Test platform construction method and device, equipment, storage medium and program product

Abstract

Embodiments of the present application disclose a test platform construction method, device, equipment, storage medium and program product. The method comprises: performing information extraction on a target document to obtain multi-modal information related to a simulation experiment, the multi-modal information comprising first simulation information in the form of text, picture information and table information; performing associated integration processing on the various information included in the multi-modal information to obtain simulation experiment information; retrieving a first knowledge base based on the simulation experiment information, and generating an experiment scheme according to the retrieved first knowledge; the first knowledge base comprises the following knowledge required for various simulation experiments: a simulation type and the following information corresponding to the simulation type: a circuit structure, a set parameter interval corresponding to a simulation index, and a simulation index calculation formula; retrieving a second knowledge base based on the experiment scheme, and generating target test platform code according to the retrieved second knowledge, the second knowledge base comprising knowledge related to a programming language.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to a method, apparatus, equipment, storage medium, and program product for constructing a testing platform. Background Technology

[0002] Currently, the process of extracting simulation experiment information from simulation experiment-related documents (such as analog circuit papers) and generating testbench (TB) code typically relies on manual extraction and conversion. This traditional method is not only time-consuming and labor-intensive, but also prone to information omissions or errors, resulting in low efficiency and insufficient accuracy. Furthermore, when processing multimodal data (such as text, images, and tables) in simulation experiment-related documents, there is often a lack of effective integration capabilities, making it difficult to fully grasp the experimental details in the paper, which significantly impacts subsequent simulation experiments. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a method, apparatus, device, storage medium, and program product for constructing a testing platform.

[0004] This application provides a method for constructing a test platform. The method includes: extracting information from a target document to obtain multimodal information related to a simulation experiment. The multimodal information includes first simulation information represented by text, image information, and table information.

[0005] The various types of information included in the multimodal information are correlated and integrated to obtain simulation experiment information;

[0006] The simulation experiment information is retrieved from the first knowledge base, and an experimental plan is generated based on the retrieved first knowledge. The first knowledge base includes the following knowledge required for simulation experiments of various simulation types: simulation type and the following information corresponding to the simulation type: circuit structure, setting parameter range corresponding to simulation index, and simulation index calculation formula.

[0007] Based on the experimental scheme, a second knowledge base is retrieved, and the target test platform code is generated according to the retrieved second knowledge. The second knowledge base includes knowledge related to programming languages.

[0008] In some optional embodiments, the step of associating and integrating multiple pieces of information in the multimodal information to obtain simulation experiment information includes: analyzing and processing the image information and table information included in the multimodal information to obtain second simulation information represented by the image information and simulation index information represented by the table information; and associating and integrating the first simulation information, the second simulation information, and the simulation index information to obtain the simulation experiment information.

[0009] In some optional embodiments, the step of analyzing and processing the image information and table information included in the multimodal information to obtain the second simulation information represented by the image information and the simulation index information represented by the table information includes: performing character recognition and feature extraction on the image information, obtaining the second simulation information represented by the image information based on the recognized characters and extracted features, wherein the simulation information includes the descriptive information of the image information and the simulation point information corresponding to the graphics in the image information; and performing semantic understanding and information extraction on the table information to obtain the simulation index information represented by the table information, wherein the simulation index information includes simulation indexes related to the simulation experiment and the corresponding numerical values ​​of the simulation indexes.

[0010] In some optional embodiments, the step of associating and integrating the first simulation information, the second simulation information, and the simulation index information to obtain the simulation experiment information includes: determining the first simulation type to which the simulation experiment belongs based on at least one of the first simulation information, the second simulation information, and the simulation index information; and determining the correspondence between the first simulation information and the second simulation information related to each simulation index based on the simulation index information included in the simulation index information, wherein the simulation experiment information includes the first simulation type and the correspondence.

[0011] In some optional embodiments, after obtaining the simulation experiment information, the method further includes: performing a completeness check on the simulation experiment information based on a preset first evaluation rule; the first evaluation rule includes at least one of the following: whether simulation indicator information is missing, whether simulation indicator information is duplicated, and whether simulation information corresponding to the simulation indicator information is missing; if the simulation experiment information is determined to be incomplete according to the check result, the simulation experiment information is corrected, and the correction process includes at least one of the following: adding missing simulation indicator information, deleting duplicate simulation indicators, and adding simulation information corresponding to the simulation indicator information.

[0012] In some optional embodiments, the step of retrieving the first knowledge base based on the simulation experiment information includes: retrieving the first knowledge base based on the first simulation type in the simulation experiment information to determine the first circuit structure corresponding to the first simulation type; retrieving the first knowledge base based on the correspondence related to each simulation index in the simulation experiment information to obtain the set parameter range and calculation formula corresponding to each simulation index; and obtaining first knowledge containing the first circuit structure, the set parameter range and calculation formula corresponding to each simulation index.

[0013] In some optional embodiments, the first knowledge includes: the first circuit structure corresponding to the first simulation type to which the simulation experiment belongs, and the set parameter range and calculation formula corresponding to each simulation index; the step of generating an experimental scheme based on the retrieved first knowledge includes: a step of building the circuit structure of the simulation experiment based on the first circuit structure; a step of setting the parameter range corresponding to each simulation index based on the set parameter range corresponding to each simulation index, and generating the parameter range corresponding to each simulation index; and a step of calculating the value of each simulation index based on the calculation formula corresponding to each simulation index.

[0014] In some optional embodiments, generating the target test platform code based on the retrieved second knowledge includes: generating initial test platform code based on the retrieved second knowledge; executing the initial test platform code and repairing the initial test platform code based on the code execution result to generate the target test platform code.

[0015] In some optional embodiments, the step of executing the initial test platform code and repairing the initial test platform code based on the code execution result to generate target test platform code includes: selecting a first circuit model corresponding to the first simulation type to which the simulation experiment belongs; running the initial test platform code on the first circuit model to obtain code execution results; determining the credibility of the initial test platform code based on the code execution results; and optimizing the initial test platform code if the credibility does not meet the preset conditions until a target test platform code whose credibility meets the preset conditions is obtained.

[0016] In some optional embodiments, determining the credibility of the initial test platform code based on the code execution result includes: determining the difference value between the code execution result and the theoretical execution result corresponding to the first circuit model, and determining the credibility of the initial test platform code based on the difference value.

[0017] This application embodiment also provides a test platform construction apparatus, the apparatus comprising:

[0018] The information extraction and processing unit is used to extract information from the target document to obtain multimodal information related to the simulation experiment. The multimodal information includes first simulation information represented by text, image information, and table information. It is also used to perform association and integration processing on the various types of information included in the multimodal information to obtain simulation experiment information.

[0019] The scheme generation unit is used to retrieve a first knowledge base based on the simulation experiment information; the first knowledge base includes the following knowledge required for simulation experiments of various simulation types: simulation type and the following information corresponding to the simulation type: circuit structure, setting parameter range corresponding to simulation index, and simulation index calculation formula; it is also used to generate an experimental scheme based on the retrieved first knowledge.

[0020] The code building unit is used to retrieve a pre-built second knowledge base based on the experimental scheme, and generate target test platform code based on the retrieved second knowledge; the second knowledge base includes knowledge related to programming languages.

[0021] This application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the test platform construction method described in this application.

[0022] This application provides a computer-readable storage medium storing a computer program. When the computer program is loaded and executed by a processor, it implements the test platform construction method of this application.

[0023] This application provides a computer program product, which includes a computer program. When the computer program is loaded and executed by a processor, it implements the test platform construction method of this application.

[0024] The technical solution of this application embodiment, in the first aspect, involves extracting information from target documents related to simulation experiments to obtain multimodal information including first simulation information represented by text, image information, and table information. The various types of information included in the multimodal information are then correlated and integrated to obtain simulation experiment information. This achieves the extraction and effective integration of multimodal data in the paper, and enables the comprehensive acquisition of experimental details in the paper. While improving the ability to extract simulation experiment information, it also improves the accuracy and efficiency of obtaining simulation experiment information.

[0025] Secondly, based on the simulation experiment information, the first knowledge is retrieved from the first knowledge base, which includes the knowledge required for simulation experiments of various simulation types. Based on the first knowledge, the experimental plan is generated, which improves the simulation experiment steps and experimental plan details in the target document.

[0026] Thirdly, based on the generated experimental scheme, second knowledge is retrieved from a second knowledge base including knowledge related to programming languages, and target test platform code is generated based on the second knowledge, thereby realizing the automatic construction of test platform code, reducing the technical threshold for simulation experiment preparation, and improving the efficiency and accuracy of experiment reproduction. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating a test platform construction method provided in an embodiment of this application;

[0028] Figure 2 This is a schematic diagram of the process for obtaining simulation experiment information provided in an embodiment of this application;

[0029] Figure 3 This is a flowchart illustrating the process of determining the target testing platform code provided in an embodiment of this application;

[0030] Figure 4 This is a schematic diagram of the processing flow of a test platform construction method provided in an embodiment of this application;

[0031] Figure 5 This is a schematic diagram of the structure of a test platform construction device provided in an embodiment of this application;

[0032] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0033] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0034] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0035] Furthermore, in the description of the embodiments of this application, the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance. Thus, features defined as "first" or "second" may explicitly or implicitly include one or more features. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Moreover, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0036] This application provides a method for building a test platform. Figure 1 This is a flowchart illustrating a test platform construction method provided in an embodiment of this application; as shown... Figure 1 As shown, the test platform construction method may include at least:

[0037] Step 101: Extract information from the target document to obtain multimodal information related to the simulation experiment. The multimodal information includes first simulation information represented by text, image information, and table information.

[0038] In this embodiment, the target document can be a document related to the simulation experiment. For example, the target document can be a paper, technical document, experimental manual, etc. related to the simulation experiment. Specifically, it can be a paper on analog circuits (i.e., a target paper for analog circuit simulation), a paper on digital circuits, etc.

[0039] In this embodiment, the extraction of target document information specifically involves extracting multimodal information from the target document. Multimodal information refers to information expressed in multiple forms. In the target document, multimodal information may include text, images, tables, and other formats. That is, information extraction from the target document may include extracting text, image, and table information related to the simulation experiment. By extracting multimodal information from the target document, multimodal information related to the simulation experiment, including first simulation information represented by text, image information, and table information, can be obtained. Of course, if the target document does not contain images or tables, the corresponding extracted image or table information will be empty.

[0040] For example, the first simulation information represented by text extracted from text may include parameter settings, experimental conditions, result analysis, and other content related to the simulation experiment; the image information extracted from images may include images related to the simulation experiment, as well as image identifiers, image titles, and text descriptions related to the images in the target document; the table information extracted from tables may include tables related to the simulation experiment, as well as table identifiers, table titles, and text descriptions related to the tables in the target document.

[0041] It should be noted that the methods used for extracting multimodal information in this application embodiment are not specifically limited. For example, a Large Language Model (LLM) can be used to extract information from the target document; that is, by leveraging the LLM's understanding of the target document, textual information related to the simulation experiment can be extracted from the target document, and first simulation information based on the generated text representation can be obtained. Alternatively, a multimodal large model can be used to extract information from the images and / or tables of the target document; that is, by leveraging the multimodal large model's understanding of the target document, image and table information related to the simulation experiment can be extracted from the target document.

[0042] The process of extracting images and / or tables from the target document using a multimodal large model can classify all images and / or tables in the target document, identifying those relevant to the simulation experiment, thereby extracting multimodal information from these relevant images and / or tables. For example, a binary classification method can be used to classify images and tables in the target document.

[0043] Step 102: Perform correlation and integration processing on the various types of information included in the multimodal information to obtain simulation experiment information.

[0044] In this embodiment, after obtaining multimodal information including first simulation information represented by text, image information, and table information, the various types of information in the multimodal information are correlated and integrated to obtain simulation experiment information related to the simulation experiment, which includes various types of information.

[0045] Step 103: Retrieve the first knowledge base based on the simulation experiment information, and generate an experiment plan based on the retrieved first knowledge; the first knowledge base includes the following knowledge required for various simulation experiments: simulation type and the following information corresponding to the simulation type: circuit structure, the set parameter range corresponding to the simulation index, and the calculation formula of the simulation index.

[0046] In this embodiment, after obtaining simulation experiment information related to the simulation experiment, the pre-built first knowledge base is retrieved based on the simulation experiment information, and the experimental scheme in text form is generated according to the retrieved first knowledge related to the simulation experiment.

[0047] The first knowledge base can be a simulation experiment knowledge base pre-built according to a preset experimental scheme template, which can include professional knowledge required for simulation experiments of various simulation types. Taking simulation experiments related to analog circuits as an example, the first knowledge base can specifically include simulation type (or circuit type), as well as the circuit structure corresponding to each simulation type, the set parameter range corresponding to the simulation index, and the calculation formula of the simulation index.

[0048] Optionally, circuit structure refers to the physical connection method and topological relationship between the components in the circuit; the set parameter range corresponding to the simulation index refers to the default parameter range range that can be set according to human experience or historical experiments; the simulation index calculation formula refers to the relevant calculation formula that involves the simulation index, and / or the relevant calculation formula used to calculate the simulation index.

[0049] It should be noted that the embodiments of this application do not impose specific restrictions on the method of constructing the first knowledge base. For example, the first knowledge base can be constructed by using LLM based on Retrieval-augmented generation (RAG) technology.

[0050] Step 104: Retrieve the second knowledge base based on the experimental scheme, and generate the target test platform code based on the retrieved second knowledge. The second knowledge base includes knowledge related to programming languages.

[0051] In this embodiment, after obtaining the experimental plan, the second knowledge used to generate the test platform code is retrieved from the pre-constructed second knowledge base, which includes programming language-related knowledge, according to the experimental plan. Based on the second knowledge, standard target test platform code is generated so as to realize the simulation experiment described in the target document by executing the target test platform code.

[0052] The second knowledge base is pre-built based on programming language-related knowledge, specifically including syntax rules, commonly used functions, and corresponding examples for various programming languages ​​(such as Python, Java, C++, etc.). The target test platform code is the test platform (TB, Testbench) code used to implement the simulation experiments described in the target document.

[0053] It should be noted that the embodiments of this application do not impose specific limitations on the method of retrieving second knowledge in the second knowledge base. For example, a code generation framework can be designed based on RAG technology, and LLM can be used to retrieve the second knowledge required to execute the experimental scheme according to the experimental scheme paradigm.

[0054] By adopting the technical solution of this application embodiment, on the one hand, information is extracted from the target documents related to the simulation experiment to obtain multimodal information including first simulation information represented by text, image information and table information. The various types of information included in the multimodal information are associated and integrated to obtain simulation experiment information. This realizes the extraction and effective integration of multimodal data in the paper, and achieves the comprehensive acquisition of experimental details in the paper. While improving the ability to extract simulation experiment information, it also improves the accuracy and efficiency of obtaining simulation experiment information.

[0055] On the other hand, based on simulation experiment information, first knowledge is retrieved from a first knowledge base containing knowledge required for simulation experiments of various simulation types. Experimental plans are then generated based on this first knowledge, refining the simulation experiment steps and details in the target document. Furthermore, based on the generated experimental plan, second knowledge is retrieved from a second knowledge base containing knowledge related to programming languages. Target test platform code is then generated based on this second knowledge, thereby achieving automatic construction of the test platform code. This lowers the technical threshold for simulation experiment preparation and improves the efficiency and accuracy of experiment reproduction.

[0056] It should be noted that the above-described test platform construction method can be executed individually or collaboratively by any computer device with large model processing capabilities. For example, the computer device may include an information extraction and processing unit, a scheme generation unit, and a code construction unit. In this case, steps 101 and 102 can be executed by the information extraction and processing unit of the computer device, step 103 can be executed by the scheme generation unit of the computer device, and step 104 can be executed by the code construction unit of the computer device.

[0057] In one feasible implementation, step 102 includes: analyzing and processing the image information and table information included in the multimodal information respectively to obtain the second simulation information represented by the image information and the simulation index information represented by the table information; and performing correlation and integration processing on the first simulation information, the second simulation information and the simulation index information to obtain the simulation experiment information.

[0058] In this embodiment, the image information and table information in the multimodal information are analyzed and extracted in depth. The second simulation information is extracted from the image information, and the simulation index information is extracted from the table information. The extracted first simulation information represented by text, the second simulation information represented by image information, and the simulation index information represented by table information are associated and integrated to obtain structured simulation experiment information related to the simulation experiment information.

[0059] In one feasible implementation, the step of analyzing and processing the image information and table information included in the multimodal information to obtain the second simulation information represented by the image information and the simulation index information represented by the table information includes: performing character recognition and feature extraction on the image information, obtaining the second simulation information represented by the image information based on the recognized characters and extracted features, wherein the second simulation information includes the descriptive information of the image information and the simulation point information corresponding to the graphics in the image information; and performing semantic understanding and information extraction on the table information to obtain the simulation index information represented by the table information, wherein the simulation index information includes simulation indexes related to the simulation experiment and the corresponding numerical values ​​of the simulation indexes.

[0060] In this embodiment, the analysis and processing of image information may specifically include: performing character recognition and feature extraction on the image information, and obtaining second simulation information based on the recognized characters and extracted features, which includes descriptive information of the image information and simulation point information corresponding to the graphics in the image information.

[0061] Specifically, character recognition can include recognizing descriptive text, numbers, and other characters in simulated images; and obtaining natural language descriptions of the simulated images based on the recognized character information. Feature extraction can specifically include recognizing and extracting features such as waveforms and simulated points from different types of simulated images (e.g., line graphs, bar charts, histograms, time series graphs, etc.). Simulated points, also known as feature points, can represent specific points in a graphic or line, or points in a feature map formed after image processing. For example, the extracted features may include the coordinate information of the simulated points; then, the features of the simulated points (such as coordinate information) are combined to generate the simulated point information corresponding to the graphic.

[0062] Furthermore, based on the identified characters, descriptive information of the image information is obtained, and based on the extracted features, simulation point information is obtained. The descriptive information and simulation point information are combined to obtain the second simulation information.

[0063] It should be noted that the methods used for image information analysis and extraction in this application embodiment are not specifically limited. For example, image information analysis and extraction can be achieved through technologies such as target detection and / or optical character recognition (OCR).

[0064] In this embodiment, the analysis and processing of table information may specifically include: semantic understanding and information extraction of table information, and the obtained simulation index information includes simulation indexes related to the simulation experiment and the corresponding values ​​of the simulation indexes.

[0065] Understandably, by semantically understanding the information in the tables, simulation indicators and their corresponding values ​​can be extracted from the tables. The values ​​corresponding to the simulation indicators can also be recorded as simulation reference values. For example, from the table of circuit experiments, the simulation indicator "device power consumption" and the corresponding value "power value" can be extracted.

[0066] It should be noted that the method used for analyzing and extracting table information in this application embodiment is not specifically limited. For example, the table information extracted by the multimodal large model can be input into the LLM, and simulation index information including simulation indexes and the corresponding values ​​of the simulation indexes can be generated through the processing of the LLM.

[0067] In one feasible implementation, the step of associating and integrating the first simulation information, the second simulation information, and the simulation index information to obtain the simulation experiment information includes: determining the first simulation type to which the simulation experiment belongs based on at least one of the first simulation information, the second simulation information, and the simulation index information; and determining the correspondence between the first simulation information and the second simulation information related to each simulation index based on the simulation index information included in the simulation index information, wherein the simulation experiment information includes the first simulation type and the correspondence.

[0068] In this embodiment, the association and integration of the first simulation information, the second simulation information, and the simulation indicators may specifically include: analyzing at least one of the first simulation information, the second simulation information, and the simulation indicator information to determine the first simulation type to which the simulation experiment belongs, where the first simulation type may also be referred to as the first experiment type or the first circuit type; and for one or more simulation indicators in the simulation indicator information, fusing the first simulation information and the second simulation information related to each simulation indicator to obtain one or more correspondences related to each simulation indicator, including multimodal information, thereby obtaining simulation experiment information including the first simulation type and the correspondences.

[0069] For example, regarding the simulation metric of device power consumption, multimodal information related to device power consumption in the first and second simulation information can be correlated and integrated. For instance, multimodal information such as the power value corresponding to device power consumption, different experimental temperature conditions, relevant simulation images, and descriptive information can be correlated and integrated to obtain a multidimensional structured correspondence of device power consumption.

[0070] In another alternative implementation, the method further includes: obtaining simulation experiment settings based on at least one of the first simulation information, the second simulation information, and the simulation index information.

[0071] The simulation experiment settings refer to global simulation experiment settings, that is, global settings for one or more experiments contained in the target document.

[0072] It should be noted that the embodiments of this application do not impose specific restrictions on the method for obtaining simulation experiment settings; for example, LLM can be used for information extraction.

[0073] In one feasible implementation, after obtaining the simulation experiment information, the method further includes: performing a completeness check on the simulation experiment information based on a preset first evaluation rule; the first evaluation rule includes at least one of the following: whether simulation indicator information is missing, whether simulation indicator information is duplicated, and whether simulation information corresponding to the simulation indicator information is missing; if the simulation experiment information is determined to be incomplete based on the check result, the simulation experiment information is corrected, and the correction process includes at least one of the following: adding missing simulation indicator information, deleting duplicate simulation indicators, and adding simulation information corresponding to the simulation indicator information.

[0074] In this embodiment, after obtaining the simulation experiment information related to the simulation experiment, it is necessary to perform a completeness check on the simulation experiment information based on the preset first evaluation rule in order to identify whether there are any missing, duplicate or inconsistent information in the simulation experiment information, thereby further improving the acquisition quality of multimodal information and simulation experiment information.

[0075] The first evaluation rule may include at least one of the following: whether simulation index information is missing, whether simulation index information is duplicated, and whether simulation information corresponding to the simulation index information is missing.

[0076] In this embodiment, the "whether simulation index information is missing" in the first evaluation rule can not only check whether the simulation indexes of the current simulation experiment extracted from the target document are missing, but also check whether there are any related indicators that are not mentioned in the target document but are involved in the current simulation experiment and are missing.

[0077] As one implementation method, the first evaluation rule may include a pre-set checklist. The checklist may include specific check items for each type of simulation experiment, including simulation settings, required simulation metrics, and other simulation experiment information. Thus, the checklist can be matched one-to-one with each item in the simulation experiment information according to the first evaluation rule. This allows for a sequential check of the simulation type, simulation settings, simulation metrics, and their corresponding multi-dimensional information to determine if there are any duplicates or omissions, thereby verifying the completeness of the simulation experiment information.

[0078] As another implementation method, the first evaluation rule can be used to check whether the simulation experiment information related to the current simulation experiment in the target document has been completely extracted. That is, check whether the simulation experiment information obtained is all the information related to the current simulation experiment in the target document, and whether there is any multi-dimensional information in the simulation settings, simulation indicators and their corresponding relationships in the target document that has not been extracted.

[0079] Optionally, if the simulation experiment information is determined to be incomplete based on the inspection results, corresponding correction processing is performed on the simulation experiment information according to different inspection results. Specifically, this may include: adding missing information (such as simulation type, simulation settings, simulation indicators and their corresponding multi-dimensional information) to the simulation indicator information, deleting duplicate simulation indicators and their corresponding relationships, and adding simulation information corresponding to the simulation indicator information. The completeness check of the corrected simulation experiment information is then performed again based on a preset first evaluation rule. Through repeated correction of the simulation experiment information, it is continued until the completeness of the simulation experiment information is determined based on the inspection results.

[0080] It should be noted that if the simulation indicators of the current simulation experiment extracted from the target document are missing, the corresponding relationship is obtained repeatedly until the simulation indicators are complete; if the simulation indicators that are not mentioned in the target document but are involved in the current simulation experiment are missing, the simulation indicator is directly generated but the corresponding value is not generated, and the generated simulation indicator is added to the simulation experiment information.

[0081] If the simulation experiment information is confirmed to be complete based on the inspection results, subsequent experimental plans can be generated based on the simulation experiment information.

[0082] It should be noted that the method used in this application embodiment to check the integrity of simulation experiment information is not specifically limited; for example, it can be implemented through the self-reflection mechanism of LLM.

[0083] In one alternative implementation, Figure 2 This is a flowchart illustrating the process of obtaining simulation experiment information provided in this application embodiment. Taking a target paper on analog circuit experiments as an example, the optional implementation process of obtaining simulation experiment information in this invention may include:

[0084] Step 201: Input the target paper of the analog circuit experiment into the information extraction and processing unit to perform the following steps.

[0085] Step 202: Extract information from the target paper to obtain multimodal information related to the simulation experiment, including first simulation information represented by text, image information, and table information.

[0086] Step 203: Perform character recognition and extraction on the image information. Based on the recognized characters and extracted features, generate natural language description information for the image information. Also, perform feature extraction on the image information. Based on the extracted features, obtain simulation point information. Integrate the recognized description information and simulation point information to obtain the second simulation information representing the image information.

[0087] Furthermore, through semantic understanding and information extraction of the table information, the simulation indicators and corresponding values ​​of the simulation experiment are extracted from the table, thereby obtaining simulation indicator information including simulation indicators related to the simulation experiment and the corresponding values ​​of the simulation indicators.

[0088] Step 204: Perform correlation and integration processing on the various information contained in the multimodal information to obtain simulation experiment information.

[0089] Specifically, the first simulation type (i.e., circuit type) of the simulation experiment is analyzed from at least one of the first simulation information, second simulation information, and simulation index information; for one or more simulation indices in the simulation index information, the first simulation information and second simulation information related to each simulation index are fused to obtain one or more correspondences related to each simulation index, including multimodal information; information is extracted from the target paper to obtain the global simulation experiment settings, thereby obtaining simulation experiment information including the first simulation type, the above correspondences, and the simulation experiment settings.

[0090] Step 205: Perform a completeness check on the simulation experiment information based on the first evaluation rule, checking whether the simulation experiment information is missing simulation index information, whether the simulation index information is duplicated, or whether the simulation information corresponding to the simulation index information is missing.

[0091] Step 206: If the simulation experiment information is determined to be incomplete, the simulation experiment information is corrected by adding missing simulation index information, deleting duplicate simulation index information, or adding simulation information corresponding to the simulation index information. Then, step 205 is executed again on the corrected simulation experiment information.

[0092] Step 207: If the simulation experiment information is complete, output the simulation experiment information.

[0093] In one optional implementation, the step of retrieving the first knowledge base based on the simulation experiment information includes: retrieving the first knowledge base based on a first simulation type in the simulation experiment information to determine a first circuit structure corresponding to the first simulation type; retrieving the first knowledge base based on the correspondence related to each simulation index in the simulation experiment information to obtain the set parameter range and calculation formula corresponding to each simulation index; and obtaining first knowledge containing the first circuit structure, the set parameter range and calculation formula corresponding to each simulation index.

[0094] In this embodiment, first knowledge corresponding to simulation experiment information is retrieved from a pre-constructed first knowledge base. Specifically, taking a simulation experiment related to analog circuits as an example, the first circuit structure corresponding to the first simulation type can be retrieved from the first knowledge base based on the first simulation type in the simulation experiment information; and the set parameter range and calculation formula related to each simulation index can be retrieved from the first knowledge base based on the correspondence related to each simulation index in the simulation experiment information, thereby obtaining first knowledge including the first circuit structure, the set parameter range and calculation formula corresponding to each simulation index.

[0095] For example, the specific process of implementing first knowledge retrieval based on the first knowledge base may include: inputting simulation experiment information into a large language model, calling the first knowledge base through a knowledge retrieval algorithm, and thus outputting the first knowledge corresponding to each piece of information in the simulation experiment information. It should be noted that the embodiments of this application do not specifically limit the design of the knowledge retrieval algorithm.

[0096] In one optional implementation, the step of generating an experimental scheme based on the retrieved first knowledge includes: a step of building a circuit structure for a simulation experiment based on the first circuit structure; a step of setting a parameter range for each simulation index based on the set parameter range corresponding to each simulation index; and a step of calculating the value of each simulation index based on the calculation formula corresponding to each simulation index.

[0097] In this embodiment, after retrieving the first knowledge related to the simulation experiment information through the first knowledge base, a specific experimental plan including multiple experimental steps is generated based on the different knowledge in the first knowledge.

[0098] Optionally, taking a simulation experiment related to analog circuits as an example, the following steps can be generated: Based on the first circuit structure in the first knowledge, the steps for building the circuit structure of the simulation experiment can be described in natural language; the parameter ranges for each simulation index in the current experiment can be determined based on the parameter ranges corresponding to each simulation index in the first knowledge, thereby generating the steps for setting the parameter ranges for each simulation index in natural language; and the calculation steps for calculating the values ​​of each simulation index can be generated based on the calculation formulas corresponding to each simulation index in the first knowledge. Based on the steps for building the circuit structure, setting the parameter ranges, and calculating the values ​​of the simulation indexes, a standardized experimental scheme corresponding to the simulation experiment information is formed according to the execution order of the experimental steps.

[0099] Understandably, during the process of retrieving the parameter ranges for each simulation indicator from the first knowledge base based on the first knowledge, if the currently obtained simulation experiment information includes parameter ranges for some or all simulation indicators, the parameter ranges for the simulation indicators in the simulation experiment information can replace the retrieved parameter ranges for the corresponding simulation indicators, thereby determining the parameter range for the simulation indicator, and generating parameter range setting steps based on the parameter range for the simulation indicator; if the currently obtained simulation experiment information does not include parameter ranges for simulation indicators, then the retrieved parameter ranges for each simulation indicator are used as the parameter ranges for the simulation indicators in the current experiment, thereby generating parameter range setting steps based on the parameter ranges for the simulation indicators.

[0100] It should be noted that the embodiments of this application do not impose specific limitations on the method of generating experimental plans based on the first knowledge. For example, the retrieved first knowledge can be input into an LLM, and the LLM can output an experimental plan described in natural language.

[0101] In some optional implementations, after obtaining the experimental scheme, the method further includes: performing a completeness check on the experimental scheme based on a preset second evaluation rule; the second evaluation rule includes at least one of the following: whether experimental steps are missing, whether the output format of the experimental scheme is correct; if the experimental scheme is determined to be incomplete based on the check results, supplementing the experimental scheme; if the output format of the experimental scheme is determined to be incorrect based on the check results, adjusting the output format of the experimental scheme.

[0102] In this embodiment, after generating the experimental plan corresponding to the simulation experiment information based on the retrieved first knowledge, the experimental plan needs to be checked based on the preset second evaluation rule to ensure that the generated experimental plan is executable and complete.

[0103] The second evaluation rule may include whether experimental steps are missing, and / or whether the output format of the experimental protocol is correct.

[0104] For example, the second evaluation rule may include a pre-set checklist, which may include specific simulation experiment plans and formats corresponding to various simulation experiments. Thus, the pre-set checklist can be matched with the experimental plans according to the second evaluation rule to check whether the experimental plans are complete.

[0105] Optionally, if the inspection results determine that the experimental plan is incomplete, i.e., the experimental plan is missing experimental steps, the missing experimental steps can be directly generated and added to the experimental plan; if the inspection results determine that the output format of the experimental plan is incorrect, the output format of the experimental plan can be directly adjusted to the correct format.

[0106] If the experimental plan is found to be complete and the output format is correct based on the inspection results, the experimental plan will be used to generate the target test platform code.

[0107] In one optional implementation, generating the target test platform code based on the retrieved second knowledge includes: generating initial test platform code based on the retrieved second knowledge; executing the initial test platform code and repairing the initial test platform code based on the code execution result to generate the target test platform code.

[0108] In this embodiment, initial test platform code can be generated first based on the second knowledge retrieved from the second knowledge base. The initial test platform code is then executed, i.e., the initial test platform code is tested. Based on the code execution result (i.e. the test result), it is determined whether the initial test platform code needs to be repaired, thereby generating target test platform code that can be executed correctly.

[0109] It should be noted that this application does not impose specific limitations on the method used to generate test platform code. Taking simulation experiments related to analog circuits as an example, the first simulation type in the experimental plan and simulation experiment information can be input into the LLM. The LLM then calls the second knowledge base, selects the second knowledge related to Python required by the experimental plan, and constructs standard test platform code based on the second knowledge.

[0110] In one optional implementation, the step of executing the initial test platform code and repairing the initial test platform code based on the code execution result to generate target test platform code includes: selecting a first circuit model corresponding to the first simulation type to which the simulation experiment belongs; running the initial test platform code on the first circuit model to obtain code execution results; determining the credibility of the initial test platform code based on the code execution results; and optimizing the initial test platform code if the credibility does not meet a preset condition until a target test platform code whose credibility meets the preset condition is obtained.

[0111] In this embodiment, a first circuit model corresponding to the current simulation type can be selected from a variety of preset test models corresponding to different experiment types. This first circuit model can also be called an ideal test model. The initial test platform code is tested by running it on the ideal test model. For example, if the current simulation experiment is related to analog circuits, the first circuit model corresponding to the first simulation type of the simulation experiment is selected, and the initial test platform code is run on the first circuit model to test the initial test platform code.

[0112] Optionally, during the generation of the target test platform code, the credibility of the initial test platform code can be determined based on the code execution results (i.e., test results). Based on preset conditions, it can be determined whether the current initial test platform code is the target test platform code. If the credibility does not meet the preset conditions, the initial test platform code can be modified once or multiple times based on the code execution results, historical modification experience, or relevant modification knowledge retrieved. This achieves continuous optimization of the initial test platform code. The modified test platform code is then retested using the first circuit model until the target test platform code with credibility meeting the preset conditions is obtained, thus completing the construction of the test platform.

[0113] It should be noted that the embodiments of this application do not impose specific restrictions on the methods used to optimize the initial test platform code. For example, the current code execution result can be input into the LLM, which analyzes the code execution result. This eliminates the need to regenerate the initial test platform code, but only modifies the initial test platform code based on the analysis results, and outputs the modified test platform code to reduce resource waste.

[0114] In one optional implementation, determining the credibility of the initial test platform code based on the code execution result includes: determining the difference value between the code execution result and the theoretical execution result corresponding to the first circuit model, and determining the credibility of the initial test platform code based on the difference value.

[0115] In this embodiment, the credibility of the initial test platform code can be determined by the difference between the code execution result and the theoretical execution result corresponding to the first circuit model previously selected for testing the initial test platform code.

[0116] The theoretical execution results corresponding to the first circuit model can be multiple, and the specific number can correspond to the number of simulation experiments in the current target document. That is, when the current target document describes multiple simulation experiments, by executing the test platform code corresponding to each simulation experiment, the corresponding code execution results can be obtained. Based on the difference between each code execution result and its corresponding theoretical execution result, the credibility of the initial test platform code of the current simulation experiment can be determined.

[0117] Optionally, if the current simulation experiment is related to analog circuits, the credibility of the initial test platform code can be determined by calculating the difference between the code execution result and the theoretical execution result corresponding to the first circuit model.

[0118] For example, the method for determining the confidence level based on the difference between the code execution result and the theoretical execution result in this application embodiment can be designed as follows:

[0119]

[0120] Where R represents credibility. y represents the code execution result, and y represents the theoretical execution result corresponding to the first circuit model. Specifically, the reliability of the initial test platform code can be obtained by comparing the difference between the code execution result and the theoretical execution result, and by calculating the difference between the absolute value of the difference and the ratio of the theoretical execution result.

[0121] Correspondingly, the preset condition for judging credibility can be: whether the credibility exceeds a preset threshold. For example, the preset threshold can be set to 0.8. It can be understood that the credibility in this embodiment represents the degree or probability of the code execution result being close to the theoretical execution result. Therefore, the credibility value is less than or equal to 1, that is, the preset range corresponding to the current preset threshold can be [0.8, 1].

[0122] In some optional embodiments, the credibility meets preset conditions, including: the credibility is greater than or equal to a preset threshold. If the credibility is less than the preset threshold or greater than 1, it indicates that the current code execution result differs significantly from the theoretical execution result of the first circuit model, that is, the current credibility does not meet the preset conditions. Even if the test platform code corresponding to the current credibility can run normally, the code still has problems and needs to be repaired. If the credibility is greater than or equal to the preset threshold and less than or equal to 1, it indicates that the current code execution result is similar to the theoretical execution result of the first circuit model, the current credibility meets the preset conditions, and the test platform code corresponding to the current credibility can be used as the target test platform code.

[0123] It should be noted that the above process of calculating and judging credibility requires that the code execution results can be output normally, that is, the test platform code runs without errors and can output code execution results. Credibility can be determined based on the output code execution results and the theoretical execution results corresponding to the first circuit model, and then the current test platform code can be judged as the target test platform code based on the credibility.

[0124] In some optional embodiments, during the execution of the initial test platform code on the first circuit model, there may be situations where the code execution results cannot be output normally. When this occurs, i.e., an error occurs during the test platform code execution, the initial test platform code needs to be modified until the modified code can output execution results. For example, the error information can be input into an LLM (Local Level Management System), which analyzes the error information and modifies the test platform code accordingly. The modified test platform code is then re-executed until it can output execution results normally.

[0125] In one alternative implementation, Figure 3 This is a flowchart illustrating the process of determining the target test platform code provided in this application embodiment. Taking a target paper on analog circuit experiments as an example, after generating the experimental scheme corresponding to the simulation experiment information through the scheme generation unit, the optional implementation process of determining the target test platform code in this invention may include:

[0126] Step 301: Input the first simulation type and experimental scheme into the code building unit; to perform the following steps.

[0127] Step 302: Based on the experimental plan, retrieve the second knowledge used to generate the test platform code from the pre-built second knowledge base, which includes programming language-related knowledge.

[0128] Step 303: Generate initial test platform code based on the second knowledge retrieved from the second knowledge base.

[0129] Step 304: Select the first circuit model corresponding to the first simulation type from the test models corresponding to the preset multiple experimental types, and obtain the corresponding code execution results by running the initial test platform code on the first circuit model.

[0130] Step 305: Determine whether the code execution result corresponding to the current initial test platform code can be output normally and whether any errors occur;

[0131] Step 306: If an error message occurs during code execution and the code execution result cannot be output normally, analyze the error message, modify the initial test platform code based on the analysis results, and re-execute steps 304 and 305 after the modification is completed.

[0132] Step 307: Under the condition of normal output code execution results, calculate the credibility of the initial test platform by calculating the code execution results and the theoretical execution results corresponding to the first circuit model.

[0133] For example, the credibility of the initial test platform can be calculated by: calculating the difference between the execution result of the computation code and the theoretical execution result corresponding to the first circuit model, and dividing the difference by the theoretical execution result to obtain the credibility of the initial test platform code.

[0134] Step 308: Determine whether the credibility of the initial test platform code meets the preset conditions.

[0135] For example, determining whether the credibility meets the preset conditions includes: determining whether the credibility is greater than or equal to a preset threshold; if the credibility is greater than or equal to the preset threshold, it indicates that the credibility does not meet the preset conditions; if the credibility is less than the preset threshold, it indicates that the credibility meets the preset conditions.

[0136] Step 309: If the judgment result is that the current credibility does not meet the preset conditions, that is, the credibility is greater than or equal to the preset threshold, then modify the initial test platform code according to the code execution result to optimize the initial test platform code, and re-execute steps 304 to 307 after the modification is completed.

[0137] Step 310: If the judgment result is that the current credibility meets the preset conditions, that is, the credibility is less than the preset threshold, then the test platform code corresponding to the current credibility is used as the target test platform code to complete the construction of the test platform.

[0138] Figure 4 This is a schematic diagram of the processing flow of the test platform construction method provided in this application embodiment. Taking a computer device including an information extraction and processing unit, a scheme generation unit, and a code construction unit as an example, as follows... Figure 4 As shown:

[0139] The information extraction and processing unit extracts and integrates (or fuses) multimodal information from the input target document related to the simulation experiment, obtaining simulation experiment information related to the simulation experiment corresponding to the target document. This simulation experiment information may include all simulation experiments of the target document for the corresponding simulation type, as well as specific simulation experiment information related to each simulation experiment. Specifically, the information extraction and processing unit can execute steps 201 to 207 above.

[0140] The scheme generation unit retrieves information from a pre-built first knowledge base based on various pieces of information in the simulation experiment information. It then generates experimental steps according to the retrieved first knowledge related to the simulation type, thereby obtaining a complete experimental scheme corresponding to the simulation experiment information.

[0141] The code construction unit, based on the complete experimental scheme, retrieves a pre-built second knowledge base and performs code construction processing according to the retrieved second knowledge related to the programming language to obtain initial test platform code that can be used to simulate the experiment described in the target document. Then, the initial test platform code is repaired to obtain the target test platform code, thus realizing the construction of the test platform. Specifically, the code construction unit can execute steps 301 to 310 above.

[0142] Based on the above description, this application also discloses a test platform construction apparatus, which can execute... Figures 1 to 4The steps in the illustrated method flow are shown below. Please refer to [link / reference]. Figure 5 This is a schematic diagram of a test platform construction device provided in an embodiment of this application. The test platform construction device may include an information extraction and processing unit 41, a scheme generation unit 42, and a code construction unit 43, wherein:

[0143] The information extraction and processing unit 41 is used to extract information from the target document to obtain multimodal information related to the simulation experiment. The multimodal information includes first simulation information represented by text, image information, and table information. It is also used to perform association and integration processing on the various types of information included in the multimodal information to obtain simulation experiment information.

[0144] The scheme generation unit 42 is used to retrieve a first knowledge base based on the simulation experiment information; the first knowledge base includes the following knowledge required for simulation experiments of various simulation types: simulation type and the following information corresponding to the simulation type: circuit structure, setting parameter range corresponding to simulation index, and simulation index calculation formula; it is also used to generate an experiment scheme based on the retrieved first knowledge.

[0145] The code construction unit 43 is used to retrieve a second knowledge base based on the experimental scheme and generate target test platform code based on the retrieved second knowledge. The second knowledge base includes knowledge related to programming languages.

[0146] In one embodiment, the information extraction and processing unit 41 is used to analyze and process the image information and table information included in the multimodal information respectively to obtain the second simulation information represented by the image information and the simulation index information represented by the table information; and to perform correlation and integration processing on the first simulation information, the second simulation information and the simulation index information to obtain the simulation experiment information.

[0147] In one embodiment, the information extraction and processing unit 41 is used to perform character recognition and feature extraction on the image information, and obtain second simulation information represented by the image information based on the recognized characters and extracted features. The second simulation information includes description information of the image information and simulation point information corresponding to the graphics in the image information. The unit also performs semantic understanding and information extraction on the table information to obtain simulation index information represented by the table information. The simulation index information includes simulation indexes related to the simulation experiment and the corresponding numerical values ​​of the simulation indexes.

[0148] In one embodiment, the information extraction and processing unit 41 is configured to determine the first simulation type to which the simulation experiment belongs based on at least one of the first simulation information, the second simulation information, and the simulation index information; and to determine the correspondence between the first simulation information and the second simulation information related to each simulation index based on the simulation indexes included in the simulation index information, wherein the simulation experiment information includes the first simulation type and the correspondence.

[0149] In one embodiment, the information extraction and processing unit 41 is further configured to, after obtaining simulation experiment information related to the simulation experiment, perform a completeness check on the simulation experiment information based on a preset first evaluation rule; the first evaluation rule includes at least one of the following: whether simulation indicator information is missing, whether simulation indicator information is duplicated, and whether simulation information corresponding to the simulation indicator information is missing; if the simulation experiment information is determined to be incomplete according to the check result, the simulation experiment information is corrected, and the correction process includes at least one of the following: adding missing simulation indicator information, deleting duplicate simulation indicators, and adding simulation information corresponding to the simulation indicator information;

[0150] In one embodiment, the scheme generation unit 42 is configured to retrieve the first knowledge base based on the first simulation type in the simulation experiment information, and determine the first circuit structure corresponding to the first simulation type; retrieve the first knowledge base based on the correspondence related to each simulation index in the simulation experiment information, and obtain the set parameter range and calculation formula corresponding to each simulation index; and obtain first knowledge containing the first circuit structure, the set parameter range and calculation formula corresponding to each simulation index.

[0151] In one embodiment, the scheme generation unit 42 is used to generate the circuit structure construction steps of the simulation experiment based on the first circuit structure; determine the parameter range corresponding to each simulation index based on the set parameter range corresponding to each simulation index, and generate the parameter range corresponding to each simulation index; and generate the value of each simulation index based on the calculation formula corresponding to each simulation index.

[0152] In one embodiment, the code construction unit 43 is configured to generate initial test platform code based on the second knowledge retrieved by the retrieval unit 42; execute the initial test platform code and repair the initial test platform code based on the code execution result to generate target test platform code.

[0153] In one implementation, the code construction unit 43 is configured to select a first circuit model corresponding to a first simulation type to which the simulation experiment belongs, run the initial test platform code on the first circuit model, and obtain code execution results; determine the credibility of the initial test platform code based on the code execution results; and optimize the initial test platform code if the credibility does not meet the preset conditions, until a target test platform code whose credibility meets the preset conditions is obtained.

[0154] In one implementation, the code construction unit 43 is used to determine the difference between the code execution result and the theoretical execution result corresponding to the first circuit model, and to determine the credibility of the initial test platform code based on the difference.

[0155] In the implementation of this invention, the information extraction and processing unit 41, the scheme generation unit 42, and the code construction unit 43 in the device can all be implemented by the central processing unit (CPU), digital signal processor (DSP), microcontroller unit (MCU), or field-programmable gate array (FPGA) in the device in practical applications.

[0156] It should be noted that the above-described test platform construction device is only illustrated by the division of the above-described program modules. In practical applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the processing described above. In addition, the test platform construction device and the test platform construction method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0157] Based on the description of the above method and apparatus embodiments, this application also provides a computer device. Please refer to... Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. The computer device in this embodiment includes a memory 52, a processor 51, and a computer program stored in the memory 52 and capable of running on the processor 51.

[0158] Optionally, when the processor 51 executes the program, it implements the steps of the test platform construction method described in the embodiments of this application.

[0159] Optionally, the various components in the computer device can be coupled together via a bus system 53. It is understood that the bus system 53 is used to implement communication between these components. In addition to a data bus, the bus system 53 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 6 The general labeled all buses as Bus System 53.

[0160] It is understood that memory 52 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), ferromagnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), 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), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).The memory 52 described in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0161] The methods disclosed in the embodiments of this application can be applied to processor 51, or implemented by processor 51. Processor 51 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 integrated logic circuit of the hardware in processor 51 or by instructions in the form of software. The processor 51 may be a general-purpose processor, DSP, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 51 can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly manifested as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium, which is located in memory 52. ​​Processor 51 reads the information in memory 52 and combines its hardware to complete the steps of the aforementioned method.

[0162] In an exemplary embodiment, the computer device may be implemented by one or more ASICs, DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), FPGAs, general-purpose processors, controllers, MCUs, microprocessors, or other electronic components to perform the aforementioned methods.

[0163] This application also provides a computer-readable storage medium on which a computer program is stored.

[0164] Optionally, the computer-readable storage medium can be applied to the test platform construction apparatus of the present application embodiment; then when the program is executed by the processor, it implements the steps of the test platform construction method of the present application embodiment.

[0165] This application also provides a computer program product, including a computer program that can be executed by the processor 51 of a network device to complete the steps of the test platform construction method described in this application.

[0166] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0167] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0168] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0169] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0170] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0171] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0172] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0173] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0174] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for constructing a test platform, characterized in that, The method includes: Information is extracted from the target document to obtain multimodal information related to the simulation experiment. The multimodal information includes first simulation information represented by text, image information, and table information. The various types of information included in the multimodal information are correlated and integrated to obtain simulation experiment information; The simulation experiment information is retrieved from the first knowledge base, and an experimental plan is generated based on the retrieved first knowledge. The first knowledge base includes the following knowledge required for various simulation experiments: simulation type and the following information corresponding to the simulation type: circuit structure, the set parameter range corresponding to the simulation index, and the calculation formula of the simulation index. Based on the experimental scheme, a second knowledge base is retrieved, and the target test platform code is generated according to the retrieved second knowledge. The second knowledge base includes knowledge related to programming languages.

2. The method according to claim 1, characterized in that, The process of associating and integrating various information from the multimodal information to obtain simulation experiment information includes: The image information and table information included in the multimodal information are analyzed and processed respectively to obtain the second simulation information represented by the image information and the simulation index information represented by the table information; The simulation experimental information is obtained by associating and integrating the first simulation information, the second simulation information, and the simulation index information.

3. The method according to claim 2, characterized in that, The step of analyzing and processing the image information and table information included in the multimodal information to obtain the second simulation information represented by the image information and the simulation index information represented by the table information includes: The image information is subjected to character recognition and feature extraction. Based on the recognized characters and extracted features, second simulation information representing the image information is obtained. The second simulation information includes description information of the image information and simulation point information corresponding to the graphics in the image information. The table information is semantically understood and information is extracted to obtain the simulation index information represented by the table information. The simulation index information includes simulation indexes related to the simulation experiment and the corresponding numerical values ​​of the simulation indexes.

4. The method according to claim 2, characterized in that, The step of associating and integrating the first simulation information, the second simulation information, and the simulation index information to obtain the simulation experiment information includes: Based on at least one of the first simulation information, the second simulation information, and the simulation index information, determine the first simulation type to which the simulation experiment belongs; Based on the simulation indicators included in the simulation indicator information, the correspondence between the first simulation information and the second simulation information related to each simulation indicator is determined, and the simulation experiment information includes the first simulation type and the correspondence.

5. The method according to any one of claims 1 to 4, characterized in that, After obtaining the simulation experiment information, the method further includes: The simulation experiment information is checked for completeness based on a preset first evaluation rule; the first evaluation rule includes at least one of the following: whether simulation index information is missing, whether simulation index information is duplicated, and whether simulation information corresponding to the simulation index information is missing. If the simulation experiment information is determined to be incomplete based on the inspection results, the simulation experiment information is corrected. The correction process includes at least one of the following: adding missing simulation index information, deleting duplicate simulation indexes, and adding simulation information corresponding to the simulation index information.

6. The method according to any one of claims 1 to 4, characterized in that, The first knowledge base retrieved based on the simulation experiment information includes: Based on the first simulation type in the simulation experiment information, the first knowledge base is retrieved to determine the first circuit structure corresponding to the first simulation type; Based on the correspondence of each simulation index in the simulation experiment information, the first knowledge base is retrieved to obtain the set parameter range and calculation formula corresponding to each simulation index. Obtain first knowledge containing the first circuit structure, the set parameter range and calculation formula corresponding to each simulation index.

7. The method according to claim 1, characterized in that, The first knowledge includes: the first circuit structure corresponding to the first simulation type of the simulation experiment and the set parameter range and calculation formula corresponding to each simulation index; the step of generating an experimental scheme based on the retrieved first knowledge includes: The steps for building the circuit structure for the simulation experiment are based on the first circuit structure. The steps for setting the parameter range for each simulation index are as follows: determine the parameter range for each simulation index based on the set parameter range for each simulation index. The calculation steps for generating the value of each simulation index based on the calculation formula corresponding to each simulation index.

8. The method according to claim 1, characterized in that, The step of generating the target test platform code based on the retrieved second knowledge includes: Generate initial test platform code based on the retrieved secondary knowledge; The initial test platform code is executed, and the initial test platform code is repaired based on the code execution results to generate the target test platform code.

9. The method according to claim 8, characterized in that, The process of executing the initial test platform code and repairing the initial test platform code based on the code execution results to generate the target test platform code includes: Select the first circuit model corresponding to the first simulation type to which the simulation experiment belongs, run the initial test platform code on the first circuit model, and obtain the code execution result; The credibility of the initial test platform code is determined based on the execution results of the code. If the credibility does not meet the preset conditions, the initial test platform code is optimized until a target test platform code whose credibility meets the preset conditions is obtained.

10. The method according to claim 9, characterized in that, The step of determining the credibility of the initial test platform code based on the code execution result includes: The difference between the code execution result and the theoretical execution result corresponding to the first circuit model is determined, and the credibility of the initial test platform code is determined based on the difference.

11. A test platform construction apparatus, characterized in that, The device includes: The information extraction and processing unit is used to extract information from the target document to obtain multimodal information related to the simulation experiment. The multimodal information includes first simulation information represented by text, image information, and table information. It is also used to perform association and integration processing on the various types of information included in the multimodal information to obtain simulation experiment information. The scheme generation unit is used to retrieve a first knowledge base based on the simulation experiment information; the first knowledge base includes the following knowledge required for simulation experiments of various simulation types: simulation type and the following information corresponding to the simulation type: circuit structure, setting parameter range corresponding to simulation index, and simulation index calculation formula; it is also used to generate an experimental scheme based on the retrieved first knowledge. The code building unit is used to retrieve a second knowledge base based on the experimental scheme and generate target test platform code based on the retrieved second knowledge. The second knowledge base includes knowledge related to programming languages.

12. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the test platform construction method according to any one of claims 1 to 10.

13. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is loaded and executed by the processor, it implements the test platform construction method according to any one of claims 1 to 10.

14. A computer program product, characterized in that, The computer program product includes a computer program that, when loaded and executed by a processor, implements the test platform construction method as described in any one of claims 1 to 10.

Images