Method and device for generating quotation sheet based on large language model

Abstract

The invention discloses a quotation generation method and device based on a large language model in the technical field of electric digital data processing, and the method comprises the steps: setting environment parameters of the large language model, and obtaining a text answer through sending a question package; reading parameters and information of the engineering bid invitation document by adopting a large language model, and extracting cue words; modifying and storing a parameter table in the project bidding document; parameters in the parameter table are calculated according to the cathode protection industry knowledge, and a price list text is generated. According to the method, the related parameters required by the project in the document are extracted by adopting the general large language model, calculation is carried out according to the extracted related parameters, the quotation list is generated, compared with manual quotation list filling, the quotation list generation time and labor cost can be effectively saved, the large language model is applied to the cathode protection industry, and the cathode protection efficiency is improved. By combining industry knowledge and a large model technology, the processing efficiency of related documents can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic digital data processing, and in particular to a method and device for generating a quotation based on a large language model. Background Art

[0002] "Cathode protection" is a common method to extend the life of key infrastructure such as bridges and pipelines. The traditional method of measuring cathodic protection is to manually read the demand documents and manually calculate the amount of cathodic protection materials used and the quotation, which often requires a lot of manpower and time costs.

[0003] According to the invention patent document with the authorization number CN118627501B, a document information extraction method, device, equipment and storage medium thereof are provided. The document information is extracted by self-supervised learning. By extracting node information and parsing the document structure of the target document, annotation consumption and computing resources can be saved. However, due to the lack of industry knowledge support in the implementation process of the scheme, the problem of cathodic protection quotation calculation cannot be solved in a targeted manner.

[0004] According to the invention patent document with publication number CN118568116A, an intelligent information processing system based on a large model agent is provided. The system is driven by large model technology, enabling people to use natural language as an interactive method to perform and process professional or complex work tasks in a highly automated manner. However, the technical solution adopted by the patent document does not involve specific tasks such as document generation and parameter calculation.

[0005] According to the invention patent document with publication number CN118312583B, a method, device and storage medium for equipment requirement matching based on LLM are provided, in which prompt word text is input into the LLM big model, and the text is searched in the vector database through the LLM big model to generate and output corresponding equipment requirement matching information, thereby completing the matching of equipment requirements in the equipment knowledge list. However, this method and device do not involve industry knowledge or parameter calculation.

[0006] Therefore, the existing processing methods based on large models have the following shortcomings: first, manually calculating the usage and quotation of cathodic protection materials often requires a lot of manpower and time costs; second, the technical solutions adopted do not involve specific tasks such as document generation and parameter calculation, resulting in insufficient accuracy of information processing; third, it does not involve industry knowledge or parameter calculation, making it difficult to provide more targeted quotations for engineering bidding documents. Summary of the invention

[0007] The purpose of the present invention is to provide a quotation generation method and device based on a large language model to solve the problem that the above-mentioned existing methods cannot solve the cathodic protection quotation calculation in a targeted manner, resulting in cumbersome quotation generation steps and inaccurate generated quotation data.

[0008] In a first aspect, to achieve the above-mentioned purpose, the present invention provides a quotation generation method based on a large language model, comprising: Set the environment parameters of the large language model and obtain text answers by sending question packages. The large language model is a general large language model. By setting this parameter of the large language model environment, the parameter weights learned during model training can be increased by setting the parameters. When a larger number of parameters is set, the capacity of the large language model will increase accordingly, so that more complex patterns and knowledge can be captured, which is beneficial to improving the reading speed of engineering bidding documents and the accuracy of the large language model in reading engineering bidding document information; The large language model is used to read the parameters and information of the engineering bidding documents and extract prompt words, which is conducive to users to quickly determine the key information of the engineering bidding documents and whether the large language model is accurately processing the engineering bidding documents, which is conducive to determining the processing speed of the engineering bidding documents. Modify and store the parameter table in the engineering bidding document. By modifying and storing the updated parameter table, the large language model can be adapted to the information extraction weights of different batches of engineering bidding documents. Calculating the parameters in the parameter table based on the cathodic protection industry knowledge and generating the quotation text is helpful to save the time for generating the quotation, shorten the quotation generation cycle, and improve the processing efficiency of the engineering bidding documents.

[0009] As a further solution of the present invention: the prompt word corresponding to each question will be written into a json format package, which contains the prompt word of the overall environment and the question corresponding to the parameter; At the same time, these packages are sent by using the client.chat.completions program in the openai api program, and the returned text is obtained. This part of the text will be displayed in the text box as the basis for modifying the variable value in the parameter table in the next step.

[0010] As a further solution of the present invention: the method for reading the parameters and information of the engineering bidding document is to use the computer language python to read the parameters and information, and the parameters and information of the engineering bidding document are stored in a database.

[0011] As a further solution of the present invention: the method for calculating the parameters in the parameter table is to use a math library to calculate the value of each parameter, and calculate the missing parameters in the parameter table according to the corresponding formula in the math library to complete the establishment of the parameter table document.

[0012] As a further solution of the present invention: the parameter table is a plain text file in a comma-separated value format, the header of the parameter table includes six attributes, and the parameter table includes all parameters that need to be used in the quotation generation process.

[0013] As a further solution of the present invention: the six attributes are respectively an "ID" attribute, a "Name" attribute, a "Value" attribute, a "Unit" attribute, a "Method" attribute and a "Describe" attribute; The "ID" attribute indicates the number of the parameter; The "Name" attribute represents the Chinese name of each parameter; The "Value" attribute indicates the value of the parameter; The "Unit" attribute indicates the unit of the parameter; The "Method" attribute indicates the type of the parameter; The "Describe" attribute represents the document content generated after the parameters are written into the parameter table; The attribute table is shown in Table 1 below: As a further solution of the present invention: the Chinese names of the various parameters are different from each other; the type of the parameter value is any one of a number, an array or a string type, the type of the parameter value is determined by the unit of the parameter, and the type of the parameter value defaults to a string type.

[0014] As a further solution of the present invention: the parameters include first-category parameters, second-category parameters, third-category parameters and fourth-category parameters, the first-category parameters are parameters that need to be searched in the engineering bidding document, the second-category parameters are numerical values ​​involved in the calculation of the quotation text, the second-category parameters are given numerical values, and the third-category parameters are calculated from the first-category parameters and the second-category parameters; the fourth-category parameters are calculated from the first-category parameters and the second-category parameters by searching the engineering bidding document, or, when the engineering bidding document is searched for no result, are calculated from the first-category parameters and the second-category parameters.

[0015] As a further solution of the present invention: the second category of parameters includes: calculation parameters of cathodic protection, industry knowledge parameters, basic parameters of cathodic protection and non-numerical parameters, the calculation parameters of cathodic protection are obtained by calculation; the industry knowledge parameters are set to fixed values ​​in the initial state of calculation; the basic parameters of cathodic protection and the non-numerical parameters are both obtained by searching the engineering bidding documents.

[0016] As a further solution of the present invention: the type numbers of the third type parameters and the fourth type parameters are both set to "Method" attributes.

[0017] As a further solution of the present invention: the basic parameters of the cathodic protection include pipeline length, pipeline diameter, protection current density, coating damage factor, soil resistivity and pipeline burial depth; the calculation parameters of the cathodic protection include protection area, protection current, sacrificial anode output current, sacrificial anode output current, sacrificial anode dosage and sacrificial anode life; the industry knowledge parameters include packing material resistivity, sacrificial anode equivalent diameter, pre-packaged sacrificial anode diameter, effective potential difference, spare coefficient and sacrificial anode net weight consumption rate; the non-numerical parameters include project overview, cathodic protection technology introduction and principle, cathodic protection technical standards and specifications and technical specifications. The setting parameters for protecting the outer surface of the pipeline are shown in Table 2 below: The setting parameters for protecting the inner surface of the tank are shown in Table 3 below: The setting parameters for protecting the outer surface of the tank using impressed current are shown in Table 4 below: The setting parameters for protection using ship impressed current are shown in Table 5 below: The setting parameters for using sacrificial anodes for wind power pile protection are shown in Table 6 below: The setting parameters for using sacrificial anodes for protection of submarine pipelines are shown in Table 7 below: The setting parameters for protection using PCCP pipeline rod zinc anodes are shown in Table 8 below: In actual use, after obtaining the project bidding document, the user first puts the project bidding document into the large model processing software and obtains the question results. Secondly, the parameter modification software is opened, and the values ​​in the parameter table are modified according to the question results, and this method is used to generate a quotation.

[0018] As a further solution of the present invention: the question package is a set of questions manually set according to different parameter tables, and each time the same parameter table is called, an automatic calling program is used to call the corresponding questions in the question set.

[0019] As a further solution of the present invention: the automatic calling program is an exe file. In the automatic calling program, the user can drag the PDF format file to be processed into the box on the right, enter the question in the question content column, and click the "Start Information Extraction" button after the input is completed. The large model will reply to the question in the question content column, and the answer text of the question will be displayed in the text box.

[0020] As a further solution of the present invention: the PyQt5.QtWidgets function in the PyQt programming language is used to obtain the text answer and save the text answer.

[0021] As a further solution of the present invention: the method for modifying and storing the parameter table includes: filling in necessary parameter values, filling in optional parameter values, saving the parameter table, and generating an updated parameter table in the same directory of the program.

[0022] In the second aspect, the present invention provides a quotation generation device based on a large language model, which is applied to the quotation generation method based on a large language model as described in the above-mentioned scheme, and the device includes a text input / output module, an information reading module, an information storage module and a parameter calculation module: the text input / output module is used to set the environmental parameters of the large language model and output the text answer; the information reading module uses the large language model to read the parameters and information of the engineering bidding document and extract prompt words; the information storage module is used to store the parameter table in the engineering bidding document; the parameter calculation module is used to calculate the parameters in the parameter table according to the cathodic protection industry knowledge and generate the quotation text.

[0023] In a third aspect, the present invention provides an electronic device, comprising: a memory and a processor coupled to the memory, the memory being used to store executable program code, the processor calling the executable program code stored in the memory, the executable program code being used to execute the quotation generation method based on a large language model described in the above scheme.

[0024] In a fourth aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store a computer program, and the computer executes the quotation generation method based on a large language model described in the above scheme by running the computer program.

[0025] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention uses a general large language model to extract relevant parameters required by the project in the document, and calculates and generates a quotation based on the extracted relevant parameters, which can effectively save the quotation generation time and labor costs compared to manually filling out the quotation.

[0026] 2. The present invention applies the big language model to the cathodic protection industry and combines industry knowledge with big model technology to improve the processing efficiency of related documents, further improve work efficiency, and effectively improve the accuracy of quotation data.

[0027] 3. The document processing method included in the present invention is applicable to documents of various formats or writing methods. There is no need to specify the format and writing method of the document. Compared with traditional document processing methods and processes, it has stronger versatility, adaptability to occasions and flexibility of use.

[0028] 4. Compared with a single document processing method based on a large language model, the present invention also involves technologies such as parameter calculation and automatic document generation, which can effectively improve the convenience of document processing, improve document processing efficiency, and provide a more targeted quotation for engineering bidding documents. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Generate a flow chart for the quotation of the present invention; Figure 2 This is a module diagram of a quotation generating device of the present invention; Figure 3 Generating a graph for the text answer of the present invention; Figure 4 Generate a graph for the parameter table of the present invention; Figure 5 Generate a graph for the quotation sheet of the present invention. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] Example: like Figure 1 As shown, the present invention provides a quotation generation method based on a large language model, comprising: S1: Set the environmental parameters of the large language model, and obtain text answers by sending question packages. Apply the large language model to the cathodic protection industry. By combining the parameters of the cathodic protection industry with the large model technology, the processing efficiency of related documents can be improved, the work efficiency can be further improved, and the accuracy of the quotation data can be effectively improved; By setting this parameter in the large language model environment, the parameter weights learned during model training can be increased. When a large number of parameters are set, the capacity of the large language model will increase, so that more complex patterns and knowledge can be captured, which is beneficial to improving the reading speed of engineering bidding documents and the accuracy of the large language model in reading engineering bidding document information. S2: Use a large language model to read the parameters and information of the project bidding document and extract prompt words; The parameters and information of the engineering bidding document are read through the large language model, and the prompt words are extracted at the same time, which is conducive to the user to quickly determine the key information of the engineering bidding document and whether the processing direction of the engineering bidding document by the large language model is accurate, which is conducive to determining the processing speed of the engineering bidding document; S3: modify and store the parameter table in the project bidding document; By modifying and storing the parameter table in the engineering bidding documents, the large language model can be adapted to the information extraction weights of different batches of engineering bidding documents; S4: Calculate the parameters in the parameter table based on cathodic protection industry knowledge and generate a quotation text.

[0032] By calculating the parameters in the parameter table based on the knowledge of the cathodic protection industry and generating the quotation text, it is helpful to save the time for generating the quotation, shorten the quotation generation cycle, and improve the processing efficiency of the engineering bidding documents.

[0033] In the above step S1, when setting the environment parameters of the large language model, a prompt word pops up on the page: "This is a document processing software. Next, I will give you a document. Answer questions based on the content of the document. For each question, first repeat the question number and the question itself in a line of text, then output the answer in a new line, and at the same time output the page number of the answer in the document. My question is as follows: ", the operator can answer the question according to the prompt and generate a corresponding text answer.

[0034] like Figure 3As shown, the prompt word corresponding to each question will be written into a JSON format package, which contains the prompt word of the overall environment and the question corresponding to the parameter. At the same time, the client.chat.completions program in the openai api program is used to send these packages and obtain the returned text. This part of the text will be displayed in the text box as the basis for modifying the variable value in the parameter table in the next step.

[0035] Preferably, the method for reading the parameters and information of the engineering bidding document is to use the computer language python to read the parameters and information, and the parameters and information of the engineering bidding document are stored in a database.

[0036] Preferably, the method for calculating the parameters in the parameter table is to use a math library to calculate the value of each parameter, and calculate the missing parameters in the parameter table according to the corresponding formula in the math library to complete the establishment of the parameter table document.

[0037] like Figure 4 As shown, the parameter table is a plain text file in comma separated value format. The header of the parameter table includes six attributes. The parameter table includes all the parameters required in the quotation generation process.

[0038] Preferably, the six attributes are "ID" attribute, "Name" attribute, "Value" attribute, "Unit" attribute, "Method" attribute and "Describe" attribute, "ID" attribute indicates the number of the parameter; "Name" attribute indicates the Chinese name of each parameter; "Value" attribute indicates the value of the parameter; "Unit" attribute indicates the unit of the parameter; "Method" attribute indicates the type of the parameter; "Describe" attribute indicates the document content generated after the parameter is written into the parameter table, and the attribute table is shown in Table 1 below: Preferably, the Chinese names of the parameters are different from each other; the type of the parameter value is any one of a number, an array or a string type, the type of the parameter value is determined by the unit of the parameter, and the type of the parameter value is a string type by default. Preferably, the parameters include first-category parameters, second-category parameters, third-category parameters and fourth-category parameters. The first-category parameters are parameters that need to be searched in the engineering bidding documents, the second-category parameters are values ​​involved in the quotation text calculation, the second-category parameters are given values, and the third-category parameters are calculated from the first-category parameters and the second-category parameters; the fourth-category parameters are calculated from the first-category parameters and the second-category parameters by searching the engineering bidding documents, or, when the engineering bidding documents are searched unsuccessfully, are calculated from the first-category parameters and the second-category parameters.

[0039] Preferably, the second category of parameters includes: calculation parameters of cathodic protection, industry knowledge parameters, basic parameters of cathodic protection and non-numerical parameters. The calculation parameters of cathodic protection are obtained by calculation; the industry knowledge parameters are set to fixed values ​​in the initial state of calculation; the basic parameters and non-numerical parameters of cathodic protection are obtained by searching the engineering bidding documents.

[0040] Preferably, the type numbers of the third type parameters and the fourth type parameters are both set to "Method" attributes.

[0041] Preferably, the basic parameters of cathodic protection include pipeline length, pipeline diameter, protection current density, coating damage factor, soil resistivity and pipeline burial depth; the calculation parameters of cathodic protection include protection area, protection current, sacrificial anode output current, sacrificial anode output current, sacrificial anode usage and sacrificial anode life; Industry knowledge parameters include packing material resistivity, sacrificial anode equivalent diameter, pre-packaged sacrificial anode diameter, effective potential difference, standby factor and net weight consumption rate of sacrificial anode; Non-numerical parameters include project overview, cathodic protection technology introduction and principles, cathodic protection technology standards and specifications and technical specifications.

[0042] The setting parameters for protecting the outer surface of the pipeline are shown in Table 2 below: The setting parameters for protecting the inner surface of the tank are shown in Table 3 below: The setting parameters for protecting the outer surface of the tank using impressed current are shown in Table 4 below: The setting parameters for protection using ship impressed current are shown in Table 5 below: The setting parameters for using sacrificial anodes for wind power pile protection are shown in Table 6 below: The setting parameters for using sacrificial anodes for protection of submarine pipelines are shown in Table 7 below: The setting parameters for protection using PCCP pipeline rod zinc anodes are shown in Table 8 below: In actual use, after obtaining the project bidding document, the user first puts the project bidding document into the large model processing software and obtains the question results. Secondly, the parameter modification software is opened, and the values ​​in the parameter table are modified according to the question results, and this method is used to generate a quotation.

[0043] Preferably, the question package is a set of questions manually set according to different parameter tables, and each time the same parameter table is called, the corresponding question in the question set is called by an automatic calling program.

[0044] Preferably, the PyQt5.QtWidgets function in the PyQt programming language is used to obtain the text answer and save the text answer.

[0045] Preferably, the automatic calling program is an exe file. In the automatic calling program, the user can drag the PDF file to be processed into the box on the right, enter the question in the question content column, and click the "Start Information Extraction" button after the input is completed. The large model will answer the question in the question content column, and the answer text of the question will be displayed in the text box; After obtaining the answer text of the question, save the answer as a text file through the "Save As" button.

[0046] The main programs of the big model are written in the API_Control function, which mainly uses the OpenAI SDK to access the online big model open platform, send questions and get return results.

[0047] Preferably, in order to output answers to multiple questions in the same inquiry, the file is stored using ContextCaching. This method can avoid repeated calculation of tokens when the file text is repeatedly called as a prompt word, which is conducive to significant cost savings. The platform used in the actual project is the moonshot open platform, and the model used is moonshot-v1-128k.

[0048] Preferably, the problem of interacting with the large model is set as follows: according to the parameter table, find all parameters whose Method value contains 0, that is, the parameters whose answers need to be found in the document. If the parameter name is "x", the corresponding question is "Based on the document content, what should the value of x be?"

[0049] In this program, users can open the .csv table file to be processed by clicking the "Read parameter table" button. The table file marks the parameter values ​​that must be filled in. Through the output results of the large model and combined with manual judgment, the corresponding data is filled in the Value unit, and the parameter values ​​to be filled in can be selected; If the output of the large model indicates that there is a corresponding answer in the document, then fill in the relevant content in the form; if there is no answer, leave it blank. After filling in all the necessary content, click the "Save Parameter Form" button to save the modified results; Then, click the "Generate Document" button to generate the final design document in the same directory of the program, and set the document format to Word format. This document is a complete quotation. The specific format of the quotation depends on the actual type of project. By using a general large language model to extract the relevant parameters required for the project in the document, and calculating and generating a quotation based on the extracted relevant parameters, it can effectively save the quotation generation time and labor costs compared to manually filling out the quotation.

[0050] The document processing method included in the above method is applicable to documents of various formats or writing methods. There is no need to specify the format and writing method of the document. Compared with traditional document processing methods and processes, it has strong versatility, adaptability to occasions and flexibility of use. On the other hand, compared with the single document processing method based on a large language model, it also involves technologies such as parameter calculation and automatic document generation, which can effectively improve the convenience of document processing and improve document processing efficiency.

[0051] like Figure 4 As shown, the method for modifying and storing the parameter table includes: filling in necessary parameter values, filling in optional parameter values, saving the parameter table, and generating an updated parameter table in the same directory of the program.

[0052] The present invention also provides a quotation generation device based on a large language model, which is applied to the quotation generation method based on a large language model as in the above-mentioned scheme, and the device includes a text input / output module, an information reading module, an information storage module and a parameter calculation module: the text input / output module is used to set the environmental parameters of the large language model and output a text answer; the information reading module uses the large language model to read the parameters and information of the engineering bidding document and extract prompt words; the information storage module is used to store the parameter table in the engineering bidding document; the parameter calculation module is used to calculate the parameters in the parameter table according to the cathodic protection industry knowledge and generate the quotation text.

[0053] The present invention also provides an electronic device, comprising: a memory and a processor coupled to the memory, the memory being used to store executable program code, the processor calling the executable program code stored in the memory, and the executable program code being used to execute the quotation generation method based on a large language model of the above-mentioned scheme.

[0054] The present invention also provides a computer-readable storage medium, which is used to store a computer program. The computer executes the quotation generation method based on a large language model of the above scheme by running the computer program.

[0055] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A quotation generation method based on a large language model, characterized in that: include: Set the environment parameters of the large language model and get text answers by sending question packages; Use a large language model to read the parameters and information of the project bidding documents and extract prompt words; Modify and store parameter tables in engineering bidding documents; Calculate the parameters in the parameter table based on cathodic protection industry knowledge and generate a quotation text.

2. The method for generating a quotation based on a large language model according to claim 1, characterized in that: The parameter table is a plain text file in a comma-separated value format. The header of the parameter table includes six attributes. The parameter table includes all parameters that need to be used in the quotation generation process.

3. The method for generating a quotation based on a large language model according to claim 2, characterized in that: The six attributes are respectively "ID" attribute, "Name" attribute, "Value" attribute, "Unit" attribute, "Method" attribute and "Describe" attribute, wherein the "ID" attribute indicates the number of the parameter; the "Name" attribute indicates the Chinese name of each parameter; the "Value" attribute indicates the value of the parameter; the "Unit" attribute indicates the unit of the parameter; the "Method" attribute indicates the type of the parameter; and the "Describe" attribute indicates the document content generated after the parameter is written into the parameter table.

4. The method for generating a quotation based on a large language model according to claim 3, characterized in that: The parameters include first-category parameters, second-category parameters, third-category parameters and fourth-category parameters. The first-category parameters are parameters that need to be searched in the engineering bidding document, the second-category parameters are numerical values ​​involved in the calculation of the quotation text, the second-category parameters are given numerical values, and the third-category parameters are calculated from the first-category parameters and the second-category parameters; the fourth-category parameters are calculated from the first-category parameters and the second-category parameters by searching the engineering bidding document, or, when the engineering bidding document is searched for no result, by searching the first-category parameters and the second-category parameters.

5. The method for generating a quotation based on a large language model according to claim 4, characterized in that: The second type of parameters includes: calculation parameters of cathodic protection, industry knowledge parameters, basic parameters of cathodic protection and non-numerical parameters, wherein the calculation parameters of cathodic protection are obtained by calculation; the industry knowledge parameters are set to fixed values ​​in the initial state of calculation; the basic parameters of cathodic protection and the non-numerical parameters are obtained by searching the engineering bidding documents, and the basic parameters of cathodic protection include pipeline length, pipeline diameter, protection current density, coating damage factor, soil resistivity and pipeline burial depth; The calculation parameters of cathodic protection include protection area, protection current, sacrificial anode output current, sacrificial anode usage and sacrificial anode life; the industry knowledge parameters include filling material resistivity, sacrificial anode equivalent diameter, pre-packaged sacrificial anode diameter, effective potential difference, spare coefficient and net weight consumption rate of sacrificial anode; the non-numerical parameters include project overview, cathodic protection technology introduction and principles, cathodic protection technical standards and specifications and technical specifications.

6. The method for generating a quotation based on a large language model according to claim 5, characterized in that: The question package is a set of questions manually set according to different parameter tables. Each time the same parameter table is called, the corresponding question in the question set is called by an automatic calling program.

7. The method for generating a quotation based on a large language model according to claim 6, characterized in that: The method for modifying and storing the parameter table includes: filling in necessary parameter values, filling in optional parameter values, saving the parameter table, and generating an updated parameter table in the same directory of the program.

8. A quotation generation device based on a large language model, characterized in that: Applied to the method for generating a quotation based on a large language model as described in any one of claims 1 to 7, the device comprises: A text input / output module, the text input / output module is used to set the environment parameters of the large language model and output the text answer; An information reading module, which uses a large language model to read parameters and information of the project bidding document and extract prompt words; An information storage module, the information storage module is used to store parameter tables in the engineering bidding documents; A parameter calculation module is used to calculate the parameters in the parameter table according to the cathodic protection industry knowledge and generate a quotation text.

9. An electronic device, characterized in that: include: A memory and a processor coupled to the memory, wherein the memory is used to store executable program codes, the processor calls the executable program codes stored in the memory, and the executable program codes are used to execute the quotation generation method based on a large language model as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store a computer program, and the computer executes the quotation generation method based on a large language model according to any one of claims 1 to 7 by running the computer program.

Images