module includes: Metadata storage unit, used to store metadata of business components, build update interface for updating and modifying metadata; The process parsing unit is used to receive executable processes, parse the calling order of the executable processes and the data flow of the business interaction channel, and generate an execution instruction sequence; The rendering execution unit parses the metadata in each business component one by one in the sequence of execution instructions, and renders the data model metadata into the corresponding interface based on the table-format metadata. Based on the process definition metadata, the logic for navigation between interfaces is generated, and user operations and changes in interface data are monitored during interface operation to trigger corresponding business rule metadata.

8. The low-code intelligent development and orchestration platform for scientific research consulting management according to claim 1, characterized in that, The integrated open module includes: The interface management unit is used to build data interaction interfaces between the management platform and external systems. The data interaction interfaces include interface authentication, authorization, and rate limiting functions. The connector unit is used to encapsulate the interface with external systems related to scientific research, generate standardized connectors, and call the executable processes of external systems.

Images