Building engineering construction disclosure text generation method and device, equipment and medium
By constructing a construction spatiotemporal knowledge graph and a large language model to generate a hierarchical text framework, the problems of low efficiency and poor consistency in construction briefing texts are solved, achieving efficient and standardized briefing text generation to meet the needs of complex construction conditions.
Patent Information
- Application Number
- CN202610941897.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-26
- Publication Date
- 2026-11-10
AI Technical Summary
In existing technologies, construction handover texts mainly rely on manual writing or template filling, which leads to low efficiency, easy omission of key elements, and difficulty in ensuring the completeness and consistency of the content, especially in complex construction conditions where it is difficult to meet standards and specifications.
By acquiring construction condition information, using the scenario element extraction module to form a semantic set of construction scenarios, constructing a spatiotemporal knowledge graph of building construction, and using a large language model to generate a hierarchical text framework, combined with knowledge paths to generate disclosure text, ensuring that regulatory constraints are accurately bound to construction scenarios.
It enables efficient generation of construction handover texts, ensuring the completeness and consistency of the content, eliminating subjective biases caused by manual writing, improving the professionalism and standardization of the handover texts, and adapting to the needs of different engineering scenarios.
Smart Images

Figure CN123070060A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of engineering construction management, and specifically relates to a method, apparatus, device and medium for generating construction disclosure texts for architectural engineering. Background Art
[0002] Construction disclosure is a key core link in architectural engineering construction management, which undertakes the important role of accurately conveying design intentions, construction schemes, technical specifications and key points of safety risks to on-site construction personnel, and is an important means to ensure standardized and normalized operations at construction sites. The completeness, normalization and logical consistency of construction disclosure texts are directly related to the engineering construction quality and on-site safety management level.
[0003] In the related art, the compilation of construction disclosure texts is generally completed by manual writing or simple filling of traditional templates. On the one hand, the construction conditions of architectural engineering are complex and variable, covering multi-dimensional and complicated information such as construction objects, operation environments, resource allocation, process activities, and risk factors. The manual compilation mode is not only inefficient, but also very likely to cause problems such as omission of key disclosure elements and incomplete content; on the other hand, the construction disclosure content must strictly follow general standards, industry specifications and enterprise management systems, and the interpretation, screening and citation of specification clauses by manpower are easily affected by personal subjective experience, making it difficult to ensure the standard uniformity and content consistency of disclosure contents for different projects and different batches. Summary of the Invention
[0004] The present invention provides a method, apparatus, device and medium for generating construction disclosure texts for architectural engineering, so as to solve the problems in the related art that construction disclosure texts mainly rely on manual writing or simple filling based on templates, resulting in low efficiency, easy omission of key elements and easy influence by subjective experience.
[0005] In a first aspect, the present invention provides a method for generating construction briefing text for building engineering projects. The method includes: acquiring input information of the text to be generated for the building engineering project, the input information being determined based on construction condition information including spatiotemporal annotations and the construction input text; extracting multiple types of scenario element information related to spatiotemporal annotations from the input information of the text to be generated using a pre-built scenario element extraction module to obtain a construction scenario semantic set; constructing a construction spatiotemporal knowledge graph for the building engineering project based on the construction scenario semantic set and a pre-determined set of normative predicates, the construction spatiotemporal knowledge graph including nodes, relational edges, and attribute information, where nodes represent various objects with independent semantics in the construction scenario, and relational edges represent the relationships between nodes; inputting the input information of the text to be generated into a pre-built large language model to output a hierarchical text framework; retrieving the knowledge paths of each level node in the construction spatiotemporal knowledge graph based on multiple level nodes in the hierarchical text framework; and generating the construction briefing text for the building engineering project from the hierarchical text framework and the knowledge paths of each level node.
[0006] The construction briefing text generation method provided by this invention obtains construction condition information containing spatiotemporal annotations and the text input information to be generated determined by the construction input text. It then utilizes a scenario element extraction module to automatically extract multiple types of scenario elements associated with spatiotemporal annotations and form a construction scenario semantic set. This allows for rapid structured organization of complex construction condition information, replacing manual compilation and effectively improving the efficiency of briefing text preparation. Simultaneously, it avoids the omission of key elements during manual writing, ensuring the completeness and comprehensiveness of the briefing information. Furthermore, based on the construction scenario semantic set and the standard predicate set, a construction engineering briefing text containing node, relation edge, and attribute information is constructed. The construction spatiotemporal knowledge graph transforms various standards and specifications into structured and interconnected graph knowledge. It then generates a hierarchical text framework through a large language model and uses the knowledge paths of nodes at each level of the graph to generate the handover text. This precisely binds regulatory constraints to construction scenarios, completely eliminating subjective biases from human interpretation of regulations. It ensures that the generated handover text strictly conforms to standard requirements, achieving consistent handover content across different engineering scenarios. Simultaneously, it aligns the generation of handover text with the construction spatiotemporal logic and industry regulatory framework, comprehensively improving the professionalism, standardization, and generation efficiency of construction handover texts, and fully compensating for various deficiencies in related technologies.
[0007] In one optional implementation, the step of constructing a spatiotemporal knowledge graph of building construction based on a construction scenario semantic set and a pre-determined set of normative predicates includes: obtaining normative texts of building construction; performing semantic parsing on the normative texts of building construction to obtain a set of normative predicates; performing semantic alignment between the construction scenario semantic set and the set of normative predicates to obtain an alignment result; and constructing a spatiotemporal knowledge graph of building construction based on the alignment result.
[0008] The method provided by this optional implementation performs semantic parsing on engineering construction normative texts to form a set of normative predicates. Then, it constructs a spatiotemporal knowledge graph by aligning the semantic set of construction scenarios with the semantic results of the normative predicates. This enables the standardized and structured transformation of scattered normative clauses, eliminating the subjective bias caused by manual compilation of norms. By leveraging semantic alignment, it achieves precise association between real-world construction information and normative constraints, allowing the constructed knowledge graph to fully carry the construction scenarios and corresponding normative logic, providing a reliable normative basis for the subsequent generation of disclosure texts.
[0009] In one optional implementation, the step of generating construction engineering handover text from a hierarchical text framework and the knowledge paths of each level node includes: compiling the knowledge paths of each level node to obtain a set of structured constraints; dividing the set of structured constraints into levels to generate hierarchical prompt information; and generating construction engineering handover text based on the hierarchical text framework, the set of structured constraints, and the hierarchical prompt information.
[0010] The method provided in this optional implementation method forms a structured constraint set by constraining and compiling the knowledge path, and obtains hierarchical prompt information by processing the constraint content in layers. It relies on a hierarchical text framework, the structured constraint set, and the hierarchical prompt information to collaboratively generate the handover text, achieving a deep integration of text structure, normative constraints, and generation guidance. This approach ensures the integrity and standardization of the handover text structure through the hierarchical framework, strictly controls hard compliance logic such as construction sequence, space, and parameters through structured constraints, and balances content professionalism and traceability through hierarchical prompts. It effectively solves the problems of rigid content in traditional template-filled materials, inconsistent standardization in manual writing, and easy compliance loopholes, significantly improving the accuracy, professionalism, and scenario adaptability of construction handover texts.
[0011] In one optional implementation, after generating the construction briefing text of the building project from the hierarchical text framework and the knowledge paths of each level node, the method further includes: using the spatiotemporal knowledge graph of building project construction to perform consistency verification on the construction briefing text of the building project, and obtaining the verification result; and optimizing the construction briefing text of the building project based on the verification result to obtain the target construction briefing text.
[0012] The method provided in this optional implementation, after the initial generation of the handover text, utilizes a construction engineering spatiotemporal knowledge graph to conduct automated consistency verification. This accurately identifies issues such as process sequence conflicts, spatial logic contradictions, and deviations in specification adaptation within the text, replacing the traditional manual review mode and avoiding subjective oversights and judgment biases inherent in manual verification. Simultaneously, based on the verification results, the text content is optimized in a targeted manner, correcting and improving flawed, non-compliant, and logically inconsistent content. This effectively ensures the standardization, logical rigor, and compliance accuracy of the final construction handover text, further enhancing the reliability of the handover text in guiding on-site construction.
[0013] In one optional implementation, the step of obtaining the text input information to be generated for a construction project includes: obtaining construction condition information and construction input text containing spatiotemporal annotations; and standardizing the construction condition information and construction input text to obtain the text input information to be generated.
[0014] The method provided by this optional implementation standardizes the construction condition information and construction input text with spatiotemporal annotations, unifies the original data format, organizes heterogeneous information, eliminates interference caused by messy original material formats and inconsistent expressions, facilitates the subsequent accurate extraction of scenario elements, construction of knowledge graphs and generation of disclosure frameworks, and improves the accuracy and processing stability of subsequent full-process data analysis.
[0015] In one optional implementation, the step of inputting the text to be generated into a pre-built large language model so that the large language model outputs a hierarchical text framework includes: inputting the text to be generated into the large language model so that the large language model outputs the construction briefing elements of the construction engineering construction briefing text, wherein the construction briefing elements include any one or more of the following: project overview, construction tasks and procedures, technical requirements, safety risk warnings, material and equipment requirements, parameter conditions, and precautions; and calling the construction engineering construction briefing standard hierarchical rule base to organize the construction briefing elements to obtain a hierarchical text framework.
[0016] The method provided by this optional implementation method automatically extracts various construction briefing elements using a large language model, and then combines it with a standard hierarchical rule base to complete the structured arrangement of elements to generate a hierarchical text framework. This saves the workload of manual sorting and chapter division, ensuring that key elements such as project overview, procedures, and safety are not omitted, and also conforms to industry standards to unify the text hierarchical structure. It avoids the defects of traditional template-based rigidity and non-standard manual arrangement of structures, and provides a regular and reliable framework foundation for subsequent constraint compilation and briefing generation based on knowledge paths.
[0017] Secondly, the present invention provides a construction engineering handover text generation device, which includes: an acquisition module for acquiring text input information to be generated for a construction project, the text input information being determined based on construction condition information including spatiotemporal annotations and construction input text; an extraction module for extracting multiple types of scenario element information associated with spatiotemporal annotations from the text input information to be generated using a pre-built scenario element extraction module, to obtain a construction scenario semantic set; a construction module for constructing a construction engineering spatiotemporal knowledge graph based on the construction scenario semantic set and a pre-determined set of normative predicates, the construction engineering spatiotemporal knowledge graph including nodes, relation edges, and attribute information, nodes representing various objects with independent semantics in the construction scenario, and relation edges representing the association relationships between nodes; a determination module for inputting the text input information to be generated into a pre-built large language model, so that the large language model outputs a hierarchical text framework; a retrieval module for retrieving the knowledge paths of each level node in the construction engineering spatiotemporal knowledge graph based on multiple level nodes in the hierarchical text framework; and a generation module for generating construction engineering handover text from the hierarchical text framework and the knowledge paths of each level node.
[0018] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the construction engineering handover text generation method described in the first aspect or any corresponding embodiment thereof.
[0019] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions, which are used to cause a computer to execute the construction engineering handover text generation method described in the first aspect or any corresponding embodiment thereof.
[0020] Fifthly, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to execute the construction engineering handover text generation method described in the first aspect or any corresponding embodiment. Attached Figure Description
[0021] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0022] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the first method for generating construction handover text for building engineering according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the second process of generating construction handover text for building engineering according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the third process of generating construction handover text for building engineering according to an embodiment of the present invention; Figure 5 This is a structural block diagram of a construction engineering handover text generation device according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation
[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0024] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.
[0025] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.
[0026] As an optional application scenario of this invention, considering the specific application environment architecture or specific hardware architecture upon which the method for generating construction handover texts for building engineering depends, the specific application environment architecture or specific hardware architecture is described here. For example... Figure 1 As shown, the architecture system may include at least one terminal device and at least one server. Figure 1 The system is illustrated in the example, which includes a computer 101, a mobile terminal 102, and a server 103, and the terminal devices such as the computer 101 and the mobile terminal 102 are connected to the server 103 through a network 110.
[0027] Specifically, the terminal device can be a smartphone, tablet, laptop, PDA, desktop computer, game console, smart TV, smart wearable device, in-vehicle terminal, VR (Virtual Reality) device, AR (Augmented Reality) device, etc. Server 103 can be a standalone physical server, a server cluster, a distributed system, or a cloud server providing cloud services. Network 110 can be a wired or wireless network, examples of which include, but are not limited to, the Internet, corporate intranet, local area network, wide area network, mobile communication network, and combinations thereof.
[0028] In related technologies, construction briefing texts mainly rely on manual writing or simple filling based on templates. On the one hand, construction conditions are complex and varied under different engineering scenarios, involving multi-dimensional information such as construction objects, working environment, resource allocation, process activities, and hazard factors. Manual writing is inefficient and prone to omitting key elements. On the other hand, the content of construction briefings must strictly comply with general standards, industry specifications, and corporate management systems. Human understanding and application of specification clauses are easily influenced by subjective experience, making it difficult to ensure continuous consistency.
[0029] In view of this, this application provides a method for generating construction briefing text for building engineering projects, which can be applied to a server to generate construction briefing text for building engineering projects. The method provided in this application obtains construction condition information containing spatiotemporal annotations and the text input information to be generated determined by the construction input text. It utilizes a scenario element extraction module to automatically extract multiple types of scenario elements associated with spatiotemporal annotations and form a construction scenario semantic set. This allows for rapid structured organization of complex construction condition information, replacing manual compilation and effectively improving the efficiency of briefing text compilation. Simultaneously, it avoids the omission of key elements during manual writing, ensuring the completeness and comprehensiveness of the briefing information. Based on this, a construction spatiotemporal framework containing node, relation edge, and attribute information is constructed based on the construction scenario semantic set and the standard predicate set. Knowledge graphs transform various standards and specifications into structured and interconnected graph knowledge. A hierarchical text framework is then generated using a large language model, and the handover text is generated based on the knowledge paths of nodes at each level of the graph. This precisely binds regulatory constraints to construction scenarios, completely eliminating subjective biases from human interpretation of regulations. It ensures that the generated handover text strictly conforms to standard requirements, achieving continuous consistency in handover content across different engineering scenarios. Simultaneously, it aligns the generation of handover text with the spatiotemporal logic of construction and the framework of industry regulations, comprehensively improving the professionalism, standardization, and generation efficiency of construction handover texts, and fully compensating for various deficiencies in related technologies.
[0030] According to an embodiment of the present invention, a method for generating construction briefing text for building engineering is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.
[0031] This embodiment provides a method for generating construction handover text for building engineering projects, which can be used in the aforementioned server. Figure 2 This is a flowchart of a construction project handover text generation method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain the text input information to be generated for the construction project. The text input information to be generated is determined based on the construction condition information containing spatiotemporal annotations and the construction input text.
[0032] For example, construction condition information includes, but is not limited to, construction plan text, construction organization design, special construction plan, construction log, process description text, BIM metadata, and on-site collected time and location data; construction input text includes draft construction briefing, construction parameter table, work instruction text, and safety technical instruction text. In this embodiment, both construction condition information and construction input text are input into the system in the form of electronic documents or structured data, and the text format includes, but is not limited to, Word, PDF, TXT, JSON, and BIM output format. In this embodiment, the construction condition information containing spatiotemporal annotations and the construction input text are preprocessed to obtain the text input information to be generated. This embodiment does not limit the specific content of the preprocessing, which can be determined by those skilled in the art according to their needs.
[0033] Step S202: Use the pre-built scenario element extraction module to extract multiple types of scenario element information related to spatiotemporal annotation from the text input information to be generated, and obtain the construction scenario semantic set.
[0034] For example, the scenario element extraction module integrates a pre-trained and domain-fine-tuned large language model. This large language model performs semantic parsing on the input text units to be generated, automatically extracts multiple types of construction scenario elements, and associates the extraction results with spatiotemporal annotations to obtain a construction scenario semantic set. The construction scenario semantic set includes at least scenario entities, time labels, spatial labels, event sequences, and confidence markers corresponding to text paragraphs, so as to align with standard semantics in subsequent steps.
[0035] In this application embodiment, the various scenario elements include at least: construction object entities, such as foundation pits, formwork, supporting structures, steel cages, scaffolding, etc.; work environment entities, such as underground spaces, nearby buildings, weather conditions, work areas, etc.; resource allocation entities, such as personnel, machinery and equipment, materials, testing tools, etc.; hazard factor entities, such as collapse, falls from heights, mechanical injuries, edge risks, etc.; and process activity entities, such as excavation, support, pouring, binding, acceptance, etc.
[0036] The scenario element extraction module will further identify the time and space information hidden in the text input information to be generated. The time information includes at least the construction stage sequence, the relationship between the procedures, the duration and the time window. The space information includes at least the spatial location of the components, the relative orientation and the division of the work area, and will be structured into event sequences and spatial nodes.
[0037] Specifically, the processing steps of the construction scenario element extraction module include: Input the text information to be generated into the zero-sample or few-sample prompt template to obtain the construction object mark, work environment mark, resource allocation mark, hazard factor mark and process activity mark output by the prompt template; The various types of markers output by the prompt template are associated with the corresponding timestamps and spatial location information to form a semantic set of construction scenarios.
[0038] Step S203: Based on the semantic set of construction scenarios and the pre-determined set of normative predicates, a spatiotemporal knowledge graph of construction engineering is constructed. The spatiotemporal knowledge graph of construction engineering includes nodes, relation edges and attribute information. Nodes are used to represent various objects with independent semantics in the construction scenario, and relation edges are used to represent the association between nodes.
[0039] For example, the set of predicates can be determined based on general standards, industry standards, local standards, and internal management documents of enterprises. Attribute information may include, but is not limited to, time attributes and spatial attributes. In this embodiment, the construction spatiotemporal knowledge graph includes at least entity nodes, relationship edges, time attributes, spatial attributes, and constraint metadata. Entity nodes include construction objects, procedures, specification clauses, risk factors, time nodes, and spatial nodes; relationship edges include logical relationships, constraint relationships, temporal sequences, and spatial dependencies; time attributes include timestamps, construction stage identifiers, and durations; spatial attributes include component locations, construction area coordinates, and spatial levels; constraint metadata includes formal predicates, source clause numbers, version information, and confidence levels. The constraint metadata is determined based on the set of predicates.
[0040] In this embodiment of the application, the process nodes and time relationship edges in the knowledge graph are used to represent the sequential order between construction steps; the spatial nodes and spatial relationship edges are used to represent the construction area, component location, and work space dependency; and the entity attribute nodes or attribute edges are used to represent parameter range, size constraints, and quantity constraints, thereby forming a structured knowledge base that can be called later.
[0041] Furthermore, the spatiotemporal knowledge graph for building construction is a dynamically updated graph, and the dynamic update includes the following steps: Step a1: Monitor changes during the construction phase of the project, updates to specifications, and changes to corporate policies.
[0042] For example, changes in the construction phase include the evolution of the project from foundation construction and main structure to decoration and finishing, capturing changes in the construction objects, procedures, resource allocation, and other scenarios; updates to the standard version include the revision, repeal, or addition of general standards and industry standards and specifications, i.e., information on changes in the validity of standard clauses; changes in corporate systems include adjustments to the company's internal construction management, safety requirements, and other systems to ensure that the blueprint aligns with the company's actual management requirements.
[0043] Step a2: Identify new entities, changed relationships, and invalid clauses based on the incremental knowledge extraction model, and write the new triples and new predicates into the spatiotemporal knowledge graph.
[0044] For example, the incremental knowledge extraction model is used to extract knowledge from the monitored changes, rather than a full reconstruction, thus reducing computational overhead. Newly added construction entities, changed relationships between entities, and invalid specifications and regulations are written into the graph in the form of triples and new predicates, completing the supplementation and updating of knowledge and ensuring that the graph content is aligned with the latest scenarios, specifications, and regulations.
[0045] Step a3 involves conflict detection and verification of the temporal order, spatial topological relationships, and logical consistency in the graph.
[0046] For example, by pre-constructing a verification rule base covering temporal sequence, spatial topology, and logical consistency, the system traverses the triplets, spatial entities, and entity attribute constraints with temporal relationships in the graph. It then uses a topological sorting algorithm to verify whether there are cyclical or reverse order conflicts between the process sequence and the construction stage. Combined with spatial reasoning rules, it verifies whether the spatial inclusion, adjacency, and separation relationships of entities such as components and work areas comply with physical and safety specifications. Finally, it uses a rule engine or knowledge reasoning methods to check for contradictions and conflicts between the same entity parameters, specification clauses, and construction requirements. The system then marks the detected conflicts and generates verification results.
[0047] Step a4: When the specification version is updated or the construction conditions change and affect the knowledge layer of the injected model, the minimum range of LoRA module is fine-tuned based on the impact analysis.
[0048] For example, the scope of knowledge involved in the update of the specification version or the change of construction conditions is first located through impact analysis to determine whether it affects the core knowledge layer already injected into the large language model. If an impact is confirmed, the smallest affected parameter area in the model is precisely located based on the granularity and scope of knowledge changes. Then, the LoRA module is triggered, using the updated graph knowledge as fine-tuning data, and incremental fine-tuning is performed only on the low-rank adaptation parameters of the corresponding area of the large language model. Knowledge adaptation is completed while keeping the main parameters of the model stable, avoiding the high cost and catastrophic forgetting problem of full fine-tuning, and finally achieving synchronous updates of model knowledge with the latest specifications and construction conditions. Among them, the knowledge injection technology is integrated into the parameters of the large language model, allowing the model to directly call these professional knowledge when generating disclosure text. This part of the professional knowledge, which is solidified into the model parameters and plays a decisive role in generating compliant and accurate text, is the "core knowledge layer".
[0049] Step S204: Input the text input information to be generated into the pre-built large language model so that the large language model outputs a hierarchical text framework.
[0050] For example, in this embodiment of the application, a construction briefing text framework generation task is performed based on a pre-built large language model. The structural components of the construction briefing text are identified based on instructional prompts, and a hierarchical framework is output, including chapter titles, task descriptions, technical requirements, safety risk warnings, material and equipment requirements, parameter conditions, and precautions. The generated content is organized into a hierarchical text skeleton. The hierarchical text framework includes at least: ① Project overview; ② Construction tasks and procedures; ③ Technical requirements; ④ Safety risk warnings; ⑤ Material and equipment requirements; ⑥ Parameter conditions; ⑦ Precautions.
[0051] In this embodiment, the hierarchical text framework can be further mapped to a finite state automaton structure, where each entry is a state, mandatory nodes are marked as mandatory, optional nodes are marked as optional, and directed edges between states are determined by temporal / sequential predicates or engineering practice rules in the knowledge graph. The finite state automaton and the formal predicates in the knowledge graph together constitute the hard constraint basis for subsequent constrained generation.
[0052] Step S205: Based on multiple hierarchical nodes in the hierarchical text framework, retrieve the knowledge paths of each hierarchical node in the construction spatiotemporal knowledge graph of the building project.
[0053] For example, in this embodiment of the application, based on each hierarchical node in the hierarchical text framework, the most relevant entities and relationship chains are retrieved from the spatiotemporal knowledge graph to form a knowledge path. The knowledge path may include, but is not limited to, construction procedure paths, specification constraint paths, temporal sequence paths, and spatially dependent paths.
[0054] Step S206: Generate construction engineering handover text from the hierarchical text framework and the knowledge paths of each level node.
[0055] For example, in this embodiment of the application, the overall chapter layout structure of the handover is defined based on the hierarchical text framework. First, the knowledge path corresponding to each level node is compiled to obtain a set of structured constraints and then split to generate hierarchical prompt information. Then, the hierarchical framework, the set of structured constraints and the hierarchical prompt information are input into the generation process. Based on the large model, the content corresponding to each node is filled in item by item according to the constraints. After integration, a complete construction handover text for building engineering is formed.
[0056] The construction briefing text generation method provided in this embodiment obtains construction condition information containing spatiotemporal annotations and the text input information to be generated determined by the construction input text. It then uses a scenario element extraction module to automatically extract multiple types of scenario elements associated with spatiotemporal annotations and form a construction scenario semantic set. This allows for rapid structured organization of complex construction condition information, replacing manual compilation and effectively improving the efficiency of briefing text preparation. Simultaneously, it avoids the omission of key elements during manual writing, ensuring the completeness and comprehensiveness of the briefing information. Based on this, a construction briefing text generation method containing node, relation edge, and attribute information is constructed based on the construction scenario semantic set and the standard predicate set. The construction spatiotemporal knowledge graph transforms various standards and specifications into structured and interconnected graph knowledge. It then generates a hierarchical text framework through a large language model and uses the knowledge paths of nodes at each level of the graph to generate the handover text. This precisely binds regulatory constraints to construction scenarios, completely eliminating subjective biases from human interpretation of regulations. It ensures that the generated handover text strictly conforms to standard requirements, achieving continuous consistency in handover content across different engineering scenarios. Simultaneously, it aligns the generation of handover text with the spatiotemporal logic of construction and the industry regulatory framework, comprehensively improving the professionalism, standardization, and generation efficiency of construction handover texts, and fully compensating for various deficiencies in related technologies.
[0057] This embodiment provides a method for generating construction handover text for building engineering projects, which can be used in the aforementioned server. Figure 3 This is a flowchart of a construction project handover text generation method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S301: Obtain the text input information to be generated for the construction project. This text input information is determined based on the construction condition information including spatiotemporal annotations and the construction input text. For details, please refer to [link to relevant documentation]. Figure 2 Step S201 of the illustrated embodiment will not be described again here.
[0058] Step S302 involves using a pre-built scenario element extraction module to extract multiple types of scenario element information related to spatiotemporal annotations from the input text to be generated, thus obtaining a construction scenario semantic set. For details, please refer to [link to relevant documentation]. Figure 2 Step S202 of the illustrated embodiment will not be described again here.
[0059] Step S303: Based on the semantic set of construction scenarios and the pre-determined set of normative predicates, a spatiotemporal knowledge graph of construction engineering is constructed. The spatiotemporal knowledge graph of construction engineering includes nodes, relation edges and attribute information. Nodes are used to represent various objects with independent semantics in the construction scenario, and relation edges are used to represent the association between nodes.
[0060] Specifically, step S303 includes: Step S3031: Obtain the engineering construction standard text.
[0061] For example, in the embodiments of this application, the normative texts for engineering construction may include, but are not limited to, general standards, industry standards, local standards, and internal management system documents of enterprises.
[0062] Step S3032: Semantic parsing is performed on the engineering construction normative text to obtain a set of normative predicates.
[0063] For example, in this embodiment of the application, a graph convolutional neural network and a semantic parser are used to jointly process the engineering construction normative text to obtain a set of normative predicates. The specific processing procedure is as follows: (1) Divide the normative clauses into the smallest semantic units; (2) Map the constraints, applicable objects, technical requirements, reference relationships and inheritance relationships in the clauses into entity nodes and relationship edges; (3) Then learn the topological structure features between the clause nodes through graph convolutional neural networks to obtain the vectorized representation of the normative clauses.
[0064] Furthermore, natural language constraints such as "sequentiality," "time window," "spatial dependency," and "numerical constraints" in the specification are translated into formal predicates or constraint expressions by a semantic parser.
[0065] Specifically, the canonical predicate set can be determined through the following steps: Step b1 involves segmenting general standards, industry specifications, and corporate management documents into clauses and standardizing terminology; Step b2 involves performing predicate parsing on phrases such as "should be before," "must not be earlier than," "located in the region of," "not less than," and "not greater than," generating temporal predicates, spatial dependency predicates, and numerical constraint predicates, respectively. Step b3 outputs the predicates along with their corresponding clause numbers, source versions, and confidence levels as a set of canonical predicates.
[0066] Step S3033: Semantic alignment is performed based on the construction scenario semantic set and the standard predicate set to obtain the alignment result.
[0067] For example, in this embodiment of the application, a Siamese network structure is used to encode the semantic representation of the construction scenario and the semantic representation of the specification, respectively, and semantic alignment is achieved through contrastive learning. The vectors output by the Siamese network are used to calculate the semantic similarity between the construction scenario entity and the specification clause entity, thereby establishing the association between the two.
[0068] Step S3034: Based on the alignment results, construct a spatiotemporal knowledge graph of building construction.
[0069] For example, in this embodiment of the application, the predicates in the canonical predicate set are attached to the nodes or edges of the knowledge graph as executable metadata, and each predicate carries the source clause number and confidence level.
[0070] Specifically, the attachment positions of predicates include: attaching temporal predicates to the temporal relationship edges between process nodes; attaching spatial dependency predicates to the spatial relationship edges between work area nodes, component nodes, or equipment nodes; attaching numerical constraint predicates to entity attribute edges or entity attribute nodes; and attaching clause source version information to the attribute fields of the corresponding predicate. For example, the specification clause "Support should be completed before excavation" can be parsed into the temporal predicate `requires_before(support, excavation)` and attached to the temporal relationship edge between the "support" process node and the "excavation" process node; the specification clause "Support depth shall not be less than design depth" can be parsed into the numerical constraint predicate `min_value(support depth, design_depth)` and attached to the attribute edge of the support entity.
[0071] Furthermore, the methods for attaching predicate sets to nodes or edges in a spatiotemporal knowledge graph include: Step c1: Attach the time sequence predicate to the time relationship edge between process nodes; Step c2: Attach the spatial dependency predicate to the spatial relationship edge between the work area node, component node, or equipment node; Step c3: Attach the numerical constraint predicate to the entity attribute edge or entity attribute node; Step c4: Attach the source version information of the terms to the attribute field of the corresponding predicate.
[0072] Step S304: Input the text input information to be generated into the pre-built large language model so that the large language model outputs a hierarchical text framework.
[0073] Specifically, step S304 includes: Step S3041: Input the text information to be generated into the large language model so that the large language model can output the construction briefing elements of the construction engineering construction briefing text. The construction briefing elements include any one or more of the following: project overview, construction tasks and process descriptions, technical requirements, safety risk warnings, material and equipment requirements, parameter conditions, and precautions.
[0074] For example, in this embodiment of the application, the standardized input text information to be generated is sent to a pre-built large language model, and the content is extracted from the input information based on the construction briefing content recognition logic learned by the model, and one or more types of construction briefing elements such as project overview, construction tasks and process descriptions are selected and output.
[0075] Step S3042: Call the construction engineering handover standard hierarchical rule library to organize the construction handover elements and obtain a hierarchical text framework.
[0076] For example, in this embodiment of the application, the standard hierarchical rule library of the pre-stored construction chapter arrangement logic is retrieved, and the extracted construction disclosure elements are classified and arranged according to the preset hierarchical order of the upper and lower chapters, dividing them into first-level and second-level hierarchical structures, thereby assembling a regular hierarchical text framework.
[0077] Furthermore, the generated text framework is mapped to a finite state automaton structure, where each entry represents a state, mandatory nodes are marked as mandatory, and optional nodes are marked as optional. Directed edges between states are determined by temporal / sequential predicates or engineering practice rules in the knowledge graph. The finite state automaton and the formal predicates in the knowledge graph together form the hard constraint basis for subsequent constrained generation.
[0078] Step S305: Based on multiple hierarchical nodes in the hierarchical text framework, retrieve the knowledge paths of each hierarchical node in the construction engineering spatiotemporal knowledge graph. For details, please refer to... Figure 2 Step S205 of the illustrated embodiment will not be described again here.
[0079] Step S306 involves generating construction handover text from the hierarchical text framework and the knowledge paths of each level node. For details, please refer to [link to relevant documentation]. Figure 2 Step S206 of the illustrated embodiment will not be described again here.
[0080] This embodiment provides a method for generating construction handover text for building engineering projects, which can be used in the aforementioned server. Figure 4 This is a flowchart of a construction project handover text generation method according to an embodiment of the present invention, such as... Figure 4 As shown, the process includes the following steps: Step S401: Obtain the text input information to be generated for the construction project. The text input information to be generated is determined based on the construction condition information containing spatiotemporal annotations and the construction input text.
[0081] Specifically, step S401 includes: Step S4011: Obtain construction condition information and construction input text containing spatiotemporal annotations.
[0082] Step S4012: Standardize the construction condition information and construction input text to obtain the text input information to be generated.
[0083] For example, in this embodiment of the application, the original text is divided into segments to be reviewed, and the construction objects, process activities, time expressions, and spatial expressions are standardized. For example, the time relationship in "excavation immediately after support is completed" is extracted as the time sequence label "support precedes excavation", "east side of the foundation pit" is extracted as the spatial area label, and "8.5m" is uniformly converted and labeled with units.
[0084] Step S402 involves using a pre-built scenario element extraction module to extract multiple types of scenario element information related to spatiotemporal annotations from the input text to be generated, thus obtaining a construction scenario semantic set. For details, please refer to [link to relevant documentation]. Figure 3 Step S302 of the illustrated embodiment will not be described again here.
[0085] Step S403: Based on the semantic set of construction scenarios and the pre-determined set of normative predicates, a spatiotemporal knowledge graph for construction engineering is constructed. This knowledge graph includes nodes, relational edges, and attribute information. Nodes represent various objects with independent semantics within the construction scenario, and relational edges represent the relationships between nodes. For details, please refer to [link to details]. Figure 3 Step S303 of the illustrated embodiment will not be described again here.
[0086] Step S404: Input the text to be generated into a pre-built large language model so that the large language model outputs a hierarchical text framework. For details, please refer to [link to details]. Figure 3 Step S304 of the illustrated embodiment will not be described again here.
[0087] Step S405: Based on multiple hierarchical nodes in the hierarchical text framework, retrieve the knowledge paths of each hierarchical node in the construction engineering spatiotemporal knowledge graph. For details, please refer to... Figure 3 Step S305 of the illustrated embodiment will not be described again here.
[0088] Step S406: Generate construction engineering handover text from the hierarchical text framework and the knowledge paths of each level node.
[0089] Specifically, step S406 includes: Step S4061: Compile the knowledge paths of each level node to obtain a set of structured constraints.
[0090] For example, in this embodiment of the application, the knowledge path is input into the constraint compilation module, which translates the knowledge path into two types of executable representations: one type is structured constraints, which are used to ensure the order of chapters and processes and the appearance of required paragraphs; the other type is a set of numerical values or predicates, which are used for parameter ranges, unit rules and spatial / temporal predicate determination.
[0091] The constraint compilation module may, but is not limited to, output machine-readable constraint tables. These constraint tables may be stored in JSON format and used by subsequent generation modules and posterior validators. The structured constraint set includes at least finite automaton constraints, sequential constraints, time window constraints, spatial constraints, and numerical constraints.
[0092] Step S4062: Divide the structured constraint set into hierarchical levels and generate hierarchical prompt information.
[0093] For example, the structured constraint set can be divided into a hard constraint layer and a soft guidance layer, generating corresponding hierarchical prompts. The hard constraint layer defines the necessary sequential, temporal, and spatial constraints, the soft guidance layer provides preferred expression methods, and the evidence fragment layer provides a summary of the normative clauses and their source numbers. The hierarchical prompts include at least: (1) Hard constraint layer hints, used to constrain chapter order, temporal relationships, spatial relationships and numerical range; (2) Soft guidance layer prompts, used to guide the standardized expression of terminology and the expression of engineering templates; (3) Evidence fragment layer prompts, which are used to provide a summary of the normative clauses, clause numbers and knowledge path numbers corresponding to the current chapter.
[0094] Specifically, the processing steps of the constraint compilation module include: Step d1: Retrieve the entity path, relation path, and normative clause path corresponding to the hierarchical text framework in the spatiotemporal knowledge graph to obtain the knowledge path; Step d2: Compile the knowledge path into a set of structured constraints. The set of structured constraints includes at least finite automaton constraints, sequence constraints, time window constraints, spatial constraints, and numerical constraints. Step d3 involves dividing the structured constraint set into a hard constraint layer and a soft guidance layer, and generating corresponding evidence fragment layer prompts.
[0095] Step S4063: Generate construction handover text for building engineering based on hierarchical text framework, structured constraint set and hierarchical prompt information.
[0096] For example, in this embodiment of the application, construction briefing text for building engineering is generated based on a hierarchical text framework, a set of structured constraints, and hierarchical prompts. During the generation process, when the candidate construction briefing text violates any sequence constraint, time window constraint, spatial constraint, or numerical constraint in the hard constraint layer, secondary generation, knowledge graph recommended value replacement, or manual review is triggered. For example, when the knowledge path has determined that "support completion" should precede "excavation," if the candidate output contains the content "excavation first, then support," the constraint decoding strategy will block the candidate branch and re-call the constraint compilation results for completion or correction.
[0097] In some optional implementations, the above method further includes: Step S407: Use the construction spatiotemporal knowledge graph of the building project to perform consistency verification on the construction briefing text of the building project, and obtain the verification result.
[0098] Step S408: Optimize the construction briefing text based on the verification results to obtain the target construction briefing text.
[0099] For example, in this embodiment of the application, the generated text is logically consistent based on a spatiotemporal knowledge graph, including verification of the sequence of procedures, time window verification, and spatial topology verification; a finely tuned natural language reasoning model is used to determine the implications, contradictions, or neutral relationships between the candidate results of the construction briefing text and the corresponding standard clauses; a rule engine is used to correct terminology errors, abnormal process order, time conflicts, and unreasonable spatial content; and the missing key technical information and safety elements in the text are supplemented through graph reasoning to obtain the target construction briefing text.
[0100] Furthermore, the system records the original inputs, the knowledge path numbers used, the triggered constraint predicates, the prompt text used by the generator, the automatic correction operations, the model version used, the adapter version, and the manual review comments, forming a traceable audit evidence chain. This evidence chain is output along with the final generated text to support compliance verification and traceability.
[0101] The method provided in this application integrates construction condition information with industry standard knowledge into a unified model, enabling standard-driven and semantically constrained processes in the generation of construction briefing texts. Through a knowledge path triggering mechanism, the large language model is guided to automatically follow process logic, temporal sequence, and spatial constraints when generating text, improving the professionalism and consistency of the generated text. Semantic enhancement and logic optimization reduce terminology errors, process anomalies, and standard conflicts, improving the accuracy and usability of construction briefing texts. The method utilizes spatiotemporal knowledge graphs and large language models to intelligently generate construction briefing texts, effectively reducing the workload of manually writing construction briefing texts and improving the informatization and intelligence level of construction management.
[0102] This embodiment also provides a construction engineering handover text generation device, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.
[0103] This embodiment provides a construction engineering handover text generation device, such as... Figure 5 As shown, it includes: The acquisition module 501 is used to acquire the text input information to be generated for the construction project. The text input information to be generated is determined based on the construction condition information containing spatiotemporal annotations and the construction input text. The extraction module 502 is used to extract multiple types of scenario element information related to spatiotemporal annotation from the text input information to be generated using a pre-built scenario element extraction module, so as to obtain a construction scenario semantic set. Module 503 is used to construct a construction spatiotemporal knowledge graph of building engineering based on the semantic set of construction scenarios and the pre-determined set of normative predicates. The construction spatiotemporal knowledge graph of building engineering includes nodes, relation edges and attribute information. Nodes are used to represent various objects with independent semantics in the construction scenario, and relation edges are used to represent the association relationship between nodes. The determination module 504 is used to input the text input information to be generated into a pre-built large language model so that the large language model outputs a hierarchical text framework. The retrieval module 505 is used to retrieve the knowledge paths of each level node in the construction spatiotemporal knowledge graph of building engineering based on multiple level nodes in the hierarchical text framework. The generation module 506 is used to generate construction engineering handover text from the hierarchical text framework and the knowledge paths of each level node.
[0104] In some alternative implementations, building module 503 includes: The first acquisition submodule is used to acquire engineering construction standard texts; The parsing submodule is used to perform semantic parsing on engineering construction normative texts to obtain a set of normative predicates; The alignment submodule performs semantic alignment between the English text based on the semantic set of construction scenarios and the set of predicates in the specifications, and obtains the alignment result. A submodule is constructed to build a spatiotemporal knowledge graph of building engineering construction based on the alignment results.
[0105] In some alternative implementations, the generation module 506 includes: The compilation submodule is used to compile the knowledge paths of nodes at each level to obtain a set of structured constraints; The sub-module is used to hierarchically divide the set of structured constraints and generate hierarchical prompts. The generation submodule is used to generate construction handover text for building projects based on a hierarchical text framework, a set of structured constraints, and hierarchical prompts.
[0106] In some alternative embodiments, the above-described apparatus further includes: The verification module is used to perform consistency verification on the construction handover text of the construction project using the spatiotemporal knowledge graph of the construction project, and obtain the verification result. The optimization module is used to optimize the construction handover text of building engineering based on the verification results, so as to obtain the target construction handover text.
[0107] In some optional implementations, the acquisition module 501 includes: The second acquisition submodule is used to acquire construction condition information and construction input text containing spatiotemporal annotations; The processing submodule is used to standardize the construction condition information and construction input text to obtain the text input information to be generated.
[0108] In some alternative implementations, the determining module 504 includes: The determination submodule is used to input the input text information to be generated into the large language model so that the large language model can output the construction disclosure elements of the construction engineering construction disclosure text. The construction disclosure elements include any one or more of the following: project overview, construction tasks and process descriptions, technical requirements, safety risk warnings, material and equipment requirements, parameter conditions, and precautions. The organization submodule is used to call the hierarchical rule library of construction engineering handover standards to organize the elements of construction handover and obtain a hierarchical text framework.
[0109] The construction engineering handover text generation device provided in this embodiment of the invention can execute the construction engineering handover text generation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.
[0110] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.
[0111] The following is a detailed reference. Figure 6This diagram illustrates a suitable structural design for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 601, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 602 or a program loaded from memory 608 into random access memory (RAM) 603. RAM 603 also stores various programs and data required for the operation of the electronic device. The processor 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.
[0112] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 608 including, for example, magnetic tapes, hard disks, etc.; and communication devices 609. Communication device 609 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.
[0113] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory 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 609, or installed from a memory 608, or installed from a ROM 602. When the computer program is executed by the processor 601, it performs the functions defined in the construction engineering handover text generation method of the embodiments of the present invention.
[0114] Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.
[0115] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the construction engineering handover text generation method shown in the above embodiments is implemented.
[0116] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.
[0117] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.
Claims
1. A method for generating construction briefing text for building engineering projects, characterized in that, The method includes: Obtain the text input information to be generated for the construction project, which is determined based on the construction condition information including spatiotemporal annotations and the construction input text; The pre-built scenario element extraction module is used to extract multiple types of scenario element information related to spatiotemporal annotation from the text input information to be generated, so as to obtain the construction scenario semantic set; Based on the construction scenario semantic set and the pre-determined set of normative predicates, a construction spatiotemporal knowledge graph for building engineering is constructed. The construction spatiotemporal knowledge graph for building engineering includes nodes, relation edges, and attribute information. Nodes are used to represent various objects with independent semantics in the construction scenario, and relation edges are used to represent the association relationships between nodes. The text input information to be generated is input into a pre-built large language model so that the large language model outputs a hierarchical text framework. Based on multiple hierarchical nodes in a hierarchical text framework, knowledge paths of each hierarchical node are retrieved in the spatiotemporal knowledge graph of building construction. The hierarchical text framework and the knowledge paths of each level node are used to generate construction engineering handover text.
2. The method according to claim 1, characterized in that, The steps for constructing a spatiotemporal knowledge graph of building construction based on the aforementioned construction scenario semantic set and a predetermined set of normative predicates include: Obtain the standard documents for engineering construction; Semantic parsing is performed on the aforementioned engineering construction normative text to obtain a set of normative predicates; Semantic alignment is performed between the construction scenario semantic set and the standard predicate set to obtain the alignment result; Based on the alignment results, a spatiotemporal knowledge graph of building construction is constructed.
3. The method according to claim 1, characterized in that, The steps of generating construction engineering handover text from the hierarchical text framework and the knowledge paths of each level node include: The knowledge paths of each level of nodes are compiled to obtain a set of structured constraints; The structured constraint set is hierarchically divided to generate hierarchical prompt information; The construction handover text for building engineering is generated based on the hierarchical text framework, the set of structured constraints, and the hierarchical prompt information.
4. The method according to claim 1, characterized in that, After generating the construction project handover text from the hierarchical text framework and the knowledge paths of each level node, the method further includes: The consistency of the construction project handover text was verified using a construction spatiotemporal knowledge graph, and the verification results were obtained. Based on the verification results, the construction briefing text for the building project is optimized to obtain the target construction briefing text.
5. The method according to any one of claims 1 to 4, characterized in that, The step of obtaining the text input information to be generated for the construction project includes: Obtain construction condition information and construction input text containing spatiotemporal annotations; The construction condition information and construction input text are standardized to obtain the text input information to be generated.
6. The method according to any one of claims 1 to 4, characterized in that, The step of inputting the text input information to be generated into a pre-built large language model, so that the large language model outputs a hierarchical text framework, includes: The input text information to be generated is input into the large language model so that the large language model outputs the construction briefing elements of the construction engineering construction briefing text. The construction briefing elements include any one or more of the following: project overview, construction tasks and procedures, technical requirements, safety risk warnings, material and equipment requirements, parameter conditions, and precautions. The construction briefing elements are organized by calling the hierarchical rule library of the construction engineering briefing standard to obtain a hierarchical text framework.
7. A construction project handover text generation device, characterized in that, The device includes: The acquisition module is used to acquire the text input information to be generated for the construction project. The text input information to be generated is determined based on the construction condition information containing spatiotemporal annotations and the construction input text. The extraction module is used to extract multiple types of scenario element information related to spatiotemporal annotation from the text input information to be generated using a pre-built scenario element extraction module, so as to obtain a construction scenario semantic set. The construction module is used to construct a spatiotemporal knowledge graph of building construction based on the semantic set of the construction scenario and the pre-determined set of normative predicates. The spatiotemporal knowledge graph of building construction includes nodes, relation edges and attribute information. Nodes are used to represent various objects with independent semantics in the construction scenario, and relation edges are used to represent the association relationship between nodes. The determination module is used to input the text input information to be generated into a pre-built large language model so that the large language model outputs a hierarchical text framework. The retrieval module is used to retrieve the knowledge paths of each level node in the spatiotemporal knowledge graph of building construction based on multiple level nodes in a hierarchical text framework. The generation module is used to generate construction engineering handover text from the hierarchical text framework and the knowledge paths of each level node.
8. An electronic device, characterized in that, include: The method includes a memory and a processor, which are interconnected and the memory stores computer instructions. The processor executes the computer instructions to perform the construction engineering handover text generation method as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the construction engineering handover text generation method according to any one of claims 1 to 6.
10. A computer program product, characterized in that, The method includes computer instructions for causing a computer to execute the construction engineering handover text generation method according to any one of claims 1 to 6.
