Engineering project report generation method, device and equipment based on AI large model

Abstract

The invention provides an engineering project report generation method, device and equipment based on an AI large model, and the method comprises the steps: receiving a prompt word inputted by a user, and generating a work report template according to the prompt word; extracting related contents of the work report template from a knowledge base by using a large language model, and filling each module in the work report template with the related contents to obtain each report module; checking and adjusting each report module by utilizing the large language model to obtain each adjusted report module; and integrating the adjusted report modules by using the large language model to generate a target work report. According to the method, the document files of the project execution condition are integrated and the work report required by the user is generated by utilizing the large language model and the knowledge base, so that the efficiency of generating the work report is improved, and meanwhile, the accuracy of the work report is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device and equipment for generating an engineering project report based on an AI big model. Background Art

[0002] Large Language Models (LLMs) are deep learning-based AI models that are trained on large amounts of text data and can understand and generate natural language, making them ideal tools for automated report generation.

[0003] Currently, the execution status report of engineering design projects is usually generated manually, which involves collecting data, collating information, analyzing data, organizing content and writing reports. However, the traditional manual report generation method still has some potential defects: it takes a lot of time and manpower costs, resulting in low efficiency in generating work reports; there is a high risk of errors, such as typing errors, analysis errors or data entry errors. Therefore, an effective method for generating work reports for engineering projects is urgently needed to solve the above problems. Summary of the invention

[0004] In view of the above problems, the present invention is proposed to provide a method, device and equipment for generating engineering project reports based on an AI big model to overcome the above problems or at least partially solve the above problems.

[0005] To achieve the above-mentioned purpose and other related purposes, the present invention provides a method for generating an engineering project report based on an AI large model, which is applied to an engineering project management system. The engineering project management system includes a large language model and a knowledge base. The method includes:

[0006] Receiving a prompt word input by a user, and generating a work report template according to the prompt word;

[0007] Extracting relevant content of the work report template from a knowledge base using a large language model, and filling the relevant content into each module in the work report template to obtain each report module;

[0008] Using the large language model to check and adjust each report module to obtain each adjusted report module;

[0009] The large language model is used to integrate the adjusted report modules to generate a target work report.

[0010] Optionally, the extracting relevant content of the work report template from the knowledge base using the large language model includes:

[0011] Extract each module of the work report template, encode each module, and generate a query text vector corresponding to each module;

[0012] Each query text vector is matched with the vectors in the knowledge base for similarity, a number of target vectors corresponding to each query text vector are determined, and a number of target text blocks corresponding to the target vectors are extracted from the knowledge base.

[0013] Optionally, before the step of receiving the prompt word input by the user and generating a work report template according to the prompt word, the step further includes:

[0014] Obtaining a document file during the execution of the engineering design project, and preprocessing the document file to obtain text in the document file; wherein the text includes basic project information, project professional information, project task information, project working time information, project progress information, and project completion information;

[0015] Using a text separator and a preset text segmentation length, the text is segmented to obtain a plurality of text blocks;

[0016] Calling a pre-built vector encoding model to encode each text block, obtaining a vector corresponding to each text block, and storing the vector corresponding to each text block in a vector database;

[0017] The metadata of each text block is extracted, and a knowledge base is constructed using each text block, the vector of the text block and the metadata of the text block.

[0018] Optionally, the preprocessing the document file to obtain text in the document file includes:

[0019] If the document file is a text format file, extract the text in the document file and remove repeated and meaningless characters in the text;

[0020] If the document file is a picture format file or contains a picture, the OCR recognition technology is used to recognize and obtain the text in the document file, and duplicate and meaningless characters in the text are removed.

[0021] Optionally, the text is segmented using a text separator and a preset text segmentation length to obtain a plurality of text blocks, including:

[0022] Performing initial text segmentation on the text using a text delimiter to obtain a plurality of text segments, and detecting whether the text length of each text segment is greater than a preset text segmentation length;

[0023] When it is detected that the text length of the text segment is greater than the preset text segmentation length, the text segment is segmented using the preset text segmentation length to obtain a plurality of text blocks.

[0024] Optionally, the project report generation method further includes:

[0025] Receiving a question text input by a user, encoding the question text to obtain a question vector;

[0026] Performing similarity matching between the question vector and the vectors in the knowledge base, determining a plurality of target vectors corresponding to the question vector, and extracting a target text block corresponding to each target vector from the knowledge base;

[0027] The question text and each target text block are input into a large language model together, and the answer to the question text is generated by using the large language model.

[0028] Optionally, the generating an answer to the question text using the large language model includes:

[0029] The large language model is used to analyze and understand the question text and each target text block, generate an answer to the question text, and visualize the answer.

[0030] In a second aspect, the present invention further provides a device for generating a project report based on an AI large model, which is applied to a project management system, wherein the project management system includes a large language model and a knowledge base, and the device includes:

[0031] A receiving module, used for receiving a prompt word input by a user, and generating a work report template according to the prompt word;

[0032] A filling module, used to extract relevant content of the work report template from a knowledge base using a large language model, and fill the relevant content into each module in the work report template to obtain each report module;

[0033] An adjustment module, used to check and adjust each report module using the large language model to obtain each adjusted report module;

[0034] A generation module is used to integrate each of the adjusted report modules using the large language model to generate a target work report.

[0035] In a third aspect, the present invention provides an electronic device, comprising: a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program stored in the memory, so that the electronic device performs the steps of the engineering project report generation method based on the AI ​​big model as described above.

[0036] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by an electronic device, implements the steps of the method for generating an engineering project report based on an AI big model as described above.

[0037] The above one or more technical solutions provided by the present invention may have the following advantages or at least achieve the following technical effects: the present invention integrates the document files of the project execution status and generates the reports required by the user by utilizing a large language model and a knowledge base, thereby improving the efficiency of generating work reports and improving the accuracy of the work reports. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Shown is a flow chart of a method for generating an engineering project report based on an AI big model in one embodiment of the present invention;

[0039] Figure 2 Shown is a schematic diagram of a process for generating a report on the execution status of an engineering design project in one embodiment of the present invention;

[0040] Figure 3 Shown is a flow chart of a method for generating a report for an engineering project in another embodiment of the present invention;

[0041] Figure 4 Shown is a flow chart of a question-and-answer dialogue in an engineering design project execution information in one embodiment of the present invention;

[0042] Figure 5 It is a schematic diagram of functional modules of a device for generating a project report based on an AI big model in one embodiment of the present invention;

[0043] Figure 6 Shown is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0045] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0046] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0047] The terms "first", "second", etc. in the specification and claims of the disclosed embodiments and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so that the embodiments of the disclosed embodiments described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions.

[0048] Unless otherwise stated, the term "plurality" means two or more.

[0049] In the embodiment of the present disclosure, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B indicates: A or B.

[0050] The term "and / or" is a description of the association relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or, A and B.

[0051] The technical solutions in the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0052] See also Figure 1 An embodiment of the present invention provides a method for generating a project report based on an AI big model, which is applied to a project management system. The project management system includes a big language model and a knowledge base. The method may include the following steps S10 to S40:

[0053] Step S10, receiving a prompt word input by a user, and generating a work report template according to the prompt word.

[0054] The prompt word may be a piece of text or instruction provided to the large language model to guide the large language model to generate a specific type of response or output.

[0055] A work report template can be a predefined work report document framework; it can usually include the following key parts: project overview, project progress, resource utilization, achievement display, risks and problems, and future planning. In addition, in actual applications, the work report template can also be adjusted and expanded according to specific needs.

[0056] The project overview includes but is not limited to: project name, project objectives, project background, project cycle, and project team members.

[0057] Project progress includes but is not limited to: completed tasks, ongoing tasks, challenges encountered and solutions, and key milestones.

[0058] Resource utilization includes but is not limited to: fund usage, manpower allocation, material usage, and third-party support.

[0059] The display of achievements includes, but is not limited to: an overview of the main achievements, quantitative indicators of the achievements, the impact of the achievements on the achievement of goals, and the way of displaying the achievements.

[0060] Risks and issues include, but are not limited to: identified risks, countermeasures, existing problems, solutions or suggestions.

[0061] Future plans include but are not limited to: next action plan, expected goals, demand support, and continuous improvement measures.

[0062] In a specific implementation, when a prompt word input by a user is received, a corresponding work report template can be generated based on the prompt word using a large language model, thereby ensuring the standardization and consistency of the work report structure.

[0063] As an example, please generate a detailed project execution report based on the information provided, ensuring that each section has an appropriate subheading and the content is detailed and logically clear; the project execution work report template is as follows:

[0064] 1. Project name: XXXX innovative technology research and development project

[0065] 2. Project goal: Develop a new technology solution for a certain industry

[0066] 3. Project background: With the development of a certain industry, existing technology cannot meet the growing demand

[0067] 4. Project period: January 2023 to December 2024

[0068] 5. Project team members: Project manager Zhang San, technical person in charge Li Si, and financial person in charge Wang Wu.

[0069] Step S20, extracting relevant content of the work report template from the knowledge base using a large language model, and filling the relevant content into each module in the work report template to obtain each report module.

[0070] Among them, the relevant content can be the text blocks of each module (ie, relevant chapters) in the work report template.

[0071] Report modules can be modules that are working report templates that have been filled with content.

[0072] In the specific implementation, the large language model can be used to extract the various chapters in the work report template, and then the vector encoding model can be used to encode each chapter to obtain the query text vector corresponding to each chapter; then the query text vector is matched with the vector in the knowledge base for similarity to determine the corresponding target vectors; then the target text block corresponding to the target vector can be extracted from the knowledge base; finally, the large language model is used to organize the target text blocks and fill them into the various modules of the work report template to obtain various report modules.

[0073] As an example, see Figure 2 , Figure 2 The flowchart of generating a report on the execution status of an engineering design project is shown; assuming that the work report template includes a project progress module, the vector encoding model encodes the project progress module into a query text vector; then the query text vector is matched with the vector in the knowledge base for similarity, and the closest vectors (e.g., usually 3) are selected as target vectors in the knowledge base based on the similarity, and corresponding target text blocks are extracted from the knowledge base based on the target vectors; the current project progress information is extracted from the target text blocks (e.g., block 1, block 2, block 3, and block 4) using a large language model, and after organizing it into coherent sentences, it is filled into the project progress module to obtain a project progress module with content filled in.

[0074] Step S30: using the large language model to check and adjust each report module to obtain each adjusted report module.

[0075] The adjusted report module may be a report module obtained by checking the text content in the report module and adjusting the order appropriately.

[0076] In the specific implementation, a large language model is used to check whether the text content of each report module is reasonable, and the text content is polished; at the same time, the order of the text content can be appropriately adjusted to make it clearer and more reasonable, and each adjusted report module is obtained.

[0077] Step S40: integrating each of the adjusted report modules using the large language model to generate a target work report.

[0078] In the specific implementation, the large language model is used to integrate the various report modules obtained in the previous step into a complete work report in a certain order, further improving the readability of the work report; and the large language model is used to generate a summary of the work report, adjust the format of the work report, add headers, report time and other contents, and generate a target work report. The target work report can also be saved to a specified path (for example, for Windows systems, the target work report can be saved to the desktop of the C drive). Since the large language model can adjust the language style and wording to ensure the professionalism and accuracy of the work report, it can reduce errors caused by improper language expression.

[0079] It should be noted that user access control permissions are pre-set in the engineering project management system; user access control permissions define the scope of permissions for each user or role in the engineering project management system; permissions can be specifically broken down into operations such as read, write, update, and delete.

[0080] For each document file, different permissions are assigned to different users, and global permission management is achieved by setting administrator roles. The corresponding permissions are obtained when the user logs in to the engineering project management system.

[0081] Therefore, through permission management, the responsibility boundaries of each user can be clearly defined to avoid information abuse and misoperation.

[0082] As an example, company-level leaders can view all project information, while project managers can only view information about their own projects and the disciplines involved.

[0083] It should be noted that different roles will result in significant differences in the format and content of the subsequently generated reports (or statements).

[0084] As an example, the project manager's report contains more detailed information, because the project manager needs to have a comprehensive understanding of the entire content of the project to effectively control the progress of the project and achieve the project goals. The generated report can be specific to each professional and task information of its project, such as project goals and scope, project plan, team management, risk management, quality management, progress monitoring, change project, project closure, etc.

[0085] For the company's leaders, they need to focus more on understanding the project level and grasping the project information; the generated reports focus more on the overall status of the project, such as project goals and progress, schedule, risk management, resource allocation, problems and solutions, project evaluation and feedback, etc., to make strategic decisions.

[0086] Furthermore, in one embodiment, step S20 may include the following sub-steps S201 to S202:

[0087] Sub-step S201, extracting each module of the work report template, and encoding each module to generate a query text vector corresponding to each module;

[0088] Sub-step S202, performing similarity matching between each query text vector and the vectors in the knowledge base, determining a number of target vectors corresponding to each query text vector, and extracting a number of target text blocks corresponding to the target vectors from the knowledge base.

[0089] The query text vector may be obtained by converting each module in the work report template into a vector representation.

[0090] The target vector may be a number of vectors (eg, usually 3 vectors) extracted from the knowledge base that are closest in similarity to the query text vector.

[0091] The target text block may be a text block corresponding to the target vector extracted from the knowledge base.

[0092] In a specific implementation, a large language model can be used to extract each module in a work report template, and then each module can be encoded using a vector encoding model to generate a query text vector corresponding to each module; then, for a single query text vector, the query text vector can be similarly calculated with a vector in a knowledge base to obtain multiple similarity scores, and the multiple similarity scores are sorted from high to low, and the most similar vectors (e.g., usually 3) are selected as target vectors; then, a number of target text blocks corresponding to the target vector are extracted from the knowledge base (corresponding to 3 text blocks).

[0093] In this embodiment, a prompt word input by a user is received, and a work report template is generated according to the prompt word; the relevant content of the work report template is extracted from the knowledge base using a large language model, and the relevant content is filled into each module in the work report template to obtain each report module; each report module is checked and adjusted to obtain each adjusted report module; each adjusted report module is integrated using the large language model to generate a target work report. The present invention integrates the document files of the project execution status and generates the work report required by the user by using a large language model and a knowledge base, thereby improving the generation efficiency of the work report and the accuracy of the work report.

[0094] Based on the above embodiments, a second embodiment of a method for generating a project report based on an AI big model of the present invention is proposed. In this embodiment, before step S10, the method for generating a project report based on an AI big model may further include the following steps A10 to A40:

[0095] Step A10, obtaining the document file during the execution of the engineering design project, and preprocessing the document file to obtain the text in the document file; wherein the text includes basic project information, project professional information, project task information, project working hours information, project progress information and project completion information.

[0096] The text may include, but is not limited to: basic project information, project professional information, project task information, project working hours information, project progress information, and project completion information.

[0097] Basic project information may include but is not limited to: project number, project name, project manager name, etc.

[0098] Project professional information refers to what disciplines are included in this project, such as piping, design, HVAC, technology, etc., as well as the specific progress of these disciplines.

[0099] Project task information refers to the information about specific tasks included in the project, such as task name, task progress, task weight, whether to postpone, etc.

[0100] Project working hours information refers to the actual working hours used and planned working hours of the project.

[0101] Project progress information refers to the current progress of the project, expressed as a percentage.

[0102] Project completion information refers to whether the project is completed.

[0103] Preprocessing may include data cleaning and OCR (Optical Character Recognition) technology.

[0104] Data cleaning includes but is not limited to: deleting meaningless special characters in the text, removing unnecessary spaces and line breaks, deduplicating text, removing stop words, etc.

[0105] OCR recognition technology can automatically identify and extract text information from images or documents through image processing and machine learning technology.

[0106] In a specific implementation, the interface of the engineering project management system can be called periodically or when there is a change in the document data of the engineering design project execution to obtain the document file during the execution of the engineering design project; the document file can then be preprocessed to identify and extract text from the document file, and remove repeated and meaningless characters in the text; at the same time, the document file can also be synchronized to a database in the engineering project management system (such as Elasticsearch) for storage.

[0107] Step A20, using a text separator and a preset text segmentation length, the text is segmented to obtain a plurality of text blocks.

[0108] The text separator may be a pre-set specific character used to separate text, such as a line break, a space, etc.

[0109] The preset text segmentation length may be a preset maximum text length, for example, the preset text segmentation length may be set to 500 words or 1000 words.

[0110] In a specific implementation, the text can be initially segmented according to text delimiters to obtain several text segments; then, when it is detected that the text length of the text segment is greater than the preset text segmentation length, the text segment can be segmented again using the preset text segmentation length until all of them meet the preset text segmentation length, and finally several text blocks are obtained.

[0111] Step A30, calling a pre-built vector encoding model to encode each text block, obtaining a vector corresponding to each text block, and storing the vector corresponding to each text block in a vector database.

[0112] The vector encoding model is a pre-trained deep model. It is a model obtained based on a large amount of training corpus. It converts fixed-length text into a vector form in numerical form, enabling the machine to process and understand language.

[0113] In the process of training the vector encoding model, a fixed-length text is encoded by calling a deep neural network, a vector is used as the encoding of the text, and the vector encoding model is output. A set of dense vectors is used to represent a text, and this vector encoding model can reflect the information of the text. Similar text paragraphs have similar vectors.

[0114] In a specific implementation, a pre-built vector encoding model is called to encode each of the above-obtained text blocks to obtain a vector corresponding to each text block; and then the vector corresponding to each text block is stored in a vector database (eg, Milvus).

[0115] It should be noted that the obtained vectors are stored in a special vector database (Milvus). Elasticsearch stores text blocks and their metadata, and Milvus stores the corresponding vectors. A text block in Elasticsearch can correspond to a fixed-length vector in Milvus. They can be matched one-to-one with a number.

[0116] When performing text search, you can use the full-text search capability of Elasticsearch; when performing vector search, you can use the similarity search capability of Milvus.

[0117] Step A40, extracting metadata of each text block, and constructing a knowledge base using each text block, the vector of the text block and the metadata of the text block.

[0118] Metadata refers to descriptive information related to a text block; it usually includes the source of the text block, such as the document file to which the text block originally belongs, the name, format, storage path of the document file, and the type attributes of the document file.

[0119] As an example, assume that the names of document files include engineering design documents, engineering management documents, engineering summary documents, reporting documents, etc.; wherein the metadata may indicate that the corresponding text block may belong to one of the engineering design documents, engineering management documents, engineering summary documents, and reporting documents.

[0120] In the specific implementation, the metadata of each text block can be extracted; then, a knowledge base is constructed using each text block, its vector and its metadata for use. The knowledge base integrates text blocks, vectors and metadata, and uses the relevant database and vector database as basic components to provide fast and accurate information retrieval services.

[0121] It should be noted that the data passed in at different times needs to be reviewed to ensure the consistency of the called vector model and the text block (e.g., paragraph length) to avoid errors in subsequent calculations.

[0122] In addition, the knowledge base needs to be regularly updated and maintained to ensure the accuracy and currency of the information. This includes adding new text blocks, updating metadata, and re-indexing.

[0123] In the user access control permissions pre-set in the engineering project management system, the user access control permissions can also be bound to the data of the knowledge base at the same time. When using the engineering project management system, different permissions are considered respectively and the data that can be viewed by the permission is returned.

[0124] Furthermore, in one embodiment, step A10 may include the following sub-steps A101 to A102:

[0125] Sub-step A101, if the document file is a text format file, extract the text in the document file and remove repeated and meaningless characters in the text;

[0126] Sub-step A102, if the document file is a picture format file or the document file contains a picture, the text in the document file is recognized and acquired using OCR recognition technology, and repeated and meaningless characters in the text are removed.

[0127] In a specific implementation, the type of the document file (such as txt, doc, docx, xml, pdf, jpg, png, etc.) can be determined first; when it is determined that the document file is a text format file (such as txt, doc, docx, pdf, etc.), the text in the document file can be directly obtained, and then the text can be cleaned to remove repeated and meaningless characters in the text; when it is determined that the document file is a picture format file (such as jpg, png, etc.) or the document file contains pictures, OCR recognition technology can be used to identify and extract the text in the document file, and then the text can be cleaned to remove repeated and meaningless characters in the text.

[0128] Furthermore, in one embodiment, step A20 may include the following sub-steps A201 to A202:

[0129] Sub-step A201, using a text delimiter to perform initial text segmentation on the text to obtain a plurality of text segments, and detecting whether the text length of each text segment is greater than a preset text segmentation length;

[0130] Sub-step A202, when it is detected that the text length of the text segment is greater than the preset text segmentation length, the text segment is segmented using the preset text segmentation length to obtain a plurality of text blocks.

[0131] In a specific implementation, the text can be preliminarily segmented using a text separator (e.g., a line break) to obtain a number of text segments; and then the text length of each text segment is detected to see whether it is greater than a preset text segmentation length. When it is determined that the text length of the text segment is greater than the preset text segmentation length, the text segment can be segmented again using the preset text segmentation length to obtain a number of text blocks. In addition, when it is determined that the text segment is less than the preset text segmentation length, two adjacent text segments can be merged to obtain a merged text block; when the length of the merged text block is greater than the preset text segmentation length, merging is not required.

[0132] In this embodiment, a document file is obtained when an engineering design project is executed, and the document file is preprocessed to obtain text in the document file; the text is segmented using a text separator and a preset text segmentation length to obtain a plurality of text blocks; a pre-built vector encoding model is called to encode each text block to obtain a vector corresponding to each text block, and the vector corresponding to each text block is stored in a vector database; metadata of each text block is extracted, and a knowledge base is constructed using each text block, the vector of the text block, and the metadata of the text block; thereby, the constructed knowledge base can be used to provide fast and accurate information retrieval services for subsequent generation of work reports.

[0133] Based on the above embodiments, a third embodiment of the method for generating an engineering project report based on an AI large model is proposed. Figure 3 In this embodiment, after step A40, the method for generating an engineering project report based on the AI ​​big model may further include the following steps S50 to S70:

[0134] Step S50, receiving the question text input by the user, encoding the question text, and obtaining a question vector.

[0135] The question text may be a natural language question input by the user.

[0136] As an example, where can I find the latest progress on XYZ engineering design project?

[0137] Question vector refers to converting question text into vector representation.

[0138] In a specific implementation, after receiving the question text input by the user, the question text can be encoded using a vector encoding model to obtain a corresponding question vector. The question text is converted into a vector representation so that similarity search can be performed in the vector database in the knowledge base.

[0139] See also Figure 4 , Figure 4 It shows a flow chart of a question-and-answer dialogue in the execution information of an engineering design project; when the user inputs the question text (such as querying the execution information of a certain engineering project), the vector encoding model can be used to convert "querying the execution information of a certain engineering project" into a vector representation to obtain the corresponding question vector.

[0140] Step S60, performing similarity matching between the question vector and the vectors in the knowledge base, determining a number of target vectors corresponding to the question vector, and extracting a target text block corresponding to each target vector from the knowledge base.

[0141] The target text block may be a text block corresponding to the target vector in the knowledge base.

[0142] In a specific implementation, the vector most similar to the question vector can be retrieved in the knowledge base, and then several corresponding target vectors (eg, 3) can be determined; for a single target vector, the corresponding target text block can be extracted from the knowledge base.

[0143] Step S70: input the question text and each target text block into a large language model, and use the large language model to generate an answer to the question text.

[0144] In a specific implementation, each target text block can be used as context and input into a large language model together with the question text. The large language model then analyzes and understands the question text and each target text to generate an answer to the question text.

[0145] Furthermore, in one embodiment, the step S70 may further include the following sub-step S701:

[0146] Sub-step S701, using the large language model to analyze and understand the question text and each target text block, generating an answer to the question text, and visually displaying the answer.

[0147] In a specific implementation, a large language model can be used to analyze and understand the question text and each target text block, thereby understanding the intention of the question text, understanding the context and integrating the information of each target text block to make it clearer and more reasonable, and generate natural language answers; the generated answers can also be visualized, and icons, infographics or other visual elements can be used to enhance the user experience, so that users can quickly obtain intuitive and accurate answers.

[0148] In this embodiment, a question text input by a user is received, the question text is encoded to obtain a question vector; the question vector is matched with the vectors in the knowledge base for similarity, a number of target vectors corresponding to the question vector are determined, and target text blocks corresponding to each target vector are extracted from the knowledge base; the question text and each target text block are input into a large language model together, and the answer to the question text is generated using the large language model. The present invention can realize retrieval-style question-answering of project execution status by using a large language model and a knowledge base, and can improve the efficiency of project management.

[0149] Based on the same inventive concept, the fourth embodiment of the present invention also provides an engineering project report generation device based on an AI large model corresponding to the engineering project report generation method based on an AI large model in the previous embodiment. Since the principle of solving the problem by the device in the fourth embodiment of the present invention is similar to the engineering project report generation method based on an AI large model in the previous embodiment of the present invention, the implementation of the system can refer to the implementation of the method, and the repeated parts will not be repeated. Please refer to Figure 5 , the engineering project report generation device based on the AI ​​big model of the present invention may include:

[0150] The receiving module 10 is used to receive the prompt word input by the user and generate a work report template according to the prompt word;

[0151] A filling module 20 is used to extract relevant content of the work report template from the knowledge base using a large language model, and fill the relevant content into each module in the work report template to obtain each report module;

[0152] An adjustment module 30, configured to use the large language model to check and adjust each report module to obtain each adjusted report module;

[0153] The generating module 40 is used to integrate each of the adjusted report modules using the large language model to generate a target work report.

[0154] In addition, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned method for generating an engineering project report based on the AI ​​big model is implemented.

[0155] Figure 6 is a schematic block diagram of an electronic device provided in an embodiment of the present application. Figure 6 As shown, the electronic device includes: at least one processor 401, a memory 402, at least one network interface 403 and a user interface 405. The various components in the electronic device are coupled together through a bus system 404. It can be understood that the bus system 404 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 404 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Figure 6 In the specification, various buses are labeled as bus systems.

[0156] The user interface 405 may include a display, a keyboard, a mouse, a trackball, a click gun, keys, buttons, a touch pad or a touch screen.

[0157] It is understood that the memory 402 can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM). The memory described in the embodiments of the present invention is intended to include but is not limited to these and any other suitable categories of memory.

[0158] The memory 402 in the embodiment of the present invention is used to store various categories of data to support the operation of the electronic device 400. Examples of these data include: any executable program for operating on the electronic device 400, such as an operating system 4021 and an application 4022; the operating system 4021 includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application 4022 may include various applications, such as a media player (Media Player), a browser (Browser), etc., for implementing various application services. The method for generating an engineering project report based on an AI large model provided in the embodiment of the present invention may be included in the application 4022.

[0159] The method disclosed in the above embodiment of the present invention can be applied to the processor 401, or implemented by the processor 401. The processor 401 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit in the processor 401 or the instruction in the form of software. The above processor 401 may be a general processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The processor 401 can implement or execute the methods, steps and logic block diagrams disclosed in the embodiment of the present invention. The general processor 401 may be a microprocessor or any conventional processor, etc. In combination with the steps of the method for generating an engineering project report based on an AI large model provided in the embodiment of the present invention, it can be directly embodied as a hardware decoding processor to execute, or it can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium, which is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0160] In an exemplary embodiment, the electronic device 400 may be implemented by one or more application specific integrated circuits (ASIC), DSP, programmable logic device (PLD), complex programmable logic device (CPLD) to execute the aforementioned method.

[0161] In summary, the present invention integrates the document files of project execution status and generates the reports required by users by utilizing a large language model and knowledge base, thereby improving the generation efficiency of work reports and the accuracy of work reports.

[0162] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for generating an engineering project report based on an AI big model, characterized in that: Applied to an engineering project management system, the engineering project management system includes a large language model and a knowledge base, and the method includes: Receiving a prompt word input by a user, and generating a work report template according to the prompt word; Extracting relevant content of the work report template from a knowledge base using a large language model, and filling the relevant content into each module in the work report template to obtain each report module; The large language model is used to check and adjust each report module to obtain each adjusted report module; the large language model is used to integrate each adjusted report module to generate a target work report.

2. The method according to claim 1, characterized in that The extracting relevant content of the work report template from the knowledge base using the large language model includes: Extract each module of the work report template, encode each module, and generate a query text vector corresponding to each module; Each query text vector is matched with the vectors in the knowledge base for similarity, a number of target vectors corresponding to each query text vector are determined, and a number of target text blocks corresponding to the target vectors are extracted from the knowledge base.

3. The method according to claim 1, characterized in that Before the step of receiving the prompt word input by the user and generating a work report template according to the prompt word, the method further includes: Obtaining a document file during the execution of the engineering design project, and preprocessing the document file to obtain text in the document file; wherein the text includes basic project information, project professional information, project task information, project working time information, project progress information, and project completion information; Using a text separator and a preset text segmentation length, the text is segmented to obtain a plurality of text blocks; Calling a pre-built vector encoding model to encode each text block, obtaining a vector corresponding to each text block, and storing the vector corresponding to each text block in a vector database; The metadata of each text block is extracted, and a knowledge base is constructed using each text block, the vector of the text block and the metadata of the text block.

4. The method according to claim 3, characterized in that The preprocessing of the document file to obtain text in the document file includes: If the document file is a text format file, extract the text in the document file and remove repeated and meaningless characters in the text; If the document file is a picture format file or contains a picture, the OCR recognition technology is used to recognize and obtain the text in the document file, and duplicate and meaningless characters in the text are removed.

5. The method according to claim 3, characterized in that: The text is segmented using a text separator and a preset text segmentation length to obtain a plurality of text blocks, including: Performing initial text segmentation on the text using a text delimiter to obtain a plurality of text segments, and detecting whether the text length of each text segment is greater than a preset text segmentation length; When it is detected that the text length of the text segment is greater than the preset text segmentation length, the text segment is segmented using the preset text segmentation length to obtain a plurality of text blocks.

6. The method according to claim 3, characterized in that The report generation method of the engineering project also includes: Receiving a question text input by a user, encoding the question text to obtain a question vector; Performing similarity matching between the question vector and the vectors in the knowledge base, determining a plurality of target vectors corresponding to the question vector, and extracting a target text block corresponding to each target vector from the knowledge base; The question text and each target text block are input into a large language model together, and the answer to the question text is generated by using the large language model.

7. The method according to claim 6, characterized in that The step of using the large language model to generate an answer to the question text includes: The large language model is used to analyze and understand the question text and each target text block, generate an answer to the question text, and visualize the answer.

8. A device for generating engineering project reports based on an AI large model, characterized in that: Applied to an engineering project management system, the engineering project management system includes a large language model and a knowledge base, and the device includes: A receiving module, used for receiving a prompt word input by a user, and generating a work report template according to the prompt word; A filling module, used to extract relevant content of the work report template from a knowledge base using a large language model, and fill the relevant content into each module in the work report template to obtain each report module; An adjustment module, used to check and adjust each report module using the large language model to obtain each adjusted report module; A generation module is used to integrate each of the adjusted report modules using the large language model to generate a target work report.

9. An electronic device, characterized in that: The electronic device comprises: a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device performs the steps of the method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, and when the program is run, it is used to implement the steps of the method described in any one of claims 1 to 7.

Images