Methods, apparatus, equipment, and media for reconstructing pharmaceutical digital teaching materials based on capability maps.

Abstract

This application discloses a method, apparatus, equipment, and medium for reconstructing pharmaceutical digital textbooks based on competency graphs, relating to the field of digital data processing technology. The method includes: first, constructing a pharmaceutical competency graph and a pharmaceutical basic knowledge graph based on job requirement data from industry-education integration enterprises and knowledge points from traditional pharmaceutical textbooks; second, performing deep feature fusion and alignment mapping on these two graphs to generate a target digital textbook knowledge graph; third, extracting matching digital media resources from an online course platform using this graph as a guide; fourth, combining these media resources with the knowledge graph for structured layout to generate a draft pharmaceutical digital textbook; and finally, configuring and reconstructing the draft for interactive functions to ultimately obtain the target pharmaceutical digital textbook. This application effectively solves the problems of outdated content, monotonous presentation methods, and inability to dynamically match the actual job competency requirements of enterprises under industry-education integration in traditional pharmaceutical textbooks.

Description

Technical Field

[0001] This application relates to the field of digital data processing technology, and in particular to a method, apparatus, equipment and medium for reconstructing pharmaceutical digital teaching materials based on capability maps. Background Technology

[0002] With the advancement of the digital transformation strategy in education, digital textbooks have become a core carrier for vocational education reform. Particularly in the field of higher vocational pharmacy, to support the development of the digital economy, there is an urgent need to establish a digital teaching resource system capable of rapidly integrating new processes, standards, and technologies. To cultivate compound technical and skilled personnel who meet the job requirements of the pharmaceutical industry, the education system needs to accurately translate the job competency requirements of enterprises into deployable digital teaching units.

[0003] Currently, the digitization of pharmacy textbooks mainly relies on the digital migration of traditional paper textbooks or the development of multimedia and loose-leaf digital textbooks. Existing technical solutions typically involve uploading text, images, videos, and virtual simulation (VR / AI) resources to online course platforms, allowing students to passively or one-way access and read these resources. In experimental teaching, digital media is primarily used to demonstrate experimental procedures and ideal experimental results to guide students in standardized operations.

[0004] However, existing methods for digitizing textbooks have significant limitations in data processing logic. First, the data carriers exhibit linear and singular characteristics, with insufficient traditional structuring, resulting in limited content coverage and low information update efficiency, making it difficult to achieve real-time synchronization between teaching data and industry technology. Second, there is a lack of semantic mapping between teaching resources and job skills. Existing textbooks are mainly constructed based on disciplinary theoretical logic, lacking effective extraction and alignment of heterogeneous data such as video logs of real-world operations in enterprises, physical space movement trajectories, and process constraint parameters in industry-education integration, leading to a disconnect between textbook content and actual production scenarios. Finally, the interactive feedback mechanism is rigid. Traditional textbook data flow is usually unidirectional, lacking dynamic interactive functions such as safety range verification based on process control parameters and abnormal state rendering, which is detrimental to cultivating students' practical thinking under complex production variables. Therefore, the outdated content, singular presentation methods, and inability to dynamically match the actual job skill requirements of enterprises under industry-education integration are urgent problems that need to be addressed. Summary of the Invention

[0005] The purpose of this application is to provide a method, device, equipment and medium for reconstructing pharmaceutical digital teaching materials based on capability maps, which aims to solve the technical problems of outdated content, single presentation mode and inability to dynamically match the actual job capability requirements of enterprises under the integration of industry and education in traditional pharmaceutical teaching materials.

[0006] To achieve the above objectives, this application proposes a method for reconstructing pharmaceutical digital teaching materials based on capability maps, the method comprising: Construct a pharmacy job competency map based on job demand data from industry-education integration enterprises; Construct a basic pharmaceutical knowledge map based on pharmaceutical knowledge points in traditional pharmaceutical textbooks; The pharmaceutical basic knowledge graph and the pharmaceutical job competency graph are fused and aligned to obtain the target digital textbook knowledge graph. Based on the target digital textbook knowledge graph, corresponding digital media resources are extracted from the online course platform to obtain the target digital textbook media resources; A draft of a pharmaceutical digital textbook is obtained by structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph; The draft pharmaceutical digital textbook is configured with interactive functions and reconstructed to generate the target pharmaceutical digital textbook.

[0007] In one embodiment, the step of constructing a pharmaceutical job competency map based on job demand data from industry-education integration enterprises includes: Obtain real-world operational video logs containing pharmaceutical manufacturing quality management standards and use them as job requirement data for industry-education integration enterprises; Extract the physical space movement trajectory of operators and workshop environmental constraint parameters from the job demand data of the industry-education integration enterprises; Map the physical space movement trajectory to the corresponding preset pharmaceutical equipment interaction events; Construct multi-dimensional job capability nodes based on the preset pharmaceutical equipment interaction events and the workshop environment constraint parameters; The multidimensional job competency nodes are connected by directed edges according to preset temporal dependency rules to obtain a pharmacy job competency map.

[0008] In one embodiment, the step of constructing a pharmaceutical basic knowledge graph based on pharmaceutical knowledge points in traditional pharmaceutical textbooks includes: To obtain pharmaceutical knowledge points from traditional pharmaceutical textbooks, including two-dimensional chemical structure images and chemical reaction equation texts; The two-dimensional chemical structure image is converted into a three-dimensional molecular conformation coordinate matrix, and the three-dimensional molecular conformation coordinate matrix is ​​encapsulated as a knowledge entity node; Based on the chemical reaction equation text, determine the reactant-product mapping relationship between the corresponding knowledge entity nodes; Calculate the reaction energy barrier values ​​between the knowledge entity nodes that have the aforementioned reactant-product mapping relationship; Using the reaction energy barrier value as the edge weight, the knowledge entity nodes that have the mapping relationship between the reactants and products are connected to obtain a pharmaceutical basic knowledge graph.

[0009] In one embodiment, the step of performing feature fusion and alignment mapping between the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph to obtain the target digital textbook knowledge graph includes: Identify the knowledge entity nodes in the pharmaceutical basic knowledge graph, and identify the multi-dimensional job competency nodes in the pharmaceutical job competency graph that correspond to the knowledge entity nodes; Calculate the thermodynamic engineering scaling factor between the knowledge entity node and the multidimensional job competency node; Extract the feature vector of the knowledge entity node, and adjust the feature vector by dimensional projection according to the thermodynamic engineering scaling factor to obtain the adjusted feature vector; The transmission cost between the adjusted feature vector and the process feature vector corresponding to the multidimensional job capability node is calculated using the optimal transmission algorithm, and the mapping node pair with the minimum transmission cost is extracted. Based on the mapping nodes, the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph are spliced ​​and reconstructed to obtain the target digital textbook knowledge graph.

[0010] In one embodiment, the step of extracting corresponding digital media resources from the online course platform based on the target digital textbook knowledge graph to obtain the target digital textbook media resources includes: Calculate the cognitive abstraction index of each node in the knowledge graph of the target digital teaching materials; Target nodes whose cognitive abstraction index is greater than a preset abstraction threshold are selected, and the structural feature vector, thermodynamic feature vector, and spatiotemporal operation feature vector of the target nodes are extracted. The structural feature vector, the thermodynamic feature vector, and the spatiotemporal operation feature vector are concatenated into a multidimensional retrieval vector. Based on the multidimensional retrieval vector, an approximate nearest neighbor retrieval is performed on the online course platform to obtain a candidate media resource set; Calculate the multimodal complementary information entropy between each candidate media resource in the candidate media resource set and the target node, and take the resource with the largest multimodal complementary information entropy as the target digital teaching material media resource.

[0011] In one embodiment, the step of obtaining a draft pharmaceutical digital textbook by performing structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph includes: The node text attributes in the target digital textbook knowledge graph are used as theoretical text resources, and the target digital textbook media resources are used as production training resources; Calculate the semantic association strength between the theoretical text resources and the production training resources, and determine the layout proximity between the two based on the semantic association strength; Based on the layout proximity, corresponding display slot positions are allocated to the theoretical text resources and the production training resources in a preset dynamic grid layout container; Analyze the key process parameters in the theoretical text resources and locate the operation frame segments containing the key process parameters in the video stream of the production training resources; An interactive floating window is generated at the theoretical text resource location of the display slot, and the operation frame fragment is associated as the default display content of the interactive floating window; Visual alignment and style rendering are performed on the preset dynamic grid layout container after filling resources to obtain a draft of a pharmaceutical digital textbook.

[0012] In one embodiment, the step of configuring and reconstructing the interactive functions of the draft pharmaceutical digital textbook to obtain the target pharmaceutical digital textbook includes: Identify the multi-dimensional job competency nodes in the draft of the pharmaceutical digital textbook, and obtain the preset parameter safety range of the multi-dimensional job competency nodes; An abnormal state rendering model composed of a conditional generative adversarial network is obtained, which is used to generate corresponding pharmaceutical physical failure morphology images based on the input process parameters. In the draft of the pharmaceutical digital textbook, a parameter adjustment component is embedded in the display area corresponding to the multi-dimensional job competency node; The preset dynamic association script, the parameter adjustment component, and the draft pharmaceutical digital textbook are compiled and encapsulated using Hypertext Markup Language to obtain the target pharmaceutical digital textbook. The preset dynamic association script is used to trigger the abnormal state rendering model to output the pharmaceutical physical failure morphology image when the parameter adjustment component receives a parameter value that exceeds the preset parameter safety range.

[0013] Furthermore, to achieve the above objectives, this application also proposes a pharmaceutical digital teaching material reconstruction device based on capability maps, the device comprising: The competency mapping module is used to construct a competency map for pharmaceutical positions based on job demand data from industry-education integration enterprises. The pharmaceutical atlas construction module is used to construct a basic pharmaceutical knowledge atlas based on pharmaceutical knowledge points in traditional pharmaceutical textbooks. The graph fusion and alignment module is used to perform feature fusion and alignment mapping between the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph to obtain the target digital textbook knowledge graph. The media resource extraction module is used to extract corresponding digital media resources from the online course platform based on the target digital textbook knowledge graph to obtain the target digital textbook media resources. The structured typesetting module is used to perform structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph to obtain a draft of the pharmaceutical digital textbook; The interactive reconstruction module is used to configure and reconstruct the interactive functions of the draft pharmaceutical digital textbook to obtain the target pharmaceutical digital textbook.

[0014] Furthermore, to achieve the above objectives, this application also proposes a pharmaceutical digital textbook reconstruction device based on capability map, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the pharmaceutical digital textbook reconstruction method based on capability map as described above.

[0015] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the pharmaceutical digital textbook reconstruction method based on capability map described above.

[0016] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the pharmaceutical digital textbook reconstruction method based on capability graphs as described above.

[0017] One or more technical solutions proposed in this application have at least the following technical effects: First, a pharmaceutical job competency map was constructed based on job demand data from industry-education integration enterprises. This transformed real-world industrial operational standards into a structured model, clarifying the core skill requirements in the production environment. Simultaneously, a pharmaceutical fundamentals knowledge map was constructed based on pharmaceutical knowledge points from traditional pharmaceutical textbooks to preserve and organize a rigorous theoretical foundation. Second, the pharmaceutical fundamentals knowledge map and the pharmaceutical job competency map were fused and aligned to obtain a target digital textbook knowledge map. This effectively broke down the underlying logical barriers between academic theory and factory practice, achieving a deep connection between the two. Finally, corresponding digital media resources were extracted from the online course platform based on the target digital textbook knowledge map. The process involves obtaining target digital textbook media resources, thus supplementing the otherwise dry theoretical framework with concrete multimedia auxiliary materials. Next, based on these resources and the target digital textbook knowledge graph, a structured layout is created to produce a draft of the pharmaceutical digital textbook, achieving automated and rational spatial arrangement and assembly of textual and graphical knowledge and video content. Finally, the draft pharmaceutical digital textbook is configured and reconstructed with interactive functions to generate the target pharmaceutical digital textbook, endowing the static textbook page with dynamic operation and real-time feedback capabilities. Through the synergistic effect of these steps, the problems of outdated content, monotonous presentation methods, and inability to dynamically match the actual job competency requirements of enterprises under the industry-education integration model are effectively solved. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating an embodiment of the pharmaceutical digital textbook reconstruction method based on capability graphs provided in this application. Figure 2 This is a schematic diagram illustrating the construction process of a pharmaceutical basic knowledge graph provided in Embodiment 1 of the pharmaceutical digital textbook reconstruction method based on capability graphs in this application. Figure 3 This is a flowchart illustrating Embodiment 2 of the pharmaceutical digital textbook reconstruction method based on capability graphs in this application; Figure 4 A simplified flowchart illustrating the pharmaceutical digital teaching material reconstruction method based on capability maps provided in Embodiment 2 of this application; Figure 5This is a schematic diagram of the module structure of the pharmaceutical digital teaching material reconstruction device based on capability map according to an embodiment of this application; Figure 6 This is a schematic diagram of the hardware operating environment involved in the pharmaceutical digital textbook reconstruction method based on capability graphs in the embodiments of this application.

[0021] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0022] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0023] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0024] It should be noted that the executing entity of this application embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or textbook reconstruction system capable of realizing the above functions. The following uses a textbook reconstruction system as an example to describe this embodiment and the following embodiments.

[0025] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0026] Based on this, embodiments of this application provide a method for reconstructing pharmaceutical digital teaching materials based on capability maps, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the pharmaceutical digital textbook reconstruction method based on capability graphs in this application.

[0027] In this embodiment, the pharmaceutical digital teaching material reconstruction method based on capability maps includes steps S10 to S60: Step S10: Construct a pharmacy job competency map based on the job demand data of industry-education integration enterprises; Step S20: Construct a basic pharmaceutical knowledge map based on pharmaceutical knowledge points in traditional pharmaceutical textbooks; Step S30: Perform feature fusion and alignment mapping between the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph to obtain the target digital textbook knowledge graph; Step S40: Extract corresponding digital media resources from the online course platform based on the target digital textbook knowledge graph to obtain the target digital textbook media resources; Step S50: Based on the target digital textbook media resources and the target digital textbook knowledge graph, perform structured typesetting to obtain a draft of the pharmaceutical digital textbook; Step S60: Configure and reconstruct the interactive functions of the draft pharmaceutical digital textbook to obtain the target pharmaceutical digital textbook.

[0028] It should be noted that the industry-education integration enterprise job demand data refers to the actual work standard information collected under the school-enterprise cooperation model, which includes the specific skill requirements, equipment operation specifications, and process control logic for each position in the real production environment of pharmaceutical companies. The pharmaceutical job competency map refers to a structured data model that uses a graph structure to computer-model pharmaceutical company production process standards and / or quality inspection procedures and / or professional ethics requirements, displaying the temporal relationships, logical dependencies, and mapping relationships between practical skills and theoretical knowledge through nodes and directed edges. Traditional pharmaceutical textbooks refer to teaching materials widely used in regular teaching in higher vocational colleges, using printed texts or simple static electronic versions as the medium, and mainly arranged according to the conventional chapters of classical discipline theories. Pharmaceutical knowledge points refer to the independent data units that constitute the academic system of the pharmaceutical major, extracted from theoretical books on pharmaceutics or pharmaceutical analysis, including basic concepts, chemical structural characteristics, pharmacological reaction mechanisms, reaction equations, and theoretical rules. Pharmaceutical basic knowledge graph refers to the use of information extraction technology to abstract various pharmacological concepts and theoretical elements into digital entity objects and establish semantic relationships, thereby forming a structured network database with basic theories as the core and reflecting the objective logic within the discipline.

[0029] The target digital textbook knowledge graph refers to a composite underlying relationship network that combines theoretical depth and practical orientation, formed by deep semantic matching and dimensional alignment of skill architecture reflecting actual industry application scenarios and basic theoretical databases reflecting academic theoretical frameworks. Digital media resources refer to general multimedia materials widely stored in various online course platforms, encompassing non-textual digital content carriers such as process operation video clips, 3D molecular virtual simulation animations, interactive charts, and digital audio. Target digital textbook media resources refer to a specific set of teaching materials selected from a massive material library based on a specific feature vector retrieval algorithm. These materials are highly adapted to each specific teaching node in the composite underlying relationship network and can provide intuitive multimodal information supplementation. The pharmaceutical digital textbook draft refers to a digital layout prototype without advanced interactive scripts, generated after initial visual layout and style rendering by allocating theoretical text and matching practical training multimedia content to corresponding interface display slots according to new loose-leaf digital typesetting rules. The target pharmacy digital textbook refers to a complete educational data product that embeds parameter control components, dynamic correlation feedback programs, and abnormal state rendering engines on the basis of a digital version prototype. It is ultimately compiled and packaged by hypertext markup language into a complete educational data product that can provide learners with virtual exercises and real-time feedback on front-end devices.

[0030] Understandably, the textbook reconstruction system first analyzes the job requirements data of industry-education integration enterprises and pharmaceutical knowledge points in traditional pharmaceutical textbooks, extracting core skill elements, disciplinary theoretical elements, and the internal connections between them. Based on this underlying logic, it constructs a pharmaceutical job competency graph and a pharmaceutical basic knowledge graph. This is done to transform the originally fragmented industrial practical standards and textbook text into a network structure data that computers can read and process. Secondly, the system compares the established pharmaceutical basic knowledge graph with the pharmaceutical job competency graph at the underlying level. By finding semantic and logical connections between the two, it performs feature fusion and alignment mapping, ultimately merging and stitching them together to obtain a comprehensive target digital textbook knowledge graph. This aims to break down the barriers between academic theory and factory practice from the underlying architecture (for example, forcibly binding the theoretical node of "chromatographic separation principle" and the practical node of "liquid chromatograph start-up specifications" in the graph). Then, the system uses the target digital textbook knowledge graph as a digital guide to initiate a targeted search in the resource library of the online course platform, extract and filter highly matching digital media resources, thereby obtaining the target digital textbook media resources. The purpose of this is to dynamically fill the dry theoretical knowledge framework with concrete multimodal materials (for example, to capture and match a real workshop operation video for a certain pharmaceutical process node).

[0031] Next, the system comprehensively analyzes the media attributes of the target digital textbook's media resources and the knowledge hierarchy of the target digital textbook's knowledge graph. Using a preset layout algorithm, it structures and typesets the text, images, and video content, assigning corresponding interface display slots to each. This renders a preliminary draft of the pharmaceutical digital textbook, replacing traditional manual image slicing and layout, and enabling automated and rapid arrangement and assembly of digital teaching content. Finally, the system embeds dynamic scripts into the initial draft of the pharmaceutical digital textbook and configures and reconstructs interactive functions, completing the code packaging and conversion from a static reading page to an interactive digital product, outputting the final target pharmaceutical digital textbook. This step is crucial, as it allows students to directly trigger interactive components on the interface while reading the digital textbook (e.g., dragging a slider to change the temperature parameters of a simulation experiment) to observe simulation results under different process conditions. This truly utilizes digital means to solve the technical pain point of traditional textbooks lacking immediate practical feedback.

[0032] As an example, the steps of constructing a pharmaceutical job competency map based on job demand data from industry-education integration enterprises include: obtaining real operation video logs containing Good Manufacturing Practice (GMP) data for pharmaceuticals, and using them as job demand data for industry-education integration enterprises; extracting the physical space movement trajectory of operators and workshop environment constraint parameters from the job demand data; mapping the physical space movement trajectory to corresponding preset pharmaceutical equipment interaction events; constructing multi-dimensional job competency nodes based on the preset pharmaceutical equipment interaction events and the workshop environment constraint parameters; and connecting the multi-dimensional job competency nodes with directed edges according to preset temporal dependency rules to obtain the pharmaceutical job competency map.

[0033] It should be noted that Good Manufacturing Practice (GMP) for pharmaceuticals refers to a set of mandatory national standards and management systems used to ensure drug quality and medication safety during the pharmaceutical manufacturing process. It encompasses strict regulatory requirements for all aspects, including personnel, facilities, equipment, materials, operating procedures, and quality monitoring. Real-world operation video logs refer to continuous audiovisual data files objectively recorded by video equipment in the actual production workshop or laboratory of a pharmaceutical company, documenting the specific pharmaceutical processes and equipment operations performed by frontline operators. Physical space movement trajectories refer to the continuous coordinate changes of an operator's body position or limb movements within the three-dimensional geographic space of the workshop over time during the execution of a specific pharmaceutical work task. Workshop environmental constraints refer to physical or chemical environmental indicators that must be strictly monitored and controlled within specific threshold ranges in the pharmaceutical work area to meet the production process and quality compliance requirements of specific drugs. These indicators include environmental cleanliness, temperature, humidity, and air pressure differentials.

[0034] Pre-defined pharmaceutical equipment interaction events refer to standardized action models pre-established in the system's underlying database. These models represent operational contacts with specific business implications between operators and pharmaceutical equipment, such as opening valves, adjusting instrument panel parameters, or adding materials. Multi-dimensional job capability nodes are independent data units representing specific comprehensive practical skills, formed in the graph database by packaging and encapsulating extracted individual equipment interaction actions with various feature dimensions such as environmental constraints. Pre-defined temporal dependency rules are logical constraints pre-set within the system based on standard pharmaceutical process specifications. These rules are specifically used to define and control the necessary sequence and causal triggering relationships between different operational skills or production steps.

[0035] Understandably, the textbook reconstruction system first acquires real-world operational video logs containing Good Manufacturing Practice (GMP) data for pharmaceuticals and uses them as job requirement data for industry-education integration enterprises. This is to ensure that the subsequently extracted skill standards meet the compliance requirements of frontline enterprises. Secondly, the system uses video image analysis technology to analyze the job requirement data frame-by-frame, identifying key points or limb movements of personnel in the footage to extract physical spatial movement trajectories in the form of spatial coordinate changes. Simultaneously, by identifying equipment dashboards or docked sensor logs in the footage, it extracts workshop environmental constraint parameters (e.g., using a pre-trained human posture estimation model to process real-world operational video logs frame-by-frame, identifying and extracting the spatial coordinates of key points in the operator's hand bones, fitting the spatial coordinates based on a time series to obtain the hand's movement trajectory in space, and the humidity displayed in the sterile workshop at that time). This is done to transform unstructured visual images into numerical indicators that can be quantified by computers.

[0036] Then, the system calculates and compares the similarity of trajectory features and coordinate offsets between the extracted physical space movement trajectory and various standard action models pre-stored in the underlying database, thereby mapping it to the corresponding preset pharmaceutical equipment interaction events (e.g., determining that a pressing trajectory with a specific arc and force direction belongs to the "starting the freeze dryer" event). This aims to give the simple physical movement trajectory actual pharmaceutical process business meaning. Next, the system concatenates the identified preset pharmaceutical equipment interaction events with the workshop environmental constraint parameters extracted at the corresponding time and encapsulates the feature vectors, deeply binding the operation action instructions with environmental constraints, thereby constructing multi-dimensional job capability nodes (e.g., generating a comprehensive skill data unit that includes the "starting the freeze dryer" action and requires "the ambient humidity must be below 50%). This is to comprehensively and accurately restore all assessment dimensions of this skill in real production.

[0037] Finally, the system strictly follows the preset temporal dependency rules, that is, according to the order and causal triggering relationship that each operation step in the national standard pharmaceutical process must follow, and uses directed edges to connect the isolated multi-dimensional job capability nodes in the actual process flow (for example, using directed arrows to point the "material mixing" node to the "fluidized bed granulation" node in one direction), and finally obtains a pharmaceutical job capability map that presents the complete business flow, thereby transforming the scattered video operation records into a network structure model that can be directly called and matched by the computer.

[0038] As an example, a pharmacy job competency map can also be constructed based on drug quality control and testing data. The steps include: acquiring drug quality control and testing data, which includes drug testing records and instrument operation logs from a laboratory information management system; extracting key process parameter offsets and sample characteristic peak vectors from the drug quality control and testing data; mapping the key process parameter offsets to corresponding standard operating procedure (SOP) quality control points; constructing multi-dimensional testing capability nodes based on the SOP quality control points and the sample characteristic peak vectors; and logically connecting the multi-dimensional testing capability nodes according to preset validation dependency rules to obtain the pharmacy job competency map.

[0039] It should be noted that a Laboratory Information Management System (LIMS) is a computerized information management system centered on a database, used in pharmaceutical testing laboratories for sample transfer, analytical data acquisition, quality standard management, and resource allocation. Pharmaceutical testing records refer to the written or electronic documentation of raw data, calculation processes, and final result determinations generated by testing personnel during the physicochemical property analysis and active ingredient content determination of pharmaceuticals, in accordance with quality standards and operating procedures. Instrument operation logs are time-series data sets generated in real-time by the equipment control software during the execution of specific testing tasks by laboratory analytical instruments, including equipment start-up and shutdown times, operating status parameters, system fault alarms, and maintenance records. Key process parameter offsets refer to the degree of numerical difference between the measured core physical variables (such as mobile phase flow rate, column temperature, or detection wavelength) in the laboratory environment or within the analytical instrument and the preset target standard values ​​during the testing process.

[0040] Sample characteristic peak vectors refer to the transformation of key fingerprint features such as retention time, peak area, or absorption intensity from chromatographic and spectral analysis results into multi-dimensional mathematical arrays using digital feature extraction technology, which objectively characterize the chemical properties of drug components. Standard Operating Procedure (SOP) quality control points refer to the key steps, critical operational nodes, or data verification nodes in the standard operating procedures for drug testing that are clearly defined as having a decisive impact on the accuracy and compliance of the test results. Multi-dimensional testing capability nodes refer to the encapsulation of extracted instrument interaction data, testing environment parameters, and corresponding personnel operation trajectories into multi-dimensional attributes, forming independent structured data units representing a specific physicochemical testing skill in a spectral database.

[0041] Pre-defined validation dependency rules refer to deterministic logic control strategies pre-set within the system based on drug testing quality control standards. This strategy transforms the execution sequence, pre-data unlocking conditions, and instrument operating status access restrictions that must be followed in the testing process into machine-readable conditional statements, which serve as the basis for determining whether logical connections can be established between multi-dimensional testing capability nodes in the graph. Regarding temporal constraints, this rule is specifically manifested as a system-mandated logic for determining the order of actions. For example, the system stipulates that the blank solvent injection node must and can only be executed before the test sample injection node; otherwise, the system will block the connection of subsequent nodes. Regarding pre-condition access restrictions, this rule is specifically manifested as cross-node data triggering conditions. For example, the system stipulates that the downstream operation node responsible for calculating the sample peak area integral will only be activated and allowed to execute when the theoretical plate number output by the system suitability test node exceeds a set numerical threshold. In terms of mutual exclusion and error correction, this rule manifests as a blocking mechanism for violations. For example, when the system detects that the data of the upstream operation node, mobile phase degassing, is empty, it will cut off the connection path of the node that generates the inspection report according to this rule.

[0042] Understandably, firstly, the textbook reconstruction system uses database interfaces to read drug testing records and instrument operation logs from the laboratory information management system as its basic data source, aiming to obtain records of personnel behavior and equipment status under real testing conditions. Secondly, the system parses the log text using regular expressions, calculates the difference between the actual instrument operating parameters and the target setpoint to obtain the offset of key process parameters (e.g., calculating the temperature difference between the actual column temperature of the liquid chromatograph and the setpoint of 30°C), and uses Principal Component Analysis (PCA) to extract the sample characteristic peak vector from the spectral or chromatographic data of the testing records. This step is to transform unstructured text and graphical testing results into numerical features that can be computed by a computer.

[0043] Then, the system uses the TF-IDF algorithm to calculate text similarity, matching the timestamps or data tags inherent in the key process parameter offsets with the standard operating procedure (SOP) text stored in the system, and mapping them to the corresponding SOP quality control points. This clarifies which specific inspection process the data deviation occurred in (e.g., mapping flow rate offset data to the "adjusting the mobile phase ratio" SOP quality control point). Next, the textbook reconstruction system concatenates and encapsulates the mapped SOP quality control point data with the sample characteristic peak vectors in the data dimension, generating multi-dimensional inspection capability nodes containing process status and result feature matrices. The aim is to combine disparate operational steps and material properties into independent structural units representing individual physicochemical inspection skills.

[0044] Finally, the system traverses all multi-dimensional verification capability nodes, calls the If-Then logical statement in the preset verification dependency rules to determine the data association between nodes. When the data status between nodes meets the requirements set by the rules, directed connections are generated between the corresponding nodes to complete the logical edge connection. In this way, the discrete skill nodes are reconstructed into a pharmaceutical job capability map with a mesh topology according to industry standards.

[0045] As an example, the steps of constructing a pharmaceutical basic knowledge graph based on pharmaceutical knowledge points in traditional pharmaceutical textbooks include: obtaining pharmaceutical knowledge points from traditional pharmaceutical textbooks, wherein the pharmaceutical knowledge points include two-dimensional chemical structure images and chemical reaction equation texts; converting the two-dimensional chemical structure images into three-dimensional molecular conformation coordinate matrices, and encapsulating the three-dimensional molecular conformation coordinate matrices into knowledge entity nodes; determining the reactant-product mapping relationship between the corresponding knowledge entity nodes based on the chemical reaction equation texts; calculating the reaction energy barrier values ​​between the knowledge entity nodes with the reactant-product mapping relationship; and connecting the knowledge entity nodes with the reactant-product mapping relationship using the reaction energy barrier values ​​as edge weights to obtain the pharmaceutical basic knowledge graph.

[0046] It should be noted that pharmaceutical knowledge points refer to independent data units that constitute the academic system of pharmacy, including basic concepts, chemical structural characteristics, pharmacological reaction mechanisms, reaction equations, and theoretical rules extracted from theoretical books on pharmaceutics or pharmaceutical analysis. Two-dimensional chemical structure images refer to static image files that visually display the types and numbers of atoms and their chemical bond topologies within a chemical molecule in a planar graphic format, as found in traditional paper textbooks or digital documents. Chemical reaction equation text refers to string-type data written using standard chemical symbols, numbers, and specific formatting to describe the transformation laws of substances during the synthesis or degradation of a specific drug. A three-dimensional molecular conformation coordinate matrix refers to a multidimensional numerical array calculated using spatial configuration algorithms, precisely recording the absolute spatial position and spatial arrangement of each atom within a chemical molecule in a three-dimensional Cartesian coordinate system.

[0047] A knowledge entity node refers to an independent data processing unit representing a specific chemical molecule, formed in a graph database after its spatial coordinate matrix data and related physicochemical properties are encapsulated in a low-level computer structure. The reactant-product mapping relationship refers to the logical matching rules established in the low-level computer logic to characterize the material transformation path and causal orientation between reactants (source material objects) and products (target material objects) before and after a specific chemical reaction. The reaction energy barrier value is an objective calculation index derived from thermodynamic theory and structural deformation calculation models, used to accurately quantify the activation energy threshold or energy difference that must be overcome during the transformation from source to target material. Edge weight refers to the specific numerical value assigned to the connection between two entity nodes in a graph network data structure; in this context, it specifically represents the energy cost required to overcome for a chemical reaction to occur, used in graph algorithms to characterize the computational resistance or difficulty of the transformation path between nodes.

[0048] Understandably, the textbook reconstruction system first uses optical character recognition and image segmentation technology to scan and parse the electronic documents of traditional pharmaceutical textbooks, extracting two-dimensional chemical structure images and chemical reaction equation texts. This is done to transform static printed content into digital source data that can be directly read by algorithms (e.g., extracting a 2D planar topological diagram of aspirin and its synthesis equation text from a page of the textbook). Secondly, the system uses optical chemical structure recognition (OCSR) technology to perform feature recognition on the extracted two-dimensional chemical structure images, converting them into standard SMILES strings describing molecular topology. Then, it calls a built-in cheminformatics algorithm library and uses distance geometry algorithms (such as the ETKDG algorithm) to parse the SMILES strings to generate a three-dimensional molecular conformation coordinate matrix of each atom in space. This matrix data is then packaged and encapsulated as knowledge entity nodes. This process aims to reconstruct planar graphics into three-dimensional digital objects that can participate in physical calculations.

[0049] Then, the system uses a chemical syntax parser to split the chemical reaction equation text into strings, using "generation arrows" or "equal signs" in the text as logical dividing lines. The chemical symbols parsed on the left are matched to reactant knowledge entity nodes, which are the starting data source, and the chemical symbols on the right are matched to product knowledge entity nodes, which are the target data source. This clearly determines the reactant-product mapping relationship between the corresponding knowledge entity nodes in the underlying database. This is to establish a clear causal transformation chain between the originally isolated substance nodes (for example, by parsing the text, a directional generation logic is established between the "salicylic acid" node and the "aspirin" node).

[0050] Next, the system calculates the reaction energy barrier required for the transformation between the preceding and following nodes that have a mapping relationship. The formula for calculating the reaction energy barrier is as follows: in, This refers to the numerical value of the reaction energy barrier; , These are the molecular topological structures of the reactants and products, respectively. It refers to the graph edit distance between reactants and products calculated based on the Maximum Common Subgraph (MCS) algorithm; It refers to the standard activation Gibbs free energy extracted or associated from the text of chemical reaction equations; is the structural energy conversion coefficient, used to harmonize the topological distance order to the energy order.

[0051] Finally, in the graph database, the system creates directional connections with reactant nodes as the starting point and product nodes as the ending point, and directly writes the calculated reaction energy barrier values ​​as edge weights into the attributes of the connections. By connecting knowledge entity nodes with mapping relationships, a network of pharmaceutical basic knowledge graphs with energy calculation scales is finally constructed.

[0052] Please refer to Figure 2 , Figure 2 This is a schematic diagram illustrating the construction process of the pharmaceutical basic knowledge map provided in Embodiment 1 of the pharmaceutical digital textbook reconstruction method based on capability maps in this application. The process begins with the input of pharmaceutical knowledge point examples, and then proceeds in two parallel paths: structure transformation and reaction analysis. In the structure transformation path, the system acquires a two-dimensional chemical structure image (such as the aspirin structure diagram), and converts it into a numerical matrix recording the three-dimensional spatial positions of atoms through two-dimensional transformation logic, namely, a three-dimensional molecular conformation coordinate matrix (shown in the figure). , , , The system uses isotropic data tables and 3D conformational coordinate visualization models, and encapsulates this matrix data into knowledge entity nodes. In the reaction analysis path, the system extracts the text of chemical reaction equations (such as acetylation reaction equations), and through reaction logic analysis, clearly determines the reactant-product mapping relationship between reactant nodes and product nodes. Next, the process enters the edge weight calculation and connection establishment stage. For node pairs with mapping relationships, the system calculates the reaction energy barrier values ​​between them and directly uses these values ​​as edge weights (represented by Greek letters in the diagram). (Representing energy cost) Establish logical connections between nodes. Finally, the processing results of these two paths converge, integrating the static three-dimensional properties of matter with dynamic transformation laws, and outputting a knowledge graph fragment example containing multiple interconnected knowledge entity nodes.

[0053] As an example, the step of extracting corresponding digital media resources from an online course platform based on the target digital textbook knowledge graph to obtain target digital textbook media resources includes: calculating the cognitive abstraction index of each graph node in the target digital textbook knowledge graph; selecting target nodes whose cognitive abstraction index is greater than a preset abstraction threshold, and extracting the structural feature vector, thermodynamic feature vector, and spatiotemporal operation feature vector of the target nodes; concatenating the structural feature vector, thermodynamic feature vector, and spatiotemporal operation feature vector into a multidimensional retrieval vector; performing an approximate nearest neighbor retrieval on the online course platform based on the multidimensional retrieval vector to obtain a candidate media resource set; calculating the multimodal complementary information entropy between each candidate media resource in the candidate media resource set and the target node, and taking the resource with the largest multimodal complementary information entropy as the target digital textbook media resource.

[0054] It should be noted that the cognitive abstraction index is a comprehensive algorithmic indicator calculated based on attributes such as the topological connectivity and semantic depth of graph nodes. It is used to quantitatively assess the difficulty and abstraction level of a specific knowledge point in the learner's subjective understanding. The preset abstraction threshold is a fixed numerical limit (e.g., 0.7) pre-set within the system, used as a criterion to determine which knowledge points in the graph are too obscure and difficult to understand, thus requiring mandatory matching of multimedia auxiliary resources. Target nodes are specific graph data units that, after cognitive abstraction index calculation and screening, exceed the preset trigger threshold and are deemed by the system to urgently require multimedia materials for intuitive explanation and teaching assistance. The structural feature vector is a set of numerical sequences generated by mathematically reducing and encoding the static disciplinary information such as the chemical molecular skeleton and basic theoretical concepts contained in a specific knowledge point, used to characterize the inherent academic theoretical attributes of that knowledge point.

[0055] Thermodynamic feature vectors refer to a set of numerical sequences characterizing the thermodynamic properties of a specific production process node by extracting and mathematically transforming industrial environmental constraints and parameters such as temperature, pressure, and mass transfer efficiency. Spatiotemporal operation feature vectors refer to a set of numerical sequences generated by mathematically modeling and encoding the three-dimensional coordinate changes of limb movements and the temporal sequence of actions during personnel operation, used to accurately describe the dynamic physical operation process. Multidimensional retrieval vectors refer to a high-dimensional mathematical array that comprehensively characterizes the features of complex skill nodes by concatenating and spatially reducing multiple single-dimensional feature vectors representing basic theory, industrial environment, and physical operation. Approximate nearest neighbor retrieval is an efficient spatial search algorithm widely used in large-scale databases. It allows the system to sacrifice a very small probability of absolute accuracy, quickly identifying a group of data objects most similar to the target retrieval vector by approximating spatial distance.

[0056] The candidate media resource set refers to a pool of alternative data composed of multiple multimedia materials that are initially matched from a massive online course platform using a high-dimensional space similarity search algorithm and have a high probability of semantic and feature relevance to the target teaching node. Candidate media resources refer to specific video clips, virtual simulation animations, or charts, or other independent digital files existing in the alternative data pool, awaiting final quantitative evaluation of their complementarity by the system. Multimodal complementary information entropy is a quantitative evaluation index calculated based on information theory models. It is specifically used to measure the extent to which the audiovisual data contained in a particular multimedia material can fill the expressive blind spots of pure text or graph nodes; a higher value indicates a stronger teaching gain and explanatory power brought by the material.

[0057] Understandably, the textbook reconstruction system first analyzes the connectivity and semantic depth of each node in the knowledge graph of the target digital textbook to calculate the cognitive abstraction index, which quantifies the difficulty of understanding. Nodes with a cognitive abstraction index greater than a preset abstraction threshold are then extracted as target nodes. This is done to locate difficult knowledge points that are difficult to understand by text alone and require multimedia assistance.

[0058] The formula for calculating the cognitive abstraction index is as follows: In the formula, This refers to the cognitive abstraction index. It refers to the in-degree of a node (i.e., the number of connections pointing to that knowledge point). This refers to the out-degree of the node (i.e., the number of connections derived from this concept). Adding 1 prevents the denominator from being zero. This refers to the current depth (level number) of the node in the semantic hierarchy tree. This refers to the maximum depth of the semantic hierarchy in the graph. , This refers to the preset weighting coefficients. .

[0059] Secondly, the system calls different underlying encoding models to extract three types of features for these target nodes: using graph embedding algorithms (such as Node2Vec) to encode the topological structure of the node's theoretical knowledge network to obtain structural feature vectors; extracting engineering constraint values ​​such as temperature and pressure recorded in the node attributes and performing matrix normalization to obtain thermodynamic feature vectors; and using sequence neural networks (such as LSTM) to reduce the dimensionality of the three-dimensional spatial coordinate changes and timestamps of the actions to obtain spatiotemporal operation feature vectors. This aims to transform complex integrated practical skills into mathematical arrays that can be accurately computed by computers (for example, transforming the "reactor feeding" node into numerical features containing three independent dimensions: material topology, ambient temperature, and human movement trajectory).

[0060] Then, the system directly concatenates the extracted structural feature vector, thermodynamic feature vector, and spatiotemporal operation feature vector along the data dimension (using a preset zero vector to fill in missing feature vector dimensions), fusing them to generate a unified high-dimensional multidimensional retrieval vector. Next, using this multidimensional retrieval vector as query input, the system performs approximate nearest neighbor retrieval in the online course platform's vector database using high-dimensional spatial indexing algorithms such as Hierarchical Navigation Small World (HNSW) or Locality Sensitive Hash (LSH). By quickly calculating the cosine similarity between the query vector and the feature vectors of the database materials, a batch of spatially closest materials is recalled to form a candidate media resource set. This is to significantly avoid the massive computational consumption caused by global traversal calculations while ensuring search matching accuracy.

[0061] Finally, the system uses a multimodal fusion algorithm to extract the visual and audio feature vectors of videos or images in the candidate media resource set, and calculates the mutual information between them and the original text features of the target node (i.e., how much new information that the new image material brings that is not in the text). This yields the multimodal complementary information entropy, which reflects the degree of complementarity between the two. The system then directly selects the resource with the largest entropy value as the target digital teaching material media resource. This is done to ensure that the final matched multimedia material can provide new visual incremental information to the maximum extent to fill the expression blind spots of theoretical text (for example, among multiple materials, the system will select a 3D simulation animation that dynamically shows the microscopic process of drug dissolution based on high information entropy, rather than selecting a static slide screenshot that simply repeats a chemical formula).

[0062] As an example, the step of obtaining a draft pharmaceutical digital textbook by performing structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph includes: using the node text attributes in the target digital textbook knowledge graph as theoretical text resources and the target digital textbook media resources as production training resources; calculating the semantic association strength between the theoretical text resources and the production training resources, and determining their typesetting proximity based on the semantic association strength; allocating corresponding display slot positions for the theoretical text resources and the production training resources in a preset dynamic grid typesetting container based on the typesetting proximity; parsing the key process parameters in the theoretical text resources and locating the operation frame segments containing the key process parameters in the video stream of the production training resources; generating an interactive floating window at the theoretical text resource at the display slot position, and associating the operation frame segments as the default display content of the interactive floating window; and performing visual alignment and style rendering on the preset dynamic grid typesetting container after filling the resources to obtain the draft pharmaceutical digital textbook.

[0063] It should be noted that node text attributes refer to data fields attached to various entity data nodes in the underlying database of the knowledge graph, specifically used to store basic teaching content such as pharmacological concepts, operational procedures, or theoretical descriptions in plain text string format. Semantic association strength refers to a numerical indicator calculated using natural language processing and multimodal matching algorithms, used to quantify the degree of relevance and business logic between a piece of theoretical text data and specific multimedia materials. Layout proximity refers to the spatial arrangement constraints set in automated layout generation algorithms based on the closeness of data associations, used to guide the interface rendering engine in determining the absolute physical spacing and relative positional relationship between text modules and multimedia modules on the final display screen. Pre-defined dynamic grid layout containers refer to pre-written, highly flexible, and reconfigurable modular user interface code frameworks in the system's front-end framework, allowing internally loaded data components to be dynamically added, deleted, dragged, or modularly rearranged according to layout parameters (such as proximity) output by the back-end algorithm.

[0064] Display slot location refers to a specific coordinate area or component placeholder planned and reserved in advance in the front-end page rendering logic, specifically used to receive and present text segments or media player streams scheduled there after being calculated by the typesetting algorithm. Key process parameters refer to specific physical quantities or chemical condition limits extracted from theoretical texts using technologies such as named entity recognition, which play a decisive role in controlling the quality and safety of the pharmaceutical process. Examples include specific temperature values, stirring speeds, or pressure difference thresholds. Operation frame segments refer to short, continuous video image sequences extracted from lengthy, continuous production training video streams using timestamp positioning and video segmentation technologies. These sequences contain only the specific actions of workers adjusting or reading the aforementioned core control parameters. Interactive floating windows are hidden, floating user interface components embedded in the front-end code of digital teaching materials. When a learner hovers the mouse over or clicks on specific text words, it temporarily pops up and floats above the main text and image content of the current page, used to instantly play related short videos or display dynamic feedback information.

[0065] Understandably, firstly, the textbook reconstruction system extracts the node text attributes from the target digital textbook knowledge graph and the acquired target digital textbook media resources. The former is marked as theoretical text resources in the system, and the latter as production training resources. This is done to distinguish the data types of basic text content and multimedia materials at the underlying logic level for separate processing. Secondly, the system uses basic word frequency statistical models (such as the TF-IDF algorithm) to calculate the keyword overlap between the theoretical text resources and the supplementary explanatory text of the production training resources, calculating a numerical value as the semantic association strength. Based on this strength score, it calculates the proximity of the two on the final interface. This aims to use an objective calculation score to determine how close the text and related videos should be on the screen.

[0066] Then, the system reads the calculated proximity values ​​and assigns corresponding display slot positions to theoretical text resources and production training resources within a preset dynamic grid layout container using a flexible grid layout. This replaces manual dragging and dropping, automatically placing the teaching content in the layout (for example, placing the operation video slot with a very high relevance score directly below the text description slot). Next, the system uses basic text extraction tools such as regular expressions to extract values ​​with units of measurement from the theoretical text resources as key process parameters. Using these parameters, it matches the timeline labels or subtitles of the production training resources, extracting the operation frame segments corresponding to the displayed values.

[0067] Subsequently, the system generates an interactive floating window at the corresponding theoretical text resource location in the front-end code, which can be triggered by mouse hover. The extracted operation frame fragment is then linked and bound to this window, setting it as the default display content. This establishes a direct link between text and video, allowing learners to easily access the corresponding short video when reading about a specific parameter. Finally, the system applies general front-end layout rules (such as setting uniform text line height and image margins) to the preset dynamic grid layout container filled with various elements for visual alignment and style rendering, completing the beautification of the layout and ultimately outputting a draft of a pharmaceutical digital textbook with basic text and image structure and initial style.

[0068] As an example, the steps of configuring and reconstructing the interactive functions of the draft pharmaceutical digital textbook to obtain the target pharmaceutical digital textbook include: identifying multi-dimensional job competency nodes in the draft pharmaceutical digital textbook and obtaining the preset parameter safety range of the multi-dimensional job competency nodes; obtaining an abnormal state rendering model composed of a conditional generative adversarial network, the abnormal state rendering model being used to generate corresponding pharmaceutical physical failure morphology images based on input process parameters; embedding a parameter adjustment component in the display area corresponding to the multi-dimensional job competency nodes in the draft pharmaceutical digital textbook; compiling and encapsulating the preset dynamic association script, the parameter adjustment component, and the draft pharmaceutical digital textbook using Hypertext Markup Language to obtain the target pharmaceutical digital textbook, wherein the preset dynamic association script is used to trigger the abnormal state rendering model to output the pharmaceutical physical failure morphology image when the parameter adjustment component receives a parameter value exceeding the preset parameter safety range.

[0069] It should be noted that the preset parameter safety range refers to the reasonable range of fluctuations in various physical and chemical indicators, pre-defined in pharmaceutical industry production standards, that ensures the normal formation of the drug and prevents quality or safety accidents. Conditional Generative Adversarial Network (cGAN) is an improved architecture that introduces conditional constraints on traditional GANs. It forces the generator to synthesize images under specific conditions by using additional auxiliary information (such as process parameters in this embodiment) as input, rather than generating disordered random data. The input process parameters refer to test values ​​representing specific production conditions, such as custom temperature, pressure, or stirring speed, which learners actively submit to the backend algorithm for simulation calculations through the interactive interface when operating digital teaching materials on the front-end device.

[0070] An abnormal state rendering model refers to a dedicated algorithm engine used in computers to simulate and calculate the visual images corresponding to specific process failures. Based on an adversarial network framework, it is trained on a large dataset of pharmaceutical defective product images with real process parameter numerical labels. During the training phase, real process parameters and noise vectors are used as input to the generator. The generated failure image and the corresponding real process parameters are input to the discriminator. Through joint optimization of the adversarial loss function and the conditional classification loss function, the model acquires the ability to generate corresponding failure visual features based on specific input parameters.

[0071] Pharmaceutical physical failure morphology images refer to visual images synthesized by underlying algorithms that intuitively demonstrate the irreversible physical damage characteristics of pharmaceuticals under conditions of non-compliant production, such as cracked tablets, layered solutions, or charred and clumped powder. Parameter adjustment components are interactive program modules embedded in the front-end user interface of digital textbooks, allowing learners to dynamically change background simulation environment variables through physical actions such as mouse dragging or keyboard input. These modules include sliders, numerical input boxes, or virtual knobs. Pre-set dynamic association scripts are pre-written logic control code at the bottom layer of the textbook system. They are specifically responsible for real-time monitoring of values ​​transmitted from the front-end interface and, upon detecting out-of-bounds data, establishing front-end and back-end communication and triggering the back-end computing engine to perform response actions such as screen updates.

[0072] Understandably, firstly, the textbook reconstruction system scans the page code to identify multi-dimensional job competency nodes representing practical steps in the draft pharmaceutical digital textbook, and simultaneously extracts the corresponding preset parameter safety ranges from the database. This is done to define reasonable numerical boundaries for the allowable fluctuations of each process operation in the background (for example, reading that the temperature of a certain reaction must be controlled between 20 and 30 degrees Celsius). Secondly, the system calls an abnormal state rendering model trained by a conditional generative adversarial network from the model library and deploys it as the background visual computing engine. The purpose is to prepare to calculate and generate pharmaceutical physical failure morphology images in real time based on various process parameters attempted by the user in subsequent processes.

[0073] Then, the system locates the multi-dimensional job competency nodes on the page and inserts a parameter adjustment component into the front-end code. This provides learners with an interactive entry point to manually change simulation environment parameters (e.g., adding a draggable temperature slider or numerical input box next to the text). Finally, the system merges the preset dynamic association script responsible for logic control, the newly inserted parameter adjustment component, and the original draft of the pharmaceutical digital textbook, using Hypertext Markup Language for unified code compilation and packaging to output the final target pharmaceutical digital textbook. During this encapsulation process, the preset dynamic association script establishes a data monitoring mechanism. When a learner inputs an incorrect value that exceeds the preset parameter safety range on the interface, the script captures this out-of-bounds action and triggers the abnormal state rendering model in the background. The model calculates and displays a pharmaceutical physical failure mode image of the failed operation to the page, thus transforming the originally one-way reading page into a dynamic digital system with operational trial and error and result feedback.

[0074] This embodiment provides a method for reconstructing pharmaceutical digital teaching materials based on competency graphs. First, a pharmaceutical competency graph is constructed based on job demand data from industry-education integration enterprises, transforming real-world industrial operational standards into a structured model and clarifying core skill requirements in the production environment. Simultaneously, a pharmaceutical fundamentals knowledge graph is constructed based on pharmaceutical knowledge points from traditional pharmaceutical textbooks to preserve and organize a rigorous theoretical framework. Second, the pharmaceutical fundamentals knowledge graph and the pharmaceutical competency graph are fused and aligned to obtain the target digital teaching material knowledge graph. This effectively breaks down the underlying logical barriers between academic theory and factory practice, achieving a deep connection between the two. Finally, the target digital teaching material knowledge graph is used in online courses... Taichung extracted relevant digital media resources to obtain target digital teaching material media resources, thus supplementing the dry theoretical framework with concrete multimedia auxiliary materials. Next, based on the target digital teaching material media resources and the target digital teaching material knowledge graph, a structured layout was performed to obtain a draft of the pharmaceutical digital teaching material, achieving automated and rational spatial arrangement and assembly of textual and graphical knowledge and video content. Finally, the interactive functions of the draft pharmaceutical digital teaching material were configured and reconstructed to generate the target pharmaceutical digital teaching material, giving the static teaching material page dynamic operation and real-time feedback capabilities. Through the synergistic effect of the above steps, the problems of outdated content, monotonous presentation, and inability to dynamically match the actual job competency requirements of enterprises under the industry-education integration model were effectively solved.

[0075] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 , Figure 3This is a flowchart illustrating the second embodiment of the pharmaceutical digital textbook reconstruction method based on capability maps according to this application. Step S30 of the pharmaceutical digital textbook reconstruction method based on capability maps includes steps S31 to S35: Step S31: Identify the knowledge entity nodes in the pharmaceutical basic knowledge graph and identify the multi-dimensional job competency nodes in the pharmaceutical job competency graph that correspond to the knowledge entity nodes; Step S32: Calculate the thermodynamic engineering scaling factor between the knowledge entity node and the multidimensional job capability node; Step S33: Extract the feature vector of the knowledge entity node, and adjust the feature vector by dimensional projection according to the thermodynamic engineering scaling factor to obtain the adjusted feature vector; Step S34: Calculate the transmission cost between the adjusted feature vector and the process feature vector corresponding to the multidimensional job capability node using the optimal transmission algorithm, and extract the mapping node pair with the minimum transmission cost; Step S35: Based on the mapping node pair, the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph are spliced ​​and reconstructed to obtain the target digital textbook knowledge graph.

[0076] It should be noted that the thermodynamic engineering scaling factor refers to the numerical multiplier or conversion coefficient used to quantify the thermodynamic energy consumption, heat and mass transfer efficiency attenuation, and parameter deviations that occur when the physicochemical parameters of pharmaceutical fundamental theories are scaled up from the ideal microscopic environment of the laboratory to the macroscopic environment of industrial production. The eigenvector refers to a set of floating-point numbers containing rich semantic features, which is converted from the original unstructured data such as molecular structural attributes or chemical reaction rules in the pharmaceutical theoretical system using low-level natural language processing or graph embedding algorithms to a set of floating-point numbers that can be directly used by computers for multidimensional spatial distance calculations. The adjusted eigenvector is a new array generated by substituting the thermodynamic engineering scaling factor as a correction weight into the original floating-point number array representing pure theoretical knowledge and performing spatial matrix transformation, thereby producing a directional scale shift in the mathematical space of the theoretical data to simulate and approximate its characteristic state under real industrial production conditions.

[0077] Optimal Transport (OT) is a classic underlying mathematical model derived from resource allocation optimization. In this embodiment, it is used to find the optimal mathematical path to transfer and match the high-dimensional feature space representing theoretical knowledge to the feature space representing factory operations with the minimum total geometric distance or probability distribution difference. The process feature vector refers to a set of multi-dimensional mathematical arrays specifically representing the actual production operation characteristics in the industrial front line, generated after digital dimensionality reduction and feature extraction of machine equipment interaction events, employee operation physical trajectories, and environmental constraint parameters in a real pharmaceutical workshop using a deep learning coding model. The transport cost is a quantified distance index obtained when matching two different data domains using the optimal transport algorithm. Its value objectively represents the semantic and logical gap or feature matching error degree that exists when forcibly associating a scaled theoretical knowledge data point with a real factory operation step. The mapping node pair refers to the optimal pair of data units containing both theoretical and application ends, selected by the system through global optimization from a massive cross-comparison matrix of theoretical entity nodes and practical capability nodes, with the minimum transport cost—that is, the pair that best matches the academic theoretical connotation and the actual industrial application scenario at the data level.

[0078] Understandably, the textbook reconstruction system first reads the text tags of each knowledge entity node by traversing the pharmaceutical basic knowledge graph, and then uses natural language processing algorithms to calculate the semantic similarity between these tags and the descriptions of each node in the pharmaceutical job competency graph. It selects multi-dimensional job competency nodes that meet the similarity criteria as preliminary matching objects. This is done in order to quickly delineate the candidate range that may have a theoretical and practical connection in the large-scale graph (for example, initially matching the "crystallization principle" node in the theoretical graph with the "controlling the cooling of the reaction vessel" node in the practical graph).

[0079] Secondly, the system calculates the thermodynamic engineering scaling factor between the two candidate nodes, using the following formula: in, It refers to the ratio of industrial production volume to laboratory theoretical volume. The logarithmic function is used because the changes in physical quantities in pharmaceutical scale-up production usually exhibit a non-linear logarithmic growth. It refers to the thermodynamic engineering shrinkage factor; It refers to the absolute value of the deviation between the heat transfer / mass transfer efficiency and the theoretical efficiency caused by factors such as equipment wall thickness and stirring power in an industrial environment; This refers to the theoretical transmission efficiency under ideal conditions; This refers to the preset process correction constant (such as the fluid viscosity correction coefficient); exponential decay term. Used to characterize the penalty weight for system non-ideality caused by mass transfer deviation.

[0080] Subsequently, the graph neural network model is invoked to read the molecular coordinates and reaction attribute data inside the knowledge entity nodes, and the data is dimensionality-reduced and encoded into a series of feature vectors in the form of a multidimensional array. Then, the calculated thermodynamic engineering scaling factor is used to construct a weight transformation matrix, and matrix multiplication is performed directly with the feature vector. This causes the coordinate position of the original vector in the multidimensional mathematical space to shift accordingly, thereby completing the dimensional projection adjustment and obtaining the adjusted feature vector. This approach aims to use mathematical transformation to virtually scale up the theoretical data in the ideal laboratory environment in order to approximate its characteristic state in the real industrial production environment.

[0081] Then, since the adjusted feature vector and the process feature vector inherent in the multi-dimensional job capability node originate from different heterogeneous data domains, their initial mathematical dimensions and distribution spaces are inconsistent. Before executing the optimal transmission algorithm calculation, the system inputs the adjusted feature vector and the process feature vector into a pre-trained shared multilayer perceptron (MLP) network, respectively, and projects them into a shared latent space of uniform dimension through nonlinear mapping to obtain standard theoretical features and standard practical features with consistent dimensions. Subsequently, the Euclidean distance between the standard theoretical features and the standard practical features in the latent space is calculated as the input for calculating the optimal transmission cost. This is then substituted into the optimal transmission algorithm to calculate the mathematical distance required to fit the distribution features of the theoretical data to the distribution of the practical data in the high-dimensional space as the transmission cost. The pair with the smallest cost value is selected from all candidate combinations and extracted as the mapping node pair. This is to find the correlation relationship with the smallest feature matching error between theoretical knowledge and actual process from the bottom layer.

[0082] The formula for calculating the transmission cost is as follows: In the formula, This refers to the adjusted feature vector; It refers to the feature vector of multi-dimensional job competency nodes; This refers to the transmission matrix; It refers to the first in the transmission matrix line, number The matrix elements of the column are used to represent the first column. The theoretical knowledge point was matched to the first... The proportion coefficient of each multi-dimensional job competency node; It refers to a probabilistically coupled set that satisfies the characteristic distribution constraints of two nodes; and In optimal transmission, this is referred to as marginal distribution. It refers to the source distribution vector, which defines each knowledge point in the pharmaceutical basics knowledge graph. The importance of This refers to the target distribution vector, which defines each skill node in the pharmacy job competency map. The degree of demand.

[0083] Finally, the system uses the extracted mapping node pairs as cross-graph connection anchors in the underlying graph database. By creating new association pointers or data edges between these paired nodes, the originally independent pharmaceutical basic knowledge graph and pharmaceutical job competency graph are merged and reconstructed into a target digital textbook knowledge graph that has both academic theory and industrial practice in a dual network structure. This opens up a path from book knowledge to workshop application in terms of data architecture.

[0084] This embodiment first identifies corresponding knowledge entity nodes and multidimensional job competency nodes in the pharmaceutical basic knowledge graph and pharmaceutical job competency graph, respectively, to clarify the initial correlation between basic theory and practical operation and avoid subsequent invalid data matching calculations. Second, it calculates the thermodynamic engineering scaling factor between these two types of nodes to quantify the parameter differences when transforming from the laboratory microenvironment to the industrial large-scale production environment, providing an objective basis for the equivalent mapping of cross-environment data. Then, it extracts the feature vectors of the knowledge entity nodes and uses the scaling factor to adjust their dimensionality projection, generating adjusted feature vectors. This simulates the characteristic distribution of theoretical data on an industrial scale in mathematical space, effectively eliminating the gap between theory and practice. The scaling bias between practical applications is addressed. Next, the optimal transmission algorithm is invoked to calculate the transmission cost between the adjusted feature vector and the process feature vector corresponding to the multi-dimensional job competency nodes. The mapping node pair with the minimum transmission cost is extracted, thereby finding the theoretical and process data combination with the highest fitting degree in the underlying high-dimensional space, maximizing the accuracy of cross-graph node matching. Finally, based directly on the extracted mapping node pairs as connection anchors across the two data domains, the originally independent pharmaceutical basic knowledge graph and job competency graph are spliced ​​and reconstructed at the underlying architecture level to generate the target digital textbook knowledge graph. This breaks down the data isolation between academic theory and front-line workshop practice, constructing a comprehensive underlying data foundation that highly integrates industry and education.

[0085] For example, to help understand the implementation process of the pharmaceutical digital teaching material reconstruction method based on capability map obtained in this embodiment combined with the above embodiment one, please refer to... Figure 4 , Figure 4 A simplified flowchart of a method for reconstructing pharmaceutical digital teaching materials based on capability maps is provided, specifically: The entire reconstruction process begins with the parallel acquisition and preprocessing of multi-source data. The system introduces job requirement data from industry-education integration enterprises into the pharmaceutical job competency graph construction module to extract front-line industrial skill standards. Simultaneously, it introduces traditional pharmaceutical textbook knowledge points into the pharmaceutical basic knowledge graph construction module to extract the discipline's theoretical framework. Subsequently, the two graph data streams converge in the feature fusion and alignment mapping module. By finding semantic and business logic-related features, dimensional alignment is performed, thus stitching isolated theoretical and practical data into a comprehensive target digital textbook knowledge graph. Next, the process enters the resource adaptation stage. Using this knowledge graph as retrieval credentials, the system calls external materials from the online course platform in the digital media resource extraction module, filtering and extracting highly matching target digital textbook media resources. Subsequently, the generated graph data stream and the extracted media resource stream enter the structured typesetting and reconstruction generation module. The system performs automated space allocation and assembly according to the layout algorithm, first generating a draft of the pharmaceutical digital textbook, and then embedding dynamic association scripts and parameter adjustment components on this draft. Through code compilation and encapsulation, the final output is a target pharmaceutical digital textbook with dynamic operation and real-time feedback functions.

[0086] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the pharmaceutical digital teaching material reconstruction method based on capability maps in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0087] This application also provides a pharmaceutical digital teaching material reconstruction device based on capability maps; please refer to [reference needed]. Figure 5 The pharmaceutical digital teaching material reconstruction device based on capability graphs includes: The competency mapping module 10 is used to construct a competency map for pharmaceutical positions based on job demand data from industry-education integration enterprises. Module 20 for constructing pharmaceutical knowledge maps based on pharmaceutical knowledge points in traditional pharmaceutical textbooks; The graph fusion and alignment module 30 is used to perform feature fusion and alignment mapping between the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph to obtain the target digital textbook knowledge graph. The media resource extraction module 40 is used to extract corresponding digital media resources from the online course platform based on the target digital textbook knowledge graph to obtain the target digital textbook media resources. The structured typesetting module 50 is used to perform structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph to obtain a draft of the pharmaceutical digital textbook; The interactive reconstruction module 60 is used to configure and reconstruct the interactive functions of the draft pharmaceutical digital textbook to obtain the target pharmaceutical digital textbook.

[0088] The pharmaceutical digital textbook reconstruction device based on capability graphs provided in this application, employing the pharmaceutical digital textbook reconstruction method based on capability graphs in the above embodiments, can solve the technical problems of outdated content, monotonous presentation methods, and inability to dynamically match the actual job capability requirements of enterprises under the integration of industry and education. Compared with the prior art, the beneficial effects of the pharmaceutical digital textbook reconstruction device based on capability graphs provided in this application are the same as those of the pharmaceutical digital textbook reconstruction method based on capability graphs provided in the above embodiments, and other technical features in the pharmaceutical digital textbook reconstruction device based on capability graphs are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0089] This application provides a pharmaceutical digital textbook reconstruction device based on capability map. The pharmaceutical digital textbook reconstruction device based on capability map includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the pharmaceutical digital textbook reconstruction method based on capability map in the first embodiment described above.

[0090] The following is for reference. Figure 6 This document illustrates a structural schematic diagram of a capability-map-based pharmaceutical digital textbook reconstruction device suitable for implementing embodiments of this application. The capability-map-based pharmaceutical digital textbook reconstruction device in this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Android Devices), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 6 The pharmaceutical digital teaching material reconstruction device based on capability map shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0091] like Figure 6As shown, the capability map-based pharmaceutical digital textbook reconstruction device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in ROM (Read Only Memory) 1002 or a program loaded from storage device 1003 into RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the capability map-based pharmaceutical digital textbook reconstruction device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, LCDs (Liquid Crystal Displays), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the capability map-based pharmaceutical digital textbook reconstruction device to communicate wirelessly or wiredly with other devices to exchange data. Although a capability map-based pharmaceutical digital textbook reconstruction device with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0092] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0093] The pharmaceutical digital textbook reconstruction device based on capability graphs provided in this application, employing the pharmaceutical digital textbook reconstruction method based on capability graphs in the above embodiments, can solve the technical problems of outdated content, monotonous presentation methods, and inability to dynamically match the actual job capability requirements of enterprises under the integration of industry and education. Compared with the prior art, the beneficial effects of the pharmaceutical digital textbook reconstruction device based on capability graphs provided in this application are the same as those of the pharmaceutical digital textbook reconstruction method based on capability graphs provided in the above embodiments, and other technical features in this pharmaceutical digital textbook reconstruction device based on capability graphs are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0094] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0095] 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.

[0096] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the pharmaceutical digital textbook reconstruction method based on capability map in the above embodiments.

[0097] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable Read Only Memory or Flash Memory), optical fibers, CD-ROM (CD-Read Only Memory), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0098] The aforementioned computer-readable storage medium may be included in a capability-map-based pharmaceutical digital textbook reconstruction device; or it may exist independently and not be assembled into a capability-map-based pharmaceutical digital textbook reconstruction device.

[0099] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by a pharmaceutical digital textbook reconstruction device based on a capability graph, the device performs the following actions: constructs a pharmaceutical job capability graph based on job demand data from industry-education integration enterprises; constructs a pharmaceutical basic knowledge graph based on pharmaceutical knowledge points in traditional pharmaceutical textbooks; performs feature fusion and alignment mapping between the pharmaceutical basic knowledge graph and the pharmaceutical job capability graph to obtain a target digital textbook knowledge graph; extracts corresponding digital media resources from an online course platform based on the target digital textbook knowledge graph to obtain target digital textbook media resources; performs structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph to obtain a draft pharmaceutical digital textbook; and configures and reconstructs the draft pharmaceutical digital textbook using interactive functions to obtain the target pharmaceutical digital textbook.

[0100] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including LAN (Local Area Network) or WAN (Wide Area Network)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0101] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0102] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0103] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described pharmaceutical digital textbook reconstruction method based on capability graphs. This addresses the technical problems of outdated content, monotonous presentation methods, and inability to dynamically match the actual job competency requirements of enterprises under the industry-education integration model in traditional pharmaceutical textbooks. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the pharmaceutical digital textbook reconstruction method based on capability graphs provided in the above embodiments, and will not be elaborated upon here.

[0104] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for reconstructing pharmaceutical digital teaching materials based on capability maps.

[0105] The computer program product provided in this application can solve the technical problems of outdated content, monotonous presentation, and inability to dynamically match the actual job competency requirements of enterprises under the integration of industry and education. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the pharmaceutical digital textbook reconstruction method based on competency maps provided in the above embodiments, and will not be repeated here.

[0106] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for reconstructing pharmaceutical digital teaching materials based on capability maps, characterized in that, The method includes: Construct a pharmacy job competency map based on job demand data from industry-education integration enterprises; Construct a basic pharmaceutical knowledge map based on pharmaceutical knowledge points in traditional pharmaceutical textbooks; The pharmaceutical basic knowledge graph and the pharmaceutical job competency graph are fused and aligned to obtain the target digital textbook knowledge graph. Based on the target digital textbook knowledge graph, corresponding digital media resources are extracted from the online course platform to obtain the target digital textbook media resources; A draft of a pharmaceutical digital textbook is obtained by structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph; The draft pharmaceutical digital textbook is configured with interactive functions and reconstructed to generate the target pharmaceutical digital textbook.

2. The method as described in claim 1, characterized in that, The steps for constructing a pharmaceutical job competency map based on job demand data from industry-education integration enterprises include: Obtain real-world operational video logs containing pharmaceutical manufacturing quality management standards and use them as job requirement data for industry-education integration enterprises; Extract the physical space movement trajectory of operators and workshop environmental constraint parameters from the job demand data of the industry-education integration enterprises; Map the physical space movement trajectory to the corresponding preset pharmaceutical equipment interaction events; Construct multi-dimensional job capability nodes based on the preset pharmaceutical equipment interaction events and the workshop environment constraint parameters; The multidimensional job competency nodes are connected by directed edges according to preset temporal dependency rules to obtain a pharmacy job competency map.

3. The method as described in claim 1, characterized in that, The steps for constructing a pharmaceutical basic knowledge map based on pharmaceutical knowledge points in traditional pharmaceutical textbooks include: To obtain pharmaceutical knowledge points from traditional pharmaceutical textbooks, including two-dimensional chemical structure images and chemical reaction equation texts; The two-dimensional chemical structure image is converted into a three-dimensional molecular conformation coordinate matrix, and the three-dimensional molecular conformation coordinate matrix is ​​encapsulated as a knowledge entity node; Based on the chemical reaction equation text, determine the reactant-product mapping relationship between the corresponding knowledge entity nodes; Calculate the reaction energy barrier values ​​between the knowledge entity nodes that have the aforementioned reactant-product mapping relationship; Using the reaction energy barrier value as the edge weight, the knowledge entity nodes that have the mapping relationship between the reactants and products are connected to obtain a pharmaceutical basic knowledge graph.

4. The method as described in claim 1, characterized in that, The step of performing feature fusion and alignment mapping between the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph to obtain the target digital textbook knowledge graph includes: Identify the knowledge entity nodes in the pharmaceutical basic knowledge graph, and identify the multi-dimensional job competency nodes in the pharmaceutical job competency graph that correspond to the knowledge entity nodes; Calculate the thermodynamic engineering scaling factor between the knowledge entity node and the multidimensional job competency node; Extract the feature vector of the knowledge entity node, and adjust the feature vector by dimensional projection according to the thermodynamic engineering scaling factor to obtain the adjusted feature vector; The transmission cost between the adjusted feature vector and the process feature vector corresponding to the multidimensional job capability node is calculated using the optimal transmission algorithm, and the mapping node pair with the minimum transmission cost is extracted. Based on the mapping nodes, the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph are spliced ​​and reconstructed to obtain the target digital textbook knowledge graph.

5. The method as described in claim 1, characterized in that, The step of extracting corresponding digital media resources from the online course platform based on the target digital textbook knowledge graph to obtain the target digital textbook media resources includes: Calculate the cognitive abstraction index of each node in the knowledge graph of the target digital teaching materials; Target nodes whose cognitive abstraction index is greater than a preset abstraction threshold are selected, and the structural feature vector, thermodynamic feature vector, and spatiotemporal operation feature vector of the target nodes are extracted. The structural feature vector, the thermodynamic feature vector, and the spatiotemporal operation feature vector are concatenated into a multidimensional retrieval vector. Based on the multidimensional retrieval vector, an approximate nearest neighbor retrieval is performed on the online course platform to obtain a candidate media resource set; Calculate the multimodal complementary information entropy between each candidate media resource in the candidate media resource set and the target node, and take the resource with the largest multimodal complementary information entropy as the target digital teaching material media resource.

6. The method as described in claim 1, characterized in that, The step of obtaining a draft pharmaceutical digital textbook by performing structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph includes: The node text attributes in the target digital textbook knowledge graph are used as theoretical text resources, and the target digital textbook media resources are used as production training resources; Calculate the semantic association strength between the theoretical text resources and the production training resources, and determine the layout proximity between the two based on the semantic association strength; Based on the layout proximity, corresponding display slot positions are allocated to the theoretical text resources and the production training resources in a preset dynamic grid layout container; Analyze the key process parameters in the theoretical text resources and locate the operation frame segments containing the key process parameters in the video stream of the production training resources; An interactive floating window is generated at the theoretical text resource location of the display slot, and the operation frame fragment is associated as the default display content of the interactive floating window; Visual alignment and style rendering are performed on the preset dynamic grid layout container after filling resources to obtain a draft of a pharmaceutical digital textbook.

7. The method according to any one of claims 1 to 6, characterized in that, The steps of configuring and reconstructing the draft pharmaceutical digital textbook to obtain the target pharmaceutical digital textbook include: Identify the multi-dimensional job competency nodes in the draft of the pharmaceutical digital textbook, and obtain the preset parameter safety range of the multi-dimensional job competency nodes; An abnormal state rendering model composed of a conditional generative adversarial network is obtained, which is used to generate corresponding pharmaceutical physical failure morphology images based on the input process parameters. In the draft of the pharmaceutical digital textbook, a parameter adjustment component is embedded in the display area corresponding to the multi-dimensional job competency node; The preset dynamic association script, the parameter adjustment component, and the draft pharmaceutical digital textbook are compiled and encapsulated using Hypertext Markup Language to obtain the target pharmaceutical digital textbook. The preset dynamic association script is used to trigger the abnormal state rendering model to output the pharmaceutical physical failure morphology image when the parameter adjustment component receives a parameter value that exceeds the preset parameter safety range.

8. A pharmaceutical digital teaching material reconstruction device based on capability maps, characterized in that, The device includes: The competency mapping module is used to construct a competency map for pharmaceutical positions based on job demand data from industry-education integration enterprises. The pharmaceutical atlas construction module is used to construct a basic pharmaceutical knowledge atlas based on pharmaceutical knowledge points in traditional pharmaceutical textbooks. The graph fusion and alignment module is used to perform feature fusion and alignment mapping between the pharmaceutical basic knowledge graph and the pharmaceutical job competency graph to obtain the target digital textbook knowledge graph. The media resource extraction module is used to extract corresponding digital media resources from the online course platform based on the target digital textbook knowledge graph to obtain the target digital textbook media resources. The structured typesetting module is used to perform structured typesetting based on the target digital textbook media resources and the target digital textbook knowledge graph to obtain a draft of the pharmaceutical digital textbook; The interactive reconstruction module is used to configure and reconstruct the interactive functions of the draft pharmaceutical digital textbook to obtain the target pharmaceutical digital textbook.

9. A pharmaceutical digital teaching material reconstruction device based on capability maps, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the pharmaceutical digital textbook reconstruction method based on capability maps as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the pharmaceutical digital teaching material reconstruction method based on capability map as described in any one of claims 1 to 7.

Images