LLM-based system and method for quickly establishing automatic test software, electronic equipment and storage medium

Abstract

The invention discloses an LLM-based automatic test software rapid establishment system and method, electronic equipment and a storage medium, and the test software system comprises a script conversion system which is used for inputting a script matched with an SCPI and converting the script into a high-level language source code; the compiling system is used for compiling the high-level language source code into a dynamic link library; the large language model interaction system is stored in the large language model server, converts a declaration file corresponding to the radio frequency test instrument into Prompt, and reads and stores data returned by the large language model; the project processing system is developed, the blank template project is created, and the stored return data is written into the blank template project to obtain the project which can be used for executing the radio frequency test. According to the method and the device, the required script file is configured once, and the relevant function corresponding to the radio frequency test instrument can be permanently called to realize a specific function; the repeated process of looking up a user manual and writing codes again due to instrument replacement during the radio frequency automatic test is avoided.

Description

Technical Field

[0001] The present invention relates to the field of radio frequency testing technology, and in particular to a system and method for quickly building an automated testing software based on LLM, an electronic device and a storage medium. Background Art

[0002] LXI (LAN eXtension for Instrumentation) is a standard for test and measurement equipment that allows instruments to be controlled via an Ethernet connection. Instrument control and interaction are mainly accomplished through SCPI (Standard Commands for Programmable Instruments, a standardized communication protocol).

[0003] In the field of RF testing, various RF test instruments are widely used in the design, development and verification of communication systems. In order to improve work efficiency and realize automated testing, RF test instruments usually follow the LXI standard, which defines the specifications for remote control over the network, so that equipment from different manufacturers can communicate through standardized interfaces. RF test software mainly completes the control and interaction of LXI instruments through SCPI.

[0004] The IVI Foundation (Interoperable Virtual Instruments Foundation, a standard organization for the virtual instrument industry) has developed standardized interfaces for various types of instruments. However, because manufacturers do not refer to standards for certain specific functions or details, different RF instruments may have different SCPIs even for the same functions.

[0005] When conducting RF testing, software tools are often required to automatically control RF instruments according to a fixed process. This type of software tool can be called automated test software. RF instruments, DUTs, and automated test software constitute the RF automated test environment. When setting up an RF automated test environment, you first need to determine the instruments and models to be used. Then, the automated test software connects, controls, and queries the instruments. After specific SCPI command issuance steps, the test data is obtained and the test is completed.

[0006] If the instruments in the RF automated test environment are completely determined, the implementation of the automated test software is relatively fixed and changes are minimal. However, in actual applications, some instruments in the RF automated test environment are often replaced with other brands and models. Since different instruments may have different SCPIs, programmers need to consult the user manual of the newly replaced instrument to obtain the SCPI and re-implement the code. Each time an instrument is replaced, this process may be repeated, greatly reducing the efficiency of RF automated testing.

[0007] For example, GPT4, DeepSeek, Qwen or similar LLM models are artificial intelligence models trained on large-scale data sets to understand and generate natural language text. LLM is widely used in the field of NLP (natural language processing) and can be used for tasks such as question-answering systems, code generation, and code quality detection. Currently, since there is no LLM specifically trained for the field of RF automated testing, it is impossible to get accurate output directly from LLM through regular questions and answers. It is necessary to optimize the specific information in the field into the Prompt and Input of LLM in order to better utilize the powerful code generation capabilities of LLM and provide an effective solution for the rapid construction of RF automated testing software.

[0008] As mentioned above, some problems of LLM may affect the quality of automated testing software code, mainly manifested in: First, hallucination output: when generating code or text, LLM may generate content that contradicts current facts; second, insufficient knowledge in specific fields: LLM is usually trained on a general corpus and may lack expertise in the field of RF testing. When performing code generation tasks in this field, LLM may not be able to generate code that meets the specifications and standards of this field; third, it is not timely: because LLM is trained based on existing data and the trained model is used for reasoning, the reasoning output content depends on the data during training and cannot continue to use the latest data for reasoning; fourth, it relies on accurate prompts: by providing LLM with appropriate prompts, LLM can be guided to output accurate content. Incorrect or unclear prompts will affect the accuracy, certainty and quality of LLM output content. Summary of the invention

[0009] The present invention provides an LLM-based rapid construction automated test software system and method, an electronic device and a storage medium to address the deficiencies of the above-mentioned prior art and to solve the problem of frequently consulting the instrument user manual and modifying the software code when replacing a radio frequency test instrument.

[0010] In order to achieve the purpose of the present invention, the following technologies are proposed: On the one hand, a method for quickly building an automated testing software system based on LLM is provided, including: A script conversion system, which is used to input the script matched by SCPI corresponding to the function of the radio frequency test instrument, and convert the input script into a high-level language source code; A compilation system that compiles the high-level language source code in the script conversion system into a dynamic link library or an executable program; The large language model interaction system is stored in the large language model server, converts the declaration file corresponding to the RF test instrument into a prompt, and transmits it to the large language model through a specific method such as Http or Cmd in combination with the input business input, and reads and stores the data returned by the large language model; Develop an engineering processing system based on a radio frequency test scenario, wherein the so-called radio frequency test scenario is a Windows system, a Qt development environment, etc. Create a blank template project, write the return data stored in the large language model interaction system to the location corresponding to the blank template project, and obtain a project that can be used to perform radio frequency testing.

[0011] On the other hand, a method for quickly building an automated test software based on LLM is provided, and the method for quickly building an automated test software system based on LLM to test the radio frequency includes the following steps: Step 01, according to the RF test instrument provided, determine the model and user manual of the RF test instrument; Step 02, configuring the functions of the RF test instrument matched in step 01 and the SCPI corresponding to the functions to the script, and entering the obtained script into the script conversion system; Step 03, loading the script entered in step 02 through the script conversion system; Step 04, converting the script entered in step 02 into high-level language source code through a script conversion system; Step 05, compiling the high-level language source code obtained in step 04 into a dynamic link library or an executable program through a compilation system; Step 06, according to the declaration file in the high-level language source code converted in step 04, create a corresponding prompt in the large language model interaction system; Step 07, inputting the input corresponding to the business into the large language model interaction system; Step 08, assemble the Prompt provided in step 06 and the Input input in step 07 into Json, and query the big language model in the big language model interaction system; Step 09, the large language model interaction system returns data according to the query in step 08, and reads and stores the returned data; Step 10, verifying the accuracy of the returned data obtained in step 09, and determining the correctness of the returned data obtained in step 09; If the returned data obtained in step 09 is correct, then return to step 011; If the returned data in step 09 is wrong, return to step 06 and recreate the optimized prompt; Step 011, creating a blank template project on the development project processing system; Step 012, the development project processing system writes the returned data obtained in step 09 to the corresponding position of the corresponding file in the blank template project created in step 011; Step 013, the development engineering processing system compiles the RF test engineering according to the dynamic link library or executable program in step 05 and the script entered in step 02 to obtain an executable RF test engineering; Step 014, calling the executable RF test project obtained in step 013 to perform the test through the RF test instrument in step 01.

[0012] Furthermore, the following steps are performed during the RF test: Step 015, during the RF test process, determine whether the RF test instrument needs to be replaced; Step 016, if it is determined in step 015 that the radio frequency test instrument needs to be replaced, then determine whether the radio frequency test instrument to be used has a dynamic link library or an executable program; If the radio frequency test instrument to be used has a dynamic link library or an executable program, return to step 06; If the radio frequency test instrument to be used does not have a dynamic link library or an executable program, return to step 01; Step 017, if it is determined in step 015 that the RF test instrument does not need to be replaced, then determine whether the SCPI script of the RF test instrument matched in step 01 is missing; When the SCPI script of the RF test instrument matched in step 01 is missing, return to step 02; When the SCPI script of the RF test instrument matched in step 01 is not missing, step 018 is executed; Step 018, determining whether the radio frequency test instrument matched in step 01 exists; If the radio frequency test instrument matched in step 01 exists, return to step 01; If the radio frequency test instrument matched in step 01 does not exist, return to step 019; Step 019, complete the RF test.

[0013] Furthermore, step 02 configures the function of the RF test instrument matched in step 01 and the SCPI corresponding to the function into the script, and the configuration method adopted is a function. The function serializes the SCPI according to the parameters input in step 02 and provides a corresponding string.

[0014] Furthermore, in step 02, when configuring the function of the RF test instrument matched in step 01 and the SCPI corresponding to the function into the script, the following steps are adopted: Step 021, create a script; Step 022, defining and setting the function matched by the RF test instrument function; Step 023, format SCPI in the function and return; Step 024, define the global module of the RF test instrument matched in step 01 and store the function.

[0015] Furthermore, the high-level language source code converted in step 04 includes a general interface code and a specific function interface code corresponding to the radio frequency test instrument in step 01.

[0016] Furthermore, in step 04, when converting the specific function interface code corresponding to the radio frequency test instrument, the following steps are performed: Step 041, calling the function corresponding to the specific function in the script entered in step 02 to obtain the formatted SCPI; Step 042, sending the SCPI obtained in step 041 to the radio frequency test instrument in step 01; Step 043, read the result after the radio frequency test instrument processes the SCPI and return it.

[0017] In the third aspect, an electronic device is also provided, comprising at least one processor and a memory; wherein the memory stores computer execution instructions; the computer execution instructions stored in the memory are executed by the at least one processor, so that the at least one processor executes the method for quickly building automated testing software based on LLM.

[0018] In a fourth aspect, a computer-readable storage medium is also provided, characterized in that a computer program is stored thereon, and when the computer program is executed by a processor, the device where the storage medium is located is controlled to execute the method for quickly building automated testing software based on LLM.

[0019] The advantages of the above technical solution are: The present invention can permanently call the relevant functions corresponding to the radio frequency test instrument to realize specific functions by configuring the required script file once, thereby avoiding the repeated process of re-referring to the user manual and writing codes due to replacing the instrument during radio frequency automated testing.

[0020] The present invention configures the script file, which has the common dynamically adjustable properties of scripts. In actual testing, the formatting process in the script function can be modified according to needs to quickly adapt to other testing needs, avoiding the need to re-modify the source code and compile conventional development tests during the testing process.

[0021] The present invention generates a specific prompt by reading a declaration file in a radio frequency test instrument, combines the business implementation described in natural language, improves the accuracy and certainty of the LLM output result, and utilizes the powerful code generation capability of LLM to realize the rapid construction of automated test software without manual code writing, thus saving manpower and equipment costs and improving test and production efficiency.

[0022] The present invention can reduce the phantom output of LLM by putting the data related to the RF test field into Prompt instead of retraining the model, and can also timely utilize the cutting-edge data in the RF test field, so as to obtain relatively accurate output results at a lower LLM training cost, and lay a foundation for detecting and optimizing the quality of the output results and further improving the accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.

[0024] Figure 1 A structural diagram of a fast-built automated testing software system based on LLM is shown.

[0025] Figure 2 A flowchart of a method for quickly building automated testing software based on LLM is shown. DETAILED DESCRIPTION

[0026] Example 1 like Figure 1 As shown, an automated test software system based on LLM for rapid construction includes: a script conversion system for entering a script matched by SCPI corresponding to a function of a radio frequency test instrument, and converting the entered script into a high-level language source code.

[0027] A compilation system compiles the high-level language source code in the script conversion system into a dynamic link library or an executable program.

[0028] The large language model interaction system is stored in the large language model server, converts the declaration file corresponding to the RF test instrument into a prompt, and transmits it to the large language model through specific methods such as Http or Cmd in combination with the input business Input, and reads and stores the data returned by the large language model.

[0029] Develop an engineering processing system based on a radio frequency test scenario, wherein the so-called radio frequency test scenario is a Windows system, a Qt development environment, etc. Create a blank template project, write the return data stored in the large language model interaction system to the location corresponding to the blank template project, and obtain a project that can be used to perform radio frequency testing.

[0030] Example 2 A method for quickly building an automated test software based on LLM is described as an embodiment in combination with an SMA100B signal source, a Lua script, a C++ language, and a Qwen2.5 model. The method tests radio frequency by quickly building an automated test software system based on LLM, including the following steps: Step 01, determine the model and user manual of the RF test instrument based on the RF test instrument provided, which generally includes a signal source, spectrum analyzer, vector network analyzer, oscilloscope, power meter, phase noise analyzer, comprehensive tester, programmable power supply, programmable temperature chamber or other LXI instruments.

[0031] Step 02, configure the functions of the RF test instrument matched in step 01 and the SCPI corresponding to the functions into a script file (such as Lua, Python, etc.), and enter the obtained script into the script conversion system. The script is used to store SCPI because the script syntax is generally concise and clear, dynamically adjustable, and can be quickly developed and iterated. It is cross-platform and highly compatible, and is suitable for automation tasks. At the same time, since it is necessary to call and execute related functions when obtaining a specific SCPI string, the medium storing SCPI information needs to be dynamically loaded and run at any time. Therefore, a script file (rather than a database or ordinary static configuration file) is used to store the abstract function information of the instrument. When entering, common LXI instrument categories and specific models should be entered into the script. The categories include signal source, spectrum analyzer, vector network analyzer, oscilloscope, power meter, phase noise analyzer, comprehensive tester, programmable power supply, programmable temperature chamber or other LXI instruments. Each function in the script should clearly annotate the relevant information of the current function (common function or specific function), such as function description, input parameter description, and return information description. Each script file corresponds to a specific RF test instrument model under each category.

[0032] Specifically, step 02 configures the function of the RF test instrument matched in step 01 and the SCPI corresponding to the function into the script. The configuration method adopted is a function. The SCPI is serialized according to the parameters input in step 02 in the function, and a corresponding string is provided.

[0033] Specifically, in step 02, when configuring the function of the RF test instrument matched in step 01 and the SCPI corresponding to the function into the script, the following steps are adopted: Step 021, create a script, for example: SMA100B.lua; Step 022, define and set the function that matches the RF test instrument function, for example: local functionsetSweFreqStop(v1), the function should provide clear and unambiguous comments, such as: -- Set the scan cutoff frequency -- v1: frequency value string, unit Hz -- eg: setSweFreqStop("1000") means setting the scan cutoff frequency to 1000Hz Step 023, format SCPI in the function and return it, for example: return string.format("SENS:SWE:FREQ:STOP %s\n", v1). This step requires reference to the user manual of the RF test instrument. Step 024, define the global module of the RF test instrument matched in step 01 and store the function, for example: ModuleSMA100B = {} ModuleSMA100B.setSweFreqStop=setSweFreqStop return ModuleSMA100B.

[0034] Step 03, load the script entered in step 02 through the script conversion system. This step can load the script through Lua's C interface function luaL_dofile; Step 04, convert the script entered in step 02 into high-level language source code (such as C++, C# and other high-level languages) through the script conversion system. The high-level language source code content includes the declaration file and implementation file of each function in the script, wherein the declaration file can be trimmed and supplemented to make it a usable prompt. The declaration file and implementation file of each function mentioned above correspond to a specific RF test instrument model.

[0035] Specifically, the high-level language source code converted in step 04 includes a general interface code and a specific function interface code corresponding to the radio frequency test instrument in step 01.

[0036] More specifically, the universal interface code includes codes for controlling the radio frequency test instrument, such as connecting and disconnecting the instrument, setting a timeout, sending a specified SCPI, and reading data.

[0037] When the conversion of the specific function interface code corresponding to the RF test instrument is performed in step 04, the following steps are performed: step 041, calling the function corresponding to the specific function in the script recorded in step 02 to obtain the formatted SCPI. step 042, sending the SCPI obtained in step 041 to the RF test instrument in step 01. step 043, reading the result after the RF test instrument processes the SCPI and returning it.

[0038] For example: First, convert and obtain two C++ source code files SMA100B.h and SMA100B.cpp; Next, the function is declared in SMA100B.h: string setSweFreqStop(string v1, ViStatus&status), The statement has completely carried the comments in the script, which is used to generate the prompt of LLM later, and adds the function execution result parameter, which is used by LLM to judge the function call result; Finally, the function is implemented in SMA100B.cpp: string code = string("return ModuleSMA100B.setSweFreqStop('%1')").arg(v1); luaL_dostring(m_lua, code.toStdString().c_str()); string scpi = lua_tostring(m_lua, -1); write(scpi.toStdString().c_str(), scpi.size(), writeCntRet); By assembling the Lua executable code string, calling the setSweFreqStop function in ModuleSMA100B.lua and passing in the parameter v1 to be set, the SCPI string returned by the function execution is obtained, and finally the SCPI string is sent to the RF test instrument. Obviously, this step should also obtain the status value and make an exception judgment, and read the response data of the RF test instrument for processing after sending the SCPI.

[0039] Step 05, compile the high-level language source code obtained in step 04 into a dynamic link library or executable program (such as .dll (or .so), .lib, .h, etc. files) through the compilation system. The dynamic link library or executable program can provide function declaration files to the outside world and store function call implementation files (such as dll, so files), thereby greatly reducing the size of the source code file in the project. The optional calling method is to not compile the dynamic link library, but to directly use the high-level language source code to complete subsequent operations.

[0040] Step 06, according to the declaration file in the high-level language source code converted in step 04, create a corresponding Prompt in the large language model interaction system. During the creation process, the above declaration file is cut and supplemented according to the needs to obtain the required Prompt. During operation, it is also necessary to consider the operating environment of the automated test software and the collection of RF test instruments. Prompt should clearly define the RF test instruments used in this RF automated test and the function information of each function of the RF test instruments, so that the high-level source code generated by LLM can accurately call these functions. At the same time, Prompt should also clearly define the format of the output code (such as pure source code output, code snippet output), working mode (such as sub-thread calling instrument function), code style preferences (such as camel case naming method), etc., to improve the certainty of the automated test software code.

[0041] For example: "I need to control the virtual instrument through C++, and have encapsulated the following classes and interfaces: Device Name:SMA100B Device interface: / / Constructor InsScpi is a packaged class, which only needs to be instantiated once: InsScpi* ins = new InsScpi(); SMA100B(InsScpi* parent); / / Open the device addr format: TCPIP0::192.168.1.1::inst0::INSTR ViStatus openDevice(QString addr); / / Close the device ViStatus closeDevice(); / / Set the device timeout msec: timeout, milliseconds ViStatus setTimeOut(ViUInt32 msec); / / Send data synchronously cnt: expected send length, retCount: actual send length ViStatus write(const char* cmd, ViUInt32 cnt, ViPUInt32retCount); / / Synchronous reading of data data: read data storage address, size: expected reading length, retCount: actual reading length ViStatus read(char* data, int size, ViPUInt32 retCount); / / Get the device address QString addr() const {return m_addr;} / / Get device session ViSession session() const {return m_session;} / / -- Set the scan cutoff frequency / / -- v1: frequency value string, unit Hz / / -- eg: setSweFreqStop("1000") means setting the scan cutoff frequency to 1000Hz string setSweFreqStop(QString v1, ViStatus&status); Please help me generate main.cpp, create a lambda function in it, and run it through std::thread, Note: Only plain text code content is output, no other instructions are given. The relevant header files are included by default, and the relevant services are implemented in the lambda function. Obviously, Prompt can also specify implementation methods, naming and specifications, boundary and exception handling, performance requirements and other related information.

[0042] Step 07: According to the use case of RF automated testing, input the input corresponding to the service to the large language model interaction system. The input should be based on the read RF test instrument model, and the content should be clear and unambiguous, and the service provided should be reasonable and achievable.

[0043] Step 08, assemble the Prompt provided in step 06 and the Input input in step 07 into Json, and query the large language model in the large language model interaction system.

[0044] For example: "Instantiate and open 1 SMA100B (IP is 192.168.1.10), set the output power to a fixed 10db; set the scanning cutoff frequency to 10GHz, hold for 10s, then set the frequency according to the start frequency 1MHz, cutoff frequency 1GHz, step 1MHz, and hold each frequency point for 100ms".

[0045] Step 09: The large language model interaction system returns data according to the inquiry in step 08, and reads and stores the returned data. The returned data can be complete code content or code snippets.

[0046] Step 10, verifying the accuracy of the returned data obtained in step 09, and determining the correctness of the returned data obtained in step 09.

[0047] If the returned data obtained in step 09 is correct, then the process returns to step 011.

[0048] If the returned data in step 09 is incorrect, return to step 06 and recreate and optimize the prompt.

[0049] Step 011, create a blank template project on the development project processing system.

[0050] Step 012, the development project processing system writes the returned data obtained in step 09 to the corresponding position of the corresponding file in the blank template project created in step 011.

[0051] Step 013, the development project processing system compiles the RF test project according to the dynamic link library or executable program in step 05 and the script entered in step 02 to obtain an executable RF test project.

[0052] Step 014, calling the executable RF test project obtained in step 013 to perform the test through the RF test instrument in step 01.

[0053] The following steps are performed during the RF test: Step 015, during the RF test process, determine whether the RF test instrument needs to be replaced; Step 016, if it is determined in step 015 that the radio frequency test instrument needs to be replaced, then determine whether the radio frequency test instrument to be used has a dynamic link library or an executable program; If the radio frequency test instrument to be used has a dynamic link library or an executable program, return to step 06; If the radio frequency test instrument to be used does not have a dynamic link library or an executable program, return to step 01; Step 017, if it is determined in step 015 that the RF test instrument does not need to be replaced, then determine whether the SCPI script of the RF test instrument matched in step 01 is missing; When the SCPI script of the RF test instrument matched in step 01 is missing, return to step 02; When the SCPI script of the RF test instrument matched in step 01 is not missing, step 018 is executed; Step 018, determining whether the radio frequency test instrument matched in step 01 exists; If the radio frequency test instrument matched in step 01 exists, return to step 01; If the radio frequency test instrument matched in step 01 does not exist, return to step 019; Step 019, complete the RF test.

[0054] Example 3 An electronic device comprises at least one processor and a memory; wherein the memory stores computer-executable instructions; the computer-executable instructions stored in the memory are executed by the at least one processor, so that the at least one processor executes the method for quickly building automated testing software based on LLM described in Example 2.

[0055] Example 4 A computer-readable storage medium, characterized in that a computer program is stored thereon, and when the computer program is executed by a processor, the device where the storage medium described in Example 2 is located is controlled to execute the method for quickly building automated testing software based on LLM.

[0056] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A fast-building automated testing software system based on LLM, characterized in that: include: A script conversion system, which is used to input the script matched by SCPI corresponding to the function of the radio frequency test instrument, and convert the input script into a high-level language source code; A compilation system that compiles the high-level language source code in the script conversion system into a dynamic link library or an executable program; The large language model interaction system is stored in the large language model server, converts the declaration file corresponding to the RF test instrument into a prompt, transmits it to the large language model in combination with the input business input, and reads and stores the data returned by the large language model; Develop an engineering processing system, which creates a blank template project based on the RF test scenario, and writes the return data stored in the large language model interaction system to the location corresponding to the blank template project to obtain a project that can be used to perform RF testing.

2. A method for quickly building automated testing software based on LLM, characterized in that: By quickly building an automated test software system based on LLM as described in claim 1, the radio frequency is tested, including the steps of: Step 01, according to the RF test instrument provided, determine the model and user manual of the RF test instrument; Step 02, configuring the functions of the RF test instrument matched in step 01 and the SCPI corresponding to the functions to the script, and entering the obtained script into the script conversion system; Step 03, loading the script entered in step 02 through the script conversion system; Step 04, converting the script entered in step 02 into high-level language source code through a script conversion system; Step 05, compiling the high-level language source code obtained in step 04 into a dynamic link library or an executable program through a compilation system; Step 06, according to the declaration file in the high-level language source code converted in step 04, create a corresponding prompt in the large language model interaction system; Step 07, inputting the input corresponding to the business into the large language model interaction system; Step 08, assemble the Prompt provided in step 06 and the Input input in step 07 into Json, and query the big language model in the big language model interaction system; Step 09, the large language model interaction system returns data according to the query in step 08, and reads and stores the returned data; Step 10, verifying the accuracy of the returned data obtained in step 09, and determining the correctness of the returned data obtained in step 09; If the returned data obtained in step 09 is correct, then return to step 011; If the returned data in step 09 is wrong, return to step 06 and recreate the optimized prompt; Step 011, creating a blank template project on the development project processing system; Step 012, the development project processing system writes the returned data obtained in step 09 to the corresponding position of the corresponding file in the blank template project created in step 011; Step 013, the development engineering processing system compiles the RF test engineering according to the dynamic link library or executable program in step 05 and the script entered in step 02 to obtain an executable RF test engineering; Step 014, calling the executable RF test project obtained in step 013 to perform the test through the RF test instrument in step 01.

3. The method for quickly building automated testing software based on LLM according to claim 2, characterized in that: The following steps are performed during the RF test: Step 015, during the RF test process, determine whether the RF test instrument needs to be replaced; Step 016, if it is determined in step 015 that the radio frequency test instrument needs to be replaced, then determine whether the radio frequency test instrument to be used has a dynamic link library or an executable program; If the radio frequency test instrument to be used has a dynamic link library or an executable program, return to step 06; If the radio frequency test instrument to be used does not have a dynamic link library or an executable program, return to step 01; Step 017, if it is determined in step 015 that the RF test instrument does not need to be replaced, then determine whether the SCPI script of the RF test instrument matched in step 01 is missing; When the SCPI script of the RF test instrument matched in step 01 is missing, return to step 02; When the SCPI script of the RF test instrument matched in step 01 is not missing, step 018 is executed; Step 018, determining whether the radio frequency test instrument matched in step 01 exists; If the radio frequency test instrument matched in step 01 exists, return to step 01; If the radio frequency test instrument matched in step 01 does not exist, return to step 019; Step 019, complete the RF test.

4. The method for quickly building automated testing software based on LLM according to claim 2, characterized in that: Step 02 configures the function of the RF test instrument matched in step 01 and the SCPI corresponding to the function into the script. The configuration method adopted is function. The function serializes the SCPI according to the parameters input in step 02 and provides the corresponding string.

5. The method for quickly building automated testing software based on LLM according to claim 4 is characterized in that: In step 02, when configuring the functions of the RF test instrument matched in step 01 and the SCPI corresponding to the functions into the script, use the following steps: Step 021, create a script; Step 022, defining and setting the function matched by the RF test instrument function; Step 023, format SCPI in the function and return; Step 024, define the global module of the RF test instrument matched in step 01 and store the function.

6. The method for quickly building automated testing software based on LLM according to claim 2, characterized in that: The high-level language source code converted in step 04 includes a general interface code and a specific function interface code corresponding to the radio frequency test instrument in step 01.

7. The method for quickly building automated testing software based on LLM according to claim 6, characterized in that: When converting the specific function interface code corresponding to the radio frequency test instrument in step 04, the following steps are performed: Step 041, calling the function corresponding to the specific function in the script entered in step 02 to obtain the formatted SCPI; Step 042, sending the SCPI obtained in step 041 to the radio frequency test instrument in step 01; Step 043, read the result after the radio frequency test instrument processes the SCPI and return it.

8. An electronic device, characterized in that: It comprises at least one processor and a memory; wherein the memory stores computer-executable instructions; the computer-executable instructions stored in the memory are executed by the at least one processor, so that the at least one processor executes the method for quickly building automated testing software based on LLM as described in any one of claims 2-7.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by the processor, the device where the storage medium is located is controlled to execute the method for quickly building automated testing software based on LLM as described in any one of claims 2 to 7.

Images